Datasets:

ccdv
/

arxiv-summarization

Dataset card Data Studio Files Files and versions Community

Subset (2)

section · 216k rows

Split (3)

train · 203k rows

article stringlengths 0 745k	abstract stringlengths 0 92.7k
in classical coding theory , the _ weight distribution _ of a code is a useful tool for measuring a linear code s performance under maximum likelihood ( ml ) decoding . for codes decoded using modern high - performance suboptimal decoding algorithms such as sum - product ( sp ) or linear - programming ( lp ) decoding , the _ pseudoweight _ is the appropriate analog of the codeword weight . there are different definitions of pseudoweight for different channels ; one of primary importance is the _ additive white gaussian noise _ ( awgn ) pseudoweight . the pseudoweight distribution considers all codewords in all codes derived from finite covers of the tanner graph , which compete with the codewords to be the best sp decoding solution . the set of pseudocodewords has a succinct characterization in terms of the so - called _ fundamental polytope _ or equivalently , the _ fundamental cone _ @xcite . also , pseudocodewords arising from finite covers of the tanner graph were shown to be equivalent to those responsible for failure of lp decoding @xcite . while much of the existing work in this area is concerned with performance characterization of particular codes , the performance of _ ensembles _ of _ low - density parity - check _ ( ldpc ) codes @xcite is also of interest . in @xcite , the growth rate of the weight distribution of irregular ldpc codes was derived , and a numerical technique was presented for its approximate evaluation . it was shown in ( * ? ? ? * corollary 50 ) that @xmath2-regular ensembles with @xmath3 have a ratio of minimum awgn - pseudoweight to block length @xmath4 which decreases to zero asymptotically as @xmath5 . apart from this result , to the author s knowledge no _ ensemble _ results exist in the literature concerning awgn - pseudoweight . in this paper , we make a first step in this direction . we define the degree-@xmath6 _ pseudoweight enumerating function _ of a linear block code , and use this concept to find an expression for the growth rate of the awgn - pseudoweight of regular ldpc code ensembles . we also present simulation results for the @xmath7-regular and @xmath8-regular ldpc code ensembles . we begin by providing some general settings and definitions . for @xmath9 , we denote the multinomial coefficient by @xmath10 for @xmath11 with @xmath12 for each @xmath13 and @xmath14 , we denote the multivariate entropy function by @xmath15 all logarithms in the paper are to the base @xmath16 . let @xmath17 be a linear block code of length @xmath4 over the binary field @xmath18 , defined by @xmath19 where @xmath20 is an @xmath21 matrix over @xmath18 called the _ parity - check matrix _ of the code @xmath17 . also denote @xmath22 , @xmath23 and for each @xmath24 @xmath25 the _ tanner graph _ of a linear block code @xmath17 over @xmath18 with parity - check matrix @xmath26 is an equivalent characterization of @xmath26 . the tanner graph @xmath27 has vertex set @xmath28 , and there is an edge between @xmath29 and @xmath30 if and only if @xmath31 . we denote by @xmath32 the set of neighbors of a vertex @xmath33 . we next define what is meant by a finite cover of a tanner graph . ( @xcite ) a graph @xmath34 is a _ finite cover _ of the tanner graph @xmath27 if there exists a mapping @xmath35 which is a graph homomorphism ( @xmath36 takes adjacent vertices of @xmath37 to adjacent vertices of @xmath38 ) , such that for every vertex @xmath39 and every @xmath40 , the neighborhood @xmath41 of @xmath42 is mapped bijectively to @xmath32 . ( @xcite ) a cover of the graph @xmath38 is said to have degree @xmath6 , where @xmath6 is a positive integer , if @xmath43 for every vertex @xmath39 . we refer to such a cover graph as an @xmath44__-cover _ _ of @xmath38 . let @xmath34 be an @xmath6-cover of the tanner graph @xmath27 representing the code @xmath17 with parity - check matrix @xmath26 . the vertices in the set @xmath45 are called _ copies _ of @xmath29 and are denoted @xmath46 , where @xmath47 . similarly , the vertices in the set @xmath48 are called _ copies _ of @xmath30 and are denoted @xmath49 , where @xmath50 . less formally , given a code @xmath17 with parity check matrix @xmath26 and corresponding tanner graph @xmath38 , an @xmath6-cover of @xmath38 is a graph whose vertex set consists of @xmath6 copies of each vertex @xmath29 and @xmath6 copies of each vertex @xmath30 , such that for each @xmath50 , @xmath51 , the @xmath6 copies of @xmath29 and the @xmath6 copies of @xmath30 are connected in an arbitrary one - to - one fashion . for any @xmath52 , an @xmath6-_cover codeword _ is a labelling of vertices of the @xmath6-cover graph with values from @xmath18 such that all parity checks are satisfied . we denote the label of @xmath53 by @xmath54 for each @xmath47 , @xmath55 , and we may then write the @xmath6-cover codeword in vector form as @xmath56 it is easily seen that @xmath57 belongs to a linear code @xmath58 of length @xmath59 over @xmath18 , defined by an @xmath60 parity - check matrix @xmath61 . to construct @xmath61 , for @xmath62 and @xmath63 , @xmath24 , we let @xmath64 , and @xmath65 it may be seen that @xmath37 is the tanner graph of the code @xmath58 corresponding to the parity - check matrix @xmath61 . we next define the concept of _ pseudocodeword _ as follows . let @xmath17 be a linear code of length @xmath4 with parity - check matrix @xmath26 . for any positive integer @xmath6 , a vector @xmath66 of length @xmath4 with nonnegative integer entries is said to be a _ degree_-@xmath6 _ pseudocodeword _ of the code @xmath17 if and only if there exists an @xmath6-cover codeword @xmath57 with @xmath67 for all @xmath63 . the _ pseudoweight _ of a degree-@xmath6 pseudocodeword @xmath68 of the code @xmath17 is equal to the vector @xmath9 , where @xmath68 has @xmath69 entries equal to @xmath70 for each @xmath13 . note that the pseudoweight as defined here corresponds to the `` type '' of a pseudocodeword in the notation of @xcite . note also that this notion of pseudoweight is applicable to different channels such as the additive white gaussian noise ( awgn ) channel , binary symmetric channel ( bsc ) or binary erasure channel ( bec ) . the awgn - pseudoweight of a pseudocodeword @xmath68 of length @xmath4 is defined by @xcite @xmath71 and its bsc - pseudoweight and bec - pseudoweight are defined in @xcite ( see also ( * ? ? ? * section 6 ) ) . the _ fundamental cone _ @xmath72 of the @xmath21 parity - check matrix @xmath26 is equal to the set of vectors @xmath73 such that @xmath74 for all @xmath63 , and @xmath75 [ def : cone ] in @xcite , it was shown that if @xmath17 is a binary linear code with an @xmath21 parity - check matrix @xmath26 , then a length-@xmath4 integer vector @xmath68 is a pseudocodeword of @xmath26 if and only if @xmath76 and @xmath77 where @xmath72 denotes the fundamental cone of @xmath26 , and the matrix @xmath26 in ( [ eq : pcw_condition_2_parity ] ) is interpreted over the integers . we next define the concept of pseudoweight enumerating function of a block code . the degree-@xmath6 _ pseudoweight enumerating function _ ( pwef ) of a block code @xmath17 of length @xmath4 is equal to @xmath78 where @xmath79 , @xmath9 and @xmath80 denotes the number of degree-@xmath6 pseudocodewords - cover codewords corresponding to a particular pseudocodeword . ] of the code with pseudoweight @xmath81 . [ prop : b ] the degree-@xmath6 pwef of the single parity - check ( spc ) code of length @xmath82 is @xmath83 - t^{(m)}({{\mbox{\boldmath $ x$ } } } ) \label{eq : b_definition}\ ] ] where @xmath84 , @xmath85 , @xmath86 , and for @xmath87 @xmath88 where the set @xmath89 in ( [ eq : t_recursion ] ) is the set of integer vectors @xmath90 satisfying @xmath91 for all @xmath92 , @xmath93 and where @xmath94 is even . in this case @xmath26 is a length-@xmath82 row vector of ones , so we have @xmath95 where ( using definition [ def : cone ] ) @xmath96 is the set of integer vectors @xmath81 which satisfy ( @xmath96@xmath97 ) : : @xmath98 ( @xmath96@xmath99 ) : : @xmath100 ( @xmath96@xmath101 ) : : if there exists @xmath102 with @xmath103 and @xmath104 for @xmath105 , then @xmath106 it is straightforward to check that in ( [ eq : b_definition ] ) , the term @xmath107\ ] ] takes into account all integer vectors @xmath81 which satisfy conditions ( @xmath96@xmath97 ) and ( @xmath96@xmath99 ) , and the term @xmath108 takes into account all those which violate the condition ( @xmath96@xmath101 ) . in particular @xmath109 , @xmath110 , and @xmath111 . also note that @xmath112 \nonumber \\ - \frac{\partial t^{(m)}({{\mbox{\boldmath $ x$}}})}{\partial x_r } \label{eq : b_derivative}\end{gathered}\ ] ] for @xmath13 . for a positive integer @xmath4 , we define a @xmath2-regular ldpc code ensemble @xmath113 as follows . the tanner graph of an ldpc code from the ensemble consists of @xmath4 variable nodes of degree @xmath114 , and @xmath115 check nodes of degree @xmath82 . the variable and check node sockets are connected by a permutation on the @xmath116 edges of the graph , each permutation being equiprobable . the concept of degree-@xmath6 assignment is defined next . this definition is a generalization of the definition of assignment in @xcite ( the definition in @xcite corresponds to that of a degree-@xmath97 assignment ) . a _ degree-@xmath6 assignment _ is a labelling of the edges of the tanner graph with numbers from the set @xmath117 . an assignment is said to have _ @xmath118 if @xmath119 edges are labelled @xmath70 for each @xmath13 . an assignment is said to be @xmath6-_check - valid _ if according to this labelling , every check node recognizes a valid local degree-@xmath6 pseudocodeword . for any positive integer @xmath6 , the growth rate of the degree-@xmath6 awgn - pseudoweight distribution of the @xmath2-regular ldpc code ensemble sequence @xmath120 is defined by @xmath121 \label{eq : growth_rate_result}\ ] ] where @xmath122 denotes the expectation operator over the ensemble @xmath113 , and @xmath123 denotes the number of degree-@xmath6 pseudocodewords of awgn - pseudoweight @xmath124 of a randomly chosen ldpc code in the ensemble @xmath113 . the limit in ( [ eq : growth_rate_result ] ) assumes the inclusion of only those positive integers @xmath4 for which @xmath125 and @xmath126 $ ] is positive ( i.e. , where the expression whose limit we seek is well defined ) . we next define a notion of _ asymptotic goodness _ of an ldpc code ensemble sequence . for each @xmath0 , let @xmath127 be the growth rate of the degree-@xmath6 pseudoweight distribution of an ldpc code ensemble sequence , and let @xmath128 . the ensemble sequence is said to be _ asymptotically good _ if and only if @xmath129 . the following theorem constitutes the main result of the paper . the growth rate of the degree-@xmath6 pseudoweight distribution of the @xmath2-regular ldpc code ensemble sequence @xmath120 is given by @xmath130 where @xmath131 , @xmath132 and @xmath133 are the solutions to the system of @xmath1 equations in @xmath1 unknowns is given by proposition [ prop : b ] . ] @xmath134 for each @xmath13 , @xmath135 - j \log x_{0,r } \\ = \lambda \left ( 2r \sum_{s=1}^{m } s q_s - \alpha r^2 \right ) \label{eq : lagrange_mult_theorem}\end{gathered}\ ] ] for each @xmath13 , and @xmath136 satisfying @xmath137 and @xmath138 for each @xmath13 . [ thm : growth_rate ] consider a degree-@xmath6 pseudocodeword @xmath66 with pseudoweight @xmath139 , where @xmath132 . this pseudocodeword naturally induces a degree-@xmath6 assignment of pseudoweight @xmath140 . using ( [ eq : awgn - pseudoweight ] ) , the awgn - pseudoweight of @xmath68 may be written as @xmath141 where @xmath142 rearranging ( [ eq : alpha ] ) , and defining @xmath143 appropriately , yields ( [ eq : g_definition ] ) . the expected number of degree-@xmath6 pseudocodewords of pseudoweight @xmath139 is then @xmath144 = \binom{n}{{{\mbox{\boldmath $ q$}}}n } \cdot p^{(m)}_{\mbox{\scriptsize c - valid}}(j { { \mbox{\boldmath $ q$ } } } ) \ ; , \label{eq : expected_over_q}\ ] ] where @xmath145 represents the probability that a randomly chosen degree-@xmath6 assignment with pseudoweight @xmath146 is @xmath6-check - valid . this probability is given by @xmath147 where @xmath148 denotes the number of @xmath6-check - valid degree-@xmath6 assignments of pseudoweight @xmath146 . the numerator of ( [ eq : prob_valid_assign ] ) may be written as be the number of ways of obtaining an outcome @xmath149 in experiment @xmath150 , and let @xmath151 be the number of ways of obtaining an outcome @xmath152 in experiment @xmath153 . also let @xmath154 be the number of ways of obtaining an outcome @xmath155 in the combined experiment @xmath156 such that @xmath157 . denoting @xmath158 , the generating functions @xmath159 , @xmath160 and @xmath161 are related by @xmath162 . ] @xmath163 \ ; .\ ] ] we next make use of the following result from ( * ? ? ? * theorem 2 ) : let @xmath164 denote a multivariate polynomial with nonnegative coefficients . for a fixed vector of positive rational numbers @xmath165 , consider the set of positive integers @xmath166 such that @xmath167 for each @xmath13 and @xmath168 . then either this set is empty , or it has infinite cardinality ; if @xmath169 is one such @xmath166 , then so is @xmath170 for every positive integer @xmath114 . in the latter case , the following limit is well defined and exists : @xmath171 \\ = \log r({{\mbox{\boldmath $ x$}}}_0 ) - \sum_{r=1}^{m } \xi_r \log x_{0,r}\end{gathered}\ ] ] where @xmath131 is the unique positive real solution to the system of equations @xmath172 for each @xmath13 . [ lemma : optimization_1d ] applying this lemma by substituting @xmath173 , @xmath174 and @xmath175 , we obtain that as @xmath5 @xmath176 where @xmath131 is the unique positive real solution to the system given by ( [ eq : bldx0_q_eqn ] ) for each @xmath13 . note that ( [ eq : bldx0_q_eqn ] ) provides an implicit definition of @xmath177 as a function of @xmath178 . using stirling s formula , the multinomial coefficients in ( [ eq : expected_over_q ] ) and ( [ eq : prob_valid_assign ] ) may be approximated as @xmath5 as @xmath179 therefore as @xmath5 @xmath180 \rightarrow \exp \bigg\ { n \bigg [ \frac{j}{k } \log b^{(m)}({{\mbox{\boldmath $ x$}}}_0 ) \\ - j \sum_{r=1}^{m } q_r \log x_{0,r } - ( j-1 ) h({{\mbox{\boldmath $ q$ } } } ) \bigg ] \bigg\}\end{gathered}\ ] ] the expected number of degree-@xmath6 pseudocodewords with awgn - pseudoweight @xmath181 is equal to the sum of the numbers of degree-@xmath6 pseudocodewords with pseudoweight @xmath178 satisfying ( [ eq : g_definition ] ) , i.e. @xmath182 = \sum_{{{\mbox{\scriptsize \boldmath $ q$}}}\ ; : \ ; g({{\mbox{\scriptsize \boldmath $ q$}}})=0 } \mathbb{e}_{{{\mathcal{m}}}_n } \left [ n^{(m)}({{\mbox{\boldmath $ q$ } } } ) \right ] \ ; .\ ] ] note that the asymptotic expression as @xmath183 is dominated by that @xmath178 satisfying ( [ eq : g_definition ] ) which maximizes the argument of the exponential function , @xmath184 . therefore as @xmath183 @xmath182 \rightarrow \exp \left\ { n \left ( \max_{{{\mbox{\scriptsize \boldmath $ q$}}}\ ; : \ ; g({{\mbox{\scriptsize \boldmath $ q$}}})=0 } f({{\mbox{\boldmath $ q$ } } } ) \right ) \right\ } \label{eq : optimization_for_awgn_pw}\ ] ] where @xmath185 and @xmath143 is given by ( [ eq : g_definition ] ) . we solve this constrained optimization problem using lagrange multipliers . at the maximum , we must have @xmath186 for all @xmath187 , where @xmath133 denotes the lagrange multiplier . this yields @xmath188 \\ - j \left ( \sum_{s=1}^{m } \frac{q_s}{x_{0,s } } \frac{\partial x_{0,s}}{\partial q_r } + \log x_{0,r } \right ) \\ - ( j-1 ) \frac{\partial h({{\mbox{\boldmath $ q$}}})}{\partial q_r } = \lambda \frac{\partial g({{\mbox{\boldmath $ q$}}})}{\partial q_r } \end{gathered}\ ] ] which is equivalent to @xmath189 \\ - j \log x_{0,r } + ( j-1 ) \log \left ( \frac{q_r}{1-\sum_{s=1}^{m } q_s } \right ) \\ = \lambda \left ( 2r \sum_{s=1}^{m } s q_s - \alpha r^2 \right ) \ ; .\end{gathered}\ ] ] the term in square brackets is equal to zero for each @xmath13 due to ( [ eq : bldx0_q_eqn ] ) ; therefore this simplifies to ( [ eq : lagrange_mult_theorem ] ) for each @xmath13 . note that for the case @xmath190 , the maximization in ( [ eq : optimization_for_awgn_pw ] ) is trivial and therefore the solution may be obtained directly from ( [ eq : f_definition ] ) as @xmath191 \\ - j \alpha \log x_{0 } - ( j-1 ) h(\alpha ) \label{eq : growth_rate_m1}\end{gathered}\ ] ] where @xmath192 is the unique positive real solution to the equation @xmath193 = \\ x_0 \left [ \left ( 1+x_0 \right)^{k-1 } - \left ( 1-x_0 \right)^{k-1 } \right ] \ ; . \label{eq : x0_m1}\end{gathered}\ ] ] note that @xmath194 is simply the growth rate of the weight distribution in this case , originally obtained in @xcite . also , this solution may be regarded as a special case of theorem [ thm : growth_rate ] where the solution for @xmath133 via ( [ eq : lagrange_mult_theorem ] ) is redundant . in this section the growth rates of the awgn - pseudoweight of two example ldpc code ensembles are evaluated using the solution of theorem [ thm : growth_rate ] . the growth rate curves for the @xmath7-regular ldpc code ensemble and for the @xmath8-regular ldpc code ensemble are shown in figures [ cap : gallager_36_ensemble ] and [ cap : gallager_48_ensemble ] respectively . note that @xmath195 for both ensembles . it is worthwhile to note some distictions between the present analysis and that of ( * ? ? ? * corollary 50 ) . in ( * corollary 50 ) , it is proved that @xmath2-regular ensembles with @xmath3 have a ratio of minimum awgn - pseudoweight to block length @xmath4 which decreases to zero asymptotically as @xmath5 . this result is not in conflict with the results of figures [ cap : gallager_36_ensemble ] and [ cap : gallager_48_ensemble ] . the detrimental pseudocodewords of ( * ? ? ? * corollary 50 ) are derived from the `` canonical completion '' ( * ? ? ? * definition 46 ) and , asymptotically , have awgn - pseudoweight _ sublinear _ in the block length therefore , these pseudocodewords do not appear in the present analysis . also , note that the analysis of ( * ? ? ? * corollary 50 ) takes the limit @xmath196 prior to ( or jointly with ) the limit @xmath5 , in contrast to the present analysis which takes the limit @xmath5 for finite @xmath6 . finally , the result of ( * ? ? ? * corollary 50 ) is concerned with _ minimum _ awgn - pseudoweight and not with the multiplicities of the corresponding pseudocodewords .	a solution is presented for the asymptotic growth rate of the awgn - pseudoweight distribution of regular low - density parity - check ( ldpc ) code ensembles for a selected graph cover degree @xmath0 . the evaluation of the growth rate requires solution of a system of @xmath1 nonlinear equations in @xmath1 unknowns . simulation results for the pseudoweight distribution of two regular ldpc code ensembles are presented for graph covers of low degree .
almost all planet - hosting stars known today are cool stars ( spectral types f - m ) . all cool stars display stellar activity a summarizing term for the occurrance of magnetic phenomena including flares , spots , and coronal high - energy emission . the magnetic activity of planet - hosting stars is an important factor to understand the evolution of exoplanets . high - energy irradiation of close - in planets can lead to atmospheric evaporation @xcite , and coronal mass ejections and the stellar wind can strip away parts of the planetary atmosphere @xcite . the stability and chemistry of exoplanetary atmospheres are both influenced by these stellar activity phenomena . recent transit observations in the uv @xcite and in x - rays @xcite have demonstrated that atmospheres of hot jupiters are strongly extended . modelling of such atmospheres suggests that time - variable phenomena like bow - shocks and variations in the stellar wind may play a role , too @xcite . stellar activity levels of stars with and without planets have received much scrutiny in recent years . activity is well - known to be a function of stellar rotation and therefore , due to magnetic breaking , of stellar age . stellar x - ray luminosity therefore declines with stellar age , as shown in fig . [ age - activity ] , left panel . instead of @xmath0 , see @xcite . ] several studies have noted that exoplanets may also have an influence on the stellar magnetic activity . the underlying idea is that massive close - in exoplanets should be able to interact with their host star magnetically and/or tidally @xcite . however , because stellar activity effectively masks signals used to detect exoplanets , one has to be careful not to interpret intrinsic biases in samples of planet - hosting stars as a physical interaction effect @xcite . a useful way to test for activity enhancements caused by hot jupiters is to look at wide stellar binaries , in which only one star hosts a planet massive and close enough for strong tidal interaction . the star without a planet then acts as a negative control , because its activity evolution should be unperturbed by any planetary effects . one can thus test if both stars display magnetic activity levels appropriate for their common stellar age ( taking into account slight differences due to spectral type ) . we are conducting an observational program using x - ray data and optical spectra to test for such activity differences . * right : * in wide stellar binaries in which one of the stars hosts an exoplanet , some of the planet - hosting stars display a much higher magnetic activity level than expected for their age . this is the case for systems with strong tidal interaction between planet and host star . @xmath1age - activity data from @xcite , @xcite , @xcite , @xcite , @xcite , @xcite . , scaledwidth=100.0% ] * right : * in wide stellar binaries in which one of the stars hosts an exoplanet , some of the planet - hosting stars display a much higher magnetic activity level than expected for their age . this is the case for systems with strong tidal interaction between planet and host star . @xmath1age - activity data from @xcite , @xcite , @xcite , @xcite , @xcite , @xcite . , scaledwidth=100.0% ] .observed x - ray luminosities of the planet - hosting primaries and the secondaries which are not known to possess planets . the x - ray luminosity of the secondary is used to estimate the age of the system . assuming that both stelar components are co - eval , we can derive an _ a priori _ estimate how x - ray luminous the primary should be ( column @xmath2 ) . the actual observed x - ray luminoities of the primaries are given in column @xmath3 . [ cols="<,^,^,^,^,^,^,^,^ " , ] we have collected observational data for five exoplanet systems with wide stellar binaries so far , using _ xmm - newton _ and _ chandra _ for x - ray observations of the stellar corona , as well as flwo s tres spectrograph for optical high - resolution spectra . we list the observed x - ray luminosities for both stellar components of each system in table [ lxtable ] . we estmate the age of the system from the x - ray brightness of the secondary ( non - planet - hostimg ) star . this yields an estimate for the x - ray luminosity of the primary ( planet - hosting ) star , assuming that both stars are co - eval . more detailed analyses of the spectral x - ray properties of some of the stars can be found in @xcite and @xcite for hd 189733 a and b , @xcite for corot-2 a , @xcite for @xmath4 boo and @xcite for @xmath5 and a. for three of the systems we find that the expected and observed x - ray luminosities for the planet - hosting stars are in agreement ; for two systems , however , the expected and observed x - ray luminosities disagree by an order of magnitude or more ; see fig . [ age - activity ] , right panel . these systems both host a hot jupiter in a very close orbit , and the host star is a late g or k dwarf . for the systems in which no discrepancy is found , the innermost planet is either small ( 55 cnc ) or the host star is an f star with a very thin outer convective envelope ( @xmath4 boo and @xmath5 and ) . we interpret our initial findings as a manifestation of tidal interaction between the planet and the host star . theoretical studies show that hot jupiters induce tidal bulges on the host star ; cool stars can dissipate the energy contained in the bulges much more effectively than hot stars due to turbulent eddies in the convective envelopes , see @xcite and @xcite . observationally , stars with substantial convective envelopes have also been found to have exoplanets in orbits with low orbital obliquity , which is also believed to be a consequence of tidal interaction @xcite . the hot jupiters in the hd 189733a and corot-2a systems may therefore have inhibited the spin - down of their host stars , by transferring angular momentum from the planetary orbit into the stellar spin . to test this hypothesis further , we are collecting high - resolution optical spectra for our sample to measure the projected rotational velocities @xmath6 of the stars as well as chromospheric activity indicators such as ca ii h and k and h @xmath7 . our initial data suggests that there is indeed a large difference in activity levels for systems in which the planet exerts a strong tidal interaction on its host star , while systems with weak tidal interaction do not show elevated activity levels . observations of a larger number of systems which can be tested for possible tidal effects from the planet are under way . the consequences of planet - induced stellar high activity are substantial for understanding exoplanetary systems and , on a larger scale , stellar evolution . for systems with hot jupiters , age estimates can no longer be derived from stellar rotation and activity ; also , the time - integrated high - energy irradiation of exoplanets will be much higher due to the prolonged high activity of the host stars , so that planetary evaporation may have had a much stronger influence on the planet zoo we observe today than current models suggest . in terms of stellar evolution , stellar population synthesis models are just beginning to include the star s spin as a parameter . first results show that the stellar rotation is a fundamental parameter for its evolution @xcite , so that it is crucial to know if hot jupiters significantly change the spin of their host stars , which comprise ca . 1% of all cool stars @xcite .	it is a long - standing question in exoplanet research if hot jupiters can influence the magnetic activity of their host stars . while cool stars usually spin down with age and become inactive , an input of angular momentum through tidal interaction , as seen for example in close binaries , can preserve high activity levels over time . this may also be the case for cool stars hosting a hot jupiter . however , selection effects from planet detection methods often dominate the activity levels seen in samples of exoplanet host stars , and planet - induced , systematically enhanced stellar activity has not been detected unambiguously so far . we have developed an approach to identify planet - induced stellar spin - up avoiding the selection biases from planet detection , by using visual proper motion binaries in which only one of the stars possesses a hot jupiter . this approach immediately rids one of the ambiguities of detection biases : with two co - eval stars , the second star acts as a negative control . we present results from our ongoing observational campaign at x - ray wavelengths and in the optical , and present several outstanding systems which display significant age / activity discrepancies presumably caused by their hot jupiters .
over the last year and a half , a number of very high redshift ( @xmath0 ) galaxies have been reported ( dey et al . 1998 , hu et al . 1998 , weymann et al . 1998 , spinrad et al . 1998 , chen et al . 1999 , hu et al . 1999 ) , and our knowledge of the @xmath8 galaxy population ( e.g. , hu et al . 1996 , petitjean et al . 1996 , fontana et al . 1996 , hu & mcmahon 1996 , trager et al . 1997 , franx et al . 1997 , hu et al . 1997 , frye & broadhurst 1998 , hu et al . 1998 , soifer et al . 1998 , pell et al . 1999 , steidel et al . 1999 , hu et al . 1999 ) has been substantially increased . these samples have recently been joined by radio galaxy identifications which reach @xmath0 ( van breugel et al . 1999a , 1999b ) . at the highest redshifts now probed , identification of galaxies basically relies on two key diagnostics : the lyman break across the continuum and the redshifted ly@xmath1 line . both characteristics have been used to select and to verify high-@xmath9 galaxy candidates . the extremely faint nature of the high-@xmath9 galaxy population in infrared continuum light and the increasing strength and frequency of nightsky emission lines at very long wavelengths makes it clear that the contrast and detection of high - redshift galaxies from the ground is challenging , even for the new generation of 8- to 10-m class telescopes . the discovery of very high equivalent width emitters at @xmath10 ( hu & mcmahon 1996 ) indicated the existance of a substantial population of strong ly@xmath1 emitters which could be observed to very high redshifts , beyond @xmath0 . in this paper we report on the status of ongoing ly@xmath1 emission - line searches combined with color - selection data and consider the prospects for future identifications at the very high redshift end . the ly@xmath1 searches are designed to address issues of detectability of field galaxies at very high redshifts , using the increased contrast of the object against the background sky when viewed in the ly@xmath1 emission line . the contrast of emission features against the neighboring continuum is also enhanced by the @xmath11 magnification of line widths , producing high equivalent width signatures . through the use of spectral regions free of strong night sky lines ( e.g. , fig . [ fig:0 ] ) , these studies are also designed to address object statistics over a range of different redshift intervals , since our ability to detect and reliably confirm very high - redshift galaxies is otherwise highly dependent on the location of emission features with respect to background nightsky lines . narrowband ly@xmath1 searches combined with deep multi - color data on the hawaii survey fields and hdf have been used as a training set ( cowie & hu 1998 ; hu 1998 ) to test emission line identifications in terms of color and equivalent width diagnostics over an extensive spectroscopic database , starting at @xmath12 , where color selection is robust , and working out to successively higher redshifts ( @xmath13 ) . this allows studying continuity in properties of genuine ly@xmath1 emitters and both color - selected galaxies and low - redshift foreground emitters at progressively longer wavelengths , where higher sky backgrounds from the increased density of strong nightsky lines and fainter continua for the more distant galaxies are issues . a variant approach is to use a fabry - perot ( calar alto deep imaging survey , meisenheimer et al.1998 ) , which requires a high - level of precision in low - level light processing and flat - fielding . the present ly@xmath1 surveys have reached @xmath14 ( hu et al . 1999 ) , and @xmath2 narrowband studies have recently been started . from the initial wide - field narrowband surveys for ly@xmath1 emitters at @xmath15 and 4.5 , hu et al . ( 1998 ) estimated a surface density of ly@xmath1 emitters of @xmath16 down to an emission flux of @xmath17 ergs @xmath18 sec@xmath19 . these estimates were found to be consistent with a complementary deep spectroscopic study , where the slit provided limited spatial sampling , but wide wavelength coverage at somewhat greater sensitivity , and recovered 4 ly@xmath1 emitters with redshifts from @xmath20 . preliminary results from the @xmath21 searches suggest number densities perhaps a factor of 6 lower than the value for the lower redshift systems given by hu et al . the surface number density estimates for ly@xmath1 emitters imply that sufficiently deep spectroscopic exposures may intercept some of these systems , at least over the lower redshift ranges . in fact , the first published @xmath0 galaxy , rd-1 , was a serendipitously discovered ly@xmath1 emitter at @xmath22 reported by dey et al . ( 1998 ) during deep long - slit observations of a neighboring object in the field , with both emission line and break identified in the spectra , and with the break confirmed by broad - band filter imaging . for rd-1 , estimates of the continuum magnitude above the lyman break are @xmath23 ab mags using the broad - band imaging data , with emission - line flux = @xmath24 ergs @xmath18 s@xmath19 , @xmath25 ( dey et al . 1998 ) . the success of the spectroscopic searches depends on the surface number density at the redshifts and fluxes being probed . equivalent width estimates , and particularly continuum magnitudes for faint objects measured in dispersed modes have larger associated errors than estimates tied to standard broadband filter photometry . higher sensitivities trade off against small area coverage , and for slit spectroscopy there is the additional issue of whether with chance superpositions sample objects fall only partly within the slit . the additional problem in using this method as a search technique is that it is difficult to assess the fraction of emitters missed because they lie in regions of strong nightsky lines , and that for such cases confirming observations can be extremely hard to make . an interesting variant of this search technique is to use slitless spectroscopy ( chen et al . 1999 ) . incomplete object sampling by a slit is removed , but the general problem of overlapping galaxy spectra must be dealt with . using extremely deep stis parallel exposures from hst , chen et al . ( 1999 ) identify a ly@xmath1 emitter at a probable redshift of 6.68 . at @xmath26 ly@xmath1 falls in regions of strong oh lines , which makes confirmation with ground - based observations extremely challenging . the strong depression of the continuum below the redshifted lyman break caused by the numerous neutral hydrogen absorbers dubbed `` the ly@xmath1 forest '' is the most notable feature in the galaxy continuum light of very high redshift systems , where the extreme faintness of these objects , typically @xmath27 mags ( ab ) above the break , precludes the use of detailed absorption features to estimate the redshift . this method has been most successfully applied ( e.g. , fernndez - soto et al . 1999 ) in the hubble deep field ( hdf ; williams et al . 1996 ) , where the high precision of the photometric measurements permits robust detection of a break signature , and where the availability of photometric data at longer wavelengths ( nicmos infrared observations ) allows discrimination of lyman break galaxies from red objects . hdf 4 - 473.0 ( weymann et al . 1998 ) , shown to be a ` @xmath28 dropout ' by the absence of detectable flux in the wfpc2 f606w filter , combined with a roughly flat @xmath29 spectrum out through 1.6@xmath30 m , was confirmed as a @xmath31 galaxy through deep lris spectra on keck . an emission line with flux @xmath32 ergs @xmath18 s@xmath19 and @xmath33 , identified as ly@xmath1 , was used to establish the redshift . the magnitude above the break is @xmath34 ab mags . a second color selected target , hdf 3 - 951.1 and 3 - 951.2 , was identified by the color break , and placed at an estimated redshift @xmath22 ( spinrad et al . 1998 ) on the basis of the break location . for this close pair , no emission features are detected ( fig . 2 of spinrad et al . 1998 ) , with upper limits on possible ly@xmath1 more than a factor of 10 below the observed flux of hdf 4 - 473.0 spinrad et al . ( 1998 ) consider the possibility that the absence of ly@xmath1 emission might be associated with its brighter continuum . because studies at this depth and precision of color measurement are only available for a small region of sky ( @xmath35 arcsec@xmath36 each for the hdf and hdf south ) , key questions one would like to address from the hdf color - break selected galaxies are : ( 1 ) how typical are these properties of the high - redshift galaxies ; what are the consequences for future detections ? ( 2 ) are the bright high - redshift galaxies devoid of ly@xmath1 emission ? figs . [ fig:1 ] and [ fig:2 ] show images and a spectral energy distribution ( sed ) for the @xmath37 galaxy , ssa22-hcm1 ( hu et al . this object is an ` @xmath38 ' dropout , and is notably absent in deep keck lris images in @xmath39 , @xmath28 , and @xmath38 , which reach @xmath40 limits of @xmath39=28.3 , @xmath28=28.2 , and @xmath41 for a @xmath42 diameter aperture . the sed shows both the strong lyman break and the ly@xmath1 emission feature ( flux=@xmath43 ergs @xmath18 s@xmath19 ; @xmath44 = 175 ) , which was confirmed with deep lris spectroscopy . the increased contrast in the appearance of the object as detected in the 8185 narrowband filter compared with the line - free @xmath4 band filter around 9200 may be noted . for these measurements we use fluxes measured in the narrowband filter , instead of values recovered from spectroscopic extractions , for greater precision . aperture 1@xmath45 errors are estimated from laying down random apertures away from identified sources , since in the deep exposures magnitude limits are set by the background faint source population ( fig . [ fig:1 ] ) . the measured continuum above the break , 25.5 mags ( ab ) , is a magnitude higher than the weymann et al . ( 1998 ) observed estimate for hdf 4 - 473.0 , and lies between our estimated continuum magnitudes of 24.9 and 25.7 ( based on our deep @xmath4 band observations of the hdf ) for the 3 - 951 pair . ssa22-hcm1 is the brightest of the @xmath46 ly@xmath1 emitters surveyed to date , and in contrast to 3 - 951 has strong emission . however , it appears that more typical emitters at these redshifts have properties like 4 - 473.0 ( hu et al . 1999 , in preparation ) . we can summarize the expected properties for ly@xmath1-emitting galaxies above @xmath47 , based on the current @xmath48 survey , as : ly@xmath1 fluxes @xmath32 ergs @xmath18 s@xmath19 , equivalent widths of a few 100 , and continuum magnitudes fainter than @xmath49 . efforts to model the high - redshift emitters have begun ( e.g. , thommes 1998 , haiman & spaans 1999 ) , and we can use these as a starting point to estimate the detection limits that will be required for working at higher redshifts . fig . [ fig:3 ] from thommes ( 1998 ) shows the expected surface number density of ly@xmath1 emitters for four redshift intervals ( @xmath50 ) corresponding to spectral regions free of strong nightsky lines . overplotted are the data from cowie & hu ( 1998 ) for @xmath51 , and other upper limits . it can be seen that , in agreement with the properties summarized above , surveys will need to exceed @xmath52 detection limits of @xmath53 ergs @xmath18 s@xmath19 ( @xmath54 w m@xmath55 ) , and preferably cover wide areas . because ssa22-hcm1 is so bright in @xmath4 it is interesting to examine color statistics for objects in ( @xmath56 ) vs. @xmath4 . [ fig:4 ] shows the color distribution of objects in the hdf and ssa22 fields over a sample area of 75 arcmin@xmath36 . magnitudes are measured on the ab system using @xmath42 diameter corrected apertures . the fiducial color boundary used here of @xmath57 lies marginally above the range of possible m star contaminants , which might be distinguished by compactness criteria . at faint magnitudes the depth of the @xmath38 exposure dominates the errors in the @xmath3 color measurement , which can be quite large . filled symbols indicate known @xmath0 galaxies ( or in the case of hdf 3 - 395.1 and 3 - 395.2 , the summed magnitudes of the object pair ) . in the case of hdf 4 - 473.0 , the colors are estimated from deep wfpc2 and nicmos photometry ( weymann et al . the @xmath52 @xmath4 band criterion is not sufficiently deep to include ssa22-hcm1 , the @xmath31 galaxy hdf 4 - 473.0 , or the @xmath22 emission - line galaxy rd1 ( based on the estimated @xmath58 ) , but the threshold is only a few tenths of a magnitude from reaching the brighter two objects . thus , it appears that the use of deep @xmath4-band filter imaging and @xmath56 color selection on the large telescopes may be a possible way to extend methods for selecting candidates to higher ( @xmath59 ) redshifts for ground - based studies , although the background sky brightness near 9200 is high . this could be used in combination with a near - ir color measurement to distinguish break objects from highly reddened galaxies . the new @xmath60 quasars discovered by the sloan survey ( fan et al . 1999 ) provide a means of evaluating the expected continuum depression below the ly@xmath1 line at high redshifts . because of the rapid increase in the gas density with redshift , it is expected that , at the higher redshifts , there will be a sizable break . in fig . [ fig:5 ] we show a spectrum of the newly discovered @xmath6 quasar , j033829.31 + 002156.3 , for which we determine a factor of 4 break across the emission line . by comparison , zhang et al . ( 1997 ) find , at @xmath61 , for @xmath62 , @xmath63 and a haardt - madau ( 1996 ) spectrum and ionizing flux , that the spectrum of a @xmath61 quasar will have an average flux in the region between @xmath64 and @xmath65 that is 10% of the continuum value which would be present in the absence of lyman alpha scattering . a plot of the continuum break across ly@xmath1 for various @xmath8 quasars ( schneider et al . 1991 ; kennefick et al . 1995 ) and galaxies is shown in fig . [ fig:6 ] . @xmath66 is the index ( oke & korycanski 1982 ) @xmath67 in the rest - frame wavelength range @xmath64 to @xmath65 . it is useful to consider possible directions for future studies . a number of @xmath8 galaxies have been found using cluster lens systems ( e.g. , trager et al . 1997 , franx et al . 1997 , frye & broadhurst 1998 , pell et al . 1999 ) , where amplification of the source has increased its detectability . in general , such discoveries do not provide information on the statistics , continuum magnitudes , and base emission - line fluxes of the high - redshift galaxies found , but for very well - studied cluster lens systems ( e.g. , bzecourt et al . 1999 ) it may be possible to recover such information . typical magnification is about 1 magnitude ( smail et al . 1999 ) , and this may be a fruitful way to extend high - redshift searches . the availability of new ir spectrographs for the 810-m . class telescopes will make it possible to study [ o ] associated with the @xmath68 systems ( e.g. egami 1997 ) . such observations may provide more information on the degree of dust extinction in the high - redshift galaxies detected by their ly@xmath1 . however , for @xmath0 systems [ o ] lies in the thermal ir and can not be readily studied from the ground . for redshifts @xmath0 , high equivalent width h@xmath1 emission from high ionization foreground extragalactic h regions is a possible contaminant which can mimic ly@xmath1 . such systems may have no detectable continuum , and [ n ] suppressed to less than 2% of h@xmath1 ( stockton & ridgway 1998 ) . the incidence of these systems is not known , but a few cases have been encountered in our @xmath21 and @xmath2 surveys and identified by wide wavelength spectra covering the corresponding [ o ] and [ o ] . extensive wavelength coverage is required to verify sources as ly@xmath1 emitters . searches for yet higher @xmath9 galaxies are of necessity forced into the near ir since the light below @xmath69 will be essentially extinguished by the strong ly @xmath1 forest blanketing or by h i gunn - peterson ( miralda - escud & rees 1998 ) at the redshifts where the igm was neutral . we can look forward to detecting these objects with the next generation space telescope . this work was supported in part by nasa grants go-6626.01 - 95a , go-7266.01 - 96a , and ar-6377.06 - 94a from space telescope science institute , which is operated by aura , inc . under nasa contract . results reported here include observations made as a visiting astronomer at the w. m. keck observatory , which is jointly operated by the california institute of technology , the university of california , and the national aeronautics and space administration . smail , i. , ivison , r. j. , blain , a. w. , & kneib , j .- 1999 , in _ after the dark ages : when galaxies were young ( the universe at @xmath70 ) _ , ed . s. s. holt , & e. p. smith ( new york : aip press ) , 312 , astro - ph/9810281	we review the properties of @xmath0 galaxies studied with _ hst _ and with the keck telescopes , and discuss the detectability of ly@xmath1 emission - line galaxies out to @xmath2 based on these data and ongoing narrowband imaging surveys . the brightest sources may show @xmath3 color breaks , although the high sky background at @xmath4 ( @xmath5 ) , makes such observations challenging for typical faint sources . keck lris observations of the @xmath6 sdss quasar and @xmath0 galaxies observed with _ hst _ in the hdf show that the strength of the lyman break is evolving more slowly than extrapolations from models at @xmath7 would predict .
the motion of quarks inside the nucleons plays an important role in some effects which are at present intensively investigated both experimentally and theoretically . actual goal of this effort is to obtain a more consistent 3-d picture of the quark - gluon structure of nucleons . for example the quark transversal momentum creates the asymmetries in particle production in polarized ( sidis ) or in unpolarized ( cahn effect ) experiments on deep inelastic scattering ( dis ) . relevant tool for the study of these effects is the set of the transverse momentum dependent distributions ( tmds ) . apparently , a better understanding of the quark intrinsic motion is also a necessary condition to clarify the role of quark orbital angular momenta in generating nucleon spin . we have paid attention to these topics in our recent studies , see @xcite and citations therein . in particular we have shown that the requirements of lorentz invariance ( li ) and the nucleon rotational symmetry in its rest frame ( rs ) , if applied in the framework of the 3-d covariant quark - parton model ( qpm ) , generate a set of relations between parton distribution functions . recently we obtained within this approach relations between usual parton distribution functions and the tmds . the wanzura - wilczek approximate relation ( ww ) and some other known relations between the @xmath0 and @xmath1 structure functions were similarly obtained in the same model before @xcite . let us remark that the ww relation has been obtained independently also in another approaches @xcite in which the li represents a basic input . the aim of the present report is to consistently apply the assumption li&rs to the kinematics of dis and to obtain the constraints on related kinematical variables . that is a complementary task to the study of above mentioned relations between distribution functions , which depend on these variables . so , the report can be considered as an addendum to our former papers related to the covariant qpm @xcite . since the present discussion is motivated and based on our earlier study of a covariant version of qpm , obtained results correspond only to the leading order of a more consistent qcd treatment . in this sense it would be interesting to compare our results with the leading order of a real qcd approach , like e.g. the recent study of perturbative qcd evolution of tmds @xcite . however such task would go beyond the scope of this short report . anyway , in general a comparison between the experimental data and the leading order predictions can be important and instructive . first , let us shortly remind the properties of the bjorken variable@xmath2 which plays a crucial role in phenomenology of lepton nucleon scattering . regardless of mechanism of the process , this invariant parameter satisfies@xmath3 this is a very well - known textbook result . a possible proof is suggested also in @xcite . now let us consider a qpm approach , where the process of lepton nucleon scattering is initiated by the lepton interaction with a quark ( see fig . [ fe1 ] ) , which obeys @xmath4 ( @xmath5 ) in initial ( final ) state , @xmath6 is initial nucleon momentum.,width=188 ] the second equality implies@xmath7 which with the use of relation ( [ k1 ] ) gives@xmath8 the basic input for the construction of qpm is the assumption @xmath9 which allows us to identify@xmath10 and to directly relate the quark momentum to the parameters of scattered lepton . moreover , if one assumes@xmath11 where @xmath12 is the nucleon mass , then one can identify@xmath13 in any reference frame in which the first axis orientation is defined by the vector @xmath14 . this relation can be proved as follows . let us consider eq . ( [ k7 ] ) in the same frame:@xmath15 in the nucleon rest frame we denote the photon momentum components by the subscript @xmath16 and using the usual symbol @xmath17 we have@xmath18 which with the use of eq . ( [ k1 ] ) gives@xmath19 it means that for @xmath20 we obtain@xmath21 ( since for @xmath22 we have also @xmath23 so the ratio @xmath24 does not contradict eq . ( [ ax7 ] ) . ) in a reference frame connected with the rest frame by the lorentz boost in the direction opposite to @xmath25 we have the corresponding ratio@xmath26 now one can easily check that eq . ( [ k7 ] ) with the use of this ratio and eq . ( [ ax7b ] ) imply @xmath27 in this way we have proved that replacement of bjorken variable by the invariant light - cone ratio in eq . ( [ k9 ] ) is valid provided the inequality ( [ k8 ] ) is satisfied . the relation ( [ k9 ] ) expressed in the nucleon rest frame reads@xmath28 which after inserting into ( [ k2 ] ) gives@xmath29 however the most important reason why we require large @xmath30 is in physics . if we accept scenario when a probing photon interact with a quark , we need sufficiently large momentum transfer @xmath30 at which the quarks can be considered as effectively free due to asymptotic freedom . at small @xmath30 the picture of quarks ( with their momenta and other quantum numbers ) inside the nucleon disappear . the rs means that the probability distribution of the quark momenta @xmath31 in the nucleon rest frame depends , apart from @xmath30 , on @xmath32 . it follows that also @xmath33 is allowed , so together with the inequality ( [ k11 ] ) we have@xmath34 the combinations of ( [ k11]),([k12 ] ) imply@xmath35 and if we again refer to rs , then further inequalities are obtained:@xmath36 and@xmath37 where@xmath38 obviously inequality ( [ k15a ] ) is satisfied also in any reference frame boosted in the directions @xmath39 . further , the above inequalities are apparently valid also for average values @xmath40 and @xmath41 . in addition , if one assumes that @xmath42 distribution is a decreasing function , then necessarily@xmath43 the above relations are valid for sufficiently high @xmath30 suggested by eqs . ( [ k6 ] ) and ( [ k8 ] ) . let us note that the on - mass - shell assumption has not been applied for obtaining these relations . these inequalities can be compared with relations obtained in @xcite , where the additional on - mass - shell condition @xmath44 had been applied . corresponding relations are more strict:@xmath45 and@xmath46 however , it is clear that in general the on - mass - shell assumption is not realistic . in the next we will assume only the off - mass - shell approach . first let us summarize more accurately what we have done in the previous section . we assumed : _ _ a ) lorentz invariance__it means that the theoretical description in terms of the standard kinematical variables ( see fig . [ fe1])@xmath47 can be boosted also to the nucleon rest frame . _ b ) inequality _ @xmath48____it means that the light - cone ratio @xmath49 satisfies the same bound ( [ k2 ] ) as the bjorken variable @xmath50 _ _ c ) rotational symmetry__the kinematical region @xmath51 of the quark intrinsic momenta @xmath31 in the nucleon rest frame has rotational symmetry ( i.e. @xmath52 , where @xmath53 is any rotation in @xmath51 ) . for example , in terms of the covariant qpm it means that probabilistic distribution of the quark momenta is controlled by some function @xmath54 . we proved these assumptions imply bounds ( [ k11])([k15a ] ) . now we will shortly comment on the obtained results : _ i ) _ the ratio @xmath49 of light - cone variables ( [ k9 ] ) has a simple interpretation in a frame , where the proton momentum is large - @xmath49 is the fraction of this momentum carried by the quark . however an interpretation of the same variable in the nucleon rest frame is more complicated . in this frame the quark transversal momentum can not be neglected and @xmath49 depends on the both , longitudinal and transversal quark momenta components . in the limit of massless quarks the connection between the variable @xmath49 in ( [ k10 ] ) and the quark momenta components is given by the relations : @xmath55 these variables were used in our recent papers on tmds @xcite . simply the value of invariant variable @xmath49 does not depend on the reference frame , but its interpretation e.g. in the rest frame differs from that in the infinite momentum frame . _ ii ) _ the relations ( [ k15 ] ) , ( [ k15a ] ) , which follow from rs , can be confronted with the experimental data on @xmath41 or @xmath56 . we have discussed the available data in @xcite and apparently relation ( [ k15a ] ) is compatible with the set of lower values @xmath41 corresponding to the leptonic data. on the other hand the second set giving substantially greater @xmath41 and denoted as the hadronic data , seems to contradict this relation . actually the conflict with the relation ( [ k15a ] ) would mean either the conflict with some of the assumptions _ a)c ) _ or simply it can be due to absence of the higher order qcd corrections . however , for possible comparison with the perturbative qcd approach @xcite let us remark the following . this approach generate evolved tmds using as input the existing phenomenological parametrizations extracted from the experimental data . for example , one of the inputs is the scale - independent gaussian fit @xcite@xmath57 } { \pi\left\langle p_{t}^{2}\right\rangle } , \label{k20}\ ] ] where @xmath58 . obviously our concept rs defined above is hardly compatible with this distribution . in fact in the rest frame this distribution gives much greater @xmath59 than the corresponding longitudinal @xmath60term @xmath61 . however rs requires @xmath62 only . let us remark that this imbalance is of the same order as a difference between the two data sets mentioned above . _ iii ) _ the first relation ( [ k15 ] ) apparently contradicts an intuitive , lorentz invariant condition@xmath63 corresponding to bound , space - like quarks . such conflict does not take place for the leading order with quarks on - mass - shell . however , for off - shell quarks the condition ( [ k20a ] ) is incompatible with the assumptions _ a)c ) , _ which imply ( [ k15 ] ) . for example it means that any approach , which is based primarily on the assumption _ _ ( [ k20a ] ) in some starting reference frame ( including the infinite momentum frame ) can not simultaneously satisfy the conditions _ a)c)_. _ iv ) _ in @xcite we explained why the rs , if applied on the level of qpm , follows from the covariant description . in fact it means the assumptions _ a)c ) _ are common for our qpm and for the approaches like @xcite where only lorentz invariance is explicitly required . the predictions of all these models are compatible with the bound ( [ k15a ] ) . this is just a consequence of the fact that general conditions _ a)c ) _ are satisfied in these approaches . another theoretical reasons for rs have been suggested in @xcite . let us remark that the rotational - symmetry properties of the nucleon state in its rest frame play an important role also in the recent studies @xcite . _ v ) _ the relation ( [ k17 ] ) is obtained for the quarks on - mass - shell . in a more general case , where only inequalities ( [ k15 ] ) , ( [ k15a ] ) hold , this relation is replaced by@xmath64 where the term @xmath65 is not a parameter corresponding to the fixed mass , but it is only a number varying in the limits defined by ( [ k15 ] ) . the last relation implies for any @xmath65:@xmath66 which is equivalent to the on - mass - shell relation ( [ k17 ] ) for @xmath67 . this general upper limit for @xmath68 depending on @xmath49 is displayed in fig . let us remark that results on @xmath69 obtained in @xcite are compatible also with the bound ( [ k22 ] ) . an equivalent form of this inequality was probably for the first time presented in @xcite . for @xmath70 ( solid line ) , @xmath71 ( dashed line ) , @xmath72 ( dotted line ) and @xmath73 ( dash - dotted line ) . , width=226 ] in the present report we studied the kinematic constraints due to the rotational symmetry of the quark momenta distribution inside the nucleon in the leading order approach . in particular , we have shown that the light - cone formalism combined with the assumption on the rotational symmetry in the nucleon rest frame imply @xmath74 . only part of existing experimental data on @xmath41 satisfies this bound , but the another part does not . in general , the existing methods for reconstruction of @xmath41 from the dis data are rather model - dependent . these are the reasons why more studies are needed to clarify these inconsistencies , since the phenomenological distributions in the @xmath75 plane at present serve also as an input for a more fundamental calculation of the qcd evolution and other effects . this work was supported by the project av0z10100502 of the academy of sciences of the czech republic . i am grateful to anatoli efremov , peter schweitzer and oleg teryaev for many useful discussions and valuable comments .	we study the kinematics of deep inelastic scattering corresponding to the rotationally symmetric distribution of quark momenta in the nucleon rest frame . it is shown that rotational symmetry together with lorentz invariance can in leading order impose constraints on the quark intrinsic momenta . obtained constraints are discussed and compared with the available experimental data .
elections are a general and widely used framework for preference aggregation in human and artificial intelligence applications . an important negative result from social choice theory , the gibbard - satterthwaite theorem @xcite , states that every reasonable election system is manipulable . however , even though every election system is manipulable , it may be computationally infeasible to determine how to manipulate the outcome . bartholdi , tovey , and trick introduced the computational study of the manipulation problem and this began an exciting line of research that explores the computational complexity of different manipulative attacks on elections ( see , e.g. , @xcite ) . the notion of single - peaked preferences introduced by black is the most important restriction on preferences from political science and economics and is naturally an important case to consider computationally . single - peakedness models the preferences of a collection of voters with respect to a given axis ( a total ordering of the candidates ) . each voter in a single - peaked election has a single most - preferred candidate ( peak ) on the axis and the farther candidates are to the left / right from her peak the less preferred they are . single - plateauedness extends this to model when each voter has multiple most - preferred candidates that appear sequentially on the axis , but are otherwise single - peaked @xcite . this standard model of single - peaked preferences has many desirable social - choice properties . when the voters in an election are single - peaked , the majority relation is transitive @xcite and there exist voting rules that are strategy - proof @xcite ( i.e. , a voter can not misrepresent her preferences to achieve a personally better outcome ) . single - peakedness for total orders can also have an effect on the complexity of many different election attack problems when compared to the general case . the complexity of manipulative attacks often decreases when the voters in an election are single - peaked @xcite , and the winner problems for kemeny , dodgson , and young elections are in p when they are @xmath0-complete in general @xcite . most of the abovementioned research on the computational complexity of manipulation of elections , both for the general case and for single - peaked electorates , has been limited to the assumption that voters have tie - free votes . in many real - world scenarios voters have votes with ties , and this is seen in the online repository preflib @xcite that contains several different preference datasets that contain ties . there are also election systems defined for votes with ties , e.g. , the kemeny rule and the schulze rule @xcite . recent work considers the complexity of manipulation for top - order votes ( votes where all of the ties are between candidates ranked last ) @xcite . fitzsimmons and hemaspaandra considered the complexity of manipulation , control , and bribery for more general votes with ties , and also the case of manipulation for a nonstandard model of single - peakedness for top - order votes . menon and larson later examined the complexity of manipulation and bribery for an equivalent ( for top orders ) model of single - peakedness . fitzsimmons and hemaspaandra use the model of possibly single - peaked preferences from lackner where a preference profile of votes with ties is said to be single - peaked with respect to an axis if the votes can be extended to tie - free votes that are single - peaked with respect to the same axis . menon and larson use a similar model for top orders that they state is essentially the model of single - peaked preferences with outside options @xcite . both fitzsimmons and hemaspaandra and menon and larson find that these notions of single - peakedness exhibit anomalous computational behavior where the complexity of manipulation can increase when compared with the case of single - peaked total orders . we are the first to study the computational complexity of manipulation for the standard model of single - peaked preferences for votes with ties , and for single - plateaued preferences for votes with ties . in contrast to the recent related work using other models of single - peakedness with ties , we find that the complexity of weighted manipulation for @xmath1-candidate scoring rules and for @xmath1-candidate copeland@xmath2 elections for all @xmath3 does not increase when compared to the cases of single - peaked total orders , and that the complexity of weighted manipulation does not increase with respect to the general case of elimination veto elections . we also compare the social choice properties of these different models , and state a surprising result on the relation between the societal axis and the complexity of manipulation for single - peaked preferences . an _ election _ consists of a finite set of candidates @xmath4 and a finite collection of voters @xmath5 . we will sometimes refer to this collection of voters as a _ preference profile_. an _ election system _ @xmath6 is a mapping from an election to a set of winners , which can be any subset of the candidate set ( the nonunique winner model , our standard model ) , or at most a single candidate ( the unique winner model ) . each voter in an election has a corresponding vote ( or _ preference order _ ) over the set of candidates . this is often assumed to be a _ total order _ , i.e. , a strict ordering of the candidates from most to least preferred . formally , a total order is a complete , reflexive , transitive , and antisymmetric binary relation . we use `` @xmath7 '' to denote strict preference between two candidates . similarly , a _ weak order _ is a total order without antisymmetry , so each voter can rank candidates as tied ( which we will sometimes refer to as indifference ) as long as their ties are transitive . we use `` @xmath8 '' to denote a tie between two candidates . a _ bottom order _ is a weak order where all ties are between top - ranked candidates and a _ top order _ is a weak order where all ties are between bottom - ranked candidates . notice that a total order is a weak order , a top order , and a bottom order , that a top order is a weak order , and that a bottom order is a weak order . throughout this paper we will sometimes refer to weak orders as votes with ties . for some of our results we consider weighted elections , where each voter has an associated positive integral weight and a voter with weight @xmath9 counts as @xmath9 unweighted voters all voting the same . our election systems include scoring rules , elimination veto , and copeland@xmath2 . we define each below and the extensions we use to properly consider votes with ties . given an election with @xmath1 candidates , a scoring rule assigns scores to the candidates using its corresponding @xmath1-candidate scoring vector of the form @xmath10 where @xmath11 and each @xmath12 . so , when the preferences of a voter are a total order , the candidate ranked in position @xmath13 receives a score of @xmath14 from that voter . below we present examples of scoring rules and their corresponding @xmath1-candidate scoring vector . * plurality : * with scoring vector @xmath15 . * veto : * with scoring vector @xmath16 . * borda : * with scoring vector @xmath17 . * triviality : * with scoring vector @xmath18 . to use a scoring rule to determine the outcome of an election containing votes with ties we must extend the above definition of scoring rules . we use the definitions of scoring - rule extensions for weak orders from our previous work @xcite , which generalize the extensions introduced for top orders from baumeister et al . and from narodytska and walsh which in turn generalizes the extensions used by emerson for borda . given a weak - order vote , we can write it as @xmath19 , where each @xmath20 is a set of tied candidates , ( so in the case of a total order vote each @xmath20 is a singleton ) . for each @xmath20 , let @xmath21 be the number of candidates strictly preferred to the candidates in @xmath20 . we now state the definitions of each of the four extensions . in example [ ex : scoring - rule - ext ] we present an example of how a given weak - order vote is scored using borda and each of the scoring - rule extensions . * min : * each candidate in @xmath20 receives a score of @xmath22 . * max : * each candidate in @xmath20 receives a score of @xmath23 . * round down : * each candidate in @xmath20 receives a score of @xmath24 . * average : * each candidate in @xmath20 receives a score of @xmath25 for top orders the scoring - rule extensions min , round down , and average are the same as round up , round down , and average used in the work by menon and larson . [ ex : scoring - rule - ext ] given the candidate set @xmath26 and the weak order vote @xmath27 we show the scores assigned to each candidate using borda and each of our extensions . we can write the vote as @xmath28 , so @xmath29 , @xmath30 , and @xmath31 , and @xmath32 , @xmath33 , and @xmath34 . recall that for total orders , the scoring vector for 5-candidate borda is @xmath35 . * borda using min : * @xmath36 , @xmath37 , and @xmath38 . * borda using max : * @xmath39 , @xmath40 , and @xmath38 . * borda using round down : * @xmath41 , @xmath37 , and @xmath38 . * borda using average : * @xmath42 , @xmath43 , and @xmath38 . for elimination veto for total orders , the veto scoring rule is used , the candidate with the lowest score is eliminated , and the rule is repeated on the remaining votes restricted to the remaining candidates until there is one candidate left @xcite . we break ties lexicographically , and for comparison with related work our results for elimination veto use the unique winner model , and for votes with ties we use the min extension . pairwise election systems are one of the most natural cases for considering votes with ties . copeland@xmath2 is an important and well - known election system that is defined using pairwise comparisons between candidates . in a copeland@xmath2 election each candidate receives one point for each pairwise majority election with each other candidate she wins and @xmath44 points for each tie ( where @xmath45 and @xmath46 ) . for votes with ties we follow the obvious extension also used by baumeister et al . and narodytska and walsh . for copeland@xmath2 elections it will sometimes be easier to refer to the _ induced majority graph _ of an election . given an election @xmath47 its induced majority graph is constructed as follows . each vertex in the induced majority graph corresponds to a candidate in @xmath4 , and for all candidates @xmath48 if @xmath49 by majority then there is an edge from @xmath50 to @xmath51 in the induced majority graph . we also will refer to the _ weighted majority graph _ of an election , where each edge from @xmath50 to @xmath51 in the induced majority graph is labeled with the difference between the number of voters that state @xmath49 and the number of voters that state @xmath52 . the computational study of the manipulation of elections was introduced by bartholdi , tovey , and trick , and conitzer , sandholm , and lang extended this to the case for weighted voters and a coalition of manipulators . we define the constructive weighted coalitional manipulation ( cwcm ) problem below . * name : * @xmath53-cwcm * given : * a set of candidates @xmath4 , a collection of nonmanipulative voters @xmath54 , a collection of manipulative voters @xmath55 , and a preferred candidate @xmath56 . * question : * does there exist a way to set the votes of @xmath55 such that @xmath57 is a winner of @xmath58 under election system @xmath53 ? for the case of cwcm for each of our models of single - peaked preferences we follow the model introduced by walsh where the societal axis is given as part of the input to the problem and the manipulators must state votes that are single - peaked with respect to this axis ( for the corresponding model of single - peakedness ) . ( see section [ sec : variants ] for all of the definitions of single - peaked preferences that we use . ) for our np - completeness results we will use the following well - known np - complete problem . * name : * partition * given : * given a collection of @xmath59 positive integers @xmath60 such that @xmath61 . * question : * does there exist a partition of @xmath60 into two subcollections @xmath62 and @xmath63 such that @xmath64 ? we consider four important models of single - peaked preferences for votes with ties . for the following definitions , for a given axis @xmath65 ( a total ordering of the candidates ) and a given preference order @xmath66 , we say that @xmath66 is strictly increasing ( decreasing ) along a segment of @xmath65 if each candidate is preferred to the candidate on its left ( right ) with respect to @xmath65 . in figure [ fig : ex - peak ] we present an example of each of the four models of single - peaked preferences , and in figure [ fig : relation ] we show how the four models relate to each other . we now give the definition of the standard model of single - peakedness from black . [ def : standard ] given a preference profile @xmath5 of weak orders over a set of candidates @xmath4 , @xmath5 is single - peaked with respect to a total ordering of the candidates @xmath65 ( an axis ) if for each voter @xmath67 , @xmath65 can be split into three segments @xmath68 , @xmath69 , and @xmath70 ( @xmath68 and @xmath70 can each be empty ) such that @xmath69 contains only the most preferred candidate of @xmath66 , @xmath66 is strictly increasing along @xmath68 and @xmath66 is strictly decreasing along @xmath70 . observe that for a preference profile of votes with ties to be single - peaked with respect to an axis , each voter can have a tie between at most two candidates at each position in her preference order since the candidates must each appear on separate sides of her peak . otherwise the preference order would not be _ strictly _ increasing / decreasing along the given axis . the model of single - plateaued preferences extends single - peakedness by allowing voters to have multiple most preferred candidates ( an indifference plateau ) @xcite . this is defined by extending definition [ def : standard ] so that @xmath69 can contain multiple candidates . lackner recently introduced another extension to single - peaked preferences , which we refer to as `` possibly single - peaked preferences '' throughout this paper . a preference profile is possibly single - peaked with respect to a given axis if there exists an extension of each preference order to a total order such that the new preference profile of total orders is single - peaked . this can be stated without referring to extensions , for votes with ties , in the following way . given a preference profile @xmath5 of weak orders over a set of candidates @xmath4 , @xmath5 is possibly single - peaked with respect to a total ordering of the candidates @xmath65 ( an axis ) if for each voter @xmath67 , @xmath65 can be split into three segments @xmath68 , @xmath69 , and @xmath70 ( @xmath68 and @xmath70 can each be empty ) such that @xmath69 contains the most preferred candidates of @xmath66 , @xmath66 is weakly increasing along @xmath68 and @xmath66 is weakly decreasing along @xmath70 . notice that the above definition extends single - plateauedness to allow for multiple indifference plateaus on either side of the peak . so for votes with ties , possibly single - peaked preferences model when voters have weakly increasing and then weakly decreasing or only weakly increasing / decreasing preferences along an axis . another generalization of single - peaked preferences for votes with ties was introduced by cantala : the model of single - peaked preferences with outside options . when preferences satisfy this restriction with respect to a given axis , each voter has a segment of the axis where they have single - peaked preferences and candidates appearing outside of this segment on the axis are strictly less preferred and the voter is tied between them . similar to how single - plateaued preferences extend the standard single - peaked model to allow voters to state multiple most preferred candidates , single - peaked preferences with outside options extends the standard model to allow voters to state multiple least preferred candidates . given a preference profile @xmath5 of weak orders over a set of candidates @xmath4 , @xmath5 is single - peaked with outside options with respect to a total ordering of the candidates @xmath65 ( an axis ) if for each voter @xmath67 , @xmath65 can be split into five segments @xmath71 , @xmath68 , @xmath69 , @xmath70 , and @xmath72 ( @xmath71 , @xmath68 , @xmath70 , and @xmath72 can each be empty ) such that @xmath69 contains only the most preferred candidate of @xmath66 , @xmath66 is strictly increasing along @xmath68 and @xmath66 is strictly decreasing along @xmath70 , for all candidates @xmath73 and @xmath74 , @xmath66 states @xmath49 , and for all candidates @xmath75 , @xmath66 states @xmath76 . menon and larson state that the model of single - peakedness for top orders that they use is similar to single - peaked preferences with outside options . it is clear from their paper that for top orders these models are the same . so for the remainder of the paper we will refer to the model used by menon and larson as `` single - peaked preferences with outside options for top orders . '' , the solid line is the single - peaked order @xmath77 , the dashed line is the single - plateaued order @xmath78 , the dashed - dotted line is the single - peaked with outside options order @xmath79 , and the dotted line is the possibly single - peaked order @xmath80 . see figure [ fig : relation ] for the relationships between the four models . ] we now state some general observations on single - peaked preferences with ties , including how the models relate to each other , as well as their social - choice properties . it is easy to see that for total - order preferences each of the four models of single - peakedness with ties that we consider are equivalent . in figure [ fig : relation ] we show how each model relates for weak orders , top orders , and bottom orders . given a preference profile of votes it is natural to ask how to determine if an axis exists such that the profile satisfies one of the above restrictions . this is referred to as the consistency problem for a restriction . bartholdi and trick showed that single - peaked consistency for total orders can be determined in polynomial time ( i.e. , is in p ) , and fitzsimmons extended this result to show that single - peaked , single - plateaued , and possibly single - peaked consistency for weak orders is in p. this leaves the consistency problem for single - peaked preferences with outside options , which we will now show to be in p. it is easy to see that given a preference profile of top orders , it is single - peaked with outside options if and only if it is possibly single - peaked . so it is immediate from the result by lackner that shows that possibly single - peaked consistency for top orders is in p , that the consistency problem for single - peaked preferences with outside options for top orders is in p. for weak orders the construction used to show that single - peaked consistency for weak orders is in p by fitzsimmons can be adapted to hold for the case of single - peaked preferences with outside options for weak orders , so this consistency problem is also in p. given a preference profile @xmath5 of weak orders it can be determined in polynomial time if there exists an axis @xmath65 such that @xmath5 is single - peaked with outside options with respect to @xmath65 . one of the most well - known and desirable properties of an election with single - peaked preferences is that there exists a transitive majority relation @xcite . a majority relation is transitive if for all distinct candidates @xmath81 if @xmath49 and @xmath82 by majority then @xmath83 by majority . when the majority relation is transitive then candidates that beat - or - tie every other candidate by majority exist and they are denoted the weak condorcet winners . this also holds when the voters have single - plateaued preferences @xcite . it is not the case that the majority relation is transitive when preferences in an election are single - peaked with outside options or are possibly single - peaked . fitzsimmons points out that for possibly single - peakedness this was implicitly shown in an entry of table 9.1 in fishburn , where a preference profile of top orders that violates single - peaked preferences was described that does not have a transitive majority relation , and this profile is possibly single - peaked ( and single - peaked with outside options ) . cantala also provides an example of a profile that is single - peaked with outside options that does not have a weak condorcet winner , and so does not have a transitive majority relation @xcite . for total - order preferences , weighted manipulation for any fixed number of candidates is known to be np for every scoring rule that is not isomorphic to plurality or triviality @xcite . for single - peaked total orders , faliszewski et al . completely characterized the complexity of weighted manipulation for 3-candidate scoring rules and this result was generalized by brandt et al . for any fixed number of candidates . these results both showed that the complexity of weighted manipulation for scoring rules often decreases when voters have single - peaked total orders . for top - order preferences , dichotomy theorems for 3-candidate weighted manipulation for scoring rules using round down , min , and average were shown by menon and larson for the general case @xcite and for the case of single - peaked preferences with outside options for top orders @xcite . menon and larson found that , counterintuitively , the complexity of weighted manipulation for single - peaked preferences with outside options often increases when moving from total orders to top orders . we also mention that for the scoring - rule extension max , we earlier found that the complexity of 3-candidate weighted manipulation for borda increases for possibly single - peaked preferences when moving from total orders to top orders @xcite . we show that for the standard model of single - peakedness and for single - plateauedness that the complexity of weighted manipulation for @xmath1-candidate scoring rules using max , min , round down , and average does not increase when moving from total orders to top orders , bottom orders , or weak orders . the following results are close analogs to the case for single - peaked total orders due to brandt et al . . [ lem : sp - first ] if @xmath57 can be made a winner by a manipulation of top - order , bottom - order , or weak - order votes that are single - peaked , single - plateaued , possibly single - peaked , or single - peaked with outside options for a scoring rule using max , min , round down , or average , then @xmath57 can be made a winner by a manipulation ( of the same type ) in which all manipulators rank @xmath57 uniquely first . suppose that a manipulator has a vote of the form @xmath84 where @xmath85 is a weak order over @xmath62 ( where @xmath62 can be empty ) , @xmath86 , each of @xmath87 are nonempty groups of tied candidates , and @xmath88 . we have the following two cases . case 1 : : : if all of the candidates in @xmath89 appear on the same side of the candidates in @xmath62 with respect to the axis ( or @xmath62 is empty ) , then we can change the vote of the manipulator to @xmath90 where @xmath91 and @xmath92 are total orders such that the vote satisfies the given notion of single - peakedness with ties with respect to the axis . case 2 : : : otherwise , we let @xmath93 be the candidates to the left of @xmath62 and @xmath94 be the candidates to the right of @xmath62 with respect to the axis . without loss of generality let @xmath95 . we can then change the vote of the manipulator to @xmath96 . notice that in all of the above cases , for all candidates @xmath97 that @xmath98 does not decrease , regardless of which scoring - rule extension is used . so if @xmath57 is a winner with the initial vote , @xmath57 is still a winner in the vote that ranks @xmath57 uniquely first . [ thm : noninc ] let @xmath99 be a scoring vector . if @xmath44-cwcm is in p for single - peaked total orders , then @xmath44-cwcm is in p for single - peaked and single - plateaued top orders , bottom orders , and weak orders for all our scoring rule extensions . if the societal axis has @xmath57 in the leftmost or rightmost position , then , by lemma [ lem : sp - first ] , we can assume that all manipulators rank @xmath57 uniquely first . there is exactly one such single - peaked or single - plateaued vote with ties , namely to rank @xmath57 first and strictly rank the remaining candidates according to their position on the axis . so , let the societal axis be @xmath100 . we first consider the p cases for single - peaked total orders described in lemma 6.6 of brandt et al . where for all @xmath101 such that @xmath102 it holds that @xmath103 . the crucial observation in the proof of this lemma is that for every set of single - peaked total order votes @xmath54 , we are in one of the following two cases : 1 . for all @xmath13 , @xmath104 , @xmath105 . 2 . for all @xmath13 , @xmath106 , @xmath107 . we will show that this is also the case for single - peaked and single - plateaued orders with ties . this then implies that the optimal vote for all manipulators is @xmath108 ( in case 1 ) or @xmath109 ( in case 2 ) . for the sake of contradiction let @xmath54 be a collection of nonmanipulators and @xmath110 and @xmath111 be integers such that @xmath112 and @xmath113 such that @xmath114 and @xmath115 . we now create an @xmath116 such that @xmath117 and @xmath118 and the argument in the proof of lemma 6.6 in brandt et al . still holds . * first , remove from @xmath54 all voters that have @xmath57 tied for first . notice that if we remove all votes with @xmath57 tied for first from @xmath54 then @xmath119 does not increase and neither @xmath120 nor @xmath121 decreases . so we can assume that @xmath57 is never tied for first in a vote ( although this does not mean that other candidates can not be tied for first in a vote when we are in the single - plateaued case ) . * we know that @xmath57 is not tied for first ( part of an indifference plateau ) , so @xmath57 is tied with at most one other candidate in each vote . so in every vote break the tie against @xmath57 . notice that this can lower only the score of @xmath57 regardless of whether we are using the min , max , round down , or average scoring - rule extension . let @xmath122 be the total weight of the voters in @xmath116 that rank some candidate in @xmath123 first and @xmath57 in the position of @xmath14 ( in every extension other than round down this means that @xmath57 is ranked @xmath13th ; for round down we mean that @xmath57 is ranked in the @xmath14 group ( i.e. , @xmath57 is followed by @xmath124 groups ) . using a similar argument to the proof of lemma 6.6 from brandt et al . with @xmath116 as described above we can reach a contradiction . we now present the remaining p cases , which can be seen as the with - ties analog of the p cases of lemma 6.7 from brandt et al . . we examine each of the cases individually . for convenience , we normalize the scoring vector so that @xmath125 . if @xmath126 then clearly the optimal action for the manipulators is to vote a total order with @xmath57 uniquely first and the remaining candidates ranked arbitrarily such that the vote is single - peaked . if @xmath127 , then , as pointed out in @xcite , lemma 6.6 applies , which has been handled above . the last p - time case is that @xmath128 , @xmath129 , and @xmath130 . notice that when the votes are single - peaked or when they are single - plateaued then @xmath131 and @xmath132 are the only candidates that can occur last in a vote , either uniquely or tied ( except for the case where all candidates are tied , but we can easily ignore these votes since they do not change the relative scores between candidates ) . since we know from lemma [ lem : sp - first ] that all of the manipulators can put @xmath57 strictly ranked first , we can have the following manipulator votes . * @xmath133 . * @xmath134 . * @xmath135 . notice that for all votes with @xmath57 first that @xmath136 . for all votes , @xmath137 or @xmath138 , or @xmath131 and @xmath132 are tied for last , in which case @xmath139 ( the worst case occurs with max where they each receive a score of @xmath140 ) . so , @xmath141 . if @xmath142 , @xmath109 is the optimal manipulator vote . otherwise , @xmath143 and @xmath108 is the optimal manipulator vote . we have proven that the above cases hold for single - peaked and for single - plateaued weak orders , and it is clear that the case for single - peaked bottom orders follows from the case of single - peaked total orders ( since they are equivalent ) and it is also clear how to adapt the above arguments to hold for single - peaked top orders , and for single - plateaued top orders and bottom orders . the most surprising result in the work by menon and larson was that the complexity of 3-candidate cwcm for elimination veto for single - peaked preferences with outside options for top orders using min is np - complete , whereas for single - peaked total orders @xcite and even for total orders in the general case @xcite it is in p. @xcite @xmath1-candidate elimination veto cwcm for total orders is in p in the unique - winner model . @xcite @xmath1-candidate elimination veto cwcm for single - peaked total orders is in p in the unique - winner model . @xcite 3-candidate elimination veto cwcm for top orders that are single - peaked with outside options is np - complete in the unique - winner model . menon and larson state this case as a counterexample to the conjecture by faliszewski et al . , which states that the complexity for a natural election system will not increase when moving from the general case to the single - peaked case . ( though menon and larson do qualify that the conjecture from faliszewski et al . concerned total orders . ) however , for the standard model of single - peaked preferences and for single - plateaued preferences , elimination veto cwcm for top orders , bottom orders , and weak orders using min is in p for any fixed number of candidates , thus the counterexample crucially relies on using a nonstandard definition . @xmath1-candidate elimination veto cwcm for single - peaked and for single - plateaued top orders , bottom orders , and weak orders using min is in p in the unique - winner model . the proof of this theorem follows from a similar argument to the proof of the theorem for single - peaked total orders from menon and larson . this is because that proof follows from the fact that for total orders the candidate eliminated after each round is on the leftmost or rightmost location of the axis , and the reverse of an elimination order is single - peaked with respect to the axis . it is easy to see that both of these statements also hold for single - peaked votes with ties , since the only candidates that can be vetoed in each round are still the leftmost and rightmost candidates on the axis . for single - plateaued preferences it is possible for voters with an indifference plateaued to veto more than two candidates ( after all of the candidates ranked below their indifference plateau have been eliminated ) , but it is easy to see that these votes do not affect which candidate is eliminated since all of the remaining candidates are vetoed by such voters . it is also clear that lemma 12 from coleman and teague @xcite still holds which states that if there exists a collection of votes that can induce an elimination order , then this elimination order can be induced by all of the manipulators voting the reverse of the elimination order . so since we know that the reverse of an elimination order is single - peaked with respect to the axis , and that there are only polynomially many possible elimination orders with @xmath57 first , the manipulators simply try each elimination order with @xmath57 first . since the elimination order is always a total order , the above argument clearly holds for single - peaked and single - plateaued top orders , bottom orders , and weak orders . it is known that 3-candidate copeland@xmath2 cwcm for all rational @xmath144 is np - complete for total orders in the nonunique - winner model @xcite , and when @xmath145 ( also known as llull ) cwcm is in p for @xmath146 @xcite and the cases for @xmath147 remain open . fitzsimmons and hemaspaandra showed that the np - completeness of the 3-candidate case holds for top orders , bottom orders , and weak orders , and menon and larson independently showed the top - order case . fitzsimmons and hemaspaandra also showed that 3-candidate llull cwcm is in p for top orders , bottom orders , and weak orders . recall that weak condorcet winners always exist when preferences are single - peaked with ties , and when they are single - plateaued . so the results that llull cwcm for single - peaked total orders is in p from brandt et al . also holds for the case of single - peaked and single - plateaued top orders , bottom orders , and weak orders . copeland@xmath2 for @xmath144 was shown to be in p by yang for single - peaked total orders . @xcite [ t : yang ] copeland@xmath2 cwcm for @xmath144 is in p for single - peaked total orders . in contrast , for top orders that are single - peaked with outside options , it is np - complete even for three candidates @xcite . @xcite 3-candidate copeland@xmath2 cwcm for @xmath144 is np - complete for top orders that are single - peaked with outside options . for single - peaked and single - plateaued weak orders , bottom orders , and top orders we again inherit the behavior of single - peaked total orders . copeland@xmath2 cwcm for @xmath144 is in p for single - peaked and single - plateaued top orders , bottom orders , and weak orders . let @xmath65 be our axis . consider a set of nonmanipulators with single - plateaued weak orders . replace each nonmanipulator @xmath66 of weight @xmath9 with two nonmanipulators @xmath148 and @xmath149 of weight @xmath9 . the first nonmanipulator breaks the ties in the vote in increasing order of @xmath65 and the second nonmanipulator breaks the ties in the vote in decreasing order of @xmath65 , i.e. , if @xmath150 and @xmath151 , then @xmath152 and @xmath153 . note that @xmath148 and @xmath149 are single - peaked total orders and that the weighted majority graph induced by the nonmanipulators after replacement can be obtained from the weighted majority graph induced by the original nonmanipulators by multiplying each weight in the graph by 2 . when we also multiply the manipulator weights by 2 , we have an equivalent copeland@xmath2 cwcm problem , where all nonmanipulators are single - peaked total orders and all manipulators have even weight . suppose @xmath57 can be made a winner by having the manipulators cast single - plateaued votes with ties . now replace each manipulator of weight @xmath154 by two weight-@xmath9 manipulators . the first manipulator breaks the ties in the vote in increasing order of @xmath65 and the second manipulator breaks the ties in the vote in decreasing order of @xmath65 . now the replaced manipulator votes are single - peaked total orders and @xmath57 is still a winner . we need the following fact from the proof of theorem [ t : yang ] : _ if @xmath57 can be made a winner in the single - peaked total order case , then @xmath57 can be made a winner by having all manipulators cast the same p - time computable vote . _ it follows that @xmath57 can be made a winner by having all replaced manipulators cast the same single - peaked total order vote . but then @xmath57 can be made a winner by having all original manipulators cast the same single - peaked total order vote . since this vote is p - time computable , it follows that copeland@xmath2 cwcm for @xmath144 is in p for single - plateaued weak orders , and this also holds for single - peaked weak orders since every single - peaked profile of weak orders is also single - plateaued . it is clear to see that similar arguments hold for single - peaked top orders , and for single - plateaued top orders and bottom orders . the case for single - peaked bottom orders follows from the case for single - peaked total orders . recall from lemma [ lem : sp - first ] that for all our single - peaked models and all our scoring rule extensions , we can assume that all manipulators rank @xmath57 uniquely first . when given an axis where the preferred candidate is in the leftmost or rightmost location , there is exactly one single - peaked total order vote that puts @xmath57 first , namely , @xmath57 followed by the remaining candidates on the axis in order . this is also the case for single - peaked and single - plateaued orders with ties , since @xmath57 is ranked uniquely first and no two candidates can be tied on the same side of the peak . it follows that the weighted manipulation problems for scoring rules for single - peaked total orders and single - peaked and single - plateaued orders with ties are in p for axes where @xmath57 is in the leftmost / rightmost position . however , this is not the case when preferences with ties are single - peaked with outside options or possibly single - peaked . in case 1 of the proof of theorem 1 in the work by menon and larson , an axis of @xmath155 is used to show np - hardness for their model . it is attractive to conjecture that for single - peaked and single - plateaued preferences , the less symmetrical ( with respect to @xmath57 ) the axis is , the easier the complexity of manipulation , but surprisingly this turns out to not be the case , even for total orders . for the theorem stated below let @xmath156 and @xmath157 denote the number of candidates to the left and to the right on the axis with respect to the preferred candidate of the manipulators . for single - peaked total orders , single - peaked weak orders , and single - plateaued weak orders , @xmath158 cwcm is in p for @xmath159 and np - complete for @xmath160 and @xmath161 for all our scoring rule extensions . @xmath162 : careful inspection of the proof of lemma 6.6 from brandt et al . proves the following refined version of that lemma that takes the axis into account . [ l : sp2p ] let @xmath163 be a scoring rule . if for all @xmath164 such that @xmath165 and @xmath166 , it holds that @xmath167 then @xmath44-cwcm for single - peaked total orders for this axis is in p. note that lemma [ l : sp2p ] applies in our case , since for all @xmath164 such that @xmath168 and @xmath169 , the following hold . [ cols="^,^,^,^,^ " , ] it follows that @xmath158 cwcm is in p for single - peaked total orders when @xmath159 . it follows from the arguments in the proof of theorem [ thm : noninc ] that @xmath158 cwcm is also in p for single - peaked and single - plateaued orders with ties for @xmath159 . to get np - completeness for the single - peaked and single - plateaued weak order cases for all our scoring rule extensions , we need to modify the proof for single - peaked total orders . careful inspection of the proof of lemma 6.4 from brandt et al . allows us to extract the following reduction from partition to @xmath158 cwcm for total orders that are single - peaked with respect to axis @xmath170 . given an instance of partition , a collection of @xmath59 positive integers @xmath60 such that @xmath61 , we are asking whether there exists a partition of this collection , i.e. , whether there exists a subcollection that sums to @xmath171 . the set @xmath54 of nonmanipulators consists of * one weight @xmath172 voter voting : @xmath173 . * one weight @xmath174 voter voting : @xmath175 . the weights of the manipulators are @xmath176 . if there exists a partition , then let the set @xmath55 of manipulators vote the following . * coalition of weight @xmath177 manipulators vote @xmath178 . * coalition of weight @xmath177 manipulators vote @xmath179 . then @xmath180 , and the scores of the other candidates are lower . for the converse , note that the optimal total order votes for the manipulators are @xmath178 and @xmath181 . simple calculation shows that in order to make @xmath57 a winner , exactly half of the manipulator weight needs to vote @xmath178 . this corresponds to a partition . for the single - peaked and single - plateaued cases for votes with ties , note that by lemma [ lem : sp - first ] , we can assume that all manipulators rank @xmath57 uniquely first . we will look at each of our scoring extensions : max , min , average , and round down . for max , if the manipulators can make @xmath57 a winner , we break the ties of the manipulators such that the resulting votes become total single - peaked orders . @xmath57 will still be a winner in the resulting election . this implies that the construction above also gives np - completeness for single - peaked votes with ties using max . for round down , the construction above also works . suppose @xmath57 can be made a winner . again , we assume that all manipulators rank @xmath57 uniquely first . now replace all votes with ties ( i.e. , @xmath182 , @xmath183 , and @xmath184 ) by @xmath181 . note that @xmath185 and @xmath186 do not decrease , and so @xmath57 is still a winner of the resulting election , in which all manipulators vote a single - peaked total order . this implies that the construction above also gives np - completeness for single - peaked votes with ties using round down . for average , the construction above also works , though the proof needs a little more work . suppose @xmath57 can be made a winner . again , we assume that all manipulators rank @xmath57 uniquely first . so , @xmath187 . the optimal manipulator votes are : @xmath178 , @xmath181 , and @xmath188 . assume that these are the only manipulator votes and let @xmath189 , @xmath190 , and @xmath191 be the total manipulator weight voting each of these three votes , respectively . since @xmath57 is a winner , @xmath192 and @xmath193 . it follows that @xmath194 , @xmath195 , and @xmath196 . standard calculation shows that this implies that @xmath197 , @xmath198 and @xmath199 . thus the weights of the voters voting @xmath178 correspond to a partition . for min , the optimal votes for the manipulators are @xmath188 and @xmath181 . in order to make the argument from the total order case work , we need to adjust the weights of the nonmanipulators a bit . namely , we let the set @xmath54 of nonmanipulators consist of * one weight @xmath200 voter voting : @xmath173 . * one weight @xmath201 voter voting : @xmath175 . the weights of the manipulators are unchanged , i.e. , @xmath202 . note that if half the manipulator weight votes @xmath188 and remaining manipulators vote @xmath181 , then @xmath203 , @xmath204 , and @xmath205 . the scores of @xmath206 and @xmath207 are lower . similarly to the previous cases , if @xmath57 can be made a winner then standard calculations show that exactly half of the manipulator weight votes @xmath188 , which corresponds to a partition . the work by menon and larson on the complexity of manipulation and bribery for single - peaked preferences with outside options for top orders is the most closely related to this paper . for manipulation , they show that for single - peaked preferences with outside options the complexity often increases when moving from total orders to top orders . they additionally considered a notion of nearly single - peakedness . we instead study the complexity of weighted manipulation for the standard model of single - peaked preferences with ties and for single - plateaued preferences with ties . the focus of our paper is on the computational aspects of models of single - peaked preferences with ties . these models can also be compared based on which social - choice properties that they have , such as the guarantee of a weak condorcet winner . barber compares such properties of the models of single - peaked , single - plateaued , and single - peaked with outside options for votes with ties . since single - peakedness is a strong restriction on preferences , in real - world scenarios it is likely that voters may only have nearly single - peaked preferences , where different distance measures to a single - peaked profile are considered . both the computational complexity of different manipulative attacks @xcite and detecting when a given profile is nearly single - peaked @xcite have been considered . an important computational problem for single - peakedness is determining the axis given a preference profile , this is known as its consistency problem . single - peaked consistency for total orders was first shown to be in p by bartholdi and trick . doignon and falmagne and escoffier , lang , and ztrk independently found faster direct algorithms . lackner proved that possibly single - peaked consistency for top orders is in p ( and for local weak orders and partial orders is np - complete ) , and fitzsimmons later showed that single - peaked , single - plateaued , and possibly single - peaked consistency for weak orders is in p. the standard model of single - peakedness is naturally defined for votes with ties , but different extensions have been considered . in contrast to recent work that studies the models of possibly single - peaked and single - peaked preferences with outside options and finds an anomalous increase in complexity compared to the tie - free case , we find that for scoring rules and other important natural systems , the complexity of weighted manipulation does not increase when moving from total orders to votes with ties in the standard single - peaked and in the single - plateaued cases . single - peaked and single - plateaued preferences for votes with ties also retain the important social - choice property of the existence of weak condorcet winners . this is not to say that possibly single - peaked and single - peaked preferences with outside options are without merit , since they both model easily understood structure in preferences . * acknowledgments : * we thank the anonymous referees for their helpful comments . this work was supported in part by nsf grant no . ccf-1101452 and a nsf graduate research fellowship under nsf grant no . dge-1102937 .	single - peakedness is one of the most important and well - known domain restrictions on preferences . the computational study of single - peaked electorates has largely been restricted to elections with tie - free votes , and recent work that studies the computational complexity of manipulative attacks for single - peaked elections for votes with ties has been restricted to nonstandard models of single - peaked preferences for top orders . we study the computational complexity of manipulation for votes with ties for the standard model of single - peaked preferences and for single - plateaued preferences . we show that these models avoid the anomalous complexity behavior exhibited by the other models . we also state a surprising result on the relation between the societal axis and the complexity of manipulation for single - peaked preferences .
symmetries are among the most important guiding principles in all of physics and mathematics . it often happens that a problem may be solved by symmetry considerations alone , and even if not , understanding the symmetries of the solution can greatly reduce the amount of work needed to find it . we study here a symmetry of functions which is known as `` self - reciprocity '' to combinatorialists and which is referred to as `` inversion relations '' in lattice statistical mechanics . our focus will be on polyomino enumeration problems which are of interest in both combinatorics and physics . we shall demonstrate that one can find examples of functional symmetry in the resulting generating functions . the inversion relation rose to prominence in statistical mechanics in the early 1980s as the most direct path to the solution of many integrable models @xcite and was soon realized to be commonplace in both solved and unsolved models @xcite . let @xmath2 be a thermodynamic quantity which depends on a collection of parameters , @xmath3 . an inversion relation is a functional equation @xmath4 where @xmath5 and @xmath6 are known functions of @xmath3 . typically @xmath5 involves taking reciprocals of one or more components of @xmath3 . the inversion relation tightly constrains the function @xmath7 . for some two - dimensional models a pair of additional conditions holds : that @xmath7 is symmetric under exchange of horizontal and vertical , and that @xmath7 is an analytic function of its arguments . very often , the three constraints taken together uniquely determine the function @xmath7 . in 1974 stanley presented a general framework for reciprocity results . he established several powerful general conditions under which a generating function will be self - reciprocal @xcite . the language and notation of stanley @xcite will be used throughout this paper . let @xmath8 be a rational function in the variables @xmath9 , with coefficients in @xmath10 . we say that @xmath11 is _ self - reciprocal _ if there exists an @xmath12-tuple of integers @xmath13 such that @xmath14 [ selfrecdef ] in what follows , we write @xmath15 and @xmath16 . thus eqn . ( [ selfrec ] ) may be concisely expressed as @xmath17 note that a rational function is self - reciprocal if and only if both its numerator and denominator are so , and that the self - reciprocity of a polynomial amounts to a certain symmetry in its coefficients . some explicit examples are given in subsection [ subsect : examples ] . let us now demonstrate the relationship between self - reciprocity and inversion relations . consider the multivariable generating function @xmath18 where @xmath19 , @xmath20 and similarly for @xmath21 and @xmath22 . the summation is over @xmath23-tuples of nonnegative integers representing the objects being enumerated . performing the summation over @xmath21 , we reexpress eqn . ( [ genfun ] ) in terms of partial generating functions , @xmath24 , @xmath25 now suppose that the partial generating functions are self - reciprocal , @xmath26 where @xmath27 is a @xmath28-tuple of elements in the set @xmath29 which characterizes the dependence of the sign on @xmath30 , and where @xmath31 depends linearly on @xmath30 : @xmath32 here , @xmath33 is a @xmath34 matrix of integers and @xmath35 is a @xmath36-tuple of integers . we can then write @xmath37 where @xmath38 is the @xmath28-tuple whose @xmath39 entry is @xmath40 . this is clearly a special case of the inversion relation ( [ invrel ] ) . a few comments are in order : * the right hand side of ( [ ginversion ] ) is zero , but in the more general situation some of the partial generating functions , @xmath41 will fail to be self - reciprocal for certain choices of @xmath30 . if we are fortunate , this will be a small , finite or otherwise controllable set of cases , and we will be able to compute the correction term we need to add to the right hand side explicitly . for many examples in statistical mechanics , this correction term depends on @xmath3 but not on @xmath22 . * in all of the cases we shall see below , the denominators of our rational functions will be a product of terms @xmath42 , which are self - reciprocal . stanley has proved that this denominator form always holds for certain classes of problems ( see theorem 4.6.11 of ref . * it might be asked which of the concepts , inversion or self - reciprocity , is the more general . on one hand , in the derivation of ( [ ginversion ] ) the dependence of the exponent @xmath43 on @xmath30 was assumed to be linear , which may not always hold , implying that reciprocity is more fundamental . on the other hand , the function @xmath5 occurring in ( [ invrel ] ) may in principle be more complicated than @xmath44 , @xmath45 . in this case , the partial generating functions might not be self - reciprocal . an example is provided in section [ sect : statmech ] by the potts model , but in the polyomino examples considered in this paper , this situation does not arise . we now present a nonexhaustive list of recipes for finding and proving reciprocity results and inversion relations . 1 . if the generating function ( or thermodynamic quantity ) is known in closed form , an inversion relation can be demonstrated directly . as an example , we treat the anisotropic perimeter generating function for directed convex polygons in this manner in section [ sect : polyinversion ] . [ item : exact ] 2 . for statistical mechanics models which admit a formulation in terms of a family of commuting transfer matrices , a transformation of parameters can often be found which inverts the transfer matrix . the commutativity property then allows the inversion relation to be derived . we review this in detail in section [ sect : statmech ] , with the two - dimensional , zero - field ising model as primary example . [ item : tm ] 3 . in non - integrable models , the transfer matrix will still be invertible and may suggest a possible inversion relation , but the required analyticity property is lacking . nevertheless , the suggested inversion relation can often be verified by inspection of the partial generating functions up to some finite order in the low - temperature expansion ( [ pargenfun ] ) . the @xmath46 potts model inversion relation discussed in section [ sect : statmech ] was derived this way in ref . some of the new results reported in the present paper were initially discovered by this method before being rederived by one of the other methods . [ item : series ] 4 . the `` temperley methodology '' @xcite can be used to obtain very general reciprocity results for many classes of column - convex polygons . the first step is to derive a functional equation for the generating function which can be interpreted as the gluing of an additional column onto the graph . step two is to show by induction that appending an additional column preserves self - reciprocity . this is detailed in section [ sect : temperley ] . [ item : temperley ] 5 . if the problem can be posed as a system of linear diophantine equations , whose solutions are subject to certain types of constraints , we may apply self - reciprocity theorems due to stanley @xcite . we have so far succeeded in applying this method only to families of directed polyominoes ( section [ sect : stanley ] ) , but it enables us to treat problems which are impossible , or at least extremely cumbersome , by the method of functional equations . [ item : stanley ] 6 . for combinatorial objects with a rational generating function of denominator @xmath47 , one can try to explain self - reciprocity _ i.e. _ , the symmetry of the numerator by interpreting the numerator combinatorially . this has been done by fdou for a family of objects related to ( but distinct from ) staircase polygons @xcite . in section [ sect : statmech ] we review the motivation for looking at inversion relations in statistical mechanics and describe the methods used to obtain them . this will be useful for making comparisons with the results obtained later , and for suggesting applications and generalizations of the inversion relations . in section [ sect : polyinversion ] , we present examples of reciprocity results and inversion relations for polyominoes , and summarize our main new results . the technical heart of the paper consists of section [ sect : temperley ] on the temperley methodology , and section [ sect : stanley ] on the application of stanley s results to polyominoes . the first use of the inversion relation in statistical mechanics was the solution by stroganov of certain two dimensional vertex models on the square lattice @xcite . generalizations of stroganov s models were later solved by the same means by schultz @xcite . shortly after stroganov , baxter used a similar method to solve the hard hexagon model @xcite and recognized its broad applicability , giving the eight - vertex and ising models as examples @xcite . subsequently , a number of authors pointed out that many known solutions to problems in two - dimensional statistical mechanics can be derived easily using the inversion relation method . among these were shankar @xcite , baxter @xcite and pokrovsky and bashilov @xcite . it is noteworthy that inversion relations hold also for models that have not been solved . prominent among such models are the two - dimensional ising model in a magnetic field whose inversion relation was found by baxter @xcite , and the three - dimensional ising model and noncritical @xmath48-state potts model , both of whose inversion relations were found by jaekel and maillard @xcite . what generally distinguishes solved and unsolved models is the growth rate in the number of poles arising in the partial generating functions in the expansion ( [ pargenfun ] ) , as a function of order . roughly speaking , a more complicated pole structure implies that the number of parameters needed to specify a given partial generating function is greater , and makes it less likely that an inversion relation can completely determine all of them . nevertheless , inversion relations are still invaluable in the study of such problems , not least because they provide an independent check on series data . the fundamental problem of statistical mechanics is to calculate the partition function . here we consider vertex models defined on a square lattice with each bond colored with one of @xmath49 possible colors . each lattice site makes a contribution to the energy of the system which depends on the colors of the adjacent bonds . this defines an @xmath50-vertex model if all possible colorings are permitted . stroganov computed the partition function per site in the thermodynamic limit of several 16- and 81-vertex models . consider first a finite lattice ( on the torus ) of @xmath51 rows and @xmath52 columns . the partition function can be expressed in terms of the transfer matrix @xmath53 as : @xmath54 \label{pf}\ ] ] ( see @xcite ) . here , @xmath53 is the @xmath55 matrix whose @xmath56th entry is the contribution to @xmath57 of a single row of @xmath52 sites connected to the row below by a set of vertical bonds in configuration @xmath58 and to the row above by a set of vertical bonds in configuration @xmath28 . it depends on the temperature , @xmath59 , and on @xmath50 parameters specifying the vertex energies . in the thermodynamic limit , the partition function per site is given by @xmath60 where @xmath61 is the largest eigenvalue of @xmath53 , assumed to be nondegenerate . for simplicity let us consider a family of models whose vertex energies are functions of a single parameter , @xmath62 . the models solved by stroganov are integrable by virtue of the commutativity of the transfer matrices at different values of this parameter . this implies that the transfer matrix eigenvectors are common to all members of the family , and that the @xmath62 dependence is only in the eigenvalues . for this reason @xmath62 is often called the _ spectral parameter_. the key observation is that the inverse of the transfer matrix in these models is itself a member of the commuting family , up to a scale factor @xmath63 ^{-1 } = \psi(b)^{-m } \ { \sst_m}\left(\phi(b)\right ) . \label{tinv}\ ] ] acting on the eigenvector corresponding to @xmath64 with both sides of eqn . ( [ tinv ] ) yields the functional equation @xmath65 it is the commutativity of the transfer matrices for all values of @xmath62 that allows the analytical continuation of the function @xmath66 from @xmath62 to @xmath67 . with knowledge of the functions @xmath68 and @xmath67 and using the analyticity of @xmath69 , stroganov finds a unique solution , thereby reproducing baxter s results for the symmetric eight vertex and homogeneous ferroelectric models , and obtaining the result for a certain 81-vertex model @xcite . as an illustrative example , we review here the derivation by baxter @xcite of onsager s expression for the partition function of the two - dimensional zero - field ising model @xcite . let the square lattice be drawn at @xmath70 to the horizontal and let the couplings between nearest neighbors along the two lattice directions be @xmath71 and @xmath72 define low temperature variables @xmath73 transfer matrices for different choices of parameters will commute provided they have the same value of @xmath74 . the transformation which inverts the transfer matrix is @xmath75 which does not modify the value of @xmath36 . define the reduced partition function per site by @xmath76 then @xmath77 obeys the inversion relation @xmath78 note that @xmath79 has an inversion relation of precisely the form ( [ ginversion ] ) . by the symmetry of the model , we have @xmath80 inspection of the low temperature expansion leads us to conjecture the form @xmath81 that the coefficient of @xmath82 is a rational function of @xmath83 is apparent from the nature of the low temperature expansion , but that the denominator has such a simple form is not expected on general grounds . presumably it is a consequence of the condition of commuting transfer matrices . here we take it as a hypothesis . then baxter has shown that the inversion relation ( [ isinginv ] ) , symmetry ( [ isingsym ] ) and the denominator form ( [ isingform ] ) determine @xmath77 completely . we present his argument in section [ subsect : computegf ] where we use it in the context of polygon enumeration . up till now we have been assuming integrability and in particular we have relied on the property that the transfer matrix and its inverse are both members of some one - parameter commuting family . what about models for which this property does nt hold ? since analyticity of @xmath69 breaks down , the step ( [ func ] ) in the above derivation is no longer valid . however , it is still possible to obtain an inversion relation by direct analysis of the low - temperature expansion of the partition function to some finite order . as an example , it was shown in ref . @xcite that the logarithm of the reduced partition function per site , @xmath84 , of the @xmath48-state potts model satisfies the inversion relation @xmath85 when @xmath86 this reduces to the ising model inversion relation ( [ isinginv ] ) . the inversion relations we will be considering in the remainder of the paper are derived by analysis of the generating function ( analogous to the low temperature expansion ) and do not depend on the models being integrable . an additional new feature is seen in this potts model example . neglecting for the moment the nonzero right - hand - side of ( [ pottsinv ] ) , which can be eliminated by a suitable redefinition of @xmath87 , we notice that when @xmath46 there is no longer an order - by - order cancellation of the partial generating functions as defined in ( [ pargenfun ] ) , but rather cancellation of combinations of partial generating functions of different orders . however , we may convert to self - reciprocal form by defining @xmath88 under which the inversion relation becomes @xmath89 in the cases we will look at in this paper , the partial generating functions turn out to be self - reciprocal in the natural variables of the problem . we have not investigated the existence of inversion relations involving more complicated changes of variables . the constructions we will consider are defined on the square lattice . all are defined only up to translation on the lattice . starting at a lattice site and moving to one of the four nearest neighbors constitutes a _ step _ which we may identify with the edge connecting the sites . a connected sequence of steps is a _ path _ or _ walk_. if no lattice site in the path occurs more than once , the path is _ self - avoiding_. if a path returns to its starting site in the final step , and otherwise does not intersect itself , the result is a _ self - avoiding polygon_. the number of steps taken is the _ perimeter _ of the polygon ; the number of steps taken in the vertical direction is the _ vertical perimeter . _ the _ horizontal perimeter _ is defined similarly . the _ area _ is the number of cells of the lattice enclosed by the polygon . + enumerating self - avoiding polygons according to perimeter or area is an unsolved problem . however , progress has been made in enumerating certain subclasses of self - avoiding polygons . _ rectangles _ coincide with the rectangles of ordinary geometry whose vertices are lattice points and whose edges lie along lattice directions . a rectangle which contains a given polygon , _ i.e. _ , all steps of the polygon lie inside or on the rectangle , is a _ bounding rectangle _ for that polygon . the smallest such rectangle is the _ minimal bounding rectangle . _ a polygon whose perimeter equals that of its minimal bounding rectangle is _ convex_. if a convex polygon contains at least one of the corners of its minimal bounding rectangle ( for concreteness say the south - west corner ) then it is a _ directed convex polygon . _ if it contains also the north - east corner , it is a _ staircase polygon , _ so called because it is bounded above and below by two staircase - like or _ directed _ paths . on the other hand , if it contains two adjacent corners , say the southwest and southeast ( northeast and southeast ) then it is a _ stack polygon _ with horizontal ( vertical ) orientation . if it contains three corners , then it is a _ ferrers graph . _ representative examples of different classes of convex polygons are shown in figure [ fig : convexpolygons ] . one way to obtain non - convex polygons is to relax the convexity condition along one direction only . a self - avoiding polygon is _ column - convex _ if the intersection of any vertical line with the polygon has at most two connected components . _ row - convex _ polygons are similarly defined . the set of convex polygons is the intersection of the sets of row- and column - convex polygons . subclasses of column - convex polygons include the _ bar - graphs _ which contain the bottom edge of the minimal bounding rectangle , and _ directed column - convex polygons _ whose bottom edge is a directed path . some examples are shown in figure [ fig : columnconvexpolygons ] . a second class of non - convex polygons is made up of four directed paths . a _ three - choice walk _ is a self - avoiding walk whose steps are taken in accordance with the three - choice rule which allows a step either to the left or the right or straight ahead after any vertical step , but forbids a right turn after any horizontal step . a polygon formed from such a walk is a _ three - choice polygon . _ when the walk returns to its starting point , we do nt specify whether the next step , _ i.e. _ , the first step , is a valid continuation of the walk . if it is , the result is a staircase polygon ; if not , it is an _ imperfect staircase polygon _ ( see figure [ subfig : thchoice ] ) . when we refer to three - choice polygons below , we include only the imperfect ones . a _ polyomino _ is a union of connected ( sharing an edge ) cells of the lattice . we shall consider one class of nonpolygon polyominoes the _ staircase polygons with a staircase hole_. the outer boundary and the hole are both staircase polygons and must not touch at any point . an example is shown in figure [ subfig : spsh ] . for each of the above classes of column - convex polygons , the _ anisotropic perimeter and area generating function , _ @xmath90 has been computed exactly ( see ref . @xcite and references therein ) . here @xmath91 is the number of polygons of the class with @xmath92 horizontal bonds , @xmath93 vertical bonds and area @xmath94 . for the classes of convex polygons , the anisotropic perimeter generating function , @xmath95 is an algebraic function of the fugacities , @xmath96 and @xmath97 , whereas the area generating function , @xmath98 is a @xmath48-series . for classes of polygons that are only column - convex , both @xmath95 and @xmath98 @xcite are algebraic , but @xmath99 involves @xmath48-series . a closed - form expression for the three - choice polygon anisotropic perimeter - area generating function is not yet known , but by means of a transfer matrix technique it can be evaluated in polynomial time @xcite . the isotropic perimeter generating function , @xmath100 is known to have a logarithmic singularity @xcite , and is therefore not algebraic , but is known to be d - finite . the generating function for staircase polygons with a staircase hole is also not known in closed form . its properties are expected to be similar in many respects to the generating function for three - choice polygons @xcite . we shall be concerned with self - reciprocity properties of the generating functions @xmath101 that count polygons of width @xmath102 . we first give two examples . 1 . the area generating function for staircase polygons of width 4 is the following rational function @xcite : @xmath103 it satisfies @xmath104 and is thus self - reciprocal . observe that the numerator is not only symmetric ( due to self - reciprocity ) , but also unimodal . the ( half-)vertical perimeter and area generating function for column - convex polygons of width 3 is the following rational function , which can be derived from the general formula of ref . @xcite : @xmath105 it satisfies @xmath106 and hence is self - reciprocal . again , observe the symmetry of the coefficients in the numerator . we shall generalize these results to polygons of any width . table [ table : summary ] summarizes the self - reciprocity properties we have established . most of them can be proved in various ways . one can for instance use a closed form expression of the generating function ( section [ subsect : exact ] ) , or a functional equation that defines it ( section [ sect : temperley ] ) ; one can also encode the polygons by a sequence of numbers constrained by linear diophantine equations and apply stanley s general results ( section [ sect : stanley ] ) . we shall see that the last two methods allow us to introduce many additional parameters and obtain self - reciprocity results that significantly generalize those of table [ table : summary ] . -5 cm [ cols="<,^,^,^",options="header " , ] when a closed form expression for the generating function of some class of polygons is known , it seems natural to use it to demonstrate an inversion relation . let us take the example of the anisotropic perimeter generating function for directed convex polygons , which is known to be @xcite : @xmath107 with @xmath108 $ ] . expanding this expression in @xmath96 gives @xmath109 which suggests that the partial generating functions , @xmath110 are self - reciprocal , and more precisely , that @xmath111 . this is equivalent to the inversion relation @xmath112 which is easily checked from the closed form of the generating function . note that an explicit expression for @xmath110 is given in @xcite . the inversion relations for convex polygons and directed column - convex polygons may also be obtained from the expression of their generating function . the partial generating functions for directed convex polygons , counted by the area , are not self - reciprocal : for instance , the generating function for width @xmath113 is @xmath114 however , many other classes of column - convex polygons have an inversion relation for the full anisotropic perimeter and area generating function . since these generating functions are also known in closed form they could be derived as above . however more can be shown , namely that there is a self - reciprocity for any parameter which is a linear function of the _ vertical heights _ in the graph . this very general result will be derived in section [ sect : temperley ] . likewise , the inversion relations for three - choice polygons and staircase polygons with a staircase hole , given in table [ table : summary ] , are also special cases of more general formulae which will be derived in section [ sect : stanley ] . as in statistical mechanics , the inversion relation and symmetry , and some general assumptions on analyticity of the generating function , are sometimes sufficient to determine the solution completely . in order to have an algorithm for computing a generating function term by term , it is necessary , but not sufficient , to have some property relating terms of different orders . for our purposes this property will always be @xmath96-@xmath97 symmetry . thus we restrict our attention to classes of graphs with @xmath96-@xmath97 symmetry , _ i.e. _ , ferrers , staircase , directed convex , convex and three - choice polygons , and staircase polygons with a staircase hole . moreover , we shall only consider the anisotropic perimeter generating function ( without area ) . for the former four classes we will show that the inversion relation provides sufficient additional information to compute the generating function , whereas for the latter two it does not . the general form of the generating function is @xmath115 where the partial generating functions , @xmath110 are rational functions @xmath116 with @xmath117 . the general form of the inversion relation is @xmath118 where @xmath119 is an integer and rhs is zero or some simple function . it is equivalent to a self - reciprocity relation of the form @xmath120 whether the inversion relation is sufficient to compute the generating function depends on the value of the exponent @xmath119 and on the degree of the denominator , @xmath121 . direct proof of the denominator form can often be obtained . for ferrers graphs , it is easily shown that @xmath122 . for staircase polygons , one finds @xmath123 . the same denominator form holds for directed convex and convex polygons also . for the three - choice polygons and staircase polygons with a staircase hole it can be shown that the denominators are @xmath124 we assume that in general we know the denominator form either empirically or by rigorous proof , and that @xmath121 is of degree @xmath125 . now we proceed inductively , following baxter @xcite . if we have already computed the coefficient functions @xmath126 in the expansion ( [ gexp ] ) and if @xmath96-@xmath97 symmetry holds , we also know the coefficients of @xmath127 in the expansion of @xmath87 . in particular , we can compute the coefficients of @xmath127 in the numerator polynomial @xmath128 . in order to obtain the unknown coefficients of @xmath128 , we must be able to express them in terms of the known ones by means of the inversion relation . writing @xmath129 , and using @xmath130 , the inversion relation fixes the value of the combinations of coefficients , @xmath131 , with @xmath132 . hence the determination of all the coefficients @xmath133 is possible if and only if the arithmetic condition holds : @xmath134 this condition is seen to hold for all the classes of convex polygons we have looked at , since @xmath135 , but not for three - choice polygons or staircase polygons with a staircase hole , since @xmath136 . we consider column - convex polygons as pairs of partially directed paths having the same endpoints , as indicated in figure [ fig : vc ] . let @xmath137 be a column - convex polygon of width @xmath102 . for @xmath138 , we denote by @xmath139 ( resp . @xmath140 ) the number of north ( resp . south ) steps in the top path @xmath141 at abscissa @xmath58 . for @xmath138 , we denote by @xmath142 ( resp . @xmath143 ) the number of north ( resp . south ) steps of the bottom path @xmath144 at abscissa @xmath58 . we choose the end points of the paths in such a way that @xmath145 . note that @xmath146 we notice that all standard statistics are _ linear _ functions of the @xmath147 and @xmath143 . for instance , the vertical perimeter of the polygon is @xmath148 the height of the @xmath39 column of the polygon is , for @xmath149 , @xmath150 and the area of the polygon is @xmath151 [ theo : reci ] let @xmath152 be one of the following sets : ferrers diagrams , stacks ( drawn as in figure [ subfig : stackh ] ) , staircase polygons , bar - graphs , column - convex polygons . let @xmath153 be the subset of @xmath152 containing all polygons of width @xmath102 . let @xmath154 be the generating function for polygons in the set @xmath153 : @xmath155 then @xmath154 is a rational function , and it is self - reciprocal : @xmath156 with @xmath157 the proof of the theorem is based on the so - called temperley approach for counting column - convex polygons @xcite , combined with the systematic use of formal power series @xcite . here we provide only the proof for column - convex polygons , since the others are very similar . we commence by showing that the partial generating functions for column - convex polygons , @xmath158 , can be computed inductively . [ functional ] let @xmath158 be the generating function for column - convex polygons of width @xmath102 . let us denote it , for the sake of simplicity , @xmath159 . then the series @xmath159 can be defined inductively by : @xmath160 and @xmath161 * proof . * the basic idea is build a polygon of width @xmath162 by adding a new column to a polygon of width @xmath102 @xcite . it is convenient to use hadamard products to establish the functional equation . let @xmath163 and @xmath164 be two formal power series in @xmath165 with coefficients in a ring @xmath166 . we denote by @xmath167 the hadamard product of @xmath168 and @xmath169 , evaluated at @xmath170 : @xmath171 in what follows , @xmath172 ( resp . @xmath173 ) will be the generating function for some column - convex polygons whose rightmost ( resp . leftmost ) column has height @xmath174 , so that @xmath175 will count polygons obtained by _ matching _ the rightmost column of a polygon of type @xmath176 with the leftmost column of a polygon of type @xmath177 . also , @xmath169 will be a rational function of @xmath165 . we shall use the following simple identity : @xmath178 the expression for @xmath179 is obvious . we build a column - convex polygon of width @xmath162 as follows : we take a polygon of width @xmath102 and match its rightmost column with the leftmost column of a column - convex polygon of width @xmath180 . this is illustrated by figure [ fig : hadamard ] , which shows that @xmath181 where @xmath182 is the generating function for column - convex polygons of width @xmath180 . in order to determine the coefficient @xmath173 of @xmath183 in @xmath169 , we expand @xmath169 in partial fractions of @xmath165 : @xmath184 note that @xmath185 . we now combine eqn . ( [ rec ] ) with the above expression for @xmath169 to obtain the announced expression for @xmath186 . * proof of theorem [ theo : reci ] . * induction on @xmath102 using the functional equation of proposition [ functional ] shows that the partial generating functions for column - convex polygons satisfy : @xmath187 we proceed similarly for the other families : the functional equation is obtained by setting some of the variables @xmath188 and @xmath189 to @xmath190 . then , an inductive argument yields the self - reciprocity result . it would be tempting to write that the self - reciprocity of @xmath191 implies the self - reciprocity of , say , the generating function for staircase polygons , obtained by setting @xmath192 and @xmath193 to @xmath190 in @xmath191 . but replacing a variable by @xmath190 in a self - reciprocal rational function might break the self - reciprocity : for instance , take @xmath194 then @xmath195 but @xmath196 is _ not _ self - reciprocal . however , the following simple lemma gives a useful stability property of self - reciprocal rational functions . [ lemma : stability ] let @xmath197 be a self - reciprocal rational function . let @xmath198 be an @xmath199 integer matrix . let @xmath200 , and define @xmath201 to be the @xmath12-tuple whose @xmath39 coordinate is @xmath202 . then the series @xmath203 , if defined , is self - reciprocal in the variables @xmath204 . more precisely , if @xmath205 then @xmath206 . from theorem [ theo : reci ] and lemma [ lemma : stability ] we immediately deduce : [ coro : linear ] for any of the sets @xmath152 listed in theorem [ theo : reci ] , and any statistics on column - convex polygons that can be expressed as _ linear _ functions of the quantities @xmath207 , the generating function for polygons in the set @xmath153 according to these statistics is a self - reciprocal rational function . this corollary allows us to complete the top part of table [ table : summary ] . let us , for instance , derive the inversion relation satisfied by the tri - variate generating function @xmath99 for column - convex polygons , taking into account the usual parameters of interest : horizontal and vertical half - perimeters ( variables @xmath96 and @xmath97 ) , and area ( variable @xmath48 ) . ( [ vperimeter ] ) and ( [ area ] ) express the vertical perimeter and the area in terms of the quantities @xmath208 and @xmath209 . they imply that the ( half ) vertical perimeter and area generating function @xmath101 for column - convex polygons of width @xmath102 is @xmath210 where @xmath211 and @xmath212 . theorem [ theo : reci ] then gives @xmath213 which implies @xmath214 note that in the first two self - reciprocity relations of table 1 , the exponent of @xmath48 depends quadratically on the width . for this reason , they only yield an inversion relation for @xmath215 . stanley has analyzed the situation where the objects to be counted correspond to integer solutions of a system of linear equations with integer coefficients ( linear diophantine system ) subject to a set of constraints . he has established certain conditions under which reciprocity relations will hold between two combinatorics problems defined by the same linear diophantine system but by different sets of constraints , and also conditions under which the solution to a given problem will be self - reciprocal @xcite . consider the linear homogeneous diophantine system ( lhd - system ) , @xmath216 in the unknowns @xmath217 with @xmath218 a matrix of integers having @xmath219 rows and @xmath220 columns and @xmath221 a @xmath219-tuple of zeros . the corank @xmath66 of the system is defined to be @xmath222 . for a linearly independent system , @xmath223 . let @xmath224 be a set of integer solutions to eqn . ( [ lhds ] ) . we define the generating function , @xmath225 , as the formal power series @xmath226 where @xmath227 is a vector of fugacities associated with the unknowns in eqn . ( [ lhds ] ) . in our applications , we find two types of constraints on the unknowns , @xmath228 . certain of the unknowns , @xmath228 , are required to be strictly positive while the rest are required to be non - negative . conveniently , precisely these kinds of constraints have been treated by stanley . let the unknowns be @xmath229 where @xmath230 is an @xmath12-tuple and @xmath231 is an @xmath232-tuple . likewise let @xmath233 . in what follows , the notation , @xmath234 , means that all coordinates of @xmath231 are positive . let @xmath235 be the set of integer solutions , @xmath236 , to a linear homogeneous diophantine system of corank @xmath66 , such that @xmath237 and @xmath234 . let @xmath238 be the set of solutions to the same system with @xmath239 and @xmath240 . if the system has an integer solution , @xmath236 such that @xmath241 and @xmath242 , then @xmath243 and @xmath244 are rational functions obeying the reciprocity relation @xmath245 [ propo : recb ] _ proof . _ this is proposition 8.3 of ref . @xcite and the proof is given there . proposition [ propo : recb ] can be specialized to obtain a self - reciprocity condition , which will be our main tool in the derivations to follow . a sufficient condition for the function @xmath243 to be self - reciprocal is that the linear homogeneous diophantine system has the solution @xmath246 . in this case @xmath247 [ coro : selfb ] _ proof . _ since the solution @xmath248 satisfies the conditions of proposition [ propo : recb ] , the reciprocity result ( [ lhdrecb ] ) holds . the result follows immediately from the shift @xmath249 which establishes a bijection between the sets @xmath235 and @xmath238 . since the conditions of the corollary are sufficient but not necessary , it is often possible to find a perfectly valid lhd - system describing a given self - reciprocal generating function , @xmath243 , which does _ not _ admit the solution @xmath248 . hence we are faced with the problem of finding a suitable lhd - system which satisfies the corollary . a useful heuristic is to start with an lhd - system in many unknowns , and selectively eliminate those unknowns whose constraints are not independent of the constraints on the other unknowns . in all the cases we will consider , the resulting system will satisfy the conditions of corollary [ coro : selfb ] . we do not justify this heuristic here . in a paper subsequent to ref . @xcite , stanley @xcite develops a more comprehensive theory which overcomes these difficulties , and which additionally gives `` correction '' terms for systems in which self - reciprocity fails to hold . we have not yet explored the ramifications of this theory . before applying the above result to staircase polygons with a staircase hole or to three - choice polygons , we use it to derive the reciprocity relation for ordinary staircase polygons of width three . this will serve to illustrate all the basic ingredients of the method . staircase polygons of width three can be characterized by the heights @xmath250 , @xmath251 , @xmath252 , @xmath253 , @xmath254 , @xmath255 , @xmath256 and @xmath257 , as shown in figure [ fig : stairthree ] . decomposing the polygon into three columns , and imposing the condition that each column be as high on the left as it is on the right , we obtain the linear homogeneous diophantine system @xmath258 all heights must be nonnegative , but the self - avoidance condition additionally requires that the @xmath259 be positive . the constraints @xmath260 and @xmath261 are actually redundant , since they follow from eqns . ( [ exlhda],[exlhdc ] ) and the constraints on the remaining unknowns , namely @xmath262 since the constraints on @xmath254 and @xmath257 play no role in the solution , we are free to eliminate these unknowns , and it turns out to be necessary to do so in order to apply corollary [ coro : selfb ] . we are left with the single equation ( [ exlhdb ] ) in the six independent unknowns @xmath263 and @xmath264 . let us associate to the unknown @xmath139 ( resp . @xmath142 , @xmath265 ) the fugacity @xmath266 ( resp . @xmath267 , @xmath268 ) . let @xmath269 be the set of solutions to eqn . ( [ exlhdb ] ) subject to the the constraints @xmath270 and @xmath271 . since @xmath272 , @xmath273 is a solution to eqn . ( [ exlhdb ] ) , corollary [ coro : selfb ] tells us that @xmath274 is self - reciprocal , @xmath275 equations ( [ exlhda],[exlhdc ] ) imply that to account for the dependent parameters @xmath254 and @xmath257 , we make the substitutions @xmath276 , @xmath277 , @xmath278 and @xmath279 . applying lemma [ lemma : stability ] , we obtain for the set @xmath235 of nonnegative solutions to ( [ exlhda],[exlhdb],[exlhdc ] ) such that @xmath280 : @xmath281 notice that reintroducing the dependent unknowns has not changed the constant factor . this feature holds as well in the more complicated models we will look at . the result ( [ exselfb ] ) may be verified by inspection of the explicit expression for the generating function @xmath282 we now apply the methods of section [ subsect : lhds ] to staircase polygons with a staircase hole and to three - choice polygons . all the essential steps have already been seen in the derivation of the reciprocity result for staircase polygons of width three . they are 1 . set up a linear homogeneous diophantine system by decomposing the polyomino into width one rectangles and imposing the condition that the left and right sides of each rectangle have equal height . 2 . sort the unknowns into three classes , @xmath230 , @xmath231 and @xmath283 , according to whether they are constrained to be nonnegative , constrained to be positive or constrained by conditions on the other unknowns . use gaussian elimination to remove the unknowns in @xmath283 . 4 . verify that the resulting system is solved by setting all members of @xmath230 equal to one and all members of @xmath231 equal to minus one . apply corollary [ coro : selfb ] to obtain the self - reciprocity result for the reduced system . reintroduce the unknowns in the set @xmath283 by means of lemma [ lemma : stability ] . we can define three widths for a staircase polygon with a staircase hole : the distance from the left edge of the figure to the left edge of the hole , @xmath36 , the distance from the left edge of the figure to the right edge of the hole , @xmath284 , and the width of the entire figure , @xmath102 . note that @xmath285 . recall that for staircase polygons the figures of width one were an exceptional case which did not obey the same reciprocity result as the general case . the staircase polygons with a hole of width one are also an exceptional case , which we must exclude . we thus impose the additional condition @xmath286 . a figure with given @xmath36 , @xmath284 and @xmath102 is specified by the following dimensions , as shown in figure [ subfig : spshb ] , 1 . heights @xmath287 and @xmath288 of the lower and upper perimeter segments of the polygon , @xmath289 , 2 . interior heights @xmath259 to the left and right of , and within , the hole , @xmath290 , 3 . heights @xmath291 and @xmath292 of the lower and upper perimeter segments of the hole , @xmath293 , 4 . interior heights @xmath294 and @xmath295 below and above the hole , @xmath296 . three - choice polygons can be regarded as staircase polygons with a hole which does nt close . the width @xmath36 has the same meaning as above , @xmath284 denotes the ultimate horizontal extent of the branch of the figure above the hole , and @xmath102 denotes the ultimate horizontal extent of the branch below the hole . note that @xmath297 and @xmath298 . again an exceptional case , @xmath299 , must be excluded . hence we impose the restriction @xmath300 . the labeling of the vertical dimensions follows , with a few obvious modifications , the pattern of staircase polygons with a staircase hole and is shown in figure [ subfig : threechoice ] . in particular , the heights @xmath259 within the hole are defined only for @xmath301 . when @xmath302 the unknowns @xmath295 and @xmath292 do not appear . this special case is treated separately . as in the case of column - convex polygons , the standard statistics are linear in these heights . the ( half-)vertical perimeter for staircase polygons with a staircase hole is given by @xmath303 and the area is given by @xmath304 in what follows , we associate to the unknowns @xmath139 ( resp . @xmath142 , @xmath305 , @xmath306 , @xmath265 , @xmath307 , @xmath308 ) the fugacities @xmath266 ( resp . @xmath309 ) . [ propo : spsh ] let @xmath310 be the generating function for staircase polygons with a staircase hole where @xmath36 , @xmath284 and @xmath102 are the widths defined above . then if @xmath286 , the generating function @xmath310 is self - reciprocal , @xmath311 @xmath312 _ proof . _ the linear homogeneous diophantine system is the union of five sets of equations which we label @xmath313@xmath314 . the regions to the left and right of the hole give @xmath313 and @xmath315 , the regions below and above the hole give @xmath316 and @xmath317 and the inside of the hole gives @xmath314 : @xmath318 \\[0.05 in ] l_2 & = \begin{cases } \mb_\ell+m_\ell+\mt_\ell+\nt_\ell - m_{\ell+1}-\nb_{\ell+1}=0 \\ m_j+\nt_j - m_{j+1}-\nb_{j+1}=0 \qquad \text{for $ \ell+1\le j\le m-2 $ } \\ m_{m-1}+\nt_{m-1}-m_m=0 \end{cases } \notag\displaybreak[0 ] \\[0.05 in ] l_3 & = \begin{cases } \mb_k-\mb_{k+1}-\nb_{k+1}=0 \\ \mb_j+\hb_j-\mb_{j+1}-\nb_{j+1}=0 \qquad \text{for $ k+1\le j\le \ell-1 $ } \end{cases } \notag\displaybreak[0 ] \\[0.05 in ] l_4 & = \begin{cases } \mt_j+\nt_j-\mt_{j+1}-\ht_{j+1}=0 \qquad \text{for $ k\le j\le \ell-2 $ } \\ \mt_{\ell-1}+\nt_{\ell-1}-\mt_\ell=0 \end{cases } \notag\displaybreak[0 ] \\[0.05 in ] l_5 & = \begin{cases } m_k - m_{k+1}-\hb_{k+1}=0 \\ m_j+\ht_j - m_{j+1}-\hb_{j+1}=0 \qquad \text{for $ k+1\le j\le \ell-2 $ } \\ m_{\ell-1}+\ht_{\ell-1}-m_\ell=0 . \end{cases}\end{aligned}\ ] ] all heights of course are nonnegative . self - avoidance imposes the additional constraint that the heights denoted @xmath259 , @xmath295 and @xmath294 be positive . the set @xmath283 , defined in step 2 above , contains six unknowns whose constraints are not independent which we eliminate as follows : @xmath254 using the first equation of @xmath313 , @xmath319 using the last equation of @xmath315 , @xmath320 using the first equation of @xmath316 , @xmath321 using the last equation of @xmath317 , and @xmath322 and @xmath323 using the first and last equations of @xmath314 . the resulting system is @xmath324 \\[0.05 in ] l'_2 & = \begin{cases } \mb_\ell+m_{\ell-1}+\ht_{\ell-1}+\mt_{\ell-1}+\nt_{\ell-1}+\nt_\ell - m_{\ell+1}-\nb_{\ell+1}=0 \\ m_j+\nt_j - m_{j+1}-\nb_{j+1}=0 \qquad \text{for $ \ell+1\le j\le m-2 $ } \end{cases } \notag\displaybreak[0 ] \\[0.05 in ] l'_3 & = \begin{cases } \mb_j+\hb_j-\mb_{j+1}-\nb_{j+1}=0 \qquad \text{for $ k+1\le j\le \ell-1 $ } \end{cases } \notag\displaybreak[0 ] \\[0.05 in ] l'_4 & = \begin{cases } \mt_j+\nt_j-\mt_{j+1}-\ht_{j+1}=0 \qquad \text{for $ k\le j\le \ell-2 $ } \end{cases } \notag\displaybreak[0 ] \\[0.05 in ] l'_5 & = \begin{cases } m_j+\ht_j - m_{j+1}-\hb_{j+1}=0 \qquad \text{for $ k+1\le j\le \ell-2$. } \end{cases}\end{aligned}\ ] ] the substitutions @xmath288 , @xmath287 , @xmath292 , @xmath325 and @xmath259 , @xmath295 , @xmath326 solve this new system of equations . one should note that when @xmath327 or @xmath328 the system is somewhat modified , but one may check that the solution still holds . therefore we may apply corollary [ coro : selfb ] to obtain a self - reciprocity condition on the generating function for the solutions of @xmath329 subject to the positivity constraints on the heights . making appropriate substitutions to restore the unknowns in set @xmath283 , and using lemma [ lemma : stability ] we obtain eqn . ( [ selfspsh ] ) . we now treat three - choice polygons . [ propo : three - choice ] let @xmath330 be the generating function for three - choice polygons where @xmath36 , @xmath284 and @xmath102 are the widths defined above . then if @xmath331 , the generating function @xmath330 satisfies a self - reciprocity condition which , when @xmath302 , takes the form @xmath332 @xmath333 and , when @xmath334 , takes the form @xmath335 @xmath336 _ proof . _ it is simpler to treat the two cases @xmath302 and @xmath334 separately . the proofs follow very closely that of proposition [ propo : spsh ] . as for column - convex polygons , the two propositions above may be extended to other statistics . [ coro : linearb ] let @xmath337 be either of the sets staircase polygons with a staircase hole or three - choice polygons . let @xmath338 be the subset of figures in @xmath337 with the widths @xmath36 , @xmath284 and @xmath102 defined as above . then the generating function in @xmath338 according to any statistics linear in the quantities @xmath339 @xmath340 , @xmath341 , @xmath342 , @xmath343 , @xmath344 , is a self - reciprocal rational function ( assuming @xmath286 for staircase polygons with a staircase hole and @xmath345 for three - choice polygons ) . the half - horizontal perimeter for either of the sets @xmath152 is given by @xmath346 . using this in combination with corollary [ coro : linearb ] , ( [ vperimeterspsh ] ) and ( [ areaspsh ] ) , we obtain the inversion relations specialized to horizontal and vertical perimeter , and area , which are listed in table [ table : summary ] . the exceptional cases ( @xmath347 and @xmath348 respectively ) can be computed explicitly by the methods of @xcite . each of the methods we have discussed for obtaining reciprocity or inversion relations has its own particular uses . for example , the method of stroganov is suitable for lattice models in statistical mechanics which are characterized by a family of commuting transfer matrices . the temperley methodology is mainly applicable to families of polygons that are column - convex or nearly so . stanley s method for obtaining reciprocity results apply to any problem defined by a system of linear homogeneous diophantine ( lhd ) equations , but the solutions to this system must be constrained by a system of simple inequalities of a certain form . it is probable that for many lattice models in statistical mechanics the low temperature expansion can be framed as an lhd - system . however , most are likely to require more general types of constraints than the simple inequalities of the directed polyomino problems we have considered . likewise , the non - directed polygon problems that we have successfully treated using the temperley methodology can be recast as lhd - systems with more complex constraints . how to handle such constraints is a worthy problem for future investigation . in recent work @xcite this statistical mechanical language has been adapted for the enumeration of lattice paths , and may apply to polyomino problems as well . it is intriguing to speculate that the inversion relations found here may be connected with this approach . we have not searched for inversion relations for any polyomino problem in variables other than the natural variables for the problem . yet the example of the potts model demonstrates that such inversion relations may exist . it is also possible that symmetries in addition to the ones presented here can be found for some problems . it is our hope that such additional symmetries might lead to the solution of currently intractable problems . for the moment , we remark that the search for inversion and symmetry relations appears to provide a new method to tackle certain combinatorial problems . the degree of applicability of this method is still unclear . we have benefited from conversations with george andrews , richard brak , jean - marie maillard , paul pearce and markus vge and from correspondence with jean - marc fdou and richard stanley . we thank iwan jensen for providing us with his series data for staircase polygons with a staircase hole . ajg and wpo acknowledge support from the australian research council . baxter , hard hexagons : exact solution , . baxter , exactly solved models . in : _ fundamental problems in statistical mechanics v ; _ proceedings of the 1980 enschede summer school , e.g.d . cohen ed . ( north holland , amsterdam , 1980 ) 109141 . baxter , two - dimensional models in statistical mechanics . in : _ statistical mechanics and field theory ; _ proceedings of the seventh physics summer school , anu , 1994 ; v.v . bazhanov and c.j . burden eds . ( world scientific , singapore , 1995 ) 129167 . r.j.baxter , _ exactly solved models in statistical mechanics , _ ( academic press , london , 1982 ) . r.j . baxter , the inversion relation method for some two - dimensional exactly solved models in lattice statistics , . m. bousquet - mlou , convex polyominoes and heaps of segments , . m. bousquet - mlou , a method for the enumeration of various classes of column - convex polygons , _ discrete math . , _ * 154 * ( 1996 ) 125 . r. brak , j.w . essam and a.l . owczarek , from the bethe ansatz to the gessel - viennot theorem , this issue . a. conway , a.j . guttmann and m. delest , on the number of three - choice polygons , _ math . comput . model . , _ * 26 * ( 1997 ) 5158 . fdou , fonctions de bessel , empilements et tresses . in : _ proceedings of the 4th conference on formal power series and algebraic combinatorics , _ p. leroux and c. reutenauer , eds . , publications du lacim , universit du qubec montral , * 11 * ( 1992 ) 189202 . guttmann , i. jensen and i.g . enting , punctured polygons and polyominoes on the square lattice , submitted to _ j. phys . jaekel and j .- m . maillard , symmetry relations in exactly soluble models , . m.t . jaekel and j .- m . maillard , inverse functional relation on the potts model , . k. y. lin and s. j. chang , rigorous results for the number of convex polygons on the square and honeycomb lattices , . maillard , the inversion relation , _ j. physique , _ * 46 * ( 1985 ) 329341 . l. onsager , crystal statistics . i. a two - dimensional model with an order - disorder transition , . s.v . pokrovsky and yu.a . bashilov , star - triangle relations in the exactly solvable statistical models , . r. shankar , simple derivation of the baxter - model free energy , . schultz , solvable @xmath48-state models in lattice statistics and quantum field theory , . stanley , combinatorial reciprocity theorems , _ adv . math . , _ * 14 * ( 1974 ) 194253 . stanley , linear diophantine equations and local cohomology , _ inventiones math . , _ * 68 * ( 1982 ) 175193 . stanley , _ enumerative combinatorics , vol . wadsworth and brooks / cole , monterey calif . 1986 ) . stroganov , a new calculation method for partition functions in some lattice models , . temperley , combinatorial problems suggested by the statistical mechanics of domains and of rubber - like molecules , .	we derive self - reciprocity properties for a number of polyomino generating functions , including several families of column - convex polygons , three - choice polygons and staircase polygons with a staircase hole . in so doing , we establish a connection between the reciprocity results known to combinatorialists and the inversion relations used by physicists to solve models in statistical mechanics . for several classes of convex polygons , the inversion ( reciprocity ) relation , augmented by certain symmetry and analyticity properties , completely determines the anisotropic perimeter generating function . keywords : inversion relations , combinatorial reciprocity theorems , polyominoes , self - avoiding polygons , convex polygons , statistical mechanics . ams subject classification : 05a15 ( 05b50 , 82b20 , 82b23 ) . * laboratoire bordelais de recherche en informatique + universit bordeaux i + * and + * department of mathematics and statistics + the university of melbourne + * june 1999 + * inversion relations , reciprocity and polyominoes + * m. bousquet - mlou@xmath0 , a.j . guttmann@xmath1 , w.p . orrick@xmath1 and a. rechnitzer@xmath1 + @xmath0labri , cnrs + universit bordeaux i + 351 cours de la libration + 33405 talence cedex + france + @xmath1department of mathematics and statistics + the university of melbourne + parkville , vic . 3052 + australia +
in the next - to - minimal supersymmetric standard model ( nmssm ) @xcite , the soft supersymmetry ( susy)-breaking higgs sector is described by the lagrangian contribution @xmath4 where @xmath5 and @xmath6 are the higgs doublet fields , @xmath7 the singlet one , @xmath8 and @xmath9 yukawa couplings while @xmath10 and @xmath11 are dimensionful parameters of order @xmath12 , the typical susy mass scale . as a result of the introduction of an extra complex singlet scalar field , which only couples to the two mssm - type higgs doublets , the higgs sector of the nmssm comprises of a total of seven mass eigenstates : a charged pair @xmath13 , three cp - even higgses @xmath14 ( @xmath15 ) and two cp - odd higgses @xmath16 ( @xmath17 ) . consequently , higgs phenomenology in the nmssm may be plausibly different from that of the mssm and extremely rich of new signals . + for a start , the sum of the squares of the two lightest scalar higgs boson masses is given by @xcite : @xmath18 with @xmath19 . the last expression can be translated into a modified upper bound of the @xmath20 mass as @xcite @xmath21 so that the upper bound on the nmssm lightest higgs boson mass is higher than the corresponding bound in the mssm . furthermore , as the higher order corrections are similar to those in the mssm , also in higher orders the upper bound on the lightest higgs boson mass remains different in the nmssm with respect to the mssm , reaching 140 gev or so , for maximal stop mixing and @xmath22 @xcite ( a configuration indeed excluded in the mssm by lep data ) . more in general , the ` little fine tuning problem ' , resulting in lep failing to detect a light cp - even higgs boson , predicted over most of the mssm parameter space , is much attenuated in the nmssm , either because a sm - like higgs can decay dominantly into @xmath23 @xcite or because the mixing among more numerous cp - even or cp - odd higgs fields enables light mass states being produced at lep or tevatron yet , they can remain undetected because of their reduced couplings to @xmath24 bosons @xcite . ( see refs . @xcite and @xcite for benchmark points with very light cp - odd higgs bosons in both the unconstrained and constrained nmssm parameter space , respectively ) . as for present accelerator machines , chiefly the tevatron at fnal and the large hadron collider ( lhc ) at cern , quite some work has been dedicated to probing the nmssm higgs sector over recent years . primarily , there have been attempts to extend the so - called ` no - lose theorem ' of the mssm stating that at least one mssm higgs boson should be observed via the usual sm - like production and decay channels at the lhc throughout the entire mssm parameter space @xcite to the case of the nmssm @xcite . from this perspective , it was realised that at least one nmssm higgs boson should remain observable at the lhc over the nmssm parameter space that does not allow any higgs - to - higgs decay . however , when the only light non - singlet ( and , therefore , potentially visible ) cp - even higgs boson , @xmath20 or @xmath25 , decays mainly to two very light cp - odd higgs bosons , @xmath26 , one may not have a higgs signal of statistical significance at the lhc @xcite . in fact , further violations to the theorem may well occur if one enables higgs - to - susy decays ( e.g. , into neutralino pairs , yielding invisible higgs signals ) . while there is no definitive evidence on whether a ` no - lose theorem ' can be proved for the nmssm , we are here also concerned with an orthogonal approach . we asked ourselves if a , so to say , ` more - to - gain theorem ' can be formulated in the nmssm . that is , whether there exist regions of the nmssm parameter space where more and/or different higgs states of the nmssm are visible at the lhc than those available within the mssm . in our attempt to overview such a possibility , we consider here the case of the di - photon and @xmath1 decay channels of a light neutral higgs boson . the first mode can successfully be probed in the mssm , but limitedly to the case of the lightest cp - even higgs boson and for masses of order 100130 gev or so . we will show that in the nmssm there exist regions of its parameter space where one can potentially have a sizable di - photon signal from a higgs state of different nature ( cp - odd instead of cp - even ) and with much smaller mass ( down to 10 gev or so ) , owing in large part to an increased br@xmath27 . such a possibility emerges in the nmssm due to the fact that such a cp - odd higgs state has a predominant singlet component and a very weak doublet one . as a consequence , all partial decay widths are heavily suppressed as they employ only the doublet component , except one : the @xmath2 partial decay width . this comes from the fact that the @xmath28 coupling is not suppressed , as it is generated through the @xmath29 lagrangian term and therefore implies no small mixing . although the direct decay @xmath0 @xmath30 @xmath31 is forbidden , the aforementioned coupling participates in the @xmath32 effective coupling . the @xmath1 channel is used in the mssm as a search channel of rather heavy and degenerate cp - even and cp - odd states and its exploitation has not been proved at very low masses , say , below @xmath33 . we will show here that the @xmath34-pair decay can be very relevant for such a mass interval , in the case of an @xmath0 state of the nmssm . the extraction of either channel for low higgs masses would then unmistakably point to the existence of non - minimal susy higgs sector . we build on the results presented in @xcite ( see also @xcite ) , where several such signals ( for rather heavy higgs states though ) were established in some ( complementary ) regions of the nmssm parameter space . the plan of this paper is as follows . in sec . [ sect : scan ] we describe the parameter space scans performed in order to isolate the combinations of nmssm inputs that can give rise to the described phenomenology . in sec . [ sect : rates ] we discuss typical event rates , highlighting the relevance of higgs production in association with bottom quark pairs as most favourable production mode . in sec . [ sect : s2b ] we describe the signal - to - background analysis performed and introduce some benchmark points where an @xmath0 signal can be extracted ( albeit limitedly to the @xmath1 decay mode ) . finally , in sec . [ sect : summa ] we summarise and conclude . for a general study of the nmssm higgs sector , we used here the up - to - date publicly available fortran package nmssmtools developed in refs . @xcite2.3.1 . ] . this package computes the masses , couplings and decay widths of all the higgs bosons of the nmssm in terms of its parameters at the electro - weak ( ew ) scale . the nmssmtools also takes into account theoretical as well as experimental constraints from negative higgs searches at lep @xcite and the tevatron , along with the unconventional channels relevant for the nmssm . for our purpose , instead of postulating unification , and without taking into account the susy breaking mechanism , we fixed the soft susy breaking terms to very high values , so that they have little or no contribution to the outputs of the parameter scans . consequently , we are left with the aforementioned six free parameters . our parameter space is in particular defined through the yukawa couplings @xmath8 and @xmath9 , the soft trilinear terms @xmath10 and @xmath11 , plus tan@xmath35 ( the ratio of the vacuum expectation values ( vevs ) of the two higgs doublets ) and @xmath36 ( where @xmath37 is the vev of the higgs singlet ) . in our numerical analyses we have taken @xmath38 gev and @xmath39 gev for the bottom- and top - quark mass , respectively . we have used the nmhdecay code to scan over the nmssm parameter space defined through the six parameters already discussed taken in the following intervals : @xmath8 : 0.0001 0.7 , @xmath9 : 0 0.65 , @xmath40 : 1.6 54 , + @xmath41 : 100 1000 gev , @xmath42 : @xmath431000 + 1000 gev , @xmath44 : @xmath4310 0 . + ( notice that our aim is exploring the parameter space which has very low @xmath45 and one way to do that is by choosing @xmath11 small , in which case its negative value is preferred @xcite ) . remaining soft terms which are fixed in the scan include : + @xmath46 tev , + @xmath47 tev , + @xmath48 tev , + @xmath49 tev . + as intimated , we have fixed soft term parameters at the tev scale to minimise their contributions to parameter space outputs but changing values of some of those parameters such as @xmath50 could decrease or increase the number of successful points emerging from the nmssmtools scans but without a significant impact on the @xmath45 distribution . also , notice that the sfermion mass parameters and the @xmath51 gaugino mass parameter , @xmath52 , play crucial roles in constraining @xmath40 . lowering values of those parameters allow less values of @xmath40 to pass experimental and theoretical constraints , however , this is a less interesting region of the nmssm parameter space for our analysis , as our higgs production mode is only relevant at large values of this parameter . the effect of heavy gaugino mass parameters on the output of parameter space , in particular @xmath53 , would be small except for @xmath52 through its effect on @xmath54 . in line with the assumptions made in @xcite , the allowed decay modes for neutral nmssm higgs bosons are to signify any of the neutral cp - even(odd ) higgs bosons of the nmssm . ] : @xmath55 ( notice that for the cp - odd higgses , the decay into vector boson pairs is not allowed due to cp - conservation ) . here , ` higgses ' refers to any final state involving all possible combination of two higgs bosons ( neutral and/or charged ) or of one higgs boson and a gauge vector . we have performed our scan over 10 million of randomly selected points in the specified parameter space . the output , as stated earlier , contains masses , branching ratios ( brs ) and couplings of the nmssm higgses , for all the points which are not forbidden by the various experimental and theoretical constraints . the points which violate the constraints are eliminated . the surviving data points are then used to determine the cross sections for nmssm higgs hadro - production by using calchep @xcite , the centre - of - mass energy at parton level , for all processes computed . ] , wherein some new modules have been implemented for this purpose @xcite . as the susy mass scale has been arbitrarily set well above the ew one ( see above ) , the nmssm higgs production modes exploitable in simulations at the lhc are those involving couplings to heavy ordinary matter only , i.e. , ( hereafter , @xmath56 and @xmath57 ) for neutral higgs production ( where the last two channels are only allowed for cp - even higgs production ) : @xmath58 @xmath59 @xmath60 @xmath61 ( these are the so - called ` direct ' higgs production modes ) . here , ` higgs ' refers to any possible neutral higgs boson . we can however anticipate that , for the purpose of extracting an @xmath62 or @xmath63 resonance , only the second channel turns out to be useful and limitedly to the case @xmath64 . this is because the associated @xmath65 pairs can be ( vertex ) tagged , thus offering a useful handle for background rejection . the case @xmath66 in fact gives too small cross section . the gluon fusion channel is instead burdened by huge standard model ( sm ) backgrounds ( @xmath67 as well as jets mis - identified as photons in the @xmath2 mode and drell - yan and di - jet production in the @xmath1 channel ) . higgs - strahlung and vbf also suffer from large sm backgrounds , though their main drawback are rather smaller production rates in comparison ( as they are ew processes ) . + in the nmssm , @xmath0 state is a composition of the usual doublet component of the cp - odd mssm higgs boson , @xmath68 , and the new singlet component , @xmath69 , coming from the singlet superfield of the nmssm . this can be written as @xcite : @xmath70 for very small values of @xmath71 , the lightest cp - odd higgs , @xmath0 , is mostly singlet - like with a tiny doublet component , i.e. the mixing angle @xmath72 is small , see top - pane of fig . 1 , which shows the relation between @xmath45 and @xmath72 . if @xmath0 is highly singlet , @xmath73 , then br@xmath74 can reach unity as shown in the bottom - pane of this figure . + to good approximation , @xmath45 can be written in the nmssm as @xcite : @xmath75 the first term of the this expression is dominant especially at large @xmath40 . also , it is clear that a combination of all tree level higgs sector parameters jointly affects @xmath45 in general . as an initial step towards the analysis of the data , we have computed @xmath45 against each of the six tree level higgs sector parameters of the nmssm . 2 presents the results of our scan , these series of plots illustrating the distribution of @xmath45 over the six parameters and as a function of br@xmath27 and br@xmath76 . in the parameter space adopted for this analysis we have noticed that the large @xmath40 and small @xmath77 ( and , to some extent , also @xmath8 ) region is the one most compatible with current theoretical and experimental constraints , though this conclusion should not be generalised to the entire parameter space . herein , it is also obvious that @xmath45 increases by increasing @xmath9 and @xmath78 whereas it decreases by increasing @xmath10 . moreover , from a closer look at fig . 2 , it is clear that the br@xmath27 can be very large , indeed dominant , and @xmath45 values in the region 50 to 100 gev maximise that rate . also , notice that br@xmath76 reaches about 10% in most of the parameter space which has @xmath79 gev , in which case the @xmath0 decay into @xmath65 is open and dominant . yet also notice that there is a small region with very low @xmath45 values , @xmath80 gev , yielding a large br@xmath76 , greater than @xmath81 , as shown in the bottom - right corner of the bottom - right pane of this figure . the latter region occurs when @xmath82 is closed , in which case br@xmath76 is dominant compared to @xmath83 , @xmath84 , etc threshold is severely constrained , see , e.g. , ref . @xcite ( and references therein ) . ] . 3 shows the distribution of event rates , @xmath85 and @xmath86 as functions of @xmath45 , br of the corresponding channel and @xmath40 . as expected , the inclusive cross section decreases with increasing @xmath45 , see the top two panes of this figure . although the br@xmath74 can be dominant over a sizable expanse of the nmssm parameter space ( middle - left pane ) , the latter does not correspond to the region that maximises the yield of @xmath85 , as the maximum of the latter occurs for brs in the region of a few @xmath87 to @xmath88 . therefore , one can not take full advantage of the phenomenon described in the introduction , with respect to the singlet nature of the @xmath0 state entering the @xmath28 coupling , as the doublet component ( necessary to enable a large @xmath89 coupling at production level ) plays a stronger role in comparison . the tension between the two trends is such that the cross section times br rates are less than 100 fb . the outlook for the @xmath1 case is much brighter , as corresponding signal rates are at nb level for br@xmath90 or even 10 nb for br@xmath91 , see middle - right pane of fig . 3 . also notice that such large rates naturally occur for any @xmath45 in the allowed interval ( see top - right pane of this figure ) . incidentally , notice in the case of both decay channels that not only the density of nmssm parameter configurations is larger as @xmath40 grows and values of the soft susy breaking parameters this would not necessarily hold . ] but also the event rates are maximal at large values of this parameter ( see bottom - left and bottom - right panes of fig . 3 ) , thereby confirming what we intimated at the beginning of this section about the relevance of the @xmath92 production mode ( whose cross section is essentially proportional to @xmath93 ) . in the nmssm , there is a large area of parameter space where one higgs state can decay into two , e.g. , @xmath94 : see fig . as it is clear from the left - pane of this figure , the majority of points generated here have @xmath95 gev and @xmath96 gev , thereby allowing the possibility of @xmath94 decays . moreover , this decay can be dominant and can reach unity as shown in the right - pane of this figure . despite this , such a decay may not give higgs signals with sufficient statistical significance at the lhc ( as discussed in previous literature and recalled here previously ) . therefore , we are here well motivated to look further at the scope of direct production of @xmath0 state in single mode at the lhc , through @xmath97 , over overlapping regions of nmssm parameter space , which we are going to do next . we perform here a partonic signal - to - background ( @xmath98 ) analysis , based on calchep , in the two channels @xmath2 and @xmath1 separately in the two forthcoming subsections . we assume @xmath99 tev throughout for the lhc energy and we will benchmark event rates on the basis of 300 inverse femtobarn of accumulated luminosity . from the output of nmssmtools , we have chosen some points which have large ( yet not maximal ) event rates as illustrative examples to test the possibility of detecting @xmath0 with different masses at the lhc . they are located in large regions of parameter space with high density . these illustrative points are as follows : * for @xmath45=9.8 gev : + @xmath8 = 0.22341 , @xmath9 = 0.41849 , tan@xmath35 = 53.82 , @xmath41 = 228.94 , @xmath10 = @xmath100 and @xmath11 = @xmath101 . * for @xmath45=20 gev : + @xmath8 = 0.07595 , @xmath9 = 0.11544 , tan@xmath35 = 51.51 , @xmath41 = 377.44 , @xmath10 = @xmath102 and @xmath11 = @xmath103 . * for @xmath45=31 gev : + @xmath8 = 0.10861 , @xmath9 = 0.46542 , tan@xmath35 = 48.06 , @xmath41 = 222.99 , @xmath10 = @xmath104 and @xmath11 = @xmath105 . * for @xmath45=46 gev : + @xmath8 = 0.14088 , @xmath9 = 0.25219 , tan@xmath35 = 50.56 , @xmath41 = 317.08 , @xmath10 = -@xmath106 and @xmath11 = @xmath107 . * for @xmath45=60.5 gev : + @xmath8 = 0.17411 , @xmath9 = 0.47848 , tan@xmath35 = 52.39 , @xmath41 = 169.83 , @xmath10 = @xmath108 and @xmath11 = @xmath109 . * for @xmath45=81 gev : + @xmath8 = 0.10713 , @xmath9 = 0.13395 , tan@xmath35 = 44.72 , @xmath41 = 331.43 , @xmath10 = @xmath110 and @xmath11 = @xmath111 . altogether , the fact that production cross section and br decay are maximised each in different regions of the nmssm parameter space makes it extremely difficult to obtain detectable event rates in this case . in fact , the signal yields for this channel are not only small in general , but also overwhelmed by the irreducible background . this is made explicit in figs . 5 - 6 , which show how the signal , despite yielding sizable peaks in the di - photon invariant mass @xmath112 ( for example , for all @xmath45 masses considered between @xmath113 and @xmath114 gev one can obtain some @xmath115 events every 300 fb@xmath116 , see top - left plot of each figure ) , is completely spoilt by the irreducible background , after using standard cuts refers to the pseudorapidity and @xmath117 to the cone distance expressed in differences of pseudorapidity and azimuthal angle @xmath118 : i.e. , @xmath119 . further , the notation @xmath120 refers to the transverse momentum . ] : @xmath121 @xmath122 @xmath123 in fact , for triggering purposes , at least one of the two photon transverse momenta ought to be increased to some 10 gev or so , which does not help to improve @xmath98 , as can be appreciated in the top - right and bottom plots in figs . 5 - 6 , which show the average transverse momentum of a photon as well as its minimum and maximum , respectively . for @xmath0 masses above 50 gev or so the signal rates are ( in general ) too poor to even pass the observability threshold of 1 event after 300 fb@xmath116 . in summary , despite the uniqueness of this signal , with the potential of clearly highlighting a possible singlet nature of the lightest cp - odd higgs of the nmssm , the latter is not extractable at the lhc . the situation is instead much rosier for the @xmath1 channel . after implementing the following standard cuts s to be on shell . ] @xmath124 @xmath125 @xmath126 we obtain the invariant masses of the @xmath1 system depicted in figs . 7 - 12 ( see their left - hand sides ) , where the signal clearly appears over the irreducible background due to @xmath127 . recalling that for low higgs masses , say below @xmath33 , in the @xmath128 channel , the dominant background is indeed the irreducible one , see , e.g. , ref . @xcite , at least in the double- and single - leptonic channels , it is clear that there exist substantial discovery potential of a very light cp - odd higgs boson of the nmssm at the lhc . for the sake of completeness , we have also shown in figs . 7 - 12 the top - antitop reducible background , i.e. , @xmath129 ( with no cuts on @xmath130 though , which could always be enforced for its suppression ) , making it clear that discovery of heavier @xmath0 states in @xmath1 pairs , even in the the double- and single - leptonic channels , becomes much more difficult , although not impossible ( also owing to the long @xmath24 tail in the irreducible background ) . the case of fully hadronic decays is further contaminated by pure qcd backgrounds , the more so the smaller the @xmath45 values , so that we have not treated it here . finally notice the stability of the aforementioned signals against the possibility of stiffer thresholds in the selection of @xmath34 decays , as is made clear by the @xmath131 dependence of the @xmath7 and @xmath132 distributions in the right - hand side of figs . 7 - 12 ( e.g. , an increase of even a factor of two in transverse momentum of the leptonic / hadronic @xmath34 decay products should not dramatically spoil the signal significances in the low @xmath0 mass region ) . overall , the @xmath1 signal yield in the low @xmath0 mass region is of order 3000 ( for @xmath45 reaching 80 gev or so ) to 30000 ( for @xmath45 starting at 10 gev or so ) signal events over a much smaller background , even assuming @xmath1 resolutions of 10 gev or so ( notice that the width of each histogram in the figures is of 1 gev and the plots are in log scale ) . in short , we have proven that there exist some regions of the nmssm parameter space where very light cp - odd higgs states , with a mixed singlet and doublet nature , could potentially be detected if @xmath133 in the @xmath134 mode if the cp - odd higgs state is produced in association with a @xmath65 pair . after a realistic @xmath98 analysis at parton level , we have in fact produced results showing that the extraction of light mass @xmath135 resonances above both the irreducible and ( dominant ) reducible backgrounds should be feasible using standard reconstruction techniques @xcite , at least in the double- and single - leptonic decay channels of the @xmath34 s . while more refined analyses , incorporating @xmath34-decay , parton shower , hadronisation and detector effects , are needed in order to delineate the true discovery potential of the lhc over the actual nmssm parameter space , we believe that our results are a step in the right direction to prove the existence of a ` more - to - gain theorem ' at the cern collider for the nmssm with respect to the mssm , as @xmath1 signals from such light higgs bosons are not possible in the latter scenario . finally , some of the parameter regions where the aforementioned signal can be detected overlap with those where @xmath136 decays could also be effective in extracting an @xmath0 signal , so that the process discussed here also offers an alternative handle to establish a ` no - lose theorem ' for the nmssm at the lhc . further notice that we have explored here the two regimes @xmath137 and @xmath138 . the former is where the @xmath63 decay rate dominates ( these are the points in the ` red island ' to the outermost right - hand side of the middle - right plot of fig . 3 ) , for which br@xmath139 , as the @xmath65 decay channel is closed . ( our mass point @xmath140 gev was representative of this situation ) . the latter is where the @xmath141 decay rate dominates ( the corresponding points are all the remaining ones of fig . 3 ) , since here br@xmath142 because the @xmath65 decay channel is open . ( our mass points @xmath143 , 31 , 46 , 60.5 and 81 gev were representative of this situation . ) unfortunately , a similar analysis in the @xmath62 channel has showed that the lhc discovery potential is hindered by an overwhelming irreducible background , despite the fact that the br@xmath27 can be equal to unity in some regions of the nmssm parameter space , a peculiarity of the nmssm with respect to the mssm . we acknowledge useful email exchanges with cyril hugonie and ulrich ellwanger . we are also very grateful to lorenzo basso for technical assistance with some of the plots and to alexander belyaev for his help with calchep . this work is supported in part by the next institute . m. m. a. acknowledges a scholarship granted to him by taibah university ( saudi arabia ) . a. djouadi _ et al . _ , jhep * 0807 * ( 2008 ) 002 . j. dai , j.f . gunion , r. vega , phys . b * 315 * ( 1993 ) 355 and phys . b * 345 * ( 1995 ) 29 ; j.r . espinosa , j.f . gunion , phys . * 82 * ( 1999 ) 1084 . u. ellwanger , j.f . gunion and c. hugonie , hep - ph/0111179 ; d.j . miller and s. moretti , hep - ph/0403137 ; c. hugonie and s. moretti , hep - ph/0110241 ; a. belyaev , s. hesselbach , s. lehti , s. moretti , a. nikitenko and c. h. shepherd - themistocleous , arxiv:0805.3505 [ hep - ph ] ; j. r. forshaw , j. f. gunion , l. hodgkinson , a. papaefstathiou and a. d. pilkington , jhep * 0804 * ( 2008 ) 090 ; a. belyaev , j. pivarski , a. safonov , s. senkin and a. tatarinov , phys . d * 81 * ( 2010 ) 075021 . m. m. almarashi , private programs . s. andreas , o. lebedev , s. r. sanchez and a. ringwald , jhep * 1008 * ( 2010 ) 003 . see , e.g. : f. sarri , preprint atl - phys - proc-2008 - 076 ( and references therein ) ; s. horvat , preprint atl - phys - proc-2009 - 063 ( and references therein ) .	in the nmssm , because of introducing a complex singlet superfield , the lightest cp - odd higgs boson , @xmath0 , can be a singlet - like state with a tiny doublet component in large regions of parameter space . in this paper , we examine the discovery potential of @xmath0 produced in association with a bottom - antibottom pair at the lhc through @xmath1 and @xmath2 decay modes . it is shown that an @xmath0 with mass @xmath3 can be extracted from the sm backgrounds by using the @xmath1 decay channel , a possibility precluded to the mssm . in contrast , the @xmath2 decay mode is overwhelmed by backgrounds despite the fact that the branching ratio of this mode can reach unity when @xmath0 is a pure singlet . preprint shep-10 - 37 + * low mass higgs signals at the lhc + in the next - to - minimal supersymmetric standard model * + +
in the integrated spectra of old stellar systems , the strength of the balmer absorption lines reflects the luminosity - weighted effective temperature , which is dominated by the main sequence turn - off . in a simple ( i.e. single - age , single - metallicity ) system , the turn - off luminosity and temperature are sensitive primarily to the age of the population . on this basis , measurements of and line strengths have been widely used to constrain the ages of early - type galaxies both in clusters and in the field ( see thomas et al . 2004b , and many references therein ) . despite extensive work , some fundamental questions remain unresolved , such as the influence of age variations in driving the `` red sequence '' of cluster ellipticals ( e.g. caldwell 2003 ; thomas et al . 2004b ; nelan et al . 2005 ) . the interpretation of the balmer lines suffers from at least three complications . first , horizontal branch morphology is not included in most stellar population models ; if blue horizontal branch stars are present , they enhance the balmer lines and mimic the effect of younger ages ( maraston & thomas 2000 ; thomas et al . 2004b ) . second , the indices are measured on low effective - resolution spectra ( usually limited by the internal velocity dispersion of the galaxy itself ) . this causes the balmer absorption and neighbouring continua to be blended with a forest of nearby metal absorption lines . thus rather than true equivalent widths , one measures `` indices '' ( such as those of the lick system , see burstein et al . 1984 ) , which have residual sensitivity to metallicity and abundance ratios . third , many early - type galaxies exhibit weak nebular emission which `` fills in '' the stellar absorption , leading to over - estimates of the population ages ( e.g. gonzlez 1993 ) . the contamination is reduced for the higher - order lines , and , as a result of the typical line ratios for nebular emission . unfortunately , however , the usual index definitions for these lines are highly contaminated by metal - line blanketing , and depend more strongly on the overall metallicity ( worthey & ottoviani 1997 ; but see vazdekis & arimoto 1999 ) . moreover , they are sensitive to the enhancement of @xmath7-elements , as shown by thomas , maraston & korn ( 2004a ) . by contrast , the low - order line , is a less ambiguous tracer of age , if the nebular emission problem can be overcome . in this paper , we explore whether these trends extend to the lowest - order balmer line , . in section [ sec : popsynth ] , we review two index definitions , and , introduced by nelan et al . ( 2005 ) ( section [ sec : ixdefs ] ) , describe the population synthesis model used ( section [ sec : synthmod ] ) , and determine the sensitivity of the indices to age and metallicity ( section [ sec : synthresults ] ) . section [ sec : agemassobs ] confronts the synthesis results with measurements for red - sequence galaxies in the noao fundamental plane survey ( smith et al . 2004 ; nelan et al . the observed @xmath8 relations for a low - emission subsample are derived in section [ sec : obsdat ] , while section [ sec : agemetvar ] discusses the implied age and metallicity variations . in section [ sec : booterrs ] , we incorporate errors in the population synthesis model , using a bootstrap method . section [ sec : nebcontam ] describes the problem of nebular emission contamination ; other potential complications are discussed in section [ sec : alternative ] . on balance , the steep slope of the observed @xmath8 relations is best reproduced by a strong gradient in age , with low - mass galaxies on average younger than larger objects . we discuss the plausibility of this picture in section [ sec : discuss ] , comparing to observations at intermediate redshifts . the main conclusions are summarized in section [ sec : concs ] -2 mm -3 mm . index definitions from nelan et al . ( 2005 ) . [ cols="<,^,^,^",options="header " , ] in this section , we compare the predicted absorption to observed values reported by nelan et al . ( 2005 ) from the nfps . we use the `` full - resolution '' ( @xmath93 fwhm ) measurements , corrected for aperture and velocity - broadening effects . the nfps galaxy sample was selected by apparent magnitude ( @xmath10 ) and colour ( @xmath11 ) , relative to the red - sequence `` ridge '' in each cluster ( smith et al . 2004 ) . there is no cut on the red side of the ridge . note especially that we do not explicitly select by galaxy morphology ; thus some of the galaxies included in this sample would not be present in a hand - picked set of bona fide ellipticals and s0s . moreover , the nfps sample explicitly selects against any massive ellipticals with relatively blue colours . it is for this reason that we refer always to red - sequence galaxies , rather than early - type galaxies . although the full nfps database includes linestrength data for @xmath94000 galaxies , is beyond the range of most of the spectra . only 689 confirmed cluster members , spanning a redshift range @xmath12 , have measurements for , , plus necessary auxiliary data ( velocity dispersions and emission - line estimates ) . the redshift range of the dataset is less than the median depth of nfps ( 15000 ) . because the nfps galaxy sample is flux limited , and is observed for only the nearby objects , the typical mass of the sample is slightly fainter than that of the whole survey . as discussed in the introduction , a major concern with is that the age - sensitive stellar absorption may be contaminated by nebular emission , both at itself , and in the neighbouring [ nii ] lines . it is critical therefore to work with an subsample of galaxies in which nebular contamination is minimized . nelan et al . ( 2005 ) determined emission equivalent widths for the [ oiii ] line at 5007 , and for , by fitting and dividing an appropriate stellar continuum . for the present analysis , we exclude 250 galaxies with emission at and/or at [ oiii ] , based on a 1@xmath13 detection limit . this rejection scheme is much stricter than the fixed thresholds imposed by nelan et al . to ensure sufficient data quality , we impose a further selection for errors less than 0.3 . after applying these cuts , the sample comprises 410 galaxies . the @xmath2 relations for this low - emission subsample are shown in figure [ fig : data ] . the data are fit using an iterative 3@xmath13 clipping , with the effect of rejecting a further four outliers in and eleven in . for the median @xmath14 , the fitted values are @xmath151.69 and @xmath151.06 . taking both data and models at face value , these indicate average ages of 1013gyr ( ) or 79gyr ( ) , if @xmath16[fe / h]@xmath17 . potential causes of this discrepancy are discussed in sections [ sec : nebcontam ] and [ sec : alternative ] . differential observations , e.g. the change in absorption with galaxy mass , are expected on general grounds to be more robust than absolute values . the observed @xmath2 relation has slope of @xmath0 per decade in @xmath13 . for the equivalent slope is @xmath18 . in the remainder of this section , we attempt to interpret these slopes first in terms of age and metallicity variations , and then consider alternative explanations . relations from nfps data of nelan et al . galaxies from the low - emission subsample ( see text ) , with errors less than 0.3 , are shown as solid points with error bars . open squares mark galaxies rejected from the final fit ( solid line ) by an iterative 3-@xmath13 clipping process . small crosses indicate galaxies failing either the low - emission selection or the low - error selection ( or both ) . in each panel , the dashed lines shows the slope predicted for @xmath8 using the age and metallicity scaling relations derived by nelan et al . from traditional lick indices . the scaling relations of thomas et al . ( 2004b ) result in predicted slopes shown by the dotted line.,title="fig:",width=321 ] -3 mm given the measured gradient of @xmath2 ( or @xmath2 ) and a pair of values summarizing the index response to age and [ fe / h ] , we can constrain the exponents ( @xmath19 ) of scaling relations such that age@xmath20 and fe / h@xmath21 . naturally , since only one gradient is measured , the viable values for @xmath7 and @xmath22 fall along a degenerate linear track , along which the metallicity and age effects combine to produce the observed slope . by setting each of ( @xmath19 ) to zero in turn , we can determine the age and metallicity trends required if only one of these parameters varies with @xmath13 . if we naively adopt simply the `` best '' parameter responses from the population synthesis of section [ sec : popsynth ] , and allow only for the errors in the observed @xmath2 relation we would require either age@xmath23 ( if metallicity is constant ) or fe / h@xmath24 ( if age is constant ) . from the equivalent slopes are ( @xmath19)=(@xmath25 ) with comparable errors . the metallicity trend required is far in excess of determinations using lick indices ( e.g. kuntschner et al . 2001 , who similarly constrained their fits by assuming no age variation , reporting fe / h@xmath26 ) . the age interpretation implies a factor of five in age over the 60300 range in velocity dispersion , which is also surprisingly strong ( but see section [ sec : discuss ] ) . a similarly strong age mass relation has been claimed by nelan et al . ( 2005 ) , based on traditional lick indices from nfps . their reported scaling relations are age@xmath27 and [ fe / h]@xmath28 . allowing _ both _ parameters to vary according to the nelan et al . relations , we would expect an slope of @xmath29 from age and an additional @xmath30 from [ fe / h ] variations , totalling @xmath31 per dex , marginally consistent with the observed @xmath0 . the nelan et al . analysis is independent of the present work , in the sense that it does not employ absorption , and uses a much larger sample of @xmath93500 galaxies . however , the sample characteristics are similar , since both works are based on the same survey . an independent study of 124 galaxies by thomas et al . ( 2004b ) , also based on traditional lick indices , and analysed using the same models as nelan et al . , yields age@xmath32 . the thomas et al . scalings predict a slope of @xmath33 ( @xmath34 from age and @xmath35 from metallicity ) , apparently quite incompatible with our observed relation . compared to the sample used here , the thomas et al . compilation is weighted towards more massive objects ( @xmath36100kms , with median of @xmath9200 ) , and also to fairly clean e or s0 morphologies . both the mass - range and the morphological mix could influence the results obtained , as could the explicit colour - magnitude selection . nelan et al . commented on an apparent steepening of the relation for low - mass systems , which could account for some of the disagreement with thomas et al . , and comparable studies . moreover , there are indications that nfps galaxies with stronger disks show a steeper age mass relation than pure ellipticals or bulge - dominated s0s . both effects act in the sense which could explain the discrepancy between nfps and thomas et al . finally , the nfps colour magnitude selection excludes any massive ` blue ellipticals ' . if included , and if present in sufficient numbers , such objects would tend to flatten the age mass relation . caldwell et al . ( 2003 ) studied a sample of 175 early - type galaxies with @xmath37 , similar to the range probed in this paper . using their favoured index combination , caldwell et al . derive a steep age mass slope of age@xmath38 ( see their figure 21 and table 9 ) , in agreement with nelan et al . and with our results . again , this seems argue for a dependence on the mass - range , with giant ellipticals favouring a flatter age mass relation , and low-@xmath13 objects yielding a steeper slope . as already hinted at , the above discussion neglects an important source of error , arising from uncertainties in the population synthesis model itself . these errors can be included in the constraints by propagating the bootstrap - response distribution of figure [ fig : bootresp ] . recall that each bootstrap - realisation of the input stellar library was used to determine a new estimate of the parameter responses . most of the bootstrap realisations yield responses close to the `` best '' values , and hence yield ( @xmath19)-constraints similar to the picture described in the previous section . on the other hand , some bootstrap points show strong [ fe / h ] responses ( e.g. around -0.5 in per dex fe / h ) . these clearly improve the chances of observing a strong @xmath2 trend with neither a strong age - trend nor an excessive [ fe / h]@xmath2 scaling . finally , the spray of `` anomalous '' bootstrap results , due to realisations which lack cool dwarfs , assigns some ( low ) probability to quite unexpected ( @xmath19 ) pairs . , the likelihood of the observed gradient , given scaling relations age@xmath39 and fe / h@xmath40 . in addition to the observational errors on the @xmath2 slope , the likelihood explicitly accounts for synthesis model errors , using the bootstrap simulations . bold contours show the resulting 1@xmath13 , 2@xmath13 and 3@xmath13 confidence regions for two parameters . for comparison , dashed contours show the equivalent limits if model errors are neglected . the labels ` n ' and ` t ' refer to the scaling relations from nelan et al . ( 2005 ) and thomas et al . ( 2004b ) respectively . the lower panel shows the equivalent results for .,title="fig:",width=321 ] -3 mm combining these simulations , figure [ fig : bootscal ] shows the likelihood function @xmath41 allowing for the bootstrap errors and also for measurement error in the observed @xmath2 slopes . as expected , the range of viable scaling relations is considerably expanded when the model errors are included . formally , for a constant - metallicity sequence , we obtain age@xmath42 from . in the constant - age case , the fe / h@xmath2 relation has a slope of @xmath43 at 1@xmath13 ( @xmath44 at 2@xmath13 ) , with no meaningful upper limits . thus if a constant age is imposed , the implied metallicity trend remains much steeper than values favoured by lick - index studies . allowing both parameters to vary , the nelan et al . ( 2005 ) scalings lie at the @xmath45 contour . the scalings of thomas et al . ( 2004b ) remain disfavoured at the @xmath93@xmath13 level . two curious aspects of figure [ fig : bootscal ] are worthy of brief comment . first , the change in shape of the likelihood contours , which now form a closed curve at 1@xmath13 suggests that allowing for the bootstrap errors has actually tightened the constraint on the metallicity - gradient taken separately . however , if we integrate over the age - gradient as a nuisance parameter , to obtain the one - parameter constraint , the vertical `` flaring '' of the likelihood function compensates for this effect . second , because the likelihoods were derived from only 1000 bootstrap realisations , the 3@xmath13 contour is not well - determined . the noisy features at the lower - left of the panel result from the anomalous bootstrap solutions , i.e. from the upper right of figure [ fig : bootresp ] . the 3@xmath13 contour is retained in the plot mainly to show this effect , and to illustrate the discrete tracks which contribute to @xmath41 . the greatest obstacle to the use of absorption as a stellar age indicator is its potential for contamination by nebular emission , both at itself , and in the neighbouring [ nii ] lines . in section [ sec : obsdat ] , we used the emission measurements at and [ oiii ] to reject the objects most strongly affected . this approach lacks precision for very low - level contamination , since the emission is weaker than . moreover , the measurements are themselves made in the presence of imperfectly - known stellar absorption , potentially leading to underestimated equivalent widths . as a result , our `` low - emission '' subsample will include some galaxies with low - level contamination . in general , weak emission will yield smaller values for , causing galaxies to appear older . in the lower panel of figure [ fig : data ] , the scatter of outliers below the mean @xmath2 relation suggests such an effect , with a few galaxies showing negative values , i.e. net emission . for the situation is more complicated , since [ nii ] in the red continuum region leads to increased index values , which may explain the absence of such dramatic outliers in the @xmath2 relation . this compensating influence depends on the ratio of [ nii ] to emission , which is strongly correlated with absolute magnitude ( phillips et al . most of the outliers in are fairly high mass objects , where the [ nii ] compensation will be most efficient . in this case however , it is surprising that they are not detected at [ oiii ] which should also be strong in such objects . although clipping of deviant points can remove a small number of outliers , a more serious concern is whether the mean relation itself might be driven by variations in emission characteristics , rather than age or metallicity . one test for this is to repeat the fits , allowing for terms in and [ oiii ] emission , as well as velocity dispersion . in these fits the @xmath2 and @xmath2 slopes are very little changed , and the emission - line coefficients are not significant . to generate the observed @xmath2 trends with emission variations would require increasing emission with increasing mass ( opposite to the behaviour at ) . the similarity of results from and is also difficult to explain if the @xmath2 slopes are mainly due to emission : because the two indices respond differently to [ nii ] , we would need either very low [ nii]/ ratio throughout , or else very little variation in this ratio as a function of mass . either option is contrary to the behaviour observed in galaxies with strong emission ( phillips et al . moreover , the amount of emission would have to be very precisely determined at any given mass , in order to retain a tight @xmath2 relation at all . a related concern is whether the -based selection itself biases the sample against low - mass weak - absorption objects . this is conceivable , since separating emission and absorption components of is more difficult at low-@xmath13 ( nelan et al . 2005 ) . the signature of such a bias would be an increased proportion of emission - rejected galaxies below the fit , at small @xmath13 . the effect is not readily apparent in figure [ fig : data ] . running the fits with a variety of alternative emission rejection schemes , including that of nelan et al . , the gradients are stable at the 15% level . for , the results are less stable , yielding steeper slopes when the looser selection criteria are used . finally , emission does not easily account for the absolute age discrepancy between and . recall that yields older ages at given @xmath13 than . the emission effect should cause both indices to yield older ages , but should be less strongly affected , as a result of the [ nii ] compensation . in conclusion , emission contamination remains a very serious concern for absorption studies . emission undoubtedly generates outliers from the observed relations , some of which can not easily be excluded using other observations . on balance however , it appears unlikely that emission variations are responsible for the strong , decreasing trends of and with velocity dispersion . as well as sensitivity to nebular emission , our result is subject to a number of other caveats , relating to assumptions and limitations of the synthesis model . the following paragraphs consider three such concerns : blue horizontal branch ( bhb ) stars , non - solar abundance ratios and imf slope variation . the balmer lines measure the luminosity - weighted temperature of a stellar population , so the trend could be generated by any hot stellar component which becomes stronger at low @xmath13 . maraston & thomas ( 2000 ) and thomas et al . ( 2004b ) have explored models with bhb populations , and shown that the strong absorption of some ellipticals might indeed arise from bhb stars rather than from younger ages . because these stars have blue continua , they contribute more strongly to the higher - order lines than to ( schiavon et al . 2004 ) , and their impact will be weakened further at . quantitative modelling of the relative responses should offer one approach to constraining the importance of bhb stars in old populations . we have argued that non - solar abundance ratios are unlikely to have a strong effect on , since there are no strong metallic features within the band - passes used by our indices . the only feature of concern is the cai 6573 line , which affects the and indices in opposite senses . the similarity of the scaling relation slopes obtained from the two indices suggests that variations of ca abundance with @xmath13 do not strongly influence the results . on the other hand , the difference in absolute ages implied by the two indices could be related to this line . in particular , if ca is under - abundant for the whole sample ( relative to stelib ) , then will be anomalously low ( leading to older ages ) , while will be larger ( yielding younger ages ) . thus the sense of the effect is as observed . however , to force and to give consistent ages at given @xmath13 , we need a large metallicity difference , of @xmath460.3dex . on balance , simple differences in flux - calibration or resolution - matching seem more likely explanations for the 0.1 offset required to resolve the absolute age inconsistency . finally , we have explored the effects of changing the assumed imf slope in the synthesis . for power - law exponents in the range 1.03.5 ( cf . default 2.35 ) , the predictions for and vary by @xmath470.3 . thus the full range of @xmath90.75 can not be generated by plausible levels of variation in the imf slope . having reviewed some of the potential systematic effects , in this section we consider whether a steep age mass relation is compatible with other constraints , focusing on observations of cluster galaxies at lookback times of a few gyr . in what follows , we will adopt the -derived relation of age@xmath48 , which assumes no metallicity variations . considering that lick index studies suggest a mix of age and metallicity effects , this slope should be regarded as an upper limit . since the absolute ages implied by the synthesis are unacceptably old , we fix the age zero - point by forcing the most massive galaxies to the age of the universe . ( note that if were used instead , this rescaling would not be necessary , and all other results would be essentially unchanged . ) in this section we adopt the wmap cosmological parameters , @xmath49 ( bennet et al . 2003 ) , yielding a maximum age 13.7gyr . in this case , mass relation implies a mean age of only 3gyr for the smallest objects , i.e. with @xmath50 , while the observed scatter around the @xmath2 relation allows for an intrinsic age range of 25gyr for these galaxies . the age mass relation is shown in figure [ fig : tform ] , together with some of the other results discussed below . in all cases , the `` age '' or `` formation redshift '' is to be understood as a luminosity - weighted mean age for the stars and does not imply the galaxy assembly time . at times earlier than this `` age '' , the galaxy may have been present in the cluster but with the bluer colours indicative of a star - forming system . the steep @xmath2 trend then appears to favour a picture in which galaxies now at the faint end of the red sequence ceased forming stars only at modest redshifts , @xmath51 . a testable implication is that the red sequence in high redshift clusters should be deficient in faint galaxies , with the depletion mass increasing with increasing redshift . recent observations of @xmath520.8 clusters have indeed suggested that the red sequences are depleted at the faint end ( de lucia et al . 2004 ; goto et al . 2005 ) . converting their observed magnitudes to a velocity dispersion ( after accounting for k - correction and passive evolution ) , the faintest non - star - forming objects in these clusters have @xmath53 , at a look - back time of @xmath97gyr . for comparison , the age mass relation derived in this paper also places the current age of such galaxies at @xmath97gyr . intriguingly , there is evidence for red - sequence truncation at much smaller lookback times . smail et al . ( 1998 ) observed clusters at @xmath54 ( lookback time of only @xmath93gyr ) ; their figure 5 shows a marked cut - off in the red sequence , at a magnitude corresponding to @xmath55110 . for comparison , our age mass relation suggests ages of @xmath95gyr for such objects . a rigorous comparison of these works is not trivial , since they do not use consistent definitions for the `` truncation magnitude '' . however , there is very suggestive agreement with our results , both in terms of the absolute age of the faint red sequence galaxies , and in the trend for fainter objects joining the red sequence at lower redshifts . additional support comes from the rapidly - evolving characteristic mass scale of e+a ( or post - starburst ) galaxies , as determined by tran et al . such galaxies can be interpreted as objects which have just ceased forming stars , and are in the process of fading onto the red sequence . the results of tran et al . suggest that e+a galaxies observed at @xmath56 ( lookback time 7gyr ) will fade to @xmath57 red - sequence objects today , while those in the e+a phase at @xmath58 ( 3.5gyr ) will fall onto the red sequence at @xmath59 . again , there is a suggestive level of agreement with our results : the slope of the e+a mass evolution itself suggests age@xmath60 , slightly steeper than our result . the e+a `` ages '' fall @xmath91gyr younger than our relation would predict . qualitatively this is expected , since the e+a phase signals the last star - formation episode , while our relation describes the mean stellar age , which is necessarily older . in conclusion , local galaxy studies which suggest rapid recent evolution in the red sequence population ( caldwell et al . 2003 ; nelan et al . 2005 ; this work ) are compatible with direct observations of galaxy populations at intermediate redshifts . it is not , of course , implied that the faint red galaxies were themselves assembled at such recent epochs , but rather that star formation ceased at that time , perhaps following expulsion of remaining gas , or its rapid consumption in a starburst . on the other hand , it may be a challenge to reconcile such recent star formation with the tightness of the colour - magnitude relation itself ( e.g. bower , lucey & ellis 1992 ) , without invoking age metallicity `` conspiracy '' models . , as suggested by the @xmath2 relation of this paper . plausible metallicity variations could flatten this relation by 1040% ( see figure [ fig : bootscal ] ) . the ages are calibrated by fixing the high - mass galaxies to the age of the universe , 13.7gyr . the results of nelan et al . ( 2005 ) , using multiple lick indices for @xmath93000 nfps galaxies are shown as the filled circles . open circles show the characteristic velocity dispersion of e+a galaxies in clusters from z=0.3 to z=0.8 ( tran et al . these results may be compared to the low - redshift age measurements if the e+a objects are interpreted as newly - quiescent objects fading onto the red sequence . finally , crosses indicate the mass scale corresponding to faint - end truncation of the red sequence in clusters at z=0.24 ( smail et al . 1998 ) , z=0.75 ( de lucia et al . 2004 ) and z=0.83 ( goto et al . we have investigated the behaviour of absorption indices in single - age , single - metallicity stellar populations , and compared to observed trends in red - sequence cluster galaxies . the principal conclusions are : 1 . the absorption strength , measured using or , appears promising as a probe of stellar ages in integrated spectra , with only weak response to metallicity variations . the indices are expected to be more robust against non - solar [ @xmath7/fe ] ratios than the high - order balmer lines . the principal drawback to the method is its sensitivity to nebular emission contamination , which can only be minimized through careful sample selection . 3 . at face value , the slope of the observed @xmath2 relation , from nfps , requires a factor of @xmath95 in age over the observed @xmath61dex range in velocity dispersion , with younger ages for lower - mass objects . a pure metallicity interpretation requires @xmath62[fe / h]@xmath63dex over the same range , inconsistent with other constraints . the combined , more modest , age and metallicity trends reported by nelan et al . ( 2005 ) for nfps reproduce the observed @xmath2 slope within the errors . by contrast , the scaling relations of thomas et al . ( 2004b ) significantly under - predict the slope . if the @xmath2 slope is due purely to age , the derived age mass relation implies ages of only a few gyr for low - luminosity red - sequence galaxies in clusters . interpreted as the time at which such galaxies ceased forming stars , the ages broadly agree with an observed deficit of faint red - sequence members in clusters at @xmath64 . in summary , absorption is a potentially powerful , but neglected , probe of early - type galaxy formation epochs . future work with improved stellar libraries should aim to constrain further the metallicity dependence of indices , and determine optimal passbands for index definitions . the analysis in this paper supports a steep age mass relation along the red sequence of cluster galaxies . although the most massive cluster ellipticals are indeed very old , our work suggests that the faint red - sequence galaxies became quiescent as recently as @xmath650.5 . qualitatively , this `` downsizing '' scenario ( cf . cowie et al . 1996 ) agrees with recent detections of a faint - end depletion of the red sequence at higher redshifts , and with the evolution of e+a galaxy masses over the past @xmath97gyr . it is a pleasure to thank mike hudson for helpful and encouraging comments on this work , and all of my collaborators in nfps , for their help in assembling the dataset used here . an anonymous referee provided several useful and constructive suggestions . all of the analysis was conducted , and all of the figures generated , using `` r '' , an open - source language and environment for statistical computing ( http://www.r-project.org ) .	we compute population synthesis models for the variation of absorption indices ( and ) , as a function of age and metallicity in old stellar systems . the models are based on the stelib spectral library of j .- f . le borgne et al . , and defined at a resolution of 3 fwhm . errors in the age and metallicity responses are derived by bootstrap resampling the input measurements on the stellar library . the indices are found to be highly sensitive to age variation , with only moderate response to metallicity . for galaxies uncontaminated by nebular emission , our index is more powerful in breaking the age metallicity degeneracy than or . using a sample of red cluster galaxies from nelan et al . , carefully selected to exclude objects with emission , we find a steep decline of with velocity dispersion ( slope @xmath0dex@xmath1 ) . the slope can be translated to constraints on age and metallicity scaling relations , incorporating measurement errors and also the model errors determined from the bootstrap method . if the @xmath2 slope is due only to age , we obtain age@xmath3 . because depends quite weakly on [ fe / h ] , a metallicity interpretation would require fe / h@xmath4 or steeper . the @xmath2 slope is consistent with the combined age and metallicity scaling relations reported by nelan et al . from classical lick indices . the relations obtained by thomas et al . significantly under - predict the observed slope . the discrepancy could arise from differences in the sample selection . in particular our sample probes a lower mass range , is not explicitly selected on morphological criteria and excludes objects significantly bluer than the red sequence . we discuss in detail the effects of emission contamination on the results , and conclude that these are unlikely to yield the observed behaviour in the @xmath2 relations . indeed , similar results are obtained using , despite its different sensitivity to and [ nii ] emission lines . the steep age mass relation supports a `` downsizing '' formation scenario : fainter red - sequence galaxies became quiescent at lower redshifts , @xmath5 . this picture accords with recent observations of truncated red sequences in clusters at @xmath6 . [ firstpage ] galaxies : elliptical and lenticular , cd galaxies : evolution galaxies : stellar content
the ` streaming motions of abell clusters ' ( smac ) project is a fundamental plane ( fp ) survey of @xmath2700 early - type galaxies in 56 local rich clusters . the cluster sample has approximately full - sky coverage , and a limiting depth of @xmath3kms@xmath1 . within each cluster , distances are based upon data for 456 e / s0 galaxies . data for the smac project is drawn from a compilation of literature sources and an extensive body of new observations , carefully combined into a homogeneous database . further details of the sample , the observations , and the data compilation procedures are reported by hudson et al . ( 1999 ) and in a forthcoming series of papers . our ( inverse ) fp analysis closely follows the methods employed , for a smaller sample of clusters , by hudson et al . the fp yields distances to a precision of @xmath4 per galaxy , so that cluster distances are determined with errors of 312% , with a median 8% . the velocity zero - point is calibrated by requiring that the sample exhibit no net radial inflow or outflow . the balanced sky - coverage of the sample ensures that there is little covariance between monopole and dipole components of the velocity field . the observed velocity field is presented in the upper panel of figure 1 . lucey et al . ( in this volume ) present some example comparisons of the smac distances / velocities with those determined from other surveys with clusters in common ( eg lauer & postman 1994 ; dale et al . 1999 ) . such comparisons generally reveal an acceptable level of agreement , in so much as the scatter is compatible with the quoted errors . however , the distance errors are such that cluster - by - cluster comparisons are rather crude tests for systematic effects in distance estimates . [ galplot ] -15 mm -10 mm the best - fitting bulk - flow model is the vector @xmath5 which minimizes @xmath6 where @xmath7 is the observed radial peculiar velocity for cluster @xmath8 , whose direction vector is @xmath9 . the weights are assigned according to the observational peculiar velocity error , @xmath10 , and a ` thermal ' velocity noise @xmath11 , here fixed at 250kms@xmath1 . in what follows , we shall refer to @xmath5 as ` the bulk - flow ' , but this terminology should be understood as a shorthand for ` the best - fitting bulk - flow model ' . as demonstrated by feldman & watkins ( 1994 ) and further illustrated by hudson et al . ( this volume ) , the incomplete cancellation of small- and intermediate - scale flows causes bulk - flow measurements to depend upon the sample geometry , characterised by the survey window function . this effect is especially important when the spatial sampling is sparse , as for cluster surveys . for the smac sample , we determine the following cmb - frame bulk - flow solution : @xmath12\,\ , { \rm km\,s}^{-1}\ ] ] in supergalactic cartesian co - ordinates ( used throughout this paper for flow vector components ) . after correcting for ` error - biasing ' , this vector has magnitude @xmath13kms@xmath1 and is directed towards @xmath14 . note that the error ellipsoid is anisotropic , as demonstrated by monte - carlo simulations in the lower panel of figure 1 . the bulk - flow errors include a contribution from ` system - matching ' errors which introduce covariance between cluster distance errors ( smith et al . 1997 ; hudson et al . 1997 ) . the large amplitude of the observed flow is initially surprising . we have tested for a wide range of possible systematic effects which might affect this result : no single cluster , or supercluster region , is responsible for the flow ; correcting fp distances for stellar population differences , based on the clusters offsets from the mg@xmath15 relation , does not significantly affect the bulk - flow ; our choice of schlegel et al . ( 1998 ) extinction maps over burstein & hieles ( 1982 ) makes little difference to the result ; a ` method ii ' analysis , which fits simultaneously the fp parameters and flow model and is insensitive to malmquist bias , yields indistinguishable results . at face value , the smac result appears inconsistent with the bulk - flow solution of lauer & postman ( 1994 ) , whose apex lies @xmath16 from the smac flow direction , as well as with the non - detection of bulk - motion in the tully fisher survey of dale et al . ( 1999 , and these proceedings ) . in fact , the inconsistencies between surveys are marginal when one accounts for differences in the survey window functions ( again see hudson in this volume ) . for the same reason , the apparently good agreement between the smac and willick ( 1999 ) solutions should not be over - interpreted : the surveys have very different window functions , and we should not _ expect _ the derived bulk - flows to agree so closely ! how to interpret the smac bulk - flow result depends upon whether the observed velocities of clusters ( individually , or combined via the best - fitting bulk - flow model ) can be understood as the expected response to the surrounding density field . here , we compare the smac data to density - velocity reconstructions from the most extensive all - sky redshift survey of iras galaxies , the pscz ( see contributions of saunders and branchini in this volume ) . the analysis presented here is somewhat preliminary , and various systematic effects remain to be investigated . first , let us consider the predictions for the bulk - flow . we compute predicted velocities for each of the smac clusters ( assuming @xmath17 for now ) , and substitute these predictions in place of the @xmath7 in equation [ chibulk ] . we again minimize @xmath18 with respect to @xmath5 , using the same weights computed from the smac measurement errors and the thermal velocity noise . the best - fitting flow vector is what pscz predicts _ for the smac bulk - flow _ , ie . given the smac errors and the smac sample geometry . specifically , we fully account for any over - representation of clusters in ( for example ) the great attractor direction . the _ expected _ bulk - flow solution for @xmath19 is then : @xmath20\,\ , { \rm km\,s}^{-1}\ , .\ ] ] a cursory comparison between solutions [ obsbulk ] and [ predbulk ] reveals that the pscz velocity field for @xmath19 , given the smac sampling geometry and measurement errors , predicts a best - fitting bulk - flow similar to that observed , ( including the large negative sgz component ) . solution [ predbulk ] has magnitude 490@xmath21140kms@xmath1 ( error - bias corrected ) and direction @xmath22 , which is only @xmath217@xmath23 from the direction of the observed flow apex ( see figure 1 ) . from monte - carlo simulations ( lower panel of figure 1 ) , we find that such a good _ directional _ agreement arises by chance in @xmath24 of realisations . alternatively , we can compare directly the predicted and observed velocity fields on a cluster - by - cluster basis , as in figure 2 . the fit yields a measurement of @xmath25 , with no individual cluster influencing the result by more than 10% . we can introduce greater freedom into the model velocity field by fitting simultaneously for @xmath26 and a ` residual ' bulk - flow vector . this residual bulk - flow will absorb any dipole signature _ not _ accounted for by the pscz galaxy distribution ( either a real streaming generated at very large depths , or a spurious flow resulting from systematic errors ) . the results of the fit are an unchanged value for @xmath27 and a residual flow vector : @xmath28\,\ , { \rm km\,s}^{-1}\ , .\ ] ] the errors are larger than previously since we are now fitting for more parameters ; however , the correlation of observed with predicted peculiar velocities remains significant at the @xmath29 level . that the residual bulk - flow is not significant is a strong indication that the observed smac flow is indeed compatible with the expected response to local density fluctuations ; and suggests that the observed flow does not result primarily from systematic distance errors . the favoured value for @xmath26 is higher than many estimates , but is consistent with other determinations based on pscz , which yield @xmath30 ( see branchini , this volume ) . note that there is some covariance between @xmath26 and @xmath31 , such that fixing @xmath32 results in a solution with larger residual flow , albeit significant at only the 1.5@xmath33 level . -8 mm -95 mm a final caveat should be noted : the preliminary analysis presented here is based on a method i approach and is therefore affected by malmquist bias . homogeneous malmquist effects have been corrected for and the effects of inhomogeneous malmquist bias ( imb ) are suppressed by the use of a cluster sample with relatively small distance errors per ` object ' . nonetheless , it may be expected that the ` spurious infall ' patterns produced by imb act to artificially inflate @xmath26 . a crude attempt to judge the effect can be made by removing the 12 clusters with @xmath34kms@xmath1 , which lie mostly in infall regions where imb will be most severe . the result has much greater uncertainties of course , since we have removed the clusters which contribute the greatest signal - to - noise : we obtain @xmath35 , with no measurable residual bulk - flow ( amplitude @xmath36kms@xmath1 ) . in a future analysis of the smac pscz comparison , we will employ a method ii approach which is free from malmquist bias effects . the smac sample includes 10 clusters within 15@xmath23 of the shapley / great attractor ( ga ) direction ( visible in figure 1 as the concentration of points at @xmath37 ) . since this region is of some historical interest , we compare in figure 3 the observed velocities with some ( over-)simplistic toy - models for the dynamics of the region . the models are based upon the simple spherical attractors of faber & burstein ( 1988 ) . we consider the effect of two such structures , one at the position of the ` traditional ' ga ( at a distance of 4500kms@xmath1 ) and a second centered on the shapley core region ( at 14500kms@xmath1 ) . the models are normalised to generate the ga - directed component of the local group s velocity in the cmb frame , viz . 500kms@xmath1 . the leftmost panel of figure 3 demonstrates that a pure ga infall model is a poor fit to the smac data , especially in the immediate ga background , where no ` far - side infall ' is observed . equally , the extreme shapley infall model dramatically over - predicts velocities beyond @xmath210000kms@xmath1 ( middle panel ) . allowing for contributions from both attractors , and fitting for their relative amplitudes as the single free parameter , we can obtain an acceptable fit to the data , as shown in the rightmost panel . the best - fitting model has shapley and the ga each generating @xmath38% of the local group velocity in this direction , proportions which accord with hudson s ( 1994 ) conclusions based on the mark ii dataset . here , as in the case of ` bulk - flow ' we caution against drawing quantitative conclusions from analyses based on unrealistically simple models of the velocity field . however , the above fits draw attention to the qualitative behaviour of the smac velocity data in the ga / shapley direction . the absence of far - side infall in the smac data ( here visualised as a retardation of one infall pattern by another more distant infall structure ) is also apparent in the pscz maps , which reveal a wealth of structure along the ga - shapley axis . from our analyses of the smac velocity field , as presented in the preceeding sections , we draw the following conclusions : 1 . the best - fitting bulk - flow model has amplitude @xmath13kms@xmath1 , towards @xmath14 . the direction of the observed bulk - flow is within @xmath215@xmath23 of the direction predicted from the iras pscz redshift survey for the smac sample , while the amplitude can be matched if @xmath39 . 3 . comparing observed and predicted velocities cluster - by - cluster yields @xmath40 , compatible with other analyses based on the pscz . the observed velocity field is fully accounted for by the local density field , with no recourse to residual flow generated beyond the limits of pscz , or to ` spurious ' flows associated with systematic errors . the shapley concentration may be responsible for a significant fraction of the infall traditionally associated with the great attractor . burstein d. , heiles c. 1982 , , 87 , 1165 dale d. a. , giovanelli r. , haynes m. p. , campusano l. e. , hardy e. , borgani s. , , 510 , l11 faber s. m. , burstein d. , 1988 , in coyne g. , rubin v. c. , eds , proceedings of the vatican study week , large scale motions in the universe . princeton univ . press , princeton , p. 135 feldman h. a. , watkins r. 1994 , , 430 , l17 hudson m. j. 1994 , mnras , 266 , 468 hudson m. j. , lucey j. r. , smith r. j. , steel j. 1997 , , 291 , 488 hudson m. j. , smith r. j. , lucey j. r. , schlegel d. j. , davies r. l. 1999 , , 512 , l79 lauer t. r. , postman m. 1994 , , 425 , 418 schlegel d. j. , finkbeiner d. p. , davis m. 1998 , , 500 , 525 smith r. j. , lucey j. r. hudson m. j. , steel j. 1997 , , 291 , 461 willick j. a. 1999 , , submitted ( astro - ph/9812470 )	the smac project is a fundamental plane peculiar velocity survey of 56 clusters of galaxies to a depth of @xmath012000kms@xmath1 . we present here some results from the analysis of the smac velocity field , focussing on three specific features : the best - fitting bulk - flow model for the smac data ; the agreement between the observed velocity field and predictions from the iras pscz redshift survey ; the role of the great attractor and shapley concentration in generating the local flows . we argue that the local mass distribution , as probed by the pscz , can fully account for the observed cluster velocities .
one - dimensional exactly solvable models ( spin chains , rings ) @xcite have been actively employed for studying various problems of spin dynamics @xcite and quantum information theory @xcite . substantial progress in our understanding of spin dynamics has been achieved on the basis of a homogeneous xy model for spin chains ( @xmath0 ) in transverse magnetic field @xcite . recently , hamiltonians of the simplest inhomogeneous systems ( alternating systems ) have been diagonalised for the ring @xcite and the linear spin chain @xcite . development of methods for exact solution of inhomogeneous spin problems has become especially urgent in conjunction with recent progress in quantum computation and quantum information theories @xcite . in particular , these methods can be used for studying the quantum state transfer from one end of the chain to another one @xcite the qubit addressing problem can be attacked by variation of the larmor frequencies of different spins @xcite in classical one - dimensional models . this immediately brings inhomogeneity into the xy model as diagonal elements of the corresponding hamiltonian are not equal . consequently , we arrive at one - dimensional spin models with hamiltonians described by three - diagonal matrices which elements on the diagonal are not - equal , and those under and over the main diagonal are not identically equal as well . low sensitivity of nuclear magnetic resonance ( nmr ) , which is widely applied in experimental implementation of quantum computations @xcite , leads to a further complication of the model described above . consideration of @xmath1-qubit systems brings us naturally to the study of @xmath2 spin chains ( @xmath3 , @xmath4 is arbitrary , @xmath5 are positive integers ) , with larmor frequencies and constants of spin - spin interactions repeating periodically with period @xmath1 . the paper suggests a new method of diagonalisation of a hamiltonian of a linear @xmath1-periodic spin chain of length @xmath2 for various values @xmath3 . special attention is given to the case @xmath6 . for a spin chain of @xmath7 sites with periodic parameters ( period 3 ) a multiple quantum dynamics @xcite is analysed and intensities of all mq coherences are calculated ( see figure 2 ) . this analysis is based on the explicit diagonalisation of the hamiltonian of a linear 3-periodic spin chain with @xmath7 sites ( theorem [ th3n+2 ] ) and the exact formulae , obtained in the paper ( theorem [ trace ] ) , for the intensities of all mq coherences developed in any nuclear spin system coupled by the nearest neighbour dipolar interactions . we conclude the paper with a discussion of the properties of eigenvalues and eigenvectors of the general @xmath1-periodic systems with @xmath8 sites ( theorem [ mainkn-1 ] ) . the hamiltonian of a spin-1/2 open chain with only nearest neighbour ( nn ) couplings has the following general form @xmath9 where @xmath10 , @xmath11 , are the larmor frequencies , and @xmath12 , @xmath13 , are the nn coupling constants . the nuclear spins are specified by the spin-1/2 operators @xmath14 at the sites @xmath11 with the projections @xmath15 . the jordan wigner transformation @xcite @xmath16 from the spin-1/2 operators @xmath14 to the creation ( annihilation ) operators @xmath17 @xmath18 of the spineless fermions takes the hamiltonian ( [ theorham ] ) into the hamiltonian @xmath19 or in the matrix notations into @xmath20 in eq . ( [ jwm ] ) we denote the row vector @xmath21 by @xmath22 , the column vector @xmath23 by @xmath24 ( the superscript @xmath25 represents the transpose ) , and specify the matrices @xmath26 and @xmath27 as @xmath28,\quad d= \left[\begin{array}{ccccc } 0 & d_1&\cdots & 0 & 0\\ d_1 & 0&\cdots & 0 & 0\\ \vdots&\vdots&\ddots&\vdots&\vdots\\ 0 & 0&\cdots & 0 & d_{n-1}\\ 0 & 0&\cdots & d_{n-1 } & 0\\ \end{array } \right].\ ] ] to diagonalise the matrix @xmath29 means to construct such unitary matrix @xmath30 and diagonal matrix @xmath31 that @xmath32 where the superscript @xmath33 represents the conjugate transpose , and columns @xmath34 of @xmath35 , @xmath36 , form an orthonormal basis of eigenvectors of @xmath37 . the new fermion operators @xmath38 and @xmath39 introduced by the relations @xmath40 bring hamiltonian ( [ jwm ] ) into the hamiltonian @xmath41 with energies @xmath42 of the free fermion waves . the problem of diagonalisation of @xmath29 with arbitrary constants @xmath10 , @xmath11 , and @xmath43 , @xmath44 , for large @xmath45 ( @xmath46 ) and computation of various functions of @xmath29 is a time consuming problem usually dealt with the help of super computers . on the other hand , there are two known exact solutions ( suitable for studying systems with large n ) for such a problem when the spin - system considered has periodic changing larmor frequencies and coupling constants . namely , the case of equal sites ( @xmath47 , @xmath48 ) has been solved in @xcite , while the case of period 2 with odd @xmath45 ( @xmath49 , @xmath50 , @xmath51 , @xmath52 ) has been solved in @xcite . the present paper concerns properties of a general periodic chain . in the following sections we shall demonstrate how to reduce the diagonalisation problem of the hamiltonian of a @xmath1-periodic chain with @xmath8 sites ( @xmath53 ) to the problem of finding roots for explicitly given polynomials of degree less or equal @xmath1 . therefore , the methods developed are particularly useful when @xmath54 . in what follows we will utilise essentially the result on diagonalisation of the homogeneous ( in other words , 1-periodic ) chain @xcite which we state now . [ jlambda ] let @xmath55 , then the matrix @xmath56\ ] ] has @xmath4-distinct eigenvalues @xmath57 and the corresponding eigenvectors are of the form @xmath58 * proof . * it is a straightforward verification . let us consider an open periodic chain of @xmath59 sites ( @xmath60 , @xmath61 ) with @xmath1-periodic non zero nn coupling constants : @xmath62 @xmath63 and @xmath1-periodic larmor frequencies @xmath64 @xmath65 ( @xmath66 , @xmath67 , @xmath68 ) . we shall demonstrate how to reduce the diagonalisation problem for the hamiltonian of such a system to the problem of diagonalising of a certain @xmath69 block matrix which entries are matrices of dimension @xmath70 , @xmath71 , @xmath72 and @xmath73 ( @xmath74 in section 5 and @xmath75 in section 7 ) . a particular case of this reduction will be used in the consecutive sections for studying @xmath1-periodic chains of length @xmath8 . before we proceed we make an agreement regarding our notations . if a gothic letter is used in the description of any matrix as its element , then this means that the corresponding place in the matrix is a matrix . the size and the structure of a matrix denoted by a gothic letter should be clear or given in the context . therefore , the original matrix is a block matrix . to underline that the matrix consists of blocks we shall often denote such matrices in bold . we shall also denote by @xmath76 the @xmath70 identity matrix . the hamiltonian of a @xmath1-periodic system with @xmath59 sites has the following form : @xmath77 where @xmath78\ ] ] and @xmath79.\ ] ] to diagonalise matrix @xmath80 one finds ( real ) eigenvectors @xmath81 and ( real ) eigenvalues @xmath82 , @xmath83 , ... @xmath59 , which satisfy the following equation @xmath84 to resolve this equation we associate to each vector @xmath85 vectors @xmath86 , @xmath87 , formed by those coordinates of @xmath88 whose numbers have residue @xmath89 modulo @xmath1 . observe that among vectors @xmath90 there are @xmath91-vectors of dimension @xmath92 and @xmath93 vectors of dimension @xmath4 . now equation ( [ eq ] ) can be rewritten as a system of @xmath4 linear equations in @xmath86 : @xmath94=\lambda_{\nu}\left[\begin{array}{c } u_{(1)}\\ \vdots\\ u_{(k)}\\ \end{array}\right]\ ] ] with matrix @xmath95 . here @xmath96,\ ] ] @xmath97 for @xmath98 , @xmath99 for @xmath100 , and with an exception of two degenerate cases @xmath101,\ ] ] with @xmath102 for @xmath103 , @xmath104 for @xmath105 , and @xmath106 given by @xmath107,\qquad { \cal d}_k=\left[\begin{array}{ccccccc } 0 & d_k & 0 & \cdots & 0&0&0\\ 0 & 0 & d_k & \cdots & 0&0&0\\ 0 & 0 & 0 & \ddots & 0&0&0\\ \vdots&\vdots & \vdots&\ddots&\ddots&\vdots&\vdots\\ 0&0&0&\cdots&0 & d_k & 0\\ 0&0&0&\cdots&0 & 0 & d_k\\ \end{array}\right].\ ] ] it is easy to see that @xmath108 where @xmath109 is the jordan @xmath110 cell : @xmath111.\ ] ] there are two degenerate cases . in the case @xmath112 matrix @xmath113 decomposes as @xmath114\ ] ] where @xmath115 : @xmath116,\ ] ] while in the case @xmath117 and @xmath118 we do not get matrix @xmath119 and matrix @xmath120 is of the form @xmath121.\ ] ] it is easy to see that this matrix satisfies @xmath122 for a diagonal matrix @xmath123 with the first @xmath124 diagonal elements equal to 1 and the last one equals to 0 . from the reduction obtained above we divide consideration of the diagonalisation problem into three different cases : period @xmath125 and @xmath126 in each case there are further reductions depending on the value of @xmath91 . in the next sections we shall work out the case @xmath127 and @xmath128 . the diagonalisation process of sections 5 and 7 will relay on some elementary facts about matrices of the form @xmath129 where for @xmath130 : @xmath131.\ ] ] these facts are mostly know and we give their proofs only for the reader s convenience . [ distinct ] if @xmath132 for all @xmath133 , then all the eigenvalues of the matrix @xmath134 are distinct . * * under the assumption that all @xmath135 , @xmath136 , are non zero , one can reconstruct every eigenvector @xmath81 of @xmath134 by its eigenvalue @xmath82 and the first coordinate @xmath137 . moreover , the expressions for all other coordinates @xmath138 , @xmath139 , of @xmath81 are linear in @xmath137 . this implies that every two eigenvectors @xmath81 and @xmath140 with the same eigenvalue @xmath82 are proportional with coefficient @xmath141 . [ adj ] non - diagonal elements of the adjoint matrix of @xmath142 are @xmath143 @xmath144 where the first index in @xmath145 denotes the row number , the second denotes the column number and @xmath146 . the diagonal elements of the adjoint matrix are @xmath147 * proof . * the first part follows from an observation that element @xmath148 situated in row @xmath149 and column @xmath150 of matrix @xmath142 does not contribute to the cofactor @xmath151 of @xmath142 , and similarly for element @xmath152 situated in row @xmath25 and column @xmath153 . the second part is due to the splitting of the complement to the element @xmath154 of @xmath142 into @xmath155 and @xmath156 . [ inverse ] if we denote elements of @xmath157 by @xmath158 ( @xmath150 row , @xmath25 column ) , then for @xmath146 @xmath159 the diagonal terms @xmath160 are @xmath161 if @xmath162 , then matrices @xmath163 and @xmath164 ( @xmath165 ) have no common eigenvalues . * first , observe that it is true for @xmath112 : @xmath166 and if @xmath167 is a root of @xmath168 , then @xmath169 , which is the root of @xmath170 only if @xmath171 . for general @xmath1 , if @xmath167 is a root of @xmath172 then for such @xmath167 : @xmath173 @xmath174 for the last polynomial we can assume that it is non zero by inductive hypothesis and because @xmath162 . from here the statement follows . to demonstrate how the reduction over the period works , first we study a partial case of the main result , namely a chain of period 3 with @xmath7 sites . the reduction of section 3 leads us to the following system of linear equations : @xmath175 \left[\begin{array}{c } u_{(1)}\\ u_{(2)}\\ u_{(3)}\\ \end{array}\right]=0.\ ] ] remind that @xmath176 , @xmath177 , are just scalar matrices , as well as the matrix @xmath178 , while @xmath179 are given below : @xmath180 , \qquad { \cal d}_3=\left[\begin{array}{ccccccc } 0 & d_3 & 0 & \cdots & 0&0&0\\ 0 & 0 & d_3 & \cdots & 0&0&0\\ 0 & 0 & 0 & \ddots & 0&0&0\\ \vdots&\vdots & \vdots&\ddots&\ddots&\vdots&\vdots\\ 0&0&0&\cdots&0 & d_3 & 0\\ 0&0&0&\cdots&0 & 0 & d_3\\ \end{array}\right].\ ] ] from the second equation of ( [ 3n+2 ] ) @xmath181 assume that an eigenvalue @xmath82 of @xmath80 is not equal to @xmath182 , then @xmath183 substituting @xmath184 into the first equation and into the third equation of ( [ 3n+2 ] ) , and using that @xmath185 we obtain a system : @xmath186 [ ass ] if @xmath187 then the following inequality holds @xmath188 * proof . * observe , that if @xmath189 , then @xmath190 and @xmath191 therefore , it is left to consider the case when @xmath192 . assume that @xmath193 then from the second equation of ( [ sys1 ] ) @xmath194 with respect to the standard scalar product @xmath195 , @xmath196 , we have : @xmath197 therefore , in order to satisfy the first equation of ( [ sys1 ] ) we necessarily have @xmath198 from here and under the assumption that @xmath199 we obtain that @xmath200 . because @xmath201 , @xmath177 , the matrix @xmath202 has rank @xmath4 and its kernel is @xmath203 . therefore , @xmath204 . finally , from ( [ u(2 ) ] ) we obtain that @xmath205 and thus , @xmath82 is not an eigenvalue for @xmath80 . this verifies the lemma . assume as in lemma [ ass ] that @xmath199 then from the first equation of ( [ sys1 ] ) we deduce that @xmath206 substituting ( [ u(1 ) ] ) into ( [ u(2 ) ] ) we obtain @xmath207 due to lemma [ ass ] @xmath208 thus , from the second equation of ( [ sys1 ] ) we have that @xmath209 therefore , because @xmath192 , @xmath210 satisfies @xmath211 @xmath212 this leads us to the following theorem : [ th3n+2 ] for @xmath213 each of the three solutions of the cubic equation @xmath214 @xmath215 is an eigenvalue for @xmath80 . the component @xmath210 of the corresponding eigenvector @xmath81 is @xmath216 the remaining components @xmath217 and @xmath184 are determined uniquely from @xmath218 and @xmath219 where @xmath220 , @xmath221 are given in ( [ d2d3 ] ) . there are two other eigenvalues of @xmath80 which satisfy the following quadratic equation @xmath222 the component @xmath210 of the corresponding eigenvector @xmath81 is necessarily zero , the component @xmath217 spans the ( one - dimensional ) kernel of @xmath223 and the component @xmath184 is @xmath224 all eigenvalues constructed are distinct and they exhaust all @xmath7 distinct eigenvalues for @xmath80 . * proof . * let us consider the case when @xmath182 is an eigenvalue for @xmath80 and @xmath81 is the corresponding eigenvector . then for the components @xmath217 and @xmath210 we have from the second equation of ( [ 3n+2 ] ) : @xmath225 from the first equation of ( [ 3n+2 ] ) : @xmath226 substituting this into the third equation of ( [ 3n+2 ] ) we obtain : @xmath227 @xmath228 thus , from lemma [ jlambda ] @xmath229 for some @xmath213 . from here it follows that @xmath230 also satisfies eq . ( [ main ] ) , moreover , the formulae for the component @xmath210 and ( as a consequence ) for @xmath217 and @xmath184 coincide with those given in the body of the theorem . we obtain that the case @xmath189 can be considered simultaneously with all the other solutions of ( [ main ] ) . therefore , every eigenvalue for @xmath80 is either a solution for @xmath231 or it is a solution of ( [ main ] ) for some @xmath213 . because @xmath80 has exactly @xmath7 eigenvalues and all of them are distinct ( see lemma [ distinct ] ) , all the solutions of the quadratic equation and of @xmath4 cubic equations are multiplicity free and pairwise distinct . now the first part of theorem [ th3n+2 ] follows from the remarks above , from ( [ ker ] ) and lemma [ jlambda ] . indeed if ( [ ker ] ) admits a nontrivial solution for some @xmath82 , then either @xmath232 or @xmath233 is an eigenvalue for @xmath234 . the first case has been already discussed , while for the second we apply lemma [ jlambda ] . it follows then that @xmath235 is necessarily of the form @xmath236 for some @xmath213 . as it was explained before , among those @xmath237 eigenvalues constructed there are no repeated . the component @xmath210 of the corresponding eigenvectors are uniquely determined by the eigenvalues due to lemma [ jlambda ] . the other two components @xmath217 and @xmath184 are also uniquely determined by ( [ u(1 ) ] ) and ( [ u(2 ) ] ) respectively . this gives us a unique eigenvector @xmath81 of @xmath80 for each solution of ( [ main ] ) . the second part of the statement follows from ( [ sys1 ] ) because @xmath202 has rank @xmath4 and its kernel is @xmath203 . information on the exact spectrum of the hamiltonian of an open spin chain provides us with the techniques for determining the multi - quantum dynamics in such a system . the mq nmr dynamics of the nuclear spins coupled by the nearest neighbour dipolar interactions was developed in @xcite . the corresponding hamiltonian is @xmath238 the hamiltonian @xmath239 ( [ mq ] ) takes exactly the form of the hamiltonian @xmath80 ( [ theorham ] ) with the larmor frequencies @xmath240 for all the sites , by making use of the unitary transformation @xcite @xmath241 acting on the even sites so that @xmath242 ( eq . ( [ theorham ] ) ; @xmath243 ) . the transformation @xmath244 brings the initial density matrix at the high temperature approximation ( see @xcite for details ) : @xmath245 to the form @xmath246 where we introduce the total polarisation @xmath247 . the liouville von neumann equation ( @xmath248 ) @xmath249\ ] ] with the hamiltonian @xmath239 ( [ mq ] ) and the initial density matrix @xmath250 ( [ ini ] ) gives us the intensities @xmath251 of @xmath252 and @xmath253 orders only with the conservation conditions @xcite @xmath254 [ trace ] the intensity @xmath251 , @xmath255 , of mq coherences of the hamiltonian @xmath239 ( [ mq ] ) are @xmath256}{n},\quad g_{\pm 2}(t)=\frac{tr \left[{\rm { sin}}^2\left(h_{mq}\cdot t\right)\right]}{2n}.\ ] ] * proof . * according to @xcite @xmath257 with @xmath258 where @xmath259 are the eigenvalues of @xmath239 , @xmath36 , and the unitary matrix @xmath260 diagonalises @xmath239 . let us rewrite ( [ gdf ] ) in the matrix form : @xmath261 where @xmath262 is the diagonal matrix with ones in odd rows and zeroes in even rows , @xmath263 is the diagonal matrix with ones in even rows and zeroes in odd rows : @xmath264 and the matrix @xmath265 is @xmath266 observe that @xmath267 therefore , @xmath268=tr\left[\left(s^\right)^m\right].\ ] ] using @xmath269 and @xmath270= tr\left[\left(b_1-b_0\right)a^a\right]= tr\left[\left(b_1-b_0\right)aa^\right]\ ] ] we deduce @xmath271=\frac{1}{4}tr\left[\left(i-\left(b_1-b_0\right)\right)a \left(i+\left(b_1-b_0\right)\right)a^\right]=\ ] ] @xmath272=\ ] ] @xmath273= -\frac{1}{8}tr\left[s^4+\left(s^\right)^4 - 2i_n\right]=\ ] ] @xmath274 = \frac{1}{2}tr\left[{\rm { sin}}^2\left(h_{mq}\cdot t\right)\right].\ ] ] from here the result follows . in @xcite formula ( [ trmq ] ) was proposed for alternating spin chains . theorem [ trace ] establishes the same formula for arbitrary spin chains coupled by the nearest neighbour dipolar interactions . theorem [ trace ] together with theorem [ th3n+2 ] allow us to calculate mq coherence intensities of the zero and the second orders , @xmath275 and @xmath276 , without performing matrix multiplication , by solving @xmath277 cubic equations ( compare with @xcite ) . let us consider a linear spin chain of length @xmath278 which consists of four - spin fragments represented in figure 1 . ( 150,80 ) ( 30,50 ) ( 27,30)@xmath279 ( 38,50)(1,0)50 ( 50,54)@xmath280 ( 96,50 ) ( 94,30)@xmath281 ( 104,50)(1,0)70 ( 125,54)@xmath282 ( 182,50 ) ( 182,30)@xmath283 ( 190,50)(1,0)90 ( 230,54)@xmath284 ( 288,50 ) ( 288,30)@xmath285 ( 296,50)(1,0)50 ( 17,4 ) ( 9,-7 ) ( 10,-18 ) the distances between neighbouring spins in the fragments are @xmath286 , @xmath287 and @xmath288 . the dipolar coupling constants @xmath289 , @xmath290 , @xmath291 are used in all numerical calculations . the intensities of mq coherences for the inhomogeneous linear chain with @xmath292 spins consisting of fragments of figure 1 are shown in figure 2 . the dynamics behaviour coincides with the one given in @xcite . now we consider the case of a @xmath1-periodic system with @xmath8-sites . according to section 3 , to diagonalise hamiltonian of such a system we have to solve the following system of linear equations @xmath293= \lambda_{\nu } \left [ \begin{array}{c } u_{(1)}\\ \vdots\\ u_{(k)}\\ \end{array}\right],\ ] ] where @xmath294,\ ] ] and @xmath295.\ ] ] observe that because all @xmath296 , @xmath297 , are diagonal @xmath110 matrices , it is also true that @xmath298 assume @xmath82 is not an eigenvalue for @xmath299 ( equivalently , is not an eigenvalue for @xmath300 ) , and consider the matrix @xmath301.\ ] ] under the above assumption on @xmath82 the eigenvector equation @xmath302=\lambda_{\nu}\left[\begin{array}{c } u_{(1)}\\ \vdots\\ u_{(k)}\\ \end{array}\right]\ ] ] is equivalent to @xmath303=0,\ ] ] or in another form to @xmath304\right ] \left[\begin{array}{c } u_{(1)}\\ u_{(2)}\\ \vdots\\ u_{(k-2)}\\ u_{(k-1)}\\ u_{(k)}\\ \end{array}\right]=0.\ ] ] let us denote elements of the matrix @xmath305 by @xmath306 ( @xmath307 row , @xmath89 column ) . using @xmath308 we derive from ( [ trinew ] ) for @xmath309 that @xmath310 substituting this expression with @xmath311 and @xmath312 into the last equation of ( [ trinew ] ) we obtain : @xmath313 @xmath314 because @xmath315 @xmath316 with the help of corollary [ inverse ] we derive that if @xmath82 is not an eigenvalue of @xmath300 , then the component @xmath317 of the corresponding eigenvector @xmath81 belongs to the kernel of @xmath318 @xmath319 since @xmath320 we can simplify the expression for @xmath321 : @xmath322 @xmath323 therefore , @xmath324 @xmath325 we are ready to state the main result of the paper . [ mainkn-1 ] each eigenvalue of the hamiltonian @xmath80 of a @xmath1-periodic system with @xmath8 sites is either an eigenvalue of @xmath300 or it is a solution of the equation @xmath326 for some @xmath327 . equation ( [ equ ] ) does not have repeated roots and all @xmath328 solutions constructed from ( [ equ ] ) are pairwise distinct and are not eigenvalues for @xmath300 . if @xmath82 is the solution of ( [ equ ] ) for some @xmath327 , then it is an eigenvalue for @xmath80 and the component @xmath317 of the corresponding eigenvector @xmath81 is @xmath329 the other components @xmath86 , @xmath330 , are determined uniquely from @xmath331 @xmath332u_{(k)},\ ] ] where @xmath333 , @xmath334 are given in ( [ dddk ] ) ( @xmath128 ) . every eigenvalue @xmath82 of @xmath300 is an eigenvalue of @xmath80 . the component @xmath317 of the corresponding eigenvector @xmath81 of @xmath80 is zero . the component @xmath217 spans the one dimensional kernel of @xmath335 the remaining components @xmath86 , @xmath336 , are @xmath337 proof . * the first part of the theorem follows from lemma [ jlambda ] because @xmath80 has exactly @xmath8 distinct eigenvalues . indeed , if an eigenvalue @xmath82 of @xmath80 is not an eigenvalue of @xmath300 , then as it was shown above , the component @xmath317 of the corresponding eigenvector @xmath81 of @xmath80 satisfies ( [ maineq ] ) . therefore , @xmath82 satisfies ( [ equ ] ) for some @xmath327 . for the second part of the theorem we again use lemma [ jlambda ] . the component @xmath317 of the corresponding eigenvector @xmath81 , satisfies ( [ maineq ] ) , and , therefore , is uniquely determined by @xmath82 as stated in lemma [ jlambda ] . the remaining components of the eigenvectors @xmath338 corresponding to the solutions of ( [ equ ] ) can be reconstructed from ( [ newu_j ] ) using corollary [ inverse ] . finally , for the last part we observe that from lemma [ adj ] and the property of the adjoint matrix @xmath339 @xmath340 @xmath341 @xmath342 @xmath343 where @xmath344 and @xmath345 is the kronecker symbol . we also have @xmath346 if @xmath82 is an eigenvalue for @xmath300 , then @xmath347 therefore , if @xmath217 spans the kernel of @xmath348 and @xmath349 , then vector @xmath81 with the other components given by ( [ theother ] ) does satisfy the eigenvalue equation @xmath350 this completes the proof . in the paper we proposed a new one - dimensional exactly solvable model for a linear @xmath1-periodic ( in space ) open spin chain with @xmath8 sites . for the diagonalisation procedure it is important that the number of sites is @xmath351 . nevertheless , the model can serve as a good approximation for any linear periodic spin system if the number of sites in it is much more than the period . the developed method of diagonalisation of the xy - hamiltonian of inhomogeneous linear spin chains can be applied to different problems of quantum information theory [ 3,7,8 ] and spin dynamics . this method allows us to avoid matrix multiplications which are time consuming operations in the systems with large numbers of spins . in some cases we can suggest analytical methods for problems of spin dynamics instead of the known numerical ones [ 9 ] . the suggested method could also be applied to different physical and technical problems which use three diagonal matrices @xcite . in particular , new numerical methods of solving such problems could be worked out on the basis of the approach proposed in the paper . the author has been supported by epsrc grant gr / s92137/01 .	a new method of diagonalisation of the xy - hamiltonian of inhomogeneous open linear chains with periodic ( in space ) changing larmor frequencies and coupling constants is developed . as an example of application , multiple quantum dynamics of an inhomogeneous chain consisting of 1001 spins is investigated . intensities of multiple quantum coherences are calculated for arbitrary inhomogeneous chains in the approximation of the next nearest interactions . _ key words : _ linear spin chain , nearest neighbour approximation , three diagonal matrices , diagonalisation , fermions , multiple quantum nmr , multiple quantum coherence , intensities of multiple quantum coherences . _ pacs numbers : _ 05.30.-d ; 76.20.+q [ section ] [ section ] [ section ] [ section ] [ section ] [ section ] [ section ]
the camassa holm equation was introduced in @xcite in the form @xmath1 and was intensively explored afterwards ( see , for example , refs . its superizations were also constructed , see @xcite . since is not an evolution equation , its integrability properties ( existence and even definition of hamiltonian structures , conservation laws , etc . ) are not standard to establish . one of the ways widely used to overcome this difficulty is to introduce a new unknown @xmath2 and transform eq . to the system @xmath3 which has _ almost _ evolutionary form . we stress this `` almost '' , because the second equation in ( that can be considered as a constrain to the first one ) disrupts the picture and , at best , necessitates to invert the operator @xmath4 . at worst , dealing with eq . as with an evolution equation may lead to fallacious results . in our approach based on the geometrical framework exposed in ref . @xcite , we treat the equation at hand as a submanifold in the manifold of infinite jets and consider two natural extensions of this equation , cf . with ref @xcite . the first one is called the _ @xmath5_-covering and serves the role of the tangent bundle . the second extension , @xmath6-covering , is the counterpart to the cotangent bundle . the key property of these extensions is that the spaces of their nonlocal ( in the sense of @xcite ) symmetries and cosymmetries contain all essential integrability invariants of the initial equation . the efficiency of the method was tested for a number of problems ( see refs . @xcite ) and we apply it to the camassa holm equation here . in section [ sec : underlying - theory ] we briefly expose the necessary definition and facts . section [ sec : matrix - version ] contains computations for the camassa holm equation in its matrix version ( computations and results are more compact in this representation ) , while in section [ sec : scalar - version ] we reformulate them for the original form and compare later the results obtained for the two alternative presentations . finally , section [ sec : conclusion ] contains discussion of the results obtained . throughout our exposition we use a very stimulating conceptual parallel between categories of smooth manifolds and differential equations proposed initially by a.m. vinogradov and in its modern form presented in table [ tab:1 ] . l\|l * manifolds * & * * equation*s + smooth manifold & infinitely prolonged equation + point & formal solution + smooth function & conservation law + vector field & higher symmetry + differential @xmath7-form & cosymmetry + de rham complex & @xmath8st line of vinogradov s @xmath9-spectral sequence + tangent bundle & @xmath5-covering + cotangent bundle & @xmath6-covering + this table is not just a toy dictionary but a quite helpful tool to formulate important definitions and results . for example , a bivector on a smooth manifold @xmath10 may be understood as a derivation of the ring @xmath11 with values in @xmath12 . translating this statement to the language of differential equations we come to the definition of variational bivectors and their description as shadows of symmetries in the @xmath6-covering ( see theorem [ thm : ellstar - shad ] below ) . another example : any vector field ( differential @xmath7-form ) on @xmath10 may be treated as a function on @xmath13 ( on @xmath14 ) . hence , to any symmetry ( cosymmetry ) there corresponds a conservation law on the space of the @xmath6-covering ( @xmath5-covering ) . this leads to the notions of nonlocal vectors and forms that , in turn , provide a basis to construct weakly nonlocal structures ( see subsections [ sec : nonlocal - forms-1 ] and [ sec : nonlocal - vectors-1 ] ) . of course , these parallels are not completely straightforward ( in technical aspects , especially ) , but extremely enlightening and fruitful . the idea of this paper arose in the discussions one of the authors had with volodya roubtsov in 2007 . we agreed to write two parallel texts on integrability of the camassa holm equation that reflect our viewpoints . the reader can now compare our results with the ones presented in @xcite . we present here a concise exposition of the theoretical background used in the subsequent sections , see refs . @xcite . let @xmath15 be a fiber bundle and @xmath16 be the bundle of its infinite jets . to simplify our exposition we shall assume that @xmath17 is a vector bundle . in all applications below @xmath17 is the trivial bundle @xmath18 . we consider _ infinite prolongations _ of differential equations as submanifolds @xmath19 and retain the notation @xmath20 for the restriction @xmath21 . any such a manifold is endowed with the _ cartan distribution _ which spans at every point tangent spaces to the graphs of jets . a _ symmetry _ of @xmath22 is a vector field that preserves this distribution . the set of symmetries is a lie algebra over @xmath23 denoted by @xmath24 . for any equation @xmath22 its _ linearization operator _ @xmath25 is defined , where @xmath26 is the module of smooth functions on @xmath22 . ] of sections of the pullback @xmath27 and @xmath28 is the module of sections of some vector bundle over @xmath22 . then @xmath24 can be identified with solutions of the equation @xmath29 for two symmetries @xmath30 , @xmath31 their commutator is denoted by @xmath32 . denote by @xmath33 the module of horizontal @xmath34-forms on @xmath22 and introduce the notation @xmath35 for any module @xmath36 . adjoint _ to @xmath37 operator @xmath38 arises and solutions of the equation @xmath39 are called _ cosymmetries _ of @xmath22 ; the space of cosymmetries is denoted by @xmath40 . let @xmath41 be the _ horizontal de rham differential_. a _ conservation law _ of the equation @xmath22 is a closed form @xmath42 . to any conservation law there corresponds its _ generating function _ @xmath43 , where @xmath44 is the differential in the @xmath45 term of vinogradov s @xmath9-spectral sequence , see @xcite . in the evolutionary case @xmath46 coincides with the euler lagrange operator . a conservation law is _ trivial _ if its generating function vanishes . in particular , @xmath47-exact conservation laws are trivial . a vector field on @xmath22 is called a _ @xmath9-field _ if it lies in the cartan distribution . a differential operator @xmath48 , @xmath28 and @xmath36 being @xmath49-modules , is called a _ @xmath9-differential operator _ if it is locally expressed in terms of @xmath9-fields . for example , @xmath37 is a @xmath9-differential operator . a @xmath9-differential operator @xmath50 is said to be a _ variational bivector _ on @xmath22 if @xmath51 this condition means that @xmath52 takes cosymmetries of @xmath22 to symmetries . if @xmath22 is an evolution eq . implies also that @xmath53 . a bivector @xmath52 is a _ _ hamiltonian structure _ on @xmath22 if @xmath54\!]}^{\mathrm{s}}=0 $ ] , where @xmath55\!]}^{\mathrm{s}}$ ] is the _ variational schouten bracket _ ( see also @xcite ) . two hamiltonian structures are _ compatible _ if @xmath56\!]}^{\mathrm{s}}=0 $ ] . a @xmath9-differential operator @xmath57 is called a _ variational @xmath58-form on @xmath22 _ if @xmath59 such operators take symmetries to cosymmetries and in evolutionary case are skew - adjoint . they are elements of the term @xmath60 of vinogradov s @xmath9-spectral sequence . a variational form is a _ symplectic structure _ on the equation @xmath22 if it is _ variationally closed _ , i.e. , @xmath61 , where @xmath62 is the corresponding differential . we shall also consider _ recursion @xmath9-differential operators _ @xmath63 and @xmath64 satisfying the conditions @xmath65 for some @xmath9-differential operators @xmath66 and @xmath67 . an operator @xmath68 satisfies the _ nijenhuis condition _ if @xmath69\!]}^{\mathrm{n}}=0 $ ] , where @xmath55\!]}^{\mathrm{n}}$ ] is the _ variational nijenhuis bracket_. a recursion operator @xmath68 is _ compatible _ with a hamiltonian structure @xmath52 is the composition @xmath70 is a hamiltonian structure as well . let @xmath22 and @xmath71 be equations and @xmath72 be a fiber bundle . denote by @xmath9 and @xmath73 the cartan distributions on @xmath22 and @xmath71 , resp . we say that @xmath74 is a _ covering _ if for any @xmath75 the differential @xmath76 isomorphically maps @xmath77 onto @xmath78 . a particular case of coverings ( the so - called _ abelian _ coverings ) is naturally associated with closed horizontal @xmath7-formswhen @xmath79 , abelian coverings are associated with conservation laws of the equation @xmath22 . ] . by definition , any @xmath9-field @xmath80 on @xmath22 can be uniquely lifted to a @xmath9-field @xmath81 on @xmath71 such that @xmath82 . consequently , any @xmath9-differential operator @xmath83 is extended to a @xmath9-differential operator @xmath84 @xmath85 being the algebra of smooth functions on @xmath71 . a _ nonlocal _ @xmath74-(co)symmetry of @xmath22 is a ( co)symmetry of the covering equation @xmath71 . they are solutions of the equations @xmath86 and @xmath87 , resp . along with these two equations one can consider the equations @xmath88 their solutions are called _ @xmath74-shadows _ of symmetries and cosymmetries , resp . a shadow of symmetry is a derivation @xmath89 that preserves the cartan distributions . for any two shadows of symmetries @xmath30 and @xmath90 their commutator @xmath32 can be defined . this commutator is a shadow in a new covering that is canonically determined by @xmath30 and @xmath90 . let @xmath19 be an equation . its _ @xmath5-covering _ @xmath91 is obtained by adding to @xmath22 the equation @xmath92 , where @xmath93 is a new variable . dually , the _ @xmath6-covering _ @xmath94 is constructed by adding the equation @xmath95 with a new variable @xmath96 . they are the exact counterparts of the tangent and cotangent bundles in the category of differential equations . by the reasons that will become clear later , we regard both @xmath93 and @xmath96 as _ odd _ variables . the main point of our method is the fundamental relation between integrability invariants of @xmath22 and shadows in @xmath97 and @xmath98 . to formulate this relation , let us give an auxiliary definition : for an arbitrary operator equation @xmath99 we say that @xmath100 is a _ trivial solutions _ if @xmath100 is of the form @xmath101 . classes of solutions modulo trivial ones will be called _ nontrivial solutions_. then the following results hold . [ thm : ell - shad ] there is a one - to - one correspondence between nontrivial solutions of the equation @xmath102 and @xmath97-shadows of symmetries linear w.r.t . the variables @xmath93 . in a similar way there is a one - to - one correspondence between nontrivial solutions of the equation @xmath103 and @xmath97-shadows of cosymmetries linear w.r.t . the variables @xmath93 . [ thm : ellstar - shad ] there is a one - to - one correspondence between nontrivial solutions of the equation @xmath104 and @xmath98-shadows of symmetries linear w.r.t . the variables @xmath96 . in a similar way there is a one - to - one correspondence between nontrivial solutions of the equation @xmath105 and @xmath98-shadows of cosymmetries linear w.r.t . the variables @xmath96 . [ thm : brackets ] let @xmath106 and @xmath107 be recursion operators for symmetries on @xmath22 and @xmath108 denote the @xmath97-shadow corresponding to @xmath68 . then @xmath109\!]}^{\mathrm{n}}=0 $ ] iff @xmath110 . similarly if @xmath111 and @xmath112 are bivectors then @xmath56\!]}^{\mathrm{s}}=0 $ ] iff @xmath113 where @xmath114 denotes the @xmath98-shadow corresponding to @xmath52 . in both cases the curly brackets denote the _ super _ bracket of shadows that arises due to oddness of the variables @xmath93 and @xmath96 . additional discussion of theorem [ thm : brackets ] the reader will find in remark [ rem : nonlocal ] . [ thm : correpondence ] to any cosymmetry of @xmath22 there canonically corresponds a conservation law of @xmath115 . dually to any symmetry of @xmath22 there canonically corresponds a conservation law of @xmath116 . let locally the equation @xmath22 be given by the system @xmath117 where @xmath118 and @xmath119 . [ [ sec : step-1 ] ] step 1 + + + + + + consists of writing out defining equations for symmetries and cosymmetries of @xmath22 . let @xmath120 be _ internal coordinates _ on @xmath22 , @xmath121 and @xmath122 being some sets of ( multi)indices and @xmath123 corresponding to @xmath124 . then any @xmath9-field on @xmath22 is a linear combination of the _ total derivatives _ @xmath125 the linearization of @xmath22 is the matrix operator with the entries @xmath126 a symmetry @xmath127 enjoys the equation @xmath128 and the corresponding field is the _ evolutionary vector field _ @xmath129 while the bracket of symmetries @xmath30 , @xmath90 is given by @xmath130 the operator adjoint to is @xmath131 and a cosymmetry @xmath132 satisfies the equation @xmath133 [ [ sec : step-2 ] ] step 2 . + + + + + + + here we look for closed @xmath7-forms and construct abelian coverings associated to them . a horizontal form @xmath134 is closed if @xmath135 such a form gives rise to a nonlocal variable @xmath136 that satisfies the equations @xmath137 these equations are compatible on due to . recall that for @xmath138 closed @xmath7-forms coincide with conservation laws . the total derivatives lifted to the covering equation @xmath71 are @xmath139 [ [ sec : step-3 ] ] step 3 . + + + + + + + at this step we compute a number of particular symmetries and cosymmetries ( using equations and , resp . ) . they are used to construct canonical nonlocal variables on the @xmath6-covering ( _ nonlocal vectors _ ) and on the @xmath5-covering ( _ nonlocal forms _ ) , resp . , at step 4 . we also use them as seed elements in series generated by recursion operators . [ [ sec : step-4 ] ] step 4 + + + + + + consists of construction of the @xmath5- and @xmath6-coverings and introduction of canonical nonlocal variables over them ( see step 3 ) . the @xmath5-covering is obtained by adding to the system of equations @xmath140 cf . with eq . , while the @xmath6-covering is given by @xmath141 that comes from . if @xmath142 is a symmetry of @xmath22 then one can introduce a covering over @xmath116 described by the system @xmath143 where @xmath144 are @xmath9-differential operators ( see theorem [ thm : correpondence ] ) . in a similar way , to any cosymmetry @xmath145 there corresponds a covering @xmath146 @xmath147 being @xmath9-differential operators as well . we omit here a general description of these operators and refer the reader to the particular case of our interest exposed in sections [ sec : nonlocal - forms ] and [ sec : nonlocal - vectors ] . [ [ sec : step-5 ] ] step 5 . + + + + + + + we now use theorems [ thm : ell - shad ] and [ thm : ellstar - shad ] to construct recursion operators and hamiltonian and symplectic structures . let @xmath148 cosymmetries of @xmath22 . let us consider the covering @xmath149 over @xmath115 with the nonlocal variables @xmath150 defined by and lift the operators @xmath37 and @xmath151 to this covering . then the following result specifies theorem [ thm : ell - shad ] : [ thm : nloc - ell ] let @xmath152 be a solution of the equation @xmath153 on @xmath149 linear w.r.t . @xmath154 and @xmath155 @xmath156 then the operator @xmath157 takes shadow of symmetries to shadows of symmetries . in a similar way to any solution @xmath158 @xmath159 there corresponds the operator @xmath160 that takes shadows of symmetries to shadows of cosymmetries . in a dual way , consider symmetries @xmath161 of the equation @xmath22 and the covering @xmath162 over @xmath116 with the nonlocal variables @xmath163 defined by . then , lifting @xmath37 and @xmath151 , we obtain a similar specification of theorem [ thm : ellstar - shad ] : [ thm : nloc - ellstar ] let @xmath152 be a solution of the equation @xmath153 on @xmath162 linear w.r.t . @xmath164 and @xmath165 @xmath166 then the operator @xmath167 takes shadow of cosymmetries to shadows of symmetries . in a similar way to any solution @xmath158 @xmath168 there corresponds the operator @xmath169 that takes shadows of cosymmetries to shadows of cosymmetries . after finding the operators @xmath68 , @xmath170 , @xmath52 and @xmath171 we check conditions and and compute necessary schouten and nijenhuis brackets . [ [ sec : step-6 ] ] step 6 + + + + + + the last step consists of establishing algebraic relations between the invariants constructed above . we consider eq . in the form @xmath172 i.e. , set @xmath173 , and , similar to , introduce a new variable @xmath174 , where @xmath175 is a new real constant . consequently , the initial equation transforms to the system @xmath176 we choose the following variables for internal local coordinates on the infinite prolongation of eq . : @xmath177 then the total derivatives in these coordinates will be of the form @xmath178 we introduce the following gradings : @xmath179 and extend them in a natural way to all polynomial functions of the internal coordinates . then all computations can be restricted to homogeneous components . in subsequent computations we shall need the following nonlocal variables arising from conservation laws and defined by the equations @xmath180 the variable @xmath181 is of grading @xmath34 and computational experiment shows that for every grading @xmath182 , @xmath183 , there exist an @xmath181 such that @xmath184 . in addition , we found conservation laws of fractional gradings : @xmath185 etc . a symmetry @xmath186 of eq . must satisfy the linearized equation @xmath187 direct computations lead to the following results . [ [ sec : x - t - independent ] ] ( x , t)-independent symmetries . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + one can observe two types of symmetries that are independent of @xmath188 and @xmath189 . the first one consists of symmetries of integer gradings : @xmath190 etc . symmetries of the second type have semi - integer gradings : @xmath191 etc . all symmetries are local and @xmath192 . if one adds to the nonlocal setting the variables @xmath193 ( see above ) then an additional series of nonlocal symmetries arises : @xmath194 etc . [ [ sec : x - t - dependent ] ] ( x , t)-dependent symmetries . + + + + + + + + + + + + + + + + + + + + + + + + + + + the first three symmetries that depend on @xmath188 and @xmath189 are @xmath195 all these symmetries , except for the first one , are nonlocal ( description of the nonlocal variable is given in subsection [ sec : nonlocal - variables-1 ] ) and , as above , the subscript denotes the grading . the defining equation for cosymmetries @xmath196 is the adjoint to the linearization of : @xmath197 similar to symmetries , we consider two types of cosymmetries . [ [ sec : x - t - independent-1 ] ] ( x , t)-independent cosymmetries . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + they are local and may be of integer and semi - integer gradings : @xmath198 etc . similar to the case of symmetries , when one adds nonlocal variables @xmath193 an additional series of nonlocal cosymmetries arises : @xmath199 etc . [ [ sec : x - t - dependent-1 ] ] ( x , t)-dependent cosymmetries . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + all them are nonlocal : @xmath200 etc . recall that nonlocal forms are nonlocal variables of a special type on the @xmath5-covering . the @xmath5-covering itself is obtained from eq . by adding two additional equations @xmath201 where @xmath202 and @xmath203 are new odd variables . the total derivatives on the @xmath5-covering are @xmath204 the nonlocal form @xmath205 associated to a cosymmetry @xmath206 ( see subsection [ sec : cosymmetries-1 ] ) is defined by the equations @xmath207 dually to nonlocal forms , nonlocal vectors arise as special nonlocal variables on the @xmath6-covering associated to symmetries of the initial equation . the @xmath6-covering is the extension of eq . by two new equations @xmath208 where @xmath209 is a new odd variable . the total deriavatives are given by @xmath210 the nonlocal vector @xmath211 associated to a symmetry @xmath186 ( see subsection [ sec : symmetries-1 ] ) is defined by the equations @xmath212 the defining equations for these operators are @xmath213 ( see theorem [ thm : nloc - ell ] ) , where the total derivatives are those described in subsection [ sec : nonlocal - forms-1 ] . the following two solutions are essential : @xmath214 the corresponding operators are of the form @xmath215 where @xmath216 , @xmath217 . all other solutions obtained in our computations corresponded to operators that are generated by the two above . symplectic structures , as it follows from theorem [ thm : nloc - ell ] , are defined by the equations @xmath218 where the total derivatives were defined in subsection [ sec : nonlocal - forms-1 ] . here are the simplest nontrivial solutions : @xmath219 with the corresponding symplectic operators @xmath220 the equations that should be satisfied by a hamiltonian operator ( see theorem [ thm : nloc - ellstar ] ) are @xmath221 where the total derivatives are from subsection [ sec : nonlocal - vectors-1 ] . in particular , we found the following solutions : @xmath222 the corresponding hamiltonian operators are @xmath223 by theorem [ thm : nloc - ellstar ] , the equation to find recursion operators for cosymmetries are @xmath224 with the total derivatives given in subsection [ sec : nonlocal - vectors-1 ] . one of solutions is presented below : @xmath225 the corresponding recursion operator is @xmath226 we expose here basic facts on structural relations between the above described invariants . the main one is the following [ thm : p - n ] recursion operator @xmath227 and hamiltonian operator @xmath228 constitute a poisson nijenhuis structure on the camassa holm equation . consequently all the operators @xmath229 are hamiltonian and pair - wise compatible . in particular @xmath230 . the proof consists of direct computations using the results and techniques of ref . @xcite is also a poisson nijenhuis structure and generates an infinite negative hierarchy of hamiltonian operators , but could not prove this fact because of computer capacity limitations . ] . a visual presentation of how symmetries are distributed over gradings is given in table [ tab:2 ] . l\|ccccccccccccccccc gradings & @xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236&@xmath7&@xmath58&@xmath237&@xmath238&@xmath239&@xmath240&@xmath241&@xmath242 + local positive & & & & & & & @xmath243 & @xmath243 & & & @xmath243&@xmath243 + local negative & @xmath243 & & @xmath243 & & & & & & + nonlocal positive & & & & & & @xmath243 & & & @xmath243 & @xmath243 & & & @xmath243 & @xmath243 + nonlocal negative & & @xmath243 & & @xmath243 & & & & & + how to prove locality of the first two series of symmetries will be discussed in section [ sec : conclusion ] . similar presentation for cosymmetries see in table [ tab:3 ] . l\|ccccccccccccccccc gradings & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath244 & @xmath7&@xmath245&@xmath58&@xmath237&@xmath238&@xmath239&@xmath240&@xmath241&@xmath242 + local positive & & & & & & & & & @xmath243 & @xmath243&&&@xmath243&@xmath243 + local negative & @xmath243 & & @xmath243 & & @xmath243 & & @xmath243 & & + nonlocal positive & & & & & & & & @xmath243 & & & @xmath243 & @xmath243 & & + nonlocal negative & & @xmath243 & & @xmath243 & & @xmath243 & & & + the action of hamiltonian and recursion operators for symmetries ( up to a constant multiplier ) is given in diagram : @xmath246_{\tens{r}_3}&\phi_{-5/2}\ar[r]_{\tens{r}_3}\ar@/{_}/[l]_{\tens{r}_{-1 } } & \phi_{-3/2}\ar[r]_{\tens{r}_3}\ar@/{_}/[l]_{\tens{r}_{-1}}&0 & \phi_1\ar[r]_{\tens{r}_3}\ar@/{_}/[l]_{\tens{r}_{-1 } } & \phi_2\ar[r]_{\tens{r}_3}\ar@/{_}/[l]_{\tens{r}_{-1 } } & \dots\ar@/{_}/[l]_{\tens{r}_{-1}}\\ \dots&\psi_{1/2}\ar[ur]^{\tens{h}_{-2}}\ar[u]^(.3){\tens{h}_{-3 } } & \psi_{3/2}\ar[ur]^{\tens{h}_{-2}}\ar[u]^(.3){\tens{h}_{-3 } } & \psi_3\ar[ur]^{\tens{h}_{-2}}\ar[u]^(.3){\tens{h}_{-3 } } & \psi_4\ar[ur]^{\tens{h}_{-2}}\ar[u]^(.3){\tens{h}_{-3 } } & \psi_7\ar[ur]^{\tens{h}_{-2}}\ar[u]^(.3){\tens{h}_{-3 } } & \dots } \ ] ] the action of recursion operators for cosymmetries and simplectic structures is similar . we shall now prove commutativity of the local hierarchies . [ lem : action ] the symmetry @xmath247 is a positive hereditary symmetry i.e. its action on local symmetries @xmath248 coincides up to a multiplier with the one of the recursion operator @xmath227 . the only symmetries that vanish under this action are @xmath249 and @xmath30 . in a similar way the symmetry @xmath250 is a negative hereditary symmetry and the only symmetry that is taken to zero under its action is @xmath30 . a direct corollary of this result is [ thm : commut ] local positive and negative symmetries form commutative hierarchies . let us consider now the camassa holm equation in its initial form with @xmath173 and , similar to the matrix case introduce a new real parameter @xmath175 : @xmath251 for the internal coordinates we choose the functions @xmath252 the total derivatives in these coordinates are of the form @xmath253 where @xmath254 . the equation becomes homogeneous if assign the following gradings : @xmath255 we introduce nonlocal variable associated to conservation laws of the equation at hand : @xmath256 a symmetry @xmath142 must satisfy the linearized equation @xmath257 we computed two types of symmetries . everywhere below the subscript was chosen in a way to correspond the enumeration taken for the matrix case . [ [ sec : x - t - independent-2 ] ] @xmath258-independent symmetries . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we present the first four of them : @xmath259 all these symmetries are local . [ [ sec : x - t - dependent-2 ] ] @xmath258-dependent symmetries . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the first three @xmath258-dependent symmetries are @xmath260 the only local symmetry in this series is @xmath30 . the defining equation for cosymmetries is @xmath261 [ [ sec : x - t - independent-3 ] ] @xmath258-independent cosymmetries . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + these cosymmetries are local . the first four of them are @xmath262 [ [ sec : x - t - dependent-3 ] ] @xmath258-dependent cosymmetries . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + these cosymmetries are nonlocal , except for the first one : @xmath263 etc . nonlocal forms arise in the @xmath5-covering which is given by the equation @xmath264 with the total derivatives @xmath265 where @xmath266 . to any cosymmetry @xmath267 we associate a nonlocal form @xmath268 defined by the equations @xmath269 nonlocal vectors arise in the @xmath6-covering . the latter is the extension of the initial equation by the equation @xmath270 with the total derivatives @xmath271 where @xmath272 . to any symmetry @xmath273 there corresponds a nonlocal vector @xmath274 defined by the relations @xmath275 the defining equation for these operators is @xmath276 where the total derivatives are those presented in subsection [ sec : nonlocal - forms ] . we consider here two nontrivial solutions : @xmath277 the operator corresponding to the second one is @xmath278 the first operator is too complicated to present it here . the equation defining symplectic structures is @xmath279 where the total derivatives are from subsection [ sec : nonlocal - forms ] . the simplest solutions are @xmath280 while the corresponding operators have the form @xmath281 hamiltonian structures are defined by the equation @xmath282 where the total derivatives are given in subsection [ sec : nonlocal - vectors ] . we present here three solutions . two of them are local and the third one is nonlocal : @xmath283 the corresponding operators are @xmath284 these recursion operators are defined by the equation @xmath285 where the total derivatives are given in subsection [ sec : nonlocal - vectors ] . one of the solutions is @xmath286 to which the operator @xmath287 corresponds . we finish this paper with a number of remarks . [ rem : what - has - been ] first of all , let us stress again what has been done above . we treated the camassa holm equation directly , without artificial assumptions about its evolution or pseudo - evolution nature . the pass from the original form to system ( which is _ not _ in evolution form ) was done by technical reasons only . on this way , we found an infinite family of pair - wise compatible hamiltonian structures , recursion operators for symmetries and cosymmetries , and symplectic operators . these structures lead to existence of two commutative series of local symmetries and conservation laws and thus the equation is integrable . [ rem : nonlocal ] several comments are worth to be made in relation to theorem [ thm : brackets ] . two cases must be distinguished : local and nonlocal . in both cases the problem reduces to _ reconstruction _ of nonlocal shadows up to symmetries and subsequent computation of the bracket . this case is simple and the result is as follows : any shadow in the @xmath5-covering corresponding to a recursion operator can be canonically lifted up to a symmetry of the covering equation and the commutator of these symmetries corresponds to the nijenhuis bracket of the operators . a similar result is valid for the shadows in the @xmath6-covering corresponding to local bivectors ( in this case the commutator of symmetries is related to the schouten bracket ) . the case of nonlocal operators is much more complicated and rests on the problem of how to commute nonlocal shadows . seemingly there exists no _ natural _ definition of such a commutator , but in ref . @xcite we proposed a procedure both to reconstruct and commute shadows . a weak point of this procedure is that it is based on rather cumbersome and intuitively not obvious notion of shadow equivalence . dealing with equivalence classes of shadows necessitates elimination of certain kind of nonlocal variables that we call _ pseudo - constants _ and that are intrinsically related to the covering at hand . this not always simple to do and we plan to simplify and clarify the procedure . [ rem : comp - two - repr ] of course , all invariants of the camassa holm equation in its initial form can be obtained from those computed for the system by simple transformations . e.g. , to obtain symmetries of eq . from those of one should substitute in the @xmath288-component the variables @xmath289 by @xmath290 . on the other way , any cosymmetry in the scalar case can be constructed directly from the corresponding cosymmetry of eq . by changing the variables @xmath289 in the @xmath291-component by @xmath290 . similar transformations are applicable to other structures . the following correspondence between the components of symmetries and cosymmetries of matrix equation is valid @xmath292^{d_t+ud_x - u_1}&\psi_k^u\ar@{\|-{>}}[r]^{\cdot(-1 ) } & \phi_{k-2}^u\ar@{\|-{>}}[r]^{\alpha - d_x^2}&\phi_{k-2}^w . } \ ] ] [ rem : local - ov - hier ] now we shall show how to prove locality of symmetry hierarchies . let us introduce a new variable @xmath293 . then eq . transforms to @xmath294 i.e. , the first equation acquires potential form , see above . ] . this means that for any symmetry @xmath295 the form @xmath296 is a conservation law of eq . . comparing gradings , it is easily checked that all conservation laws of the form @xmath297 are trivial and consequently @xmath298 lies in the image of @xmath299 for any symmetry @xmath142 . from this fact it follows that the action of recursion operators is local . [ sec : relations - n - m ] finally , we indicate relations between the camassa holm equation and the equation describing short capillary - gravity waves ( the neveu manna equation , see @xcite ) @xmath300 we strongly believe that there exists a deformation connecting these two equations and their integrable structures are closely related to each other . we intend to discuss this relation elsewhere . aratyn h. , gomes j.f . , zimerman a.h . , on a negative flow of the akns hierarchy and its relation to a two - component camassa holm equation , sigma ( symmetry , integrability and geometry : methods and applications ) , 2 * , 070 ( 2006 ) , arxiv : hep - th/0610182 . bocharov , a.v . , chetverikov , v.n . , duzhin , s.v . , et al . , symmetries and conservation laws for differential equations of mathematical physics , xiv+333 pp . amer . soc . , providence , ri ( 1999 ) . edited and with a preface by i. krasilshchik and a. vinogradov . igonin s. , verbovetsky a. , vitolo r. , variational multivectors and brackets in the geometry of jet spaces , v int . conf . on on symmetry in nonlinear mathematical physics , kyiv 2003 inst . of mathematics , acad . of sciences of ukraine , pp . 1335 - 1342 . see also http://www.math.utwente.nl/publicaties/2002/1641abs.html . kersten p. , krasilshchik i. , verbovetsky a. , ( non)local hamiltonian and symplectic structures , recursions , and hierarchies : a new approach and applications to the @xmath302 supersymmetric kdv equation , j. phys . , * 37 * , pp . 50035019 ( 2004 ) , arxiv : nlin/0305026 .	using geometrical approach exposed in refs . @xcite , we explore the camassa holm equation ( both in its initial scalar form , and in the form of @xmath0-system ) . we describe hamiltonian and symplectic structures , recursion operators and infinite series of symmetries and conservation laws ( local and nonlocal ) .
a lattice @xmath1 is _ biatomic _ , if it is atomic ( i.e. , every element of @xmath2 lies above some atom of @xmath1 ) and whenever @xmath3 , @xmath4 , and @xmath5 are elements of @xmath1 such that @xmath3 is an atom , @xmath4 and @xmath5 are nonzero , and @xmath6 , there are atoms @xmath7 and @xmath8 such that @xmath9 , see definition [ d : atatbiat ] . in our first result of this paper , theorem [ t : allfinlatt ] , we prove that any finite lattice can be easily embedded atom - preservingly into a finite biatomic one . biatomicity arises naturally in geometric lattices such as lattices of subspaces of a vector space or , more generally , projective geometries . it was also noticed by m.k . bennett @xcite that in geometric lattices , biatomicity is equivalent to modularity . biatomicity is probably even more common among _ convex geometries_. the lattice theoretical facet of these at first finite and purely combinatorial structures was studied in p.h . edelman @xcite and p.h . edelman and r. jamison @xcite . in @xcite , these structures , now generalized to the infinite case , were considered as an antithesis of geometric lattices , in terms of the properties of the closure operators that define them . we mention the lattices of convex subsets of a given affine space and the lattice of subsemilattices of a given meet - semilattice as a few examples of biatomic convex geometries . still , not all convex geometries are biatomic , thus to describe the biatomic members within a given class of such structures would be of great interest . convex geometries are closely connected with the class of join - semidistributive lattices . a lattice @xmath1 is called _ join - semidistributive _ , if @xmath10 it is proved in @xcite that every finite join - semidistributive lattice can be embedded , in an atom - preserving way , into a finite , atomistic , join - semidistributivelattice , or , equivalently , into a finite atomistic convex geometry . this convex geometry is not generally biatomic . the other construction in @xcite embeds any finite join - semidistributive lattice into the lattice @xmath11 of algebraic subsets of some ( generally infinite ) algebraic and dually algebraic lattice @xmath12 . this is also a convex geometry with the additional properties that it is atomistic , biatomic , and join - semidistributive . this , together with ( * ? ? ? * theorem 1.4 ) , implies that _ every join - semidistributivelattice @xmath1 can be embedded into an atomistic , biatomic , join - semidistributive lattice @xmath13 _ , see also the proof of ( * ? ? ? * theorem 3.26 ) . it is asked in problem 4 of @xcite whether @xmath13 can be taken to be finite whenever @xmath1 is finite . in the present paper , we solve this problem in the negative , by showing a quasi - identity @xmath14 that is satisfied by all finite , atomistic , biatomic , join - semidistributive lattices but not by all finite atomistic join - semidistributive lattices , see theorem [ t : biatquasid ] . this result is inspired by a geometrical example of finite convex geometry that in general produces non - biatomic join - semidistributive lattices . in contrast with this , we prove that every finite atomistic join - semidistributive lattice @xmath1 can be @xmath15-embedded , in an atom - preserving way and with the congruence extension property , into a finite , atomistic , join - semidistributivelattice @xmath13 such that all biatomicity problems of @xmath1 can be solved in @xmath13 , see theorem [ t : partsd+ ] . we also study the case of finite _ lower bounded _ lattices , an important subclass of join - semidistributive lattices . the first author s earlier work @xcite provides an embedding of any finite lower bounded lattice into some finite biatomic convex geometry , which implies corollary [ c : embsubp ] ( see also theorem [ t : partsd+ ] ) . still , we do not know whether such an embedding can be done atom - preservingly . this contributes to the list of open problems that concludes the paper . for a lattice @xmath1 with zero ( i.e. , least element ) , we denote by @xmath16 the set of atoms of @xmath1 . the following definition recalls classical notions , related to their counterparts in @xcite . [ d : atatbiat ] a lattice @xmath1 with zero is * _ atomic _ , if every element of @xmath1 is above an atom of @xmath1 ; * _ atomistic _ , if every element of @xmath1 is a join of atoms of @xmath1 ; * _ biatomic _ , if @xmath1 is atomic and for every atom @xmath3 of @xmath1 and all nonzero @xmath4 , @xmath17 , if @xmath6 , then there are atoms @xmath7 and @xmath8 such that @xmath9 . we observe that every finite lattice is atomic , and @xmath1 is atomistic iff for all @xmath4 , @xmath17 such that @xmath18 , there exists @xmath19 such that @xmath20 and @xmath21 . the following lemma is trivial . [ l : charactbiat ] let @xmath1 be an atomic lattice . then @xmath1 is biatomic iff for every atom @xmath3 of @xmath1 and all @xmath4 , @xmath22 such that @xmath23 , @xmath21 , and @xmath6 , there exists an atom @xmath24 of @xmath1 such that @xmath25 . for a lattice @xmath1 with zero and @xmath26 , let @xmath27 be the statement that @xmath28 is an atom of @xmath1 . we can prove right away the following easy embedding result . [ t : allfinlatt ] let @xmath1 be a finite lattice . then @xmath1 has a @xmath29-embedding into some finite , atomistic , biatomic lattice @xmath30 . put @xmath31 . for each @xmath32 , let @xmath33 and @xmath34 be new distinct elements , and put @xmath35 we define a partial ordering on @xmath30 , extending the partial ordering of @xmath1 , by making all the elements of @xmath36 mutually incomparable , and by saying that @xmath37 for all @xmath32 and @xmath26 . then it is straightforward to verify that @xmath38 is a lattice ordering on @xmath30 and that the inclusion map from @xmath1 into @xmath30 is a @xmath29-embedding . since @xmath39 for all @xmath32 , the lattice @xmath30 is atomistic . to prove that @xmath30 is biatomic , it is convenient to use lemma [ l : charactbiat ] . so let @xmath40 , let @xmath4 , @xmath41 such that @xmath23 , @xmath21 , and @xmath6 ( in particular , @xmath4 and @xmath5 are incomparable , thus they are nonzero ) , we find an atom @xmath42 of @xmath30 such that @xmath24 and @xmath25 . if @xmath4 is an atom of @xmath30 then @xmath43 works , so suppose from now on that @xmath32 . if @xmath17 , then @xmath44 , thus @xmath45 is as required . if @xmath46 , say , @xmath47 for some @xmath48 such that @xmath49 , then @xmath50 , so @xmath45 is as required again . we say that a partially ordered set is _ ntherian _ , if it has no infinite strictly increasing chain . equivalently , every nonempty subset has a maximal element . of course , every finite partially ordered set is ntherian . in this paragraph we will prove that if an atomistic lattice can be embedded into some ntherian , join - semidistributive , biatomic lattice , then it also can be embedded into an _ atomistic _ such lattice . we will work toward the proof of this statement in corollary [ c : equivbae2 ] . an immediate application of ntherianity gives the following result . [ l : finrep ] let @xmath1 be a ntherian lattice and let @xmath51 be a subset of @xmath1 . then every join of elements of @xmath51 is a finite join of elements of @xmath51 . we recall the following elementary property of join - semidistributive lattices that follows immediately from ( * ? ? ? * theorem 1.21 ) . [ l : xcapy ] let @xmath52 be a join - semidistributive lattice with zero , let @xmath53 and let @xmath54 and @xmath55 be finite sets of atoms of @xmath52 . if @xmath56 , then @xmath57 . for a subset @xmath58 of a join - semilattice @xmath1 , we denote the set of all joins of nonempty finite subsets of @xmath58 by @xmath59 . [ p : genat ] let @xmath1 be a ntherian , biatomic , lattice and let @xmath60 . we put @xmath61\cap\operatorname{at}l$ ] . then @xmath62 is a ntherian , biatomic , , atomistic lattice . by definition , @xmath63 is a @xmath64-subsemilattice of @xmath1 . since @xmath1 is ntherian , it follows from lemma [ l : finrep ] that @xmath63 is a lattice in its own right , and it is ntherian . it is obviously atomistic , with @xmath65 . let @xmath66 and let @xmath28 , @xmath67 such that @xmath9 . since @xmath1 is biatomic , there are atoms @xmath68 and @xmath69 of @xmath1 such that @xmath70 . from @xmath28 , @xmath71 , it follows that @xmath72 , @xmath73 . this shows that @xmath63 is biatomic . now we prove that @xmath63 is join - semidistributive provided @xmath1 is . let @xmath5 , @xmath28 , @xmath74 such that @xmath75 . by the definition of @xmath63 , there are finite subsets @xmath54 and @xmath55 of @xmath58 such that @xmath76 and @xmath77 . it follows from lemma [ l : xcapy ] that @xmath78 . since @xmath79 , we get @xmath80 , thus completing the proof that @xmath63 is join - semidistributive . for subsets @xmath81 and @xmath82 in a lattice @xmath1 , we say that @xmath81 _ separates the elements of @xmath82 _ , if for all @xmath28 , @xmath83 with @xmath84 , there exists @xmath85 such that @xmath86 and @xmath87 . observe , in particular , that @xmath1 is atomistic iff @xmath16 separates the elements of @xmath1 . [ c : equivbae ] let @xmath1 be a lattice . then the following are equivalent : 1 . there is a lattice embedding from @xmath1 into some finite resp . , ntherian , join - semidistributive , biatomic lattice @xmath30 such that @xmath88 separates the elements of @xmath1 . 2 . there is a lattice embedding from @xmath1 into some finite resp . , ntherian , join - semidistributive , biatomic , atomistic lattice . the results above also hold for `` @xmath89-lattice embedding '' instead of `` lattice embedding '' . we prove the nontrivial direction ( i)@xmath90(ii ) , for `` ntherian''the proof for `` finite '' is similar . let @xmath30 be a ntherian , biatomic , join - semidistributive lattice containing @xmath1 as a sublattice such that @xmath88 separates the elements of @xmath1 . we put @xmath91 and we observe that @xmath58 separates the elements of @xmath1 . now we put @xmath92 . by proposition [ p : genat ] , @xmath63 is a ntherian , atomistic , biatomic , join - semidistributive lattice , with @xmath65 and @xmath93 . now we shall investigate further the interaction of the lattices @xmath63 , @xmath30 , @xmath1 . define a map @xmath94 by the rule @xmath95 in view of lemma [ l : finrep ] , @xmath96 is well - defined . we observe that @xmath97 , for all @xmath26 . [ cl : fmemb ] the map @xmath96 is a meet - embedding from @xmath1 into @xmath63 . if @xmath98 , then @xmath99 . it is clear that @xmath96 is order - preserving and that if @xmath98 , then @xmath99 . let @xmath28 , @xmath100 such that @xmath84 . since @xmath58 separates the elements of @xmath1 , there exists @xmath66 such that @xmath86 and @xmath87 . thus @xmath101 ( by the definition of @xmath96 ) and @xmath102 ( because @xmath103 ) , whence @xmath104 . so @xmath96 is an order - embedding . now let @xmath28 , @xmath100 . we prove that @xmath105 . let @xmath66 such that @xmath106 . since @xmath97 and @xmath103 , this implies that @xmath86 and @xmath107 . thus , since @xmath1 is a sublattice of @xmath30 , the inequality @xmath108 holds , whence @xmath109 . therefore , since @xmath96 is order - preserving , @xmath96 is a meet - homomorphism . [ cl : joinhom ] the map @xmath96 is a join - homomorphism from @xmath1 to @xmath63 . let @xmath28 , @xmath100 and let @xmath66 such that @xmath110 . suppose that @xmath111 . in particular , @xmath28 and @xmath112 are nonzero in @xmath1 , thus in @xmath30 . by the definition of @xmath96 , this means that @xmath113 . hence , since @xmath30 is biatomic and @xmath28 and @xmath112 are nonzero in @xmath30 , there are atoms @xmath68 and @xmath69 of @xmath30 such that @xmath114 . observe that @xmath72 , @xmath73 . moreover , @xmath115 and @xmath116 , whence @xmath117 , a contradiction . therefore , we have proved that @xmath118 . since @xmath96 is order - preserving , the converse inequality holds , which concludes the proof of the claim . the proof of corollary [ c : equivbae ] is completed . [ c : equivbae2 ] let @xmath1 be an atomistic lattice . then the following are equivalent : 1 . @xmath1 has a @xmath119-embedding into some finite resp . , ntherian , join - semidistributive , biatomic lattice . @xmath1 has a @xmath119-embedding into some finite resp . , ntherian , join - semidistributive , biatomic , atomistic lattice . we prove the nontrivial direction ( i)@xmath90(ii ) , for `` ntherian '' ( the proof for `` finite '' is similar ) . let @xmath30 be a ntherian , join - semidistributive , biatomic lattice such that @xmath1 is a @xmath89-sublattice of @xmath30 . by corollary [ c : equivbae ] , it is sufficient to prove that the atoms of @xmath30 separate the elements of @xmath1 . so let @xmath28 , @xmath100 such that @xmath84 . since @xmath1 is atomistic , there exists @xmath19 such that @xmath86 and @xmath87 . since @xmath30 is atomic , there exists an atom @xmath42 of @xmath30 below @xmath3 . then @xmath120 . furthermore , @xmath121 , and hence @xmath122 , that is , @xmath123 . this proves our assertion . for lattices @xmath52 and @xmath1 , a lattice homomorphism @xmath124 is _ lower bounded _ , if the preimage under @xmath96 of any principal dual ideal of @xmath1 is either empty or has a least element . a lattice @xmath1 is _ lower bounded _ , if every homomorphism from a finitely generated lattice to @xmath1 is lower bounded . we refer the reader to @xcite for more details . for a finite meet - semilattice @xmath81 , we denote by @xmath125 the lattice of all subsemilattices of @xmath81 ( @xmath126 included ) . we state here the following result from the first author s paper @xcite . [ t : embsubp ] a finite lattice @xmath1 is lower bounded iff it can be embedded into @xmath125 for some finite meet - semilattice @xmath81 . as @xmath125 is lower bounded , atomistic , and biatomic , this implies immediately the following result . [ c : embsubp ] any finite lower bounded lattice can be embedded into some finite , atomistic , biatomic , lower bounded lattice . for a finite , atomistic , lower bounded lattice @xmath1 , theorem [ t : embsubp ] says that there exists an embedding from @xmath1 into @xmath125 for some finite meet - semilattice @xmath81 . this embedding can be chosen to preserve the zero , however , it may not preserve atoms . the reason for this is that all lattices of the form @xmath125 have the property that for all atoms @xmath28 and @xmath112 , there are at most three atoms below @xmath127 , while there are finite , atomistic , lower bounded lattices that fail this property . we start with the following definition . [ d : extpair ] let @xmath1 be a finite atomistic lattice . extension pair _ of @xmath1 is a pair @xmath128 , where the following are satisfied : 1 . @xmath129 ; 2 . @xmath30 is a meet - subsemilattice of @xmath1 that contains @xmath130 $ ] . for any @xmath60 , we put @xmath131 $ ] . for an extension pair @xmath128 , we put @xmath132 endowed with the componentwise ordering . by definition , a _ closure operator _ of @xmath1 is a map @xmath133 such that @xmath134 , @xmath135 , and @xmath136 implies that @xmath137 , for all @xmath28 , @xmath100 . fix an extension pair @xmath128 of a finite atomistic lattice @xmath1 . let @xmath96 be the closure operator of @xmath1 associated with @xmath30 , that is , @xmath133 is given by the rule @xmath138 we observe that @xmath139 is a meet - subsemilattice of @xmath140 ( where @xmath141 ) that contains both @xmath142 and @xmath143 as elements . hence it is a lattice in its own right . for @xmath144 , we denote by @xmath145 the least element of @xmath139 above @xmath146 . this element can easily be calculated , by the rule @xmath147 we leave to the reader the straightforward proof of the following lemma . [ l : basiclam ] let @xmath1 be a finite atomistic lattice and let @xmath128 be an extension pair of @xmath1 . then the lattice @xmath139 is finite atomistic , and the map @xmath148 defined by @xmath149 for all @xmath26 is a @xmath29-embedding from @xmath1 into @xmath139 . furthermore , @xmath150 . hence , @xmath139 is an atomistic extension of @xmath1 by exactly one atom , here @xmath151 . in the sequel , the only properties of @xmath152 $ ] , where @xmath153 is the new atom , that will be used are the ones listed below : @xmath154 for all @xmath28 , @xmath100 . furthermore , @xmath155 . from now on we shall use the more wieldy description of @xmath139 given by , , and . it is not difficult to verify that conversely , every @xmath156-extension of @xmath1 by exactly one atom below @xmath157 is , up to isomorphism above @xmath1 , of the form @xmath139 for exactly one extension pair @xmath128 of @xmath1 . however , we shall not need this fact . our next result describes when @xmath139 is join - semidistributive . for a subset @xmath54 of @xmath1 , we denote by @xmath158 the set of all maximal elements of @xmath54 . [ l : lamjsd ] let @xmath1 be a finite , atomistic , join - semidistributive lattice and let @xmath128 be an extension pair of @xmath1 with associated closure operator @xmath96 . then @xmath139 is join - semidistributive iff the following conditions are satisfied : 1 . @xmath159 ; 2 . @xmath160 implies that @xmath115 , for all @xmath26 and all distinct atoms @xmath72 and @xmath161 of @xmath1 . we put @xmath162 . suppose first that @xmath13 is join - semidistributive . let @xmath163 . suppose that @xmath164 . since @xmath1 is atomistic , there exists an atom @xmath72 of @xmath1 such that @xmath115 while @xmath165 . from @xmath166 and the maximality of @xmath28 in @xmath131 $ ] , it follows that @xmath167 , whence @xmath168 . so , since @xmath169 , we obtain that @xmath170 . moreover , @xmath171 , thus @xmath172 , and so , by the join - semidistributivity of @xmath13 , @xmath68 , a contradiction . therefore , @xmath159 . now let @xmath26 and @xmath72 , @xmath161 be distinct atoms of @xmath1 such that @xmath160 . it follows from that @xmath173 , whence , by the join - semidistributivity of @xmath13 , @xmath174 , and therefore , again by , @xmath115 . conversely , suppose that both conditions ( i ) and ( ii ) are satisfied . to prove the join - semidistributivity of @xmath13 , if suffices to prove that @xmath175 can not happen , for all @xmath176 and all distinct atoms @xmath72 and @xmath161 of @xmath13 ( see ( * ? ? ? * lemma 1.2 ) ) . since @xmath1 is join - semidistributive , this holds if @xmath28 , @xmath72 , @xmath177 . now suppose that @xmath26 and @xmath178 , so @xmath179 . hence , by using , we obtain that @xmath180 , thus @xmath181 , that is , by , @xmath182 . moreover , @xmath183 , so @xmath184 . hence there exists @xmath185 such that @xmath136 . by assumption , @xmath186 , and consequently @xmath187 , a contradiction since @xmath182 and @xmath188 . since @xmath175 , the last case to consider is @xmath189 for some @xmath100 ( see ) . it follows again from that @xmath190 , and consequently , by assumption , @xmath191 , and so @xmath192 , a contradiction . by definition , a `` biatomicity problem '' in a lattice @xmath1 is a formal expression of the form @xmath6 , where @xmath19 , @xmath4 , @xmath193 , and the inequality @xmath6 holds while @xmath23 , @xmath21 . solution _ of the problem above in @xmath1 consists of atoms @xmath7 and @xmath8 of @xmath1 such that @xmath9 . recall that a lattice embedding @xmath194 has the _ congruence extension property _ , if every congruence of @xmath52 is the inverse image under @xmath96 of some congruence of @xmath1 . the present section will be mainly devoted to proving the following results . [ t : partsd+ ] every finite , atomistic , join - semidistributive resp . , lower bounded lattice @xmath1 admits a @xmath29-embedding with the congruence extension property into some finite , atomistic , join - semidistributive resp . , lower bounded lattice @xmath13 such that all biatomicity problems in @xmath1 can be solved in @xmath13 . it will turn out that the embedding from @xmath1 into @xmath13 in theorem [ t : partsd+ ] preserves more than the congruences , it is in fact an embedding for the transitive closure @xmath195 of the join - dependency relation @xmath196 . this is equivalent to @xmath13 being a congruence - preserving extension of @xmath1 in the finite , lower bounded case , but not in general . the core of the difficulty underlying theorem [ t : partsd+ ] consists of solving very special sorts of biatomicity problems . in lemma [ l : onebb ] to corollary [ c : f , mcep ] , we let @xmath1 be a finite , atomistic , join - semidistributivelattice , and @xmath3 , @xmath42 , @xmath60 such that @xmath3 and @xmath42 are distinct atoms , @xmath129 , @xmath197 , and @xmath198 for all @xmath199 in @xmath1 . furthermore , we let @xmath133 be the map defined by the rule @xmath200 for all @xmath26 . [ l : onebb ] the following assertions hold . 1 . the map @xmath96 is a closure operator of @xmath1 . if we denote by @xmath30 the range of @xmath96 , then @xmath128 is an extension pair of @xmath1 . @xmath139 is join - semidistributive . 4 . denote by @xmath201 the unique atom of @xmath202 . then @xmath203 and @xmath204 . hence , @xmath139 is a join - semidistributive extension of @xmath1 in which the biatomicity problem @xmath197 has a solution . the assertion ( i ) is straightforward . furthermore , it is obvious that @xmath205 is contained in @xmath30 . now let @xmath206 $ ] , we prove that @xmath207 . this is obvious if @xmath208 , so suppose that @xmath209 . from @xmath197 , @xmath209 , and the join - semidistributivity of @xmath1 , it follows that @xmath210 , whence @xmath211 . this completes the proof of ( ii ) . now let @xmath163 . we prove that @xmath207 . this is trivial if @xmath208 , so suppose that @xmath209 . if @xmath212 , then , by the maximality assumption on @xmath28 , @xmath213 thus , from @xmath197 , @xmath214 , and the join - semidistributivity of @xmath1 , it follows that @xmath215 . thus , since @xmath209 and by the join - semidistributivity of @xmath1 , we obtain that @xmath86 , a contradiction . therefore , @xmath86 , so @xmath216 . this proves that @xmath159 . let @xmath26 and @xmath72 , @xmath161 be distinct atoms of @xmath1 such that @xmath160 . we prove that @xmath115 . if @xmath217 , then @xmath218 . otherwise we would have @xmath219 thus @xmath220 , a contradiction . thus @xmath221 , whence , by the join - semidistributivity of @xmath1 , @xmath222 . suppose now that @xmath223 . by the previous paragraph , @xmath224 , and thus @xmath225 . hence , by the join - semidistributivity of @xmath1 , we have @xmath226 , and so @xmath227 . therefore , @xmath228 . by lemma [ l : lamjsd ] , this completes the proof of assertion ( iii ) . the assertion ( iv ) follows immediately from @xmath229 . from lemma [ l : pdlu ] to corollary [ c : f , mcep ] , we let @xmath30 and @xmath201 be as in the statement and proof of lemma [ l : onebb ] . for a finite lattice @xmath52 , we let @xmath230 denote the relation of join - dependency on the set of join - irreducible elements of @xmath52 . observe that for _ atoms _ @xmath28 and @xmath112 of @xmath52 , the relation @xmath230 takes the following simple form : @xmath231 further , we denote by @xmath232 the binary relation on @xmath233 defined by @xmath234 iff either @xmath235 or @xmath236 . then we let @xmath237 denote the transitive closure of @xmath230 and @xmath238 denote the reflexive , transitive closure of @xmath230 . furthermore , since @xmath1 is finite , atomistic , and join - semidistributive , it follows from lemma [ l : xcapy ] that every element @xmath4 of @xmath1 has a minimal decomposition , that is , a least ( with respect to containment ) subset @xmath54 of @xmath16 such that @xmath239 . we denote this set of atoms by @xmath240 ( `` extreme boundary of @xmath4 '' ) , or @xmath241 if @xmath1 is understood . note that @xmath241 is also the unique irredundant decomposition of @xmath4 . observe that @xmath241 consists exactly of the elements which are join - prime in the interval @xmath242 $ ] . first it is convenient to prove the following lemma . [ l : pdlu ] for any @xmath243 , the following relations hold : 1 . @xmath244 ; 2 . @xmath245}u$ ] . for any @xmath243 , we put @xmath246 . from the fact that @xmath243 , it follows that @xmath247 , and so @xmath248 by the minimality assumption on @xmath4 . however , @xmath249 while @xmath250 ( because @xmath23 ) , whence @xmath244 . furthermore , @xmath251 while , since @xmath247 , we have by that @xmath252 , and consequently @xmath245}u$ ] . [ l : drelsd ] for all @xmath28 , @xmath253 , the following assertions hold : 1 . @xmath254}y$ ] implies that @xmath255 ; 2 . @xmath254}p^$ ] implies that @xmath256 ; 3 . @xmath245}x$ ] implies that there exists @xmath243 such that @xmath257 ; 4 . @xmath258}y$ ] iff @xmath255 ; 5 . @xmath259}p^$ ] iff there exists @xmath243 such that @xmath260 . \(i ) by assumption , @xmath261 and there exists @xmath262 $ ] such that @xmath263 and @xmath264 . suppose that the relation @xmath265 does _ not _ hold . so @xmath266 , and therefore there exists @xmath267 such that @xmath268 , hence , by , @xmath269 and @xmath270 . since the relation @xmath265 does not hold , we obtain that @xmath271 . hence @xmath272 , from where we obtain @xmath273 ( so @xmath274 ) . then , since @xmath275 and @xmath271 , we obtain that @xmath276 if @xmath277 , then , since @xmath267 , the equalities @xmath278 holds by the definition of @xmath96 , whence @xmath279 , and so @xmath280 , a contradiction . hence @xmath281 , but @xmath282 , so we obtain the relation @xmath283 finally , since @xmath197 and @xmath23 , we obtain that @xmath284 , therefore , from and , it follows that @xmath255 . \(ii ) there exists @xmath262 $ ] such that @xmath285 and @xmath264 . thus @xmath286 , so @xmath287 and @xmath288 . from the relation @xmath264 , it follows that @xmath289 , so @xmath290 while @xmath264 , and so @xmath256 . \(iii ) there exists @xmath291 $ ] such that @xmath292 and @xmath293 . thus @xmath177 , and @xmath294 while @xmath295 . from the second relation , it follows that there exists @xmath243 such that @xmath296 . however , @xmath297 , and so @xmath257 . \(iv ) from the fact that the natural embedding from @xmath1 into @xmath298 $ ] is atom - preserving , it follows that @xmath255 implies that @xmath258}y$ ] for all @xmath28 , @xmath253 . conversely , for any @xmath28 , @xmath253 , the relation @xmath258}y$ ] means that there are a positive integer @xmath299 and atoms @xmath300 , @xmath301 of @xmath298 $ ] such that @xmath302}z_{i+1}$ ] for all @xmath303 . we prove by induction on @xmath299 that this implies that @xmath255 . for @xmath304 , the conclusion follows from item ( i ) above . suppose that @xmath305 . if @xmath306 , then it follows from the induction hypothesis that @xmath307 , while , by item ( i ) above , @xmath308 , so @xmath255 . suppose now that @xmath309 . then @xmath310 . thus , by the induction hypothesis , @xmath311 ( the equality may hold , _ e.g _ , for @xmath312 ) . furthermore , it follows from items ( ii ) and ( iii ) above that @xmath313 and @xmath314 for some @xmath243 . but from lemma [ l : pdlu](i ) , it follows that @xmath244 , and so @xmath315 . therefore , @xmath255 . \(v ) there exists @xmath316 such that @xmath259}z{\vartriangleleft}_{l[p^]}p^$ ] . from ( ii ) , ( iii ) , and ( iv ) , it follows that @xmath317 , for some @xmath243 , and @xmath318 , whence @xmath319 , but @xmath320 ( because @xmath23 ) , and so @xmath260 . conversely , let @xmath243 such that @xmath260 . thus we also have @xmath321}p$ ] . since @xmath322 by lemma [ l : onebb ] ( iv ) and since @xmath3 , @xmath201 , and @xmath42 are distinct atoms , the relation @xmath323}p^$ ] holds . from lemma [ l : pdlu](ii ) , it follows that @xmath245}u$ ] , so @xmath245}u{\vartriangleleft}_{l[p^]}p{\mathbin{d}}_{l[p^]}p^$ ] , whence @xmath259}p^$ ] . [ c : f , mcep ] 1 . the canonical embedding from @xmath1 into @xmath298 $ ] has the congruence extension property ; in fact , it is an embedding for the @xmath195 relation on atoms . 2 . if @xmath1 is lower bounded , then @xmath298 $ ] is lower bounded . \(i ) by theorem 2.30 and lemma 2.36 in @xcite , it is sufficient to prove that @xmath324 iff @xmath325}y$ ] , for all atoms @xmath28 and @xmath112 of @xmath1 , which follows immediately from the stronger statement lemma [ l : drelsd](iv ) . \(ii ) it is well - known that a finite lattice @xmath52 is lower bounded iff it has no @xmath326-cycle , that is , the relation @xmath237 is _ irreflexive _ , see ( ? ? ? * corollary 2.39 ) . suppose that @xmath1 is lower bounded . it follows from lemma [ l : drelsd](iv ) that the relation @xmath258}x$ ] holds for no @xmath327 . suppose that @xmath259}p^$ ] . it follows from lemma [ l : drelsd](v ) that there exists @xmath243 such that @xmath260 . by lemma [ l : pdlu](i ) , @xmath328 , whence @xmath1 has a @xmath329-cycle , a contradiction . therefore , the relation @xmath258}x$ ] holds for no atom @xmath28 of @xmath298 $ ] . we present the proof for `` join - semidistributive '' , the proof for `` lower bounded '' is similar . since @xmath1 is finite , it suffices to prove that every biatomicity problem @xmath6 in @xmath1 can be solved in some finite , atomistic , join - semidistributive @xmath29-extension of @xmath1 in which @xmath1 has the congruence extension property . we argue by induction on @xmath330 , where @xmath331 denotes the minimal size of a subset @xmath54 of @xmath16 such that @xmath76 , for all @xmath26 . if @xmath332 then the biatomicity problem @xmath6 is already solved in @xmath1 , by @xmath333 and @xmath334 . now suppose , for example , that @xmath335 , for some @xmath336 and some atom @xmath42 such that @xmath337 . let @xmath338 be minimal such that @xmath339 . by lemma [ l : onebb ] , there exists a finite join - semidistributive@xmath29-extension @xmath340 of @xmath1 , in which @xmath1 has the congruence extension property , such that there exists an atom @xmath341 with @xmath342 . so @xmath343 in @xmath340 and @xmath344 . thus , arguing as above , we obtain a finite join - semidistributive@xmath29-extension @xmath345 of @xmath340 , in which @xmath340 has the congruence extension property , with atoms @xmath7 and @xmath346 such that @xmath347 . so @xmath348 . thus , again by lemma [ l : onebb ] , there exists a finite join - semidistributive@xmath29-extension @xmath349 of @xmath345 , in which @xmath345 has the congruence extension property , with an atom @xmath350 such that @xmath9 . observe that @xmath351 . let @xmath14 be the following quasi - identity in the language @xmath352 of lattice theory : @xmath353\\ \longrightarrow u\leq a.\end{gathered}\ ] ] the present section will be mainly devoted to proving the following result . [ t : biatquasid ] every ntherian , atomistic , biatomic join - semidistributive lattice with zero satisfies @xmath14 . let @xmath30 be a ntherian , atomistic , biatomic join - semidistributive lattice with zero . observe that @xmath30 is a _ complete _ lattice . let @xmath4 , @xmath5 , @xmath354 , @xmath72 , and @xmath161 be elements of @xmath30 satisfying the premise of @xmath14 , that is , the statement @xmath355 suppose that @xmath356 , and put @xmath357 $ ] . since @xmath30 is atomistic , there exists @xmath66 such that @xmath358 . [ l : u0v0 ] there are elements @xmath359 , @xmath360 of @xmath59 such that the following inequalities hold : @xmath361 suppose that @xmath362 . then @xmath363 , thus @xmath364 , a contradiction . hence @xmath365 . put @xmath366 ; observe that @xmath367 . so @xmath368 , with @xmath161 nonzero ( because @xmath365 ) . thus , since @xmath30 is biatomic , there exists an atom @xmath369 of @xmath30 such that @xmath370 and @xmath371 . if @xmath372 , then @xmath373 , but @xmath374 , and so @xmath375 , a contradiction . hence @xmath376 . proceeding the same way with the inequality @xmath377 and then inductively , we obtain elements @xmath378 and @xmath379 , for @xmath380 , of @xmath58 such that @xmath381 , @xmath382 , @xmath383 , and @xmath384 , for all @xmath380 . then @xmath385 and @xmath386 ( these are , by lemma [ l : finrep ] , finite joins ) are as required . now , for @xmath380 , suppose we have constructed @xmath387 , @xmath388 that satisfy the following inequalities : @xmath389 since @xmath390 and @xmath30 is biatomic , for every @xmath391 $ ] , there exists an atom @xmath392 of @xmath30 such that @xmath393 . if @xmath394 , then @xmath395 . however , @xmath396 , and so @xmath397 , a contradiction since @xmath398 . hence , @xmath399 , so that @xmath400\}}$ ] belongs to @xmath401 $ ] and @xmath402 . proceeding in a similar fashion with the inequality @xmath403 , we obtain elements @xmath404 and @xmath405 of @xmath59 such that the following inequalities hold , see the right half of figure 1 : @xmath406 we verify that all the inequalities are satisfied with @xmath299 replaced by @xmath407 . @xmath408 , and , similarly , @xmath409 . * @xmath410 , and , similarly , @xmath411 . therefore , the values @xmath359 and @xmath360 obtained in lemma [ l : u0v0 ] can be extended to sequences @xmath412 and @xmath413 of elements of @xmath59 that satisfy the inequalities listed in and for all @xmath380 . a straightforward application of the last two inequalities in yields immediately the following lemma . [ l : bcuv ] the sequence @xmath414 is increasing . since @xmath30 is ntherian , there exists @xmath380 such that @xmath415 . therefore , by using the last two inequalities in , we also obtain the following equality : @xmath416 for any @xmath380 , we let @xmath417 and @xmath418 be finite subsets of @xmath58 such that @xmath419 and @xmath420 . the existence of such sets is ensured by lemma [ l : finrep ] . [ l : uvdisj ] @xmath421 , for all @xmath422 , @xmath423 . . then @xmath425 and @xmath426 , thus @xmath427 , which contradicts the fact that @xmath398 . now can be written as @xmath428 . but from lemma [ l : xcapy ] ( applied to @xmath429 ) and lemma [ l : uvdisj ] , it follows that @xmath430 , and consequently @xmath431 . however , @xmath432 , so @xmath433 , a contradiction since @xmath434 . this completes the proof of theorem [ t : biatquasid ] . [ c : ljtrnumber6 ] there exists a finite , atomistic , join - semidistributive lattice @xmath1 that can not be embedded into any finite or even ntherian atomistic biatomic join - semidistributive lattice . put @xmath435 where @xmath4 , @xmath5 , @xmath354 , @xmath72 , @xmath161 are as on the left half of figure 1 , the lattice of all intersections with @xmath436 of all convex subsets of @xmath437 , see @xcite . it is well - known that all lattices of that form are join - semidistributive . this configuration is obtained , for example , with @xmath438 , @xmath439 , @xmath440 , @xmath441 , and @xmath442 . then the premise of @xmath14 holds in @xmath1 for these elements , although @xmath356 : hence @xmath1 does not satisfy @xmath14 . a closer look at the proof of theorem [ t : biatquasid ] shows that @xmath1 can not be embedded into any finite , biatomic , join - semidistributive lattice @xmath30 such that there is an atom of @xmath30 below either @xmath72 or @xmath161 but not below @xmath4 . the lattice @xmath1 can be embedded into an algebraic , atomistic , biatomic convex geometry ( see @xcite ) , namely , the lattice @xmath443 of all convex subsets of @xmath437 . in fact , @xmath1 can be embedded into a _ join - semidistributive _ atomistic biatomic sublattice of @xmath443 , namely , the lattice of convex polytopes of @xmath437 , that is , finitely generated convex subsets of @xmath437 . of course , this lattice is neither ntherian , nor complete . [ rk : dcc ] we recall that a partially ordered set is _ well - founded _ , if every nonempty subset has a minimal element . then one can prove that the quasi - identity @xmath14 is satisfied by every well - founded , biatomic , join - semidistributive lattice @xmath1 , _ under the additional assumption that @xmath72 is an atom of @xmath1_. we do not know whether the latter assumption can be eliminated , see problem [ pb : biatquasid ] . [ rkloco ] it is proved in @xcite that every finite join - semidistributive lattice has a zero - preserving lattice embedding into a lattice of the form @xmath11 ( the lattice of all algebraic subsets of @xmath12 ) , for a lattice @xmath12 that is both algebraic and dually algebraic . in particular , @xmath11 is biatomic and lower continuous . it follows that @xmath14 is not satisfied by all lower continuous , atomistic , biatomic join - semidistributive lattices . we observe the following immediate consequence of corollary [ c : equivbae2 ] and theorem [ t : biatquasid ] . [ c : ljtrnb0 ] the lattice @xmath435 of corollary can not be embedded into any ntherian , biatomic , join - semidistributive lattice . we denote by @xmath445 the class of all atomistic , biatomic , join - semidistributivelattices , and by @xmath446 the class of all lower bounded lattices . we use standard notation for the basic operators defined on @xmath352-structures , in particular , for a class @xmath447 of @xmath352-structures , we define * the class @xmath448 of all _ finite _ structures from @xmath447 , * the class @xmath449 of all structures that are embeddable into some structure of @xmath447 , * the class @xmath450 of all _ finite _ direct products of structures of @xmath447 , * the class @xmath451 of all ultraproducts of structures of @xmath447 , * the quasivariety @xmath452 generated by @xmath447 . it was proved in @xcite that @xmath453 , both containments being proper . the class of biatomic join - semidistributive lattices provides a new element in this hierarchy . our interest in this section will be focused on the quasivariety generated by @xmath454 . first we state that the finite members of this quasivariety are those embeddable into lattices from @xmath454 . [ p : qbi ] the finite members of @xmath444 are exactly the lattices that are embeddable into some finite , atomistic , biatomic , join - semidistributive lattice . in formula , @xmath455 . it follows from results of the algebraic theory of quasivarieties that the equality @xmath456 holds for any class @xmath447 , see , for example , ( * ? ? ? * corollary 2.3.4(3 ) ) . we wish to prove that any finite member @xmath1 of @xmath444 is embeddable into some member of @xmath454 . since the class @xmath454 is closed under finite direct products , that is , @xmath457 , it follows from the formula above that there exists a lattice embedding @xmath458 where @xmath459 , that is , @xmath13 is an ultraproduct of members of @xmath454 . since @xmath1 is a finite system in a finite first - order language , a standard argument about ultraproducts shows that @xmath1 can be embedded into some system from @xmath454 . evidently , the proof above can be extended to any finite first - order language , in particular the language @xmath119 if we want to deal with lattices with zero , and so on . [ p : qbif ] the following proper containments hold : @xmath460 . it follows from corollary [ c : embsubp ] that @xmath461 . furthermore , the finite members of @xmath462 are exactly the finite lower bounded lattices while the lattice @xmath463 of all order - convex subsets of a four - element chain is finite , atomistic , biatomic , join - semidistributive , and not lower bounded , which shows that the containment above is proper . the containment @xmath464 holds by definition , and corollary [ c : ljtrnumber6 ] provides an example of a finite join - semidistributive lattice which , by proposition [ p : qbi ] , does not belong to @xmath444 . according to corollary [ c : ljtrnumber6 ] , there exists a finite atomistic join - semidistributive lattice that can not be embedded into any finite , atomistic , biatomic , join - semidistributive lattice . however , it is not hard to prove that for all finite atomistic lattices @xmath52 and @xmath1 such that @xmath52 has a @xmath465-embedding into @xmath1 , if @xmath1 is join - semidistributive(resp . lower bounded ) , then so is @xmath52 . thus , in view of theorem [ t : partsd+ ] , the following question is natural . [ pb : vee0 ] let @xmath1 be a finite , atomistic , join - semidistributive ( resp . , lower bounded ) lattice . does @xmath1 have a @xmath465-embedding into some finite , atomistic , biatomic , join - semidistributive ( resp . , lower bounded ) lattice ? [ pb : embdec ] for a finite join - semidistributive lattice @xmath1 , is it _ decidable _ whether @xmath1 can be embedded into some finite atomistic biatomic join - semidistributive lattice ? a variant of problem [ pb : embdec ] is the following . [ pb : finax ] is the quasivariety @xmath444 ( see section [ s : quasiv ] ) finitely based ? that is , is the set of all quasi - identities satisfied by all finite atomistic biatomic join - semidistributive lattices equivalent to one of its finite subsets ? by proposition [ p : qbi ] , a positive answer to problem [ pb : finax ] would imply a positive answer to problem [ pb : embdec ] . nevertheless we conjecture that problem [ pb : finax ] has a negative solution . [ pb : sdat ] let @xmath1 be a finite lattice . if @xmath1 has a @xmath119-embedding into some finite , atomistic , biatomic , join - semidistributive lattice , then does @xmath1 have an _ atom - preserving _ such embedding ? [ pb : lbat ] does every finite lower bounded lattice @xmath1 have an _ atom - preserving _ embedding into some finite , biatomic , lower bounded lattice ? we have seen that every finite lower bounded lattice admits a zero - preserving lattice embedding into some finite atomistic biatomic lower bounded lattice , see corollary [ c : embsubp ] . our final problem asks for extensions of theorem [ t : biatquasid ] . [ pb : biatquasid ] does any complete , _ upper continuous _ _ well - founded _ ) , atomistic , biatomic , join - semidistributive lattice satisfy the quasi - identity @xmath14 defined in section [ s : ljtr ] ? if we replace `` join - semidistributive '' by `` convex geometry '' then the answer to the corresponding problem is _ no _ , as , for example , @xmath443 does not satisfy @xmath14 while it is a complete , algebraic ( thus upper continuous ) , atomistic , biatomic convex geometry . however , @xmath443 is not join - semidistributive , see @xcite . see also remarks [ rk : dcc ] and [ rkloco ] . the work on this paper was initiated in february 2001 during both authors visit of vanderbilt university , arranged by prof . r. mckenzie , to whom we wish to express our gratitude . the authors are grateful to the referees for careful reading of the paper . gorbunov , `` algebraic theory of quasivarieties '' , ( algebraicheskaya teoriya kvazimnogoobrazij ) ( russian ) sibirskaya shkola algebry i logiki . 5 . novosibirsk : nauchnaya kniga , 1999 . xii+368 p. english translation by plenum , new york , 1998 . xii+298 p.	for a class @xmath0 of finite lattices , the question arises whether any lattice in @xmath0 can be embedded into some atomistic , _ biatomic _ lattice in @xmath0 . we provide answers to the question above for @xmath0 being , respectively , * the class of all finite lattices ; * the class of all finite _ lower bounded _ lattices ( solved by the first author s earlier work ) . * the class of all finite _ join - semidistributive lattices _ ( this problem was , until now , open ) . we solve the latter problem by finding a quasi - identity valid in all finite , atomistic , biatomic , join - semidistributive lattices but not in all finite join - semidistributivelattices .
the aim of any notation for software system design must be to faithfully capture the structure and behaviour of the proposed system . the aim of any tooling that supports such a notation must be that it is reliable and provides useful functionality . what kind of functionality ? certainly , the notation must be a good vehicle for communication - other people must be able to understand your ideas expressed in the notation . however , communication is not sufficient ; the notation must support other useful tasks that contribute to the development of a system . a design notation must allow a tool to provide feedback on whether the system is likely to work , perhaps even allow a tool to animate parts of the design . in addition , it should be possible to generate parts of a system from its design . in recent years there has been interest in model driven architecture @xcite , software factories @xcite , tool definition languages and domain specific languages . all of these initiatives aim to provide design notations that achieve the aims outlined above . these technologies fall broadly into two categories : standard vs. extensible languages . the design notation is standard . uml 2.x is an example of this category whereby the notation is standard with some limited scope for extension points . in the case of uml there is a very large number of different types of design element . each type of element can be stereotyped by defining properties and changing the iconization . this is similar to _ annotations _ recently added to java . the advantages of this approach are that the design notation is completely standard in terms of how it is presented to the user , therefore tools mature quickly , the expertise required to use the technology is relatively low , and information is interoperable between tools . the disadvantage is that the notation designers must preempt all the element types that will be required , leading to notational bloat , and the scope for notational extension is very limited . the design notation is user - defined . there are a number of technologies including microsoft visual studio dsl tools @xcite , gmf @xcite and metaedit+ @xcite that allow a user to define new notation . the extent to which the notation can be extended differs between technologies : some tooling limits the definition through wizards and some allows arbitrary extension through program - level interfaces . in most cases , the underlying model for the design notation ( the so - called meta - model ) is defined by the user and therefore the data representation is tailored to the application domain . an advantage of this approach is that both the notation and underlying data representation are a good fit for the application domain , therefore tooling can take advantage of this by supporting the use of the notation . the main disadvantages of this approach are that the skill levels required to work with technology are relatively high and that the graphical tooling must be developed ( even model - based ) for each new language . is it possible to achieve the advantages of each approach without the disadvantages ? analysis of the use of many design notations leads to the conclusion that many features recur while others differ . in most cases there are package - like and class - like elements with association - like relationships between them . variations occur due to different categories of these basic features with their own specific properties . this paper describes an approach to domain specific languages that allows arbitrary extension of a small collection of underlying modelling concepts via access to a self - describing meta - model . it shows that by adding the notion of _ meta - packages _ to the meta - language , tooling can be designed that offers domain specific languages without wholesale definition of a new language ( such as that required by emf and gmf ) . the need for dsl tooling is discussed further by fowler @xcite.the result is an extensible modelling notation that reuses the same tooling . the use of meta - package based technology does not require skills in complete language definition and is not limited to the simple property - based extensions of the standards based approach described above . in addition , by extending the notion of a package slightly , many new language features can be added . meta - packages are implemented in xmf - mosaic and have been used on a number of projects including generating code for telecomms applications @xcite . the screen - shots and code in this paper are taken from a tutorial example that is implemented in xmf - mosaic . the book @xcite provides a good introduction to meta - modelling concepts . the rest of this paper is structured as follows : section [ sec : example ] provides the motivation for meta - packages using a simple example ; section [ sec : metapackages ] defines meta - packages and how associated tooling accommodates them ; section [ sec : beandsl ] implements the example dsl using meta - packages ; finally , section [ sec : review ] reviews meta - packages , describes some extensions and discusses related systems . design notations should support the generation of source code where this is sensible . the generation of source code from standard modelling notations raises a problem . one of the merits of a modelling notation is that it abstracts away from implementation details ; however , implementation details are required in order to generate the code . in effect , an arbitrary number of sub - categories of modelling element needs to be identified and represented in the design . each category can be defined in terms of properties , structure or behaviour . standard modelling notations often allow categories to be defined using simple properties . however , this is not addressing the real issue which is to be able to define new categories of modelling element with their own structure and behaviour . meta - packages support the definition of new element categories by allowing a standard modelling language to be extended at the meta - level . this section motivates the definition of meta - packages using a simple example . java has recently been extended with annotations which can be used to add properties to standard java program elements such as classes . the motivation for annotations has been the need to _ mark - up _ java components with static information that can be used by tooling . an example use of annotations is in the implementation of enterprise information systems where java class use annotations to define a mapping to relational database tables . for example the simple model defined in figure [ fig : orderprocessing1 ] may give rise the to following java code : .... @entity @table(name="order_table " ) public class order { private int i d ; private string address ; @id @column(name="order_id " ) public int getid ( ) { return i d ; } public void setid(int i d ) { this.id = i d ; } @column(name="shipping_address " ) public string getaddress ( ) { return address ; } public void setaddress(string address ) { this.address = address ; } } .... in which instances of the class order are to be persisted in a relational database as rows in the table named order_table with columns order_id and shipping_address . the primary key of this table is order_id . the tool used to represent the model in figure [ fig : orderprocessing1 ] is standard since it provides tools in the palette to create basic modelling elements : package , class , attribute etc . given a model using this tool , there is no way to distinguish between those model element that will become entity beans and those that will become basic java classes . figure [ fig : beanmodelling ] shows a tool that provides a new category of modelling element : beans . the model for the order processing system now includes two different categories of modelling element : classes and attributes , entitybeans and beanattributes . it is now possible to distinguish between the elements in terms of code generation . in addition , entitybean will have a property that defines the name of the relational database table used to represent its instances . in addition to defining new modelling entities , a dsl has semantics . part of the semantics of a language are the rules that govern correct model formation . these are expressed at the meta - package level , for example an entitybean is correctly formed when it has at most one bean attribute that is designated as an i d ( the primary key in the table ) . figure [ fig : dslsemantics ] shows the result of running the well - formedness checks for the meta - package over the instance shown in figure [ fig : beanmodelling ] . leaves of the tree are constraints that have been applied to the model elements . triangles represent constraints that are satisfied . crosses are constraints that have failed . therefore , in the example model the customer element has oneid ( a single attribute designated as a primary key ) but does not specify a persistent name ( a table in the database ) . meta - packages provide a mechanism for extending the basic language for class - based modelling shown in figure [ fig : orderprocessing1 ] with _ domain specific _ features such as those shown in figure [ fig : beanmodelling ] . tooling for standard modelling detects meta - package automatically and extends the functionality appropriately . meta - packages are a way of defining semantically rich dsls without having to specify associated tooling . the idea is that there is a single basic meta - package that defines a language . the concepts in the base meta - package can be extended to produce new meta - packages . each meta - package is a language and tooling is written against the base meta - package . since all new meta - packages are extensions of the base , the existing tooling will work with any new language . since languages are written at the meta - level , executable meta - modelling allows any new language to have a rich semantics . this section defines meta - packages and is structured as follows : section [ sub : xcore ] defines the base meta - package called xcore ; section [ sub : tools ] defines a tool model for languages over xcore ; section [ sub : diagrams ] defines a model for tool diagrams ; section [ sub : displays ] defines a model for the display elements on diagrams ; finally , section [ sub : mappings ] defines mappings that ensure the diagrams and model elements are synchronized in the tools . meta - packages are defined in a meta - circular language defined in figure [ fig : xcore ] . this language is part of the basic package , called xcore , of the xmf - mosaic modelling tool . in xmf - mosaic , everything is an element . classes are elements with names , constraints , attributes and parents . note that every element has a type ( _ of _ ) which is a class . a package is a class that contains elements and which has a meta - package . the idea of the meta - package association is that each element contained by a package is _ of _ some class contained by its meta - package . by default , the meta - package of all packages is xcore . the semantics of xcore is defined using a collection of constraints . since xcore is self defining these constraints are defines with respect to xcore itself . the following is an example which states that all elements of an enumerated type must be instances of that type and no other instances exist : .... context enum elements->forall(e \| e.of = self ) context element of.of = enum implies of.elements->includes(self ) .... xcore supports executable meta - modelling via xmf . the following operations are used in the rest of this paper : .... context class @operation allparents():set(class ) parents->iterate(p p = set { } \| p + p.allparents ( ) ) end context package @operation modellingelements():set(class ) elements->select(e \| e.iskindof(class ) ) end context element @operation tag(expected : class):string if of = expected then " " else of.name end end .... modelling in xcore - based languages is supported by tooling . figure [ fig : xcore ] shows an example of a modelling tool that consists of a palette on the left - hand side and a diagram containing modelling elements on the right . the palette contains buttons for the modelling elements defined in the modelling language . the tool palette depends on the modelling language s meta - package ; however there is only one tool engine which is parameterized with respect to the meta - package . figure [ fig : toolmodels ] shows the key elements involved in defining a modelling tool . a tool associates a package of modelling elements , a diagram and a palette of button groups . buttons are used to select creation modes . events on the diagram modify diagram elements and events to the package modify modelling elements . daemons on both the diagram and the package ensure that events from the model are propagated to the diagram and vice versa . figure [ fig : toolpalette ] shows the structure of a tool palette . the palette consists of groups of buttons which are selected in order to determine the creation mode on a diagram . each group corresponds to a meta - package . all packages have xcore as a meta - package by default and therefore have the xcore and xmap groups . if the meta - package of the package associated with a tool is p which inherits from xcore then the palette will have groups named xcore , xmap and p. this leads to the following constraint : .... context tool palette.groups.name = package.metapackage.allparents().name .... the buttons provided by each group are determined by the language elements defined in the meta - package . a modelling element is a sub - class of package , class or attribute as defined in xcore . therefore : .... context tool package.metapackage.allparents ( ) ->forall(p \| palette.groups ->exists(g \| g.name = p.name and p.modellingelements().name = g.buttons.name ) ) .... each tool manages a diagram that consists of nodes and edges . each node has a display that is used to render the node on the diagram . each edge has a source and a target node and a collection of labels . the diagram model is shown in figure [ fig : diagrammodels ] . the tool requires that the diagram is always synchronized with the package . therefore the following constraint must always hold : .... context tool package.classes()->size = diagram.nodes()->size .... the constraint on edges is less easy to define since attributes may be shown on a diagram within classes or as edges . in addition , inheritance is shown as edges . to make matters more complex the types of attributes and the target of inheritance edges may be imported from other packages . figure [ fig : displays ] shows some of the display types that can be used to render nodes on diagrams . two basic types of display are shown : boxes and text . package - based diagrams use specializations of these display types to render classes . a class box consists of a name box and several attribute boxes . note that attboxes and namebox are derived attributes since they are just specifically identified components of the displays . both attrbute boxes and name boxes have some text for the name of the element . they also both have a tag which is used to identify the meta - type of the element if it not class or attribute respectively . an attribute box has an additional text field for the type of the attribute . diagram - model synchronization is handled by a collection of mappings associated with a tool . figure [ fig : mapping ] shows the mapping model for package diagrams . a mapping has the form a_x_b which associates instances of a with instances of b. each mapping class has a constraint that defines the synchronization requirements . many of these constraints state commutative properties of the model . for example : .... context tool mapping.package = package and mapping.diagram = diagram .... each of the mappings inherit from a class onetoone ( not shown ) which requires that all maplets are unique . a mapping instance can reference all of the elements in the domain and range of the mapping and also reference all of the associations ( one of which is itself ) . the following constraints require the onetoone mapping to have unique associations : .... context onetoone domain->forall(d \| range->exists(r1 r2 \| map(d , r1 ) and map(d , r2 ) implies r1 = r2 ) ) context onetoone range->forall(r \| domain->exists(d1 d2 \| map(d1,r ) and map(d , r2 ) implies d1 = d2 ) ) .... all of the classes in the package must be shown on the diagram and all class nodes on the diagram must correspond to some class in the package : .... context diagram_x_package package.classes = classbox_x_classes.class context diagram_x_package diagram.classboxes = classbox_x_classes.classbox .... all the attribute edges must be related and must correspond to an attribute . note that not all attributes need to be shown as edges on the diagram : .... context diagram_x_package attedge_x_attributes.attedge = diagram.attedges context diagram_x_package attedge_x_attributes.attribute ->subset(package.classes.attributes ) .... the name of a class on a diagram must always be synchronized with the name of the corresponding model element . if the element is directly an instance of class then its tag is empty otherwise the tag must be the name of the meta - type : .... context classbox_x_class classbox.name = class.name context classbox_x_class classbox.tag = class.tag .... all of the attributes in a class box must correspond to some attribute of the associated class . however , the not all attributes need to be displayed in a class box : .... context classbox_x_class attbox_x_attributes.attribute ->subset(class.attributes ) context classbox_x_class attbox_x_attributes.attbox = classbox.attboxes .... attribute boxes offer a name , a type and an optional meta - tag . the following constraints require that these are always synchronized ( attribute edge tags are specified using similar constraints ) : .... context attbox_x_attribute attribute.name = attbox.name context attbox_x_attribute attribute.type.name = attbox.type context attbox_x_attribute attribute.tag = attbox.tag .... finally , any attribute edges that are shown on the diagram must correspond to attributes . the source of the edge must be a class box that corresponds to the owner of the attribute and the target of the edge must be a class that corresponds to the type of the attribute . the predicate isclass is not defined but is satisfied isclass(b , c ) when the box b is associated with the class c in the tool : .... context attedge_x_attribute isclass(attedge.sourcebox , attribute.owner ) context attedge_x_attribute isclass(attedge.targetbox , attribute.type ) .... all of the attributes must be shown as either edges or in boxes : .... context diagram_x_package package.classes.attributes = attedge_x_attributes.attribute + classbox_x_classes.attbox_x_attributes.attribute .... other constraints can be defined to require that an attribute edge can not be shown as a boxed attribute . the previous section has defined the meta - package relationship that allows new class - based dsls to be defined by extending meta - concepts . existing tooling mechanisms can accommodate the new dsls without any new code being necessary . this section shows how a new language to support beans for enterprise systems can be defined using the meta - package relationship . the result of the definition is a new tool as shown in figure [ fig : beanmodelling ] . this section is defined as follows : section [ sub : abstractsyntax ] describes the abstract syntax for the dsl ; section [ sub : concretesyntax ] describes how concrete syntax can be defined in a number of ways ; section [ sub : constraintchecking ] describes well - formedness semantics for the dsl ; section [ sub : codegeneration ] defines how java code is generated from the dsl . the first step in dsl definition is to define a new meta - package whose modelling concepts extend those of an existing meta - package . figure [ fig : beans ] shows the definition of the package named beans . the concepts are defined as follows : 00.00.0000 a bean container is a specialization of package . the contents of a bean container may be entity beans and the class beancontainer provides us with a container for bean specific constraints . an entity bean is a class that has a persistsas property that is used to name the relational database table used to contain all the instances of the class . a bean attribute is a special type of attribute that names the column in the database table used to contain its values . a bean attribute can be tagged as being a primary key in the relational table by setting its isid attribute . a bean attribute may also require a java accessor and updater . these properties are set via the attribute modifiers of the bean attribute . the beans package must be designated as a meta - package . this is done by making it _ inherit from _ xcore as shown in figure [ fig : metapackage2 ] . by default , a package inherits from an empty package and is therefore not a meta - package . by making a package inherits from a meta - package , the new package also becomes a meta - package . the package orderprocessing is to be written in the language defined by beans . by default the meta - package of a new package is xcore . figure [ fig : metapackage2 ] shows the meta - package of orderprocessing has been changed to beans . note that the super - package of orderprocessing has not been changed , so it stays as the default empty package and therefore orderprocessing is not itself a meta - package . a dsl does not require a concrete syntax , but it is generally important for usability to have one . dsls that are defined via meta - packages get a concrete diagram syntax for free . the tooling that is specified in section [ sec : metapackages ] detects the new modelling concepts and provides new palette buttons and labelling as appropriate . in addition to diagram syntax for a language , textual syntax is often useful . xmf @xcite has a text language for package definition that supports the declaration of meta - package information . this is exactly the same as the diagram tooling : the same text processing engine is used even though the language has changed via meta - package extension . the following shows how the orderprocessing bean container can be defined using xmf package definition : .... @package orderprocessing metaclass beancontainer metapackage beans @class namedelement isabstract @attribute name : string end end @class order metaclass entitybean @attribute identifier metaclass beanattribute : integer end @attribute address metaclass beanattribute : string end @attribute customer metaclass beanattribute : customer end @attribute product metaclass beanattribute : product end end @class customer metaclass entitybean extends namedelement end @class product metaclass entitybean extends namedelement @attribute amount metaclass beanattribute : integer end end end .... a problem with the basic text definition shown above is that it does not provide support for setting the new meta - properties such as persistas . these can be set independently , but it is much better if an entity can be defined as a modular unit . in addition , the above syntax exposes implementation details about the modelling concepts by requiring their meta - classes to be specified . xmf supports extensible syntax via syntax - classes @xcite which are normal classes that define grammars for processing text . this mechanism can be used to define a new concrete syntax for a dsl that translates into the basic definitions given above . the following is a dsl for beans that has been used to define the order processing system : .... @beancontainer orderprocessing entity namedelement name ( name ) : string end entity order(order_table ) [ namedelement ] * identifier(order_id ) : integer address ( shipping_address ) : string customer ( customer_ref ) : customer product ( product_ref ) : product end entity customer extends namedelement entity product extends namedelement amount ( amount ) : integer end end .... the example above hides all of the implementation detail . this is achieved by defining a new syntax class for beancontainer as shown below : .... @class beancontainer @grammar beancontainer : : = n = name es = entity * { [ \| let p = @package < n > metaclass beancontainer metapackage beans end in < es->iterate(e x = [ \| p \| ] \| [ \| p.add(<e > ) ; < x > \| ] ) > end \| ] } . entity : : = n = name p = persist s = super as = att * { [ \| let c = @class < n > metaclass entitybean extends < s > end in c.persistas : = < p.lift ( ) > ; < as->iterate(a x = [ \| c \| ] \| [ \| < x>.add(<a > ) \| ] ) > end \| ] } . persist : : = ' ( ' name ' ) ' . super : : = ' [ ' type ' ] ' . att : : = i = isid n = name p = persist ' : ' t = type { [ \| let a = beanattribute(<n.lift()>,<t > ) in a.persistas : = < p.lift ( ) > end \| ] } . end end .... constraint checking involves executable modelling ( being able to attach executable predicate expressions to classes as shown in figure [ fig : xcore ] ) . elements are well - formed when they satisfy all of the constraints defined by their class . in addition , containers , such as packages , are well formed when their contents are well - formed . in order for a bean container to be well - formed , all of the persistent elements must specify a name that can be used in the relational database : .... context persistent @constraint hasname persistas < > " " fail " must specify a persistent name . " end .... an entity bean is well formed when there is at most one bean attribute that designates a primary key : .... context entitybean @constraint oneid not @exists a1,a2 in attributes - > a1 < > a2 and a1.isid and a2.isid end fail " can not have multiple ids . " end .... code generation involves a mapping from a model to source code . this is the essence of mda in which uml models are translated to code . however , uml does not allow access to the meta - level and therefore the scope for extending the modelling language with sophisticated translation mechanisms is limited . to translate from elements in figure [ fig : beans ] to the source code shown in section [ sec : example ] we can also use executable meta - modelling technology . a code export operation is defined for each class that is to be translated . xmf provides _ code - template _ technology that makes it easy to generate code @xcite . the following template : .... @java(out , leftmargin ) class < n > { } end .... writes an empty class definition to the output channel , newlines tab to leftmargin . within the java code template all text is faithfully written to the output except for expressions delimited by < and > . such expressions are evaluated and then written to the output channel . in the example above , the generated class will have a name that is the value of the variable n. within expressions , the use of [ and ] delimiters switch back to literal code and nesting of [ , ] and < , > is permitted . entity beans are translated to code as follows : .... context entitybean @operation code(out : outputchannel ) @java(out,7 ) @entity @table(name=" < persistas > " ) public class < name > { < @for a in attributes when a.iskindof(beanattribute ) do [ private < a.typename ( ) > < a.name > ; ] end ; @for a in attributes when a.iskindof(beanattribute ) do a.code(out ) end > } end end .... each bean attribute is translated to code as follows : .... context beanattribute @operation code(out : outputchannel ) let name = name.tostring ( ) then name = name.uppercaseinitialletter ( ) in @java(out,7 ) < if isid then [ @id ] end > @column(name=" < persistas > " ) < if self.canget ( ) then self.getcode(out,name,name ) end > < if self.canset ( ) then self.setcode(out,name,name ) end > end end end context beanattribute @operation getcode(out , name , name ) @java(out,5 ) public < self.typename ( ) > get < name > ( ) { return < name > ; } end end context beanattribute @operation setcode(out , name , name ) @java(out,5 ) public void set < name>(<self.typename ( ) > < name > ) { this.<name > = < name > ; } end end .... this paper has identified two broad approaches to dsls for design notations : standards based and user defined . standards , notably the uml family , offer mature tooling and interchangeable models , but can be bloated and lack mechanisms for sophisticated extensibility . technologies for user defined notations offer arbitrary flexibility but at a cost of complexity and starting from scratch each time . meta - packages is an meta - modelling based approach to defining languages that allows tooling to be developed that does not require significant modification each time a new language is developed . since the approach uses true meta - modelling , object - oriented techniques can be used to define semantics for the new language features . meta - packages have been specified and a simple example dsl for enterprise systems has been described . meta - packages are an approach rather than a single technology . the key features are a single meta - circular meta - model ( xcore defined here ) , executable modelling , and the meta - package relationship . this combination of features guarantees that any tooling based on the base meta - language will work with any new language that is defined . the xcore language as defined in this paper is rather small . meta - packages are supported by xmf - mosaic where xcore is much larger , however the principles are the same . one important feature supported by xmf - mosaic is the ability for diagrams to render element - nodes and slot - edges . since we advocate a true meta - modelling approach , everything is ultimately an instance of the class element . packages can contain elements . if diagrams can render elements and represent slot - values via edges then _ any _ package element can be represented on a diagram and related to their owner . this feature guarantees that any language can be supported by tooling that is parameterized with respect to the base language , even if the element is not a specialization of the basic meta - concepts ( package , class , attribute etc ) . for example , classes could be extended to represent components with ports . a port can be represented on a diagram as a basic element with a slot - value edge from the component to the port . xmf - mosaic is available open - source under epl from the ceteva web - site ( http://www.ceteva.com ) . virtually all other tools for dsl definition do not implement the golden braid ( meta - model - instance ) @xcite . for example , gmf models are instances of ecore but can not themselves have instances . the same is true of uml and therefore uml tooling and of visual studio dsl tools . the golden braid is a key feature in the meta - package approach since it allows tooling to be defined that is reusable with models that can be extended with their own semantics . n. georgalas , m. azmoodeh , a. clark , a. evans , p. sammut , and j.willans . mda - driven development of standard - compliant oss components : the oss / j inventory case - study . in 2nd european workshop on mda , kent , uk , 2004 . n. georgalas , m. azmoodeh . using mda in technology - independent specifications of ngoss architectures . first european workshop on model driven architecture with emphasis on industrial application ( mda - ia 2004 ) , enschede , the netherlands ( march 2004 ) . georgalas n. , azmoodeh m. and ou s. model driven integration of standards based oss components . proceedings of the eurescom summit 2005 on ubiquitous services and application , heidelberg , germany ( april 2005 ) . pierre cointe . metaclasses are first class : the objvlisp model . in norman meyrowitz , editor , proceedings of the conference on object - oriented programming systems , languages and applications , pages 156165 . acm , october 1987 .	domain specific languages are used to provide a tailored modelling notation for a specific application domain . there are currently two main approaches to dsls : standard notations that are tailored by adding simple properties ; new notations that are designed from scratch . there are problems with both of these approaches which can be addressed by providing access to a small meta - language based on packages and classes . a meta - modelling approach based on _ meta - packages _ allows a wide range of dsls to be defined in a standard way . the dsls can be processed using standard object - based extension at the meta - level and existing tooling can easily be defined to adapt to the new languages . this paper introduces the concept of meta - packages and provides a simple example .
the ability to determine an unknown quantum state produced by some source is central for many applications in quantum information science . the procedure of reconstructing the quantum state , known as quantum state tomography , has therefore been under intense investigations and continues to attract a lot of attention @xcite . in the continuous variable regime , and in particular its quantum optical realizations , there are two commonly used approaches to quantum tomography . in optical homodyne tomography , the set of rotated quadratures is measured using balanced homodyne detection , thus allowing one to scan the phase space of the system @xcite . the alternative method uses the husimi @xmath0-function which can be measured using a double homodyne detection scheme , and has the advantage that the reconstruction requires the measurement of only a single observable @xcite . both of the above instances fall under the class of gaussian measurements , i.e. , measurements which yield a gaussian measurement outcome distribution whenever the system is initially in a gaussian state @xcite . the purpose of this paper is to present a method for investigating whether or not a given set of such gaussian observables is informationally complete @xcite , i.e. , allows the reconstruction of an unknown quantum state of the system from the statistics . we consider an @xmath1-mode electromagnetic field , whose phase space is therefore @xmath2-dimensional . we show that by measuring a gaussian observable one obtains the values of the weyl transform of the state on a linear subspace of the phase space . therefore , for a set of observables the union of these subspaces needs to be `` sufficiently large '' in order for unique state determination to be possible . in particular , we show that if one does not have access to a single informationally complete gaussian observable , then one necessarily needs infinitely many observables . after these general results we focus on two specific instances . firstly , we investigate single informationally complete gaussian observables in more detail . we show that if we restrict to the smallest possible dimension of the outcome space , then the set of informationally complete gaussian observables is exhausted , up to linear transformations of the measurement outcomes , by gaussian observables which are covariant with respect to phase space translations . furthermore , we show that in a suitable symplectic coordinatization of the phase space , any informationally complete gaussian observable with a minimal outcome space is a postprocessing of the @xmath0-function . secondly , we study commutative gaussian observables which then include projection valued ( also called sharp ) gaussian observables as special cases . we show that no finite set of such observables is informationally complete . for an arbitrary set of generalized field quadratures , i.e. , sharp gaussian observables with one - dimensional outcome space , we prove a further characterization for informational completeness . we also find an interesting connection between the generalized quadratures and informationally complete gaussian phase space observables . finally , we consider a more general scenario where the measurement coupling is represented by a general linear bosonic channel . also in the general case we obtain a characterization for informational completeness and deal explicitly with general covariant phase space observables . the hilbert space of an electromagnetic field consisting of @xmath1 bosonic modes is the @xmath1-fold tensor product @xmath3 , where each single mode hilbert space is spanned by the number states @xmath4 . the creation and annihilation operators related to the @xmath5 mode are denoted by @xmath6 and @xmath7 . in the coordinate representation @xmath8 where the number states are represented by the hermite functions . the hilbert space of the entire field is then @xmath9 . the states of the field are represented by positive trace one operators acting on @xmath10 , and the observables are represented by positive operator valued measures ( povms ) defined on a @xmath11-algebra of subsets of some measurement outcome set . in this paper , we will only consider observables taking values in @xmath12 . each observable is thus represented by a map @xmath13 , where @xmath14 is the borel @xmath11-algebra of @xmath12 and @xmath15 denotes the set of bounded operators on @xmath10 , and which satisfies positivity @xmath16 , normalization @xmath17 and @xmath11-additivity @xmath18 for any sequence of pairwise disjoint sets where the series converges in the weak operator topology . when a measurement of @xmath19 is performed on the system initially prepared in a state @xmath20 , the measurement outcomes are distributed according to the probability measure @xmath21}$ ] . the phase space of the @xmath1-mode em - field is @xmath22 and we use the notation @xmath23 for the canonical coordinates . the phase space translations are represented in @xmath10 by the weyl operators @xmath24 where the matrix @xmath25 determines the symplectic form @xmath26 and @xmath27 , with @xmath28 and @xmath29 being the canonical quadrature ( i.e. , position and momentum ) operators acting on the @xmath5 mode . whenever appropriate , we will emphasize the number of modes in question with a subscript so that , e.g. , in the above case we would have @xmath30 . due to the stone - von neumann theorem , the weyl operators are determined by the relation @xmath31 up to unitary equivalence . this unitary freedom corresponds to the choice of the canonical coordinates : let @xmath32 denote the group of symplectic transformations of @xmath22 , i.e. , real invertible @xmath33-matrices @xmath34 satisfying @xmath35 . if we change the coordinates as @xmath36 , then the original weyl operators in the new coordinates are given by @xmath37 . since @xmath38 is symplectic , the operators @xmath39 indeed also satisfy the weyl relation , and there exists a unitary operator @xmath40 such that @xmath41 for all @xmath42 . summarizing , a change of canonical coordinates is effected by the transformation @xmath43 where @xmath44 . a set of observables @xmath45 is called informationally complete @xcite if an arbitrary unknown state @xmath20 is uniquely determined by the collective measurement outcome statistics @xmath46 , or equivalently , by the corresponding fourier transforms @xmath47 for our purposes , the crucial fact is that any state @xmath20 is uniquely determined by its weyl transform @xmath48},\ ] ] which in our case turns out to be directly related to the fourier transforms of the outcome distributions . the weyl transform is a continuous function , so that @xmath20 is already determined if we know the values of @xmath49 on a dense subset of @xmath22 . the following lemma will then turn out to play a central role . it expresses precisely the intuitive idea that in order to uniquely determine the state , one has to `` scan '' the phase space sufficiently well . here @xmath50 refers to the set of trace class operators on @xmath10 . [ lemma : open_set ] let @xmath51 be a nonempty open set . then there exists a nonzero @xmath52 such that @xmath53 for all @xmath54 . let @xmath55 be such that @xmath56 for all @xmath57 ( e.g. , the vacuum state ) . for any @xmath58 define the operator - valued convolution @xmath59 @xcite via @xmath60 so that @xmath61 if and only if @xmath62 . we can now choose @xmath63 to be a nonzero compactly supported @xmath64-function such that @xmath62 for all @xmath54 . we then obtain @xmath65 via the inverse fourier transform , and our choice of @xmath63 guarantees that @xmath58 . thus , @xmath66 is of the required type . we will next recall the relevant concepts of gaussian states , channels , and observables , as well as the concept of postprocessing of a measurement . for more details on continuous variable quantum information , we refer the readers to @xcite and ( * ? ? ? a state @xmath20 is called a gaussian state , if the weyl transform @xmath49 is of a gaussian form . more specifically , the weyl transform of a gaussian state is given by @xmath67 where @xmath68 is the displacement vector whose components are given by the first moments @xmath69}$ ] and @xmath70 is the covariance matrix whose elements are @xmath71}$ ] where @xmath72 denotes the anticommutator . the covariance matrix is a real symmetric matrix which satisfies the uncertainty relation @xcite @xmath73 a gaussian channel is a completely positive trace preserving map @xmath74 which maps gaussian states into gaussian states . written in terms of the weyl transforms of the input and output states , any gaussian channel can be characterized by @xmath75 where @xmath76 is a @xmath77-matrix and @xmath78 is a @xmath79-matrix which must satisfy the complete positivity condition @xmath80 and @xmath81 is a fixed vector @xcite . equivalently , we may use the dual channel @xmath82 , whose gaussian character is captured in its action on the weyl operators @xmath83 we say that an observable @xmath13 is gaussian , if the measurement outcome distribution is gaussian whenever the system is initially in a gaussian state . this can be conveniently expressed in terms of the characteristic function of the povm @xmath84 in which case @xmath19 is gaussian whenever @xmath85 where @xmath86 is a @xmath87-matrix and @xmath88 is a @xmath89-matrix which satisfy the positive definiteness condition @xmath90 and @xmath91 . note that our definition differs from the one used , e.g. , in @xcite where the outcome space of a gaussian observable is assumed to be a symplectic space , i.e. , even dimensional . the physical motivation for this definition will become apparent in the next section where measurement schemes realizing gaussian observables are discussed . a postprocessing of a measurement is a fixed transformation performed on the measurement outcome distribution . this is relevant for informational completeness , because information can only get lost in such a process ; if a postprocessing is informationally complete , then so is the original measurement . here we consider two different kinds of postprocessings : linear postprocessings and smearings . to this end , let @xmath13 be a gaussian observable parametrized by @xmath92 . given a @xmath93 matrix @xmath94 we define the observable @xmath95 via @xmath96 where @xmath97 . then @xmath98 is clearly gaussian , and it is called a linear postprocessing of @xmath19 . if @xmath94 is invertible , @xmath99 is called a bijective linear postprocessing of @xmath19 . a straightforward computation shows that the triple of parameters characterizing @xmath98 is @xmath100 . it is clear from the definition that a bijective linear postprocessing of @xmath19 is informationally complete if and only if @xmath19 is . for any probability measure @xmath101 $ ] , we define the observable @xmath102 via @xmath103 and say that @xmath104 is a smearing of @xmath19 . note that for any state @xmath20 the corresponding probability distribution is just a convolution of the original one with the measure @xmath105 , i.e. , @xmath106 . if @xmath105 is gaussian , i.e , @xmath107 with @xmath108 , then the smeared observable is also gaussian and , using the fact that @xmath109 , we find that the parameters of the smeared observable are @xmath110 . since @xmath111 is nonzero everywhere , the smearing @xmath104 is informationally complete if and only if @xmath19 is . we will next construct a unitary measurement dilation for an arbitrary gaussian observable . this is done by first showing that any gaussian observable can be measured by applying a gaussian channel to the field and then performing homodyne detection on the output of the channel . since unitary dilations of gaussian channels are known @xcite , this then allows us to construct the desired measurement dilation . suppose first that we have a gaussian channel @xmath112 determined by the parameters @xmath113 . let @xmath114 be the canonical spectral measure , i.e. , @xmath115 corresponds to multiplication by the indicator function of the set @xmath116 , and define the observable @xmath117 as @xmath118 . the characteristic function of @xmath19 is then @xmath119 where @xmath120 . for any @xmath121 we denote @xmath122 so that @xmath123 , and we obtain @xmath124 by eq . . finally , we define the @xmath87-matrix @xmath86 , the @xmath89-matrix @xmath88 and the vector @xmath91 by @xmath125 so that the characteristic function can be expressed as @xmath126 in other words , the observable is gaussian . conversely , suppose that we have a gaussian povm @xmath19 determined by the parameters @xmath127 . we need to find parameters @xmath113 of a gaussian channel such that eqs . hold . to this end , first define the matrix @xmath128 by setting @xmath129 and @xmath130 otherwise , and then set @xmath131 so that @xmath132 . define also the matrix @xmath76 via @xmath133 and @xmath134 otherwise , and the vector @xmath135 in a similar manner as @xmath136 and @xmath137 otherwise . in order to prove the validity of the complete positivity condition , we note that by our construction this reduces to showing that @xmath138 which is an immediate concequence of and the definitions of the matrices in question . in order to reach the desired unitary dilation , we recall that any gaussian channel can be realized by coupling the @xmath1-mode system to @xmath139 auxiliary gaussian modes , and then applying two types of unitary operators to the total system : symplectic transformations @xmath40 with @xmath140 , and displacements @xmath141 with @xmath142 @xcite ( see fig . [ fig : dilation ] ) . the number of auxiliary modes can always be chosen so that @xmath143 , though the optimal choice depends on the details of the channel @xcite . the rest of this section is devoted to demonstrating how the parameters of the gaussian channel , and hence the corresponding observable , are determined by the dilation of the channel . input modes are coupled to @xmath139 auxiliary modes by two unitary couplings : a displacement @xmath141 and a symplectic unitary @xmath40 . after the coupling , homodyne detection is performed on @xmath144 output modes , while the other @xmath145 modes are discarded.,width=340 ] now suppose that the auxiliary @xmath139-mode field is in some gaussian state @xmath11 with covariance matrix @xmath70 and displacement vector @xmath68 . since we are interested only in @xmath146 modes , we must discard the other @xmath147 modes after the interaction . the resulting channel is then given by @xmath148\ ] ] where @xmath149 $ ] denotes partial trace over the hilbert space @xmath150 of the discarded modes . without loss of generality we may assume the ordering of the hilbert spaces to be fixed as @xmath151 the weyl transform of the output state is then @xmath152}\ ] ] so that by denoting @xmath153 we have @xmath154 if we now write @xmath34 in block form as @xmath155 where @xmath156 is a @xmath77-matrix and @xmath157 is a @xmath158-matrix , and define @xmath159 we have @xmath160 where @xmath161 , @xmath162 , and @xmath163 . finally , the parameters @xmath127 of the corresponding gaussian observable are obtained from these via eqs . . we will next show a general method for determining the informational completeness of a given set of gaussian observables . to this end , let @xmath13 be a gaussian povm with parameters @xmath164 . for an arbitrary state @xmath20 , the fourier transform of the probability measure @xmath165}$ ] then reads @xmath166 } = \widehat{\rho}\ , ( { \bf a}_0{\bf p } ) e^{-\frac{1}{4}{\bf p}^t { \bf b}_0 { \bf p } - i { \bf v}^t_0{\bf p}},\ ] ] and by linearity , also holds with @xmath20 replaced by a general trace class operator @xmath167 in which case the measure is complex valued . since the gaussian term on the right - hand - side of eq . is non - zero , we can divide both sides by it . hence , by measuring @xmath19 we are able to determine the values @xmath168 for all @xmath169 in the set @xmath170 which is clearly a subspace of @xmath22 . we are now ready to prove the main general result of this paper . [ prop : ic_set_of_gaussian_povms ] a set of gaussian observables @xmath171 is informationally complete if and only if @xmath172 is dense in @xmath22 . if @xmath172 is dense in @xmath22 , then by measuring the observables one can determine the values of @xmath49 on a dense set , so that by the continuity of the weyl transform , @xmath20 is uniquely determined . conversely , assume that @xmath172 is not dense . then there exists a nonempty open set @xmath173 in the complement of the closure of @xmath172 , so by lemma [ lemma : open_set ] , there exists a nonzero @xmath52 whose weyl transform vanishes outside @xmath173 . by and the injectivity of the fourier transform we then have @xmath174 } = 0 $ ] for each @xmath175 , and all @xmath176 . since the povm elements are positive , the same holds with @xmath167 replaced by either of the traceless selfadjoint operators @xmath177 and @xmath178 . at least one of these is nonzero , and hence a constant times the difference of two distinct density operators . by definition , this implies that the set @xmath45 is not informationally complete . the following two results now follow immediately . [ cor : ic_finite_gaussian ] if @xmath179 is a finite set , then @xmath45 is informationally complete if and only if @xmath180 is informationally complete for some @xmath181 . the " if-part is trivial since each @xmath182 is closed and @xmath179 is finite , also the union @xmath172 is closed . hence , @xmath45 is informationally complete only if @xmath183 . but this is possible only if @xmath184 for some @xmath181 , because otherwise each subspace @xmath182 would have zero lebesgue measure and by subadditivity , so would also @xmath22 . an application of theorem [ prop : ic_set_of_gaussian_povms ] completes the proof . [ cor : ic_gaussian_povm ] a gaussian observable @xmath19 is informationally complete if and only if @xmath185 . being a subspace , @xmath186 is closed , so @xmath19 is informationally complete if and only if @xmath187 , i.e. , @xmath185 . we now proceed to investigate the informational completeness of a single gaussian observable more carefully . [ prop : minimal ] let @xmath188 be a linear subspace and denote @xmath189 . then there exists a gaussian observable @xmath190 such that @xmath191 . in particular , for each @xmath1 there exist both informationally complete an informationally incomplete gaussian observables . let @xmath192 be an orthonormal basis of @xmath116 and define @xmath193 so that @xmath86 is a @xmath87 -matrix . since @xmath194 is hermitian , there exists a matrix @xmath88 , such that is satisfied . then for any @xmath91 the parameters @xmath127 define a gaussian observable @xmath19 which satisfies @xmath191 . the last claim then follows by choosing @xmath195 and , e.g. , @xmath196 . [ prop : minimal ] shows that if the dimension of @xmath186 is strictly smaller then the dimension of the measurement outcome space of @xmath19 , then from the state reconstruction point of view there is some redundancy in the measurement setup . indeed , the same information about the state can be obtained with a smaller outcome space and hence , with fewer homodyne measurements . in particular , for an @xmath1-mode field it is necessary but also sufficient to couple the field to @xmath1 auxiliary modes , thus giving the outcome space dimension @xmath2 . in this case , we say that an observable satisfying the rank condition of cor . [ cor : ic_gaussian_povm ] is a minimal informationally complete gaussian observable . clearly , this condition is equivalent to @xmath86 being invertible , i.e. , @xmath197 . one immediate consequence of the determinant condition is worth noting explicitly : a triple @xmath164 with @xmath33-matrices @xmath198 and @xmath199 , if drawn randomly from the subset given by , according to any nonsingular probability density , defines a minimal informationally complete gaussian observable with probability one . hence , almost all gaussian observables with minimal outcome space are informationally complete . we have seen that a minimal informationally complete gaussian observable has outcome space @xmath22 , which we can identify with the phase space , whose translations and symplectic transformations act as unitary transformations on the range of the observable . we now look at the consequences of this identification . an observable @xmath200 is a covariant phase space observable if it satisfies @xmath201 a direct computation using the characteristic function and the weyl relations shows that a gaussian phase space observable @xmath19 , parametrized by @xmath92 , is covariant if and only if @xmath202 according to cor . [ cor : ic_gaussian_povm ] , every covariant gaussian phase space observable is thus informationally complete . this result was previously obtained , e.g. , in @xcite . as the next proposition shows , any other minimal informationally complete gaussian observable is connected to a covariant one via linear postprocessing . [ prop : infocomplete_covariant ] a gaussian observable @xmath200 is informationally complete if and only if it is a bijective linear postprocessing of a covariant gaussian observable . let @xmath203 be a covariant gaussian observable , parametrized by @xmath204 , and @xmath94 any invertible matrix . the postprocessing @xmath205 is then parametrized by @xmath206 , and is thus informationally complete by corollary [ cor : ic_gaussian_povm ] . conversely , suppose that @xmath19 is an informationally complete gaussian observable , parametrized by @xmath164 . then @xmath86 is invertible , so we can define an invertible matrix @xmath207 , and further define @xmath208^t$ ] , @xmath209 . then the triple @xmath210 determines a covariant gaussian observable @xmath211 for which @xmath212 . next we take into account the symplectic structure of the phase space . the change of the canonical coordinates via a symplectic matrix @xmath34 transforms a phase space observable @xmath200 into the phase space observable @xmath213 given by @xmath214 in particular , a covariant gaussian observable parametrized by @xmath204 transforms into the covariant gaussian observable given by @xmath215 , because @xmath216 . since the positivity condition reduces to @xmath217 , we can use williamson s theorem @xcite to choose the symplectic matrix @xmath34 such that it diagonalizes @xmath218 , i.e. , @xmath219 , where the symplectic eigenvalues of @xmath218 satisfy @xmath220 . letting @xmath221 $ ] be the gaussian probability measure such that @xmath222 , the transformed covariant observable is thus @xmath223 where @xmath224 is the is the @xmath0-function of the state , i.e. , the covariant gaussian observable with @xmath225 . to summarize , each covariant gaussian observable is a gaussian smearing of the @xmath0-function , up to the choice of the canonical coordinates of the phase space . combining this with prop . [ prop : infocomplete_covariant ] , we have the following result : [ prop : minimal_infocomplete ] a gaussian observable @xmath200 is informationally complete if and only if there exist an invertible matrix @xmath94 , a symplectic matrix @xmath34 , and a gaussian probability measure @xmath105 , such that @xmath226 in other words , any minimal informationally complete gaussian observable coincides with a postprocessing of @xmath227 , up to the choice of the canonical coordinates . notice that the any covariant gaussian observable can be measured with the following setup : the signal @xmath1-mode field is coupled to @xmath1 auxiliary parameter modes in a gaussian state by means of 50:50 beam splitters , and then rotations of @xmath228 are performed on half of the output modes ( the case of a single mode is illustrated in fig . [ fig : eightport ] ) . by varying the gaussian state of the parameter field , one can obtain an arbitrary covariant gaussian observable . in particular , a measurement of the @xmath0-function is obtained by choosing the parameter field to be the vacuum ( i.e. , * v=i * , * m=0 * ) . is performed on one of the output modes . homodyne detection is then performed on the two output modes.,width=302 ] recall that an observable @xmath13 is commutative , if @xmath229 for all @xmath230 , and sharp if @xmath231 . it is well - known that an observable is sharp if and only if it is projection valued , i.e. , a spectral measure . clearly , sharp observables are always commutative . conversely , every commutative observable is obtained from some sharp observable by postprocessing with a markov kernel ( see ( * ? ? ? * thm . 4.4 ) ) . smearing with a convolution kernel is a special case , which however turns out to be sufficient to get all commutative gaussian observables from sharp ones . in terms of characteristic functions , @xmath19 is sharp if and only if the map @xmath232 is a unitary representation of @xmath12 , i.e. , @xmath233 similarly , @xmath19 is commutative if and only if @xmath234 this follows by standard approximation arguments from the fact that @xmath235 holds for any integrable @xmath236 , where @xmath63 is the ( suitably normalized ) fourier transform . in the case of gaussian observables , the relationship between commutativity and sharpness can now be analyzed explicitly . by , @xmath19 is a commutative if and only if all the weyl operators @xmath237 commute , which happens exactly when @xmath238 note that this condition is exactly the one that makes @xmath239 a unitary representation , and thus a characteristic function of a sharp gaussian observable which we denote by @xmath240 . now if holds , then @xmath241 by the positivity condition . hence , there exists a gaussian probability measure @xmath101 $ ] , such that @xmath242 , and we have @xmath243 in conclusion , we have proved that a gaussian observable is commutative exactly when it is a gaussian smearing of a sharp gaussian observable . since sharp observables are commutative , we can use the along with condition to characterize sharp gaussian observables . indeed , since now @xmath244 and the operators @xmath245 are unitary , we have that @xmath19 is sharp if and only if @xmath246 for all @xmath247 . but since @xmath105 is gaussian , this reduces to the condition @xmath248 . this implies @xmath249 , and since @xmath241 by , we have @xmath250 . hence , sharp gaussian observables are parametrized by @xmath251 with . note that this can be obtained from the unbiased sharp gaussian observable @xmath252 by shifting the outcomes with the vector @xmath253 . here `` unbiased '' refers to the fact that for any state with zero expectation for all canonical quadratures @xmath254 , @xmath255 , also the expectation of @xmath252 vanishes . it is known that a single commutative observable is never informationally complete ( * ? ? ? for commutative gaussian observables with minimal outcome space @xmath256 this follows immediately from cor . [ cor : ic_gaussian_povm ] , because we have @xmath257 . for a general commutative gaussian observable this can still be seen directly by noting that the rank of @xmath86 is at most @xmath1 , as can be seen by applying the frobenius inequality ( * ? ? ? * eq . 4.3.3(9 ) , p. 61 ) : @xmath258 cor . [ cor : ic_finite_gaussian ] then tells that no finite set of commutative gaussian observables is informationally complete . it is worth noting that in the case of a single mode ( @xmath259 ) , and minimal outcome space ( @xmath260 ) we have @xmath261 so that @xmath19 is informationally complete if and only if @xmath19 is noncommutative . in other words , informational incompleteness is equivalent to the observable being a smearing of a sharp observable . note that this is not true for multiple modes . indeed , using the above result that @xmath262 for commutative observables , we see that if @xmath263 , any given @xmath33-matrix @xmath86 with @xmath264 , together with a @xmath88 chosen to satisfy , determine a gaussian observable which is informationally incomplete and noncommutative . before proceeding to more specific cases , we make one more observation . it follows from that a set of commutative gaussian observables with matrices @xmath265 , @xmath266 is informationally complete if and only if the set @xmath267 of the corresponding sharp observables is such . in other words , a set of commutative observables can always be reduced to a set of sharp observables , without affecting informational completeness . accordingly , we concentrate mostly on sharp observables in the concrete examples . consider now the special case of a one dimensional outcome space . then for any gaussian observable @xmath19 , the matrix @xmath86 is , in fact , a vector in @xmath22 , and we let @xmath268 . the condition is automatically satisfied , so we conclude that every gaussian observable @xmath269 is commutative , and thus a gaussian smearing of the corresponding sharp observable @xmath252 . furthermore , if @xmath269 is a gaussian observable with matrix @xmath86 , then @xmath19 is a gaussian smearing of @xmath270 with @xmath271 , exactly when @xmath272 . the observable @xmath252 has the characteristic function @xmath273 so that @xmath252 is simply the spectral measure of the selfadjoint operator @xmath274 these operators are sometimes referred to as a generalized field quadratures , and they have also previously been considered in the context of quantum tomography @xcite . [ prop : ic_set_of_gaussian_povms ] now reduces to the following result : [ prop : one_dim_case ] let @xmath275 , @xmath266 be gaussian observables with corresponding vectors @xmath276 . then @xmath277 is informationally complete if and only if @xmath278 is dense in the surface of the unit ball of @xmath22 . we now look at two particular examples in the case of a single mode ( @xmath259 ) . by setting @xmath279 and @xmath280 we obtain the well known rotated quadrature operators @xmath281 , the corresponding observables being denoted by @xmath282 . from prop . [ prop : one_dim_case ] we immediately see that we can restrict our attention to the values @xmath283 , and that a set of the form @xmath284 where @xmath285 is informationally complete if and only if @xmath179 is dense in @xmath286 . this therefore sharpens the result of , e.g. , @xcite by proving that density is indeed also necessary . explicitly , the line on which the weyl transform can be determined from the measurement of a quadrature @xmath282 , is given by @xmath287 as a slight modification , we set @xmath288 and @xmath289 so that we obtain the squeezed rotated quadratures @xmath290 . in this case we have @xmath291 so that for @xmath292 any change in the value of the squeezing parameter @xmath293 causes a change in the slope of the line . for instance , we can fix only two values @xmath294 and consider a set @xmath295 of squeezing parameters . by prop . [ prop : one_dim_case ] , the set @xmath296 is informationally complete if and only if @xmath297 is dense in the unit circle , which happens exactly when @xmath179 is dense in @xmath298 . indeed , for each of the four choices of signs , the map @xmath299 bijectively parametrizes the part of the unit circle lying in the interior of the @xmath300-quadrant . in a similar manner we can consider any finite number of values for @xmath301 . the benefit of adding more rotations is that for a fixed @xmath301 one only has to find a set of squeezing parameters which is dense in some interval @xmath302 $ ] in order to guarantee informational completeness . we could also have looked at smearings of sharp quadratures . as already mentioned , this does not add any more structure from the point of view of informational completeness . however , smearings often appear more naturally than sharp observables ; we close this section with a particular case involving postprocessing . consider the linear postprosessing @xmath98 of a phase space observable @xmath19 with matrix @xmath86 , given by a @xmath303 matrix @xmath94 . such a postprocessing is called a ( nonnormalized ) marginal of @xmath19 . now @xmath98 has one - dimensional outcome space , and is a commutative gaussian smearing of the generalized quadrature given by the vector @xmath304 . using this equality , together with cor . [ cor : ic_gaussian_povm ] , we now get the following characterization of informational completeness of phase space observables in terms of quadratures . [ prop : info_complete_connection ] let @xmath19 be a gaussian phase space observable . then the following are equivalent : * @xmath19 is informationally complete ; * there is a one - to - one correspondence between generalized quadratures @xmath270 , @xmath271 , and linear maps @xmath305 , such that @xmath99 is a gaussian smearing of @xmath270 . as we have noted before , the physical motivation for the definition of gaussian observables comes from their measurement realizations : any gaussian povm can be measured by first applying a gaussian channel to the system , and then performing homodyne detection . now gaussian channels are special cases of what are known as linear bosonic channels , that is , channels @xmath306 which map weyl operators according to @xmath307 where @xmath76 is a @xmath77 matrix and @xmath65 is a complex valued function which must again satisfy suitable complete positivity conditions @xcite . the corresponding generalization for gaussian observables is then given by povms @xmath13 satisfying @xmath308 again with some restrictions on @xmath86 and @xmath309 . we immediately notice that the only difference in the treatment of informational completeness when compared to the gaussian case is that the function @xmath65 may be zero at some points . hence , by measuring @xmath19 we can always determine the values of @xmath49 on the set @xmath310 where @xmath186 is the subspace related to any gaussian observable with the parameter @xmath86 . in particular , we have @xmath311 but this is typically not a subspace . the proof of the following characterization of informational completeness is identical to the proof of thm . [ prop : ic_set_of_gaussian_povms ] and is therefore omitted . let @xmath312 , @xmath313 , be observables satisfying for some set of parameters @xmath314 . then @xmath315 is informationally complete if and only if @xmath316 is dense in @xmath22 . note that in this more general scenario , a finite set of observables may be informationally complete even though no single observable is such . for instance , suppose that @xmath19 and @xmath317 are two observables for which @xmath318 but @xmath319 . then it may happen that @xmath320 and @xmath321 for some open sets @xmath173 and @xmath322 for which @xmath323 . in that case neither observable is informationally complete but the pair @xmath324 is since @xmath325 . as an example , consider covariant phase space observables , i.e. , povms @xmath200 satisfying @xmath201 any covariant phase space observable is generated by a unique positive trace one operator @xmath11 , giving the observable the explicit form @xcite @xmath326 in particular , if @xmath11 is a gaussian state , then @xmath327 is a covariant gaussian observable . the characteristic function of @xmath327 can be calculated using eq . and the fact that @xmath328}\ , { { \rm d}}{\bf x } = { \mathrm{tr}\left [ { t } \right]}{\mathrm{tr}\left [ { s } \right]}\ ] ] for any @xmath329 ( see , e.g. , ( * ? ? ? * lemma 1 ) ) . as a result we obtain @xmath330 so that compared to eq . we have @xmath331 , in accordance with eq . , and @xmath332 . using the fact that @xmath333 if and only if @xmath334 , we find that @xmath335 in particular , the covariant phase space observable @xmath327 is informationally complete if and only if @xmath336 , that is , the weyl transform of the generating operator @xmath11 has full support . this result was obtained also in @xcite . notice that any covariant phase space observable can be measured with the setup depicted in sec . [ sec : covariant ] and fig . [ fig : eightport ] by choosing the state of the parameter field appropriately ( for technical details , see @xcite ) . this example actually demonstrates how these more general observables corresponding to linear bosonic channels arise . indeed , if we take the dilation of a gaussian channel as presented in sec . [ sec : dilation ] and replace the state of the auxiliary field by an arbitrary one , then we obtain a linear bosonic channel , and thus an observable , which is typically not gaussian but still of the form . another occasion where these observables may arise is any gaussian measurement which is subject to classical noise . indeed , if @xmath337 is a gaussian observable parametrized by @xmath92 and @xmath338 $ ] is a probability measure , then @xmath339 which corresponds to eq . with @xmath340 . hence , @xmath341 in particular , if @xmath342 is an informationally complete phase space observable , then @xmath104 is informationally complete if and only if @xmath343 . we have proved a characterization for the informational completeness of an arbitrary set of gaussian observables . as a consequence , we have shown that unless one has access to a single informationally complete gaussian observable , then one needs infinitely many observables . we have characterized informationally complete gaussian observables which are minimal in the sense of having the smallest possible dimension of the outcome space , as the observables which are bijective linear postprocessing of covariant gaussian phase space observables . we have then developed this further and shown that any minimal informationally complete gaussian observable is actually a postprocessing of the observable @xmath227 for which the outcome distribution is the @xmath0-function of the state , given a suitable symplectic coordinatization of the phase space . we have also treated commutative gaussian observables separately , and shown that infinitely many such observables are needed in order to reach informational completeness . as a special case we have characterized informationally complete sets of generalized field quadratures , i.e. , sharp gaussian observables with one - dimensional outcome space , and proved a connection between informationally complete gaussian phase space observables and generalized field quadratures . since gaussian observables can be measured by combining gaussian channels with homodyne detection , we have also studied the natural generalization to the case where the gaussian channel is replaced by a linear bosonic channel . also in this case we have obtained necessary and sufficient conditions for the informational completeness of any set of such observables . js acknowledges support from the academy of finland ( grant no . 138135 ) and the italian ministry of education , university and research ( firb project rbfr10coaq ) . jk acknowledges support from the european chist - era / bmbf project cqc , and the project ep / j009776/1 .	we prove necessary and sufficient conditions for the informational completeness of an arbitrary set of gaussian observables on continuous variable systems with finite number of degrees of freedom . in particular , we show that an informationally complete set either contains a single informationally complete observable , or includes infinitely many observables . we show that for a single informationally complete observable , the minimal outcome space is the phase space , and the observable can always be obtained from the quantum optical @xmath0-function by linear postprocessing and gaussian convolution , in a suitable symplectic coordinatization of the phase space . in the case of projection valued gaussian observables , e.g. , generalized field quadratures , we show that an informationally complete set of observables is necessarily infinite . finally , we generalize the treatment to the case where the measurement coupling is given by a general linear bosonic channel , and characterize informational completeness for an arbitrary set of the associated observables .
the galactic rotation curve has been determined for the inner parts of the galactic disk , interior to the solar annulus , from hi - measurements using the tangential point method ( burton & gordon 1978 ) , whereas the outer rotation curve has been determined using radial velocities of objects with individually known distances , e.g. ob stars ( fich et al . 1989 ) , planetary nebulae ( schneider & terzian 1983 ) , young open clusters ( hron 1987 ) and carbon stars ( metzger & schechter 1994 ) . an alternative method is based on the vertical thickness of the galactic hi - layer ( merrifield 1992 ) . recently brand & blitz ( 1993 ) have rederived the outer rotation curve from co radial velocities of ob stars associated with hii regions , and pont et al . ( 1994 ) have used new radial velocity measurements of classical cepheids for this purpose . all these methods rely on the assumption of circular orbits around the galactic centre , so that radial velocities can be converted to circular velocities . obviously proper motions of the objects may provide independent information on the rotation curve . the ppm catalogue which has recently been compiled at the astronomisches rechen - institut ( rser & bastian 1991 , bastian et al . 1993 , rser et al . 1994 ) is well suited as a broad data base of proper motions of high accuracy for such purposes . brand & blitz ( 1993 ) give a list containing radial velocities of hii regions and their associated molecular cloud complexes which is the basis of our analysis . the exciting ob stars in the northern hemisphere are tabulated by georgelin ( 1975 ) and others ( see blitz et al . 1982 for references ) and can be identified by cross reference numbers . for the exciting ob stars of the hii regions in the southern hemisphere finding charts are given by brand ( 1986 ) . using cross reference numbers , spectral types , magnitudes or accurate stellar positions , which we derived in an intermediate identification step using the hst guide star catalog ( lasker et al . 1990 ) , we have identified as many stars from the source lists in the ppm star catalogue as possible . in total we have been able to identify 228 ob stars associated with 71 hii regions . radial velocities and distances of classical cepheids are taken from the recent work of pont et al . using again accurate stellar positions and cross reference numbers , which we obtained from the hipparcos input catalogue ( turon et al . 1992 ) and from kholopov et al . ( 1985 - 1987 ) , we were able to identify 152 cepheids in the ppm star catalogue . the ppm star catalogue is tied into the fk5 system ( fricke et al . lindegren et al . ( 1995 ) have shown by a systematic comparison of the fk5 system with preliminary hipparcos data that the fk5 positional system has still deficiencies , in particular in a strip at about @xmath7 , corresponding to @xmath8 . unfortunately a number of hii regions and cepheids fall into this area , which we have omitted from our analysis because of this systematic uncertainty , leaving us with 197 ob stars associated with 59 hii regions and 144 cepheids with well - determined proper motions for our final analysis . tables [ obtab ] and [ ceptab ] give the identification numbers of these stars in the ppm catalogue . using the positions and proper motions of the stars given in the ppm star catalogue and the distances and the radial velocities given by brand & blitz ( 1993 ) or pont et al . ( 1994 ) , respectively , we have calculated the positions and the velocities of the stars in galactic coordinates . fig.[xyz ] shows the projection of the positions of the objects onto the galactic plane . @xmath9 and @xmath10 are the cartesian spatial coordinates of the stars relative to the sun . the @xmath9-axis points towards the galactic centre and the @xmath10-axis into the direction of galactic rotation . @@xmath11lr & + s8 & 296225 + & 296206 + s11 & 296337 + s25 & 267779 + & 267782 + & 267822 + & 267797 + s29 & 267950 + s31 & 267973 + s32 & 267965 + & 267970 + & 267964 + s27 & 231915 + s41 & 234340 + & 234237 + & 234243 + & 234273 + & 234280 + & 234293 + & 234337 + & 234362 + s44 & 234275 + s45 & 234399 + & 234410 + s46 & 234023 + s49 & 234332 + & 234333 + & 234345 + & 234348 + s54 & 234320 + & 234299 + & 234306 + & 234321 + & 234336 + & 234352 + & 234357 + s86 & 109578 + & 109560 + & 109575 + s101 & 83984 + s112 & 60147 + s117 & 60726 + s119 & 61260 + s124 & 39725 + s126 & 88031 + s129 & 39225 + s132 & 40485 + s134 & 40324 + s137 & 23309 + & 23333 + @@xmath11lr & + s140 & 23638 + & 23640 + s142 & 41076 + & 41098 + s154 & 23983 + s155 & 24016 + & 24060 + & 24070 + & 24074 + & 24030 + & 24072 + & 23979 + & 24019 + & 24021 + & 24036 + & 24043 + & 24049 + & 24098 + & 24093 + & 24117 + & 24148 + & 24150 + & 24162 + & 24169 + & 24185 + & 24186 + s157 & 24306 + s161 & 24309 + & 24329 + s162 & 24384 + s170 & 11863 + s173 & 12143 + s184 & 25788 + & 25791 + s190 & 13731 + & 13702 + & 13720 + & 13721 + & 13723 + & 13712 + & 13718 + s199 & 28316 + & 28326 + s202 & 14174 + & 14185 + & 14227 + s206 & 28874 + s220 & 68961 + s232 & 70744 + s234 & 70403 + @@xmath11lr & + s234 & 70368 + & 70374 + s236 & 70266 + & 70271 + & 70273 + s238 & 119824 + s252 & 95519 + s263 & 148818 + s264 & 149155 + & 149166 + s273 & 151033 + & 151058 + & 151028 + & 151013 + & 151053 + & 151030 + & 151073 + & 151050 + s275 & 150705 + & 150694 + & 150692 + & 150670 + & 150682 + & 150706 + s277 & 188303 + s279 & 188224 + s281 & 187839 + & 175945 + & 188455 + & 175888 + & 188218 + & 188223 + & 188225 + & 188149 + & 215598 + s292 & 218092 + s293 & 218026 + s295 & 218051 + s296 & 218372 + & 218096 + & 218138 + & 218164 + & 218171 + & 218242 + & 218262 + & 218324 + s297 & 218121 + s310 & 252114 + s311 & 253358 + & 253404 + [ cols= " < , > " , ] @r@lr & + @xmath12 & aql & 168843 + u & aql & 202954 + sz & aql & 167110 + tt & aql & 167242 + ff & aql & 135550 + fm & aql & 135861 + fn & aql & 167384 + v336 & aql & 167002 + v496 & aql & 202574 + v600 & aql & 167670 + y & aur & 48141 + rt & aur & 71665 + rx & aur & 69838 + sy & aur & 47840 + ry & cma & 218422 + rz & cma & 713898 + ss & cma & 252386 + tv & cma & 713601 + tw & cma & 713915 + vz & cma & 727576 + rw & cam & 28771 + rx & cam & 28898 + @xmath13 & car & 357533 + u & car & 339614 + v & car & 356744 + y & car & 339156 + uw & car & 339046 + ux & car & 339090 + xx & car & 358358 + xy & car & 358418 + xz & car & 358450 + yz & car & 339070 + er & car & 339816 + gi & car & 339882 + it & car & 358563 + ry & cas & 42304 + @r@lr & + su & cas & 13918 + sw & cas & 41496 + sz & cas & 27633 + xy & cas & 12507 + dl & cas & 12258 + fm & cas & 25159 + v636 & cas & 13051 + v & cen & 342919 + uz & cen & 358897 + xx & cen & 342137 + bb & cen & 359040 + bk & cen & 358996 + v378 & cen & 360023 + v381 & cen & 342310 + v419 & cen & 340152 + @xmath14 & cep & 40731 + cr & cep & 41067 + ax & cir & 361050 + r & cru & 359375 + s & cru & 341409 + t & cru & 359350 + x & cru & 341282 + vw & cru & 359483 + ag & cru & 341197 + bg & cru & 341066 + x & cyg & 85419 + su & cyg & 109630 + sz & cyg & 60114 + tx & cyg & 60815 + vx & cyg & 60748 + vy & cyg & 85963 + vz & cyg & 62131 + cd & cyg & 84139 + dt & cyg & 86036 + v386 & cyg & 61152 + v532 & cyg & 61309 + @r@lr & + v1334 & cyg & 86379 + @xmath15 & dor & 354837 + @xmath16 & gem & 96982 + w & gem & 122711 + v & lac & 41123 + x & lac & 41132 + y & lac & 40284 + z & lac & 40952 + rr & lac & 40959 + bg & lac & 62336 + gh & lup & 343690 + t & mon & 150465 + sv & mon & 150342 + r & mus & 359566 + s & mus & 371417 + rt & mus & 358941 + uu & mus & 359025 + s & nor & 344832 + u & nor & 344085 + y & oph & 201153 + bf & oph & 266398 + rs & ori & 122354 + sv & per & 47393 + sx & per & 46949 + vx & per & 27145 + aw & per & 69651 + v440 & per & 27565 + x & pup & 252643 + rs & pup & 284941 + vx & pup & 252632 + vz & pup & 727776 + wx & pup & 252971 + wz & pup & 728158 + aq & pup & 253563 + at & pup & 284926 + s & sge & 137241 + @r@lr & + u & sgr & 234663 + w & sgr & 267817 + x & sgr & 267303 + y & sgr & 234421 + wz & sgr & 234281 + xx & sgr & 234503 + yz & sgr & 235075 + ap & sgr & 268062 + ay & sgr & 719227 + bb & sgr & 268971 + v350 & sgr & 268853 + rv & sco & 295723 + ry & sco & 296879 + kq & sco & 762230 + v482 & sco & 296427 + v500 & sco & 296821 + v636 & sco & 322879 + x & sct & 234646 + y & sct & 201965 + z & sct & 202059 + ru & sct & 202038 + ss & sct & 202078 + ev & sct & 707164 + st & tau & 121352 + sz & tau & 120081 + r & tra & 361350 + s & tra & 361800 + u & tra & 361877 + @xmath17 & umi & 431 + v & vel & 337875 + sv & vel & 339405 + bg & vel & 337630 + t & vul & 112020 + u & vul & 109337 + x & vul & 110094 + sv & vul & 109871 + ob stars ccccc & & & & + & & & & + 0.0 & 7.19 & 1.26 & -2.99 & 3.24 + -1.0 & 6.77 & 0.59 & -2.14 & 2.22 + -2.0 & 6.37 & -0.06 & -1.30 & 1.30 + -3.0 & 5.96 & -0.72 & -0.46 & 0.85 + -4.0 & 5.55 & -1.38 & 0.40 & 1.44 + -2.9 & 5.98 & -0.68 & -0.50 & 0.84 + cepheids ccccc & & & & + & & & & + 0.0 & 6.86 & -0.43 & -1.96 & 2.01 + -1.0 & 6.43 & -0.93 & -1.21 & 1.53 + -2.0 & 6.02 & -1.41 & -0.47 & 1.49 + -3.0 & 5.59 & -1.90 & 0.29 & 1.92 + -4.0 & 5.18 & -2.39 & 1.04 & 2.61 + -1.5 & 6.23 & -1.17 & 0.84 & 1.44 + in the fk5 system the velocities of the stars have a component of rigid rotation due to the motion of the lsr around the galactic centre . assuming now that the stars move on circular orbits around the galactic centre and that the rotation curve is flat , one may fit a model of the form @xmath18 to the space velocities of the stars . @xmath19 and @xmath20 correspond to the directions of the @xmath9- and @xmath10-axis of fig.[xyz ] and denote the space velocities of the stars which have been corrected for the solar motion with respect to the lsr ( cf . section [ solar motion ] ) . @xmath21 is the distance of the sun from the galactic centre which we assume as 8.5kpc throughout the present study , @xmath22 denotes the galactic azimuthal angle of a star . since some of the stars are fairly distant from the sun we do not use oort s approximation but adopt rigorous formulae describing a flat rotation curve . deviations of the stars from the midplane are ignored . from a @xmath23-fit of eq.([model ] ) to the space velocities of the 197 ob stars we find @xmath24 for the angular velocity of the lsr which corresponds to a circular velocity of 223@xmath2514km / s in good agreement with the iau 1985 value of 220km / s . from the data of the 144 cepheids we obtain @xmath26 in good agreement with the ob stars . the estimated error of @xmath27 has been calculated from the errors given individually for the proper motions of each star in the ppm catalogue and the errors of the radial velocities and the distances given in our reference lists . furthermore we took into account the velocity dispersion of the stars relative to the circular orbits by an additional error of 6.5km / s in both velocity components for the ob stars ( brand & blitz 1993 ) and of 13km / s in the u- and 9km / s in the v - direction for the cepheids ( pont et al . 1994 ) , respectively . judging from the numerical value of @xmath23 and the statistical distribution of the velocity residuals we found that the error estimates obtained this way were too low and have increased the errors of the proper motions by 30% . this is apparently due to systematic errors of the proper motions in the ppm catalogue defined in the reference frame of the fk5 system , which are not fully represented by the individual errors given there . moreover there is a further source of potentially serious errors of the proper motions which is related to non - physical global rotations of the reference system . there is now a general agreement that the iau ( 1976 ) value of the constant of precession , proposed by fricke ( 1977 ) and used in the construction of the fk5 , has to be corrected by about @xmath28mas / a ( williams et al . 1994 ) . this corresponds to a rigid rotation with a spin vector pointing to the ecliptical pole . the immediate effects on our results are illustrated in table [ precession ] . to each star the correction has been applied and both data sets have been reduced again according to the procedure described above . @xmath27 is again the angular velocity of the lsr wiht respect to the galactic centre . @xmath29 and @xmath30 are angular velocities of the mean rotation of all stars about axes lying in the galactic plane , pointing towards the galactic centre ( @xmath31 ) and the direction of galactic rotation ( @xmath32 ) , respectively . as can be seen from table [ precession ] the introduction of the correction of the constant of precession leads to an unacceptably large apparent tilting of the galactic plane . following fricke ( 1977 ) we have therefore considered a simultaneous correction of the motion of the vernal equinox @xmath33 . this corresponds to a rigid rotation of the reference frame about an axis pointing towards the celestial pole . table [ precession ] shows that the apparent tilting of the galactic plane can be minimized by adopting a value of @xmath34mas / a for the ob stars or of @xmath35mas / a for the cepheids . nevertheless the remaining tilting is significantly higher than in the case where no correction at all was applied , i.e. 0.35mas / a for ob stars with distances greater than 800pc and 1.10mas / a for cepheids . in all these cases @xmath27 agrees within 1mas / a with the values given above . this discrepancy can not be solved on the basis of the present material . the corrections @xmath33 of the motion of the equinox suggested here are of the same order of magnitude as those proposed by miyamoto and sma ( 1993 ) and by wielen ( unpublished ) . they do not agree exactly mainly because of different samples of stars used . miyamoto and sma derived @xmath36mas / a from k giants . wielen obtained @xmath37mas / a from 512 fk5 stars which were used by fricke ( 1977 ) for deriving the constant of precession and the motion of the equinox and which were re - investigated by schwan ( 1988 ) after the fk5 was completed . we conclude that the spurious , non - physical rotations of the fk5 system seem to be in total of the order of @xmath38mas / a in each component of the rotation vector . this is consistent with the claimed accuracy of the fk5 system with respect to non - physical rotations of about @xmath39mas / a ( fricke 1977 , schwan 1988 ) . finally we note that the vla measurement of the proper motion of the sgra@xmath6 radio source in the galactic centre of @xmath40mas / a ( backer 1996 ) , implying after substraction of the peculiar velocity of the sun an angular velocity of the lsr of @xmath5mas / a , deviates only insignificantly from the values found here . before applying eq.([model ] ) the space velocities have to be corrected for the solar motion with respect to the lsr . the standard values for this motion are @xmath41km / s and @xmath42km / s ( delhaye 1965 ) . we introduced @xmath43 and @xmath44 as additional free parameters in eq.([model ] ) and checked the goodness of the fit for various combinations of @xmath43 and @xmath44 . the results are shown in fig.[grid ] . the @xmath23-fit to the ob star data rejects the standard values of 9 and 12km / s at a 4@xmath45 confidence level , whereas the fit to the cepheid data is in good agreement with the classical values . the minimum of the numerical value of @xmath23 for the cepheids however is quite flat and the errors are larger than those for the ob stars , so we decided to use values of @xmath46km / s and @xmath47km / s , which correspond within the errors simultaneously to the minima of @xmath23 of both data sets . note that the value of @xmath44 is closely related to the asymmetric drift of the sample of stars under consideration . jahrei and wielen ( 1983 ) found @xmath48km / s relative to the youngest stars in the solar neighbourhood in close agreement with the value used here . the low value of @xmath43 reflects the apparent inward motion of stars relative to the lsr which was noted previously by fich et al . this effect is usually interpreted as an outward motion of the lsr ( blitz & spergel 1991 ) . we have determined oort s constants @xmath49 and @xmath50 using stars within a circle of 1kpc around the sun , assuming a constant radial gradient of the rotation curve near the sun . by means of a @xmath23-fit to the circular velocity @xmath51 we find for the ob stars @xmath52km / s / kpc and @xmath53km / s / kpc , which agree nicely with the flat shape of the rotation curve over larger distance intervals . the cepheids give @xmath54km / s / kpc and @xmath55km / s / kpc . obviously oort s constants only reflect the local behaviour of the rotation curve . over larger distance intervals the rotation curves show a flatter shape than one would assume from the above values . once the velocities have been corrected for the solar motion , it is straightforward to determine the circular velocity @xmath56 for according to the formula @xmath57 where @xmath58 is the distance of the star from the galactic centre . it is now no longer necessary to assume that the stars move on circular orbits around the galactic centre , because the full space velocities are used . alternatively , assuming again circular orbits , one may derive rotation curves based solely on radial or tangential velocity components of the stars , @xmath59 where @xmath60 denotes the distance of the star from the sun , @xmath61 the radial velocity , and @xmath62 the proper motion in the direction of galactic longitude @xmath63 . both velocities have to be corrected again for the solar motion . @xmath60 is the distance of the star from the sun . the resulting rotation curves determined using either the full space velocities or deprojected radial or tangential velocity components are shown in figs.[ob1 ] to [ cep2 ] . they agree very closely and are all consistent with a flat shape of the rotation curve @xmath64 . stars with unsuitable projection angles onto the supposed circular velocities have been excluded . in the case where only radial velocity data are used stars with @xmath65 have been rejected . if only proper motion data are used stars with @xmath66 have been rejected . therefore the overlap between these two samples is small and we can not correlate directly the rotation curve derived from radial velocities with the rotation curve based on proper motions or both . instead , we show in fig.[diff ] the deviations of circular velocities derived from the space velocities from circular velocities based solely on deprojected radial velocities of the stars illustrating again that both methods give consistent results . the orientations of the velocity residuals after subtraction of the systematic velocity components due to a flat rotation curve from the space velocities of the stars are shown in fig.[illu ] projected onto the galactic plane . the velocity residuals are dominated by the errors of the proper motions . most of the velocity residuals can be shown to be randomly orientated with the exception of the perseus spiral arm . there is a trend of coherent motion along the spiral arm with the effect that the ob stars in the spiral arm tend to lag behind the general rotation of the disk . exactly such a behaviour is predicted by the density wave theory of spiral structure for stars recently born in the shock front of the interstellar gas included by the spiral density wave ( shu et al . this is not observed for the cepheids . they are about 100 times older than the ob stars and their systematic flow pattern is already dissolved ( wielen 1979 ) . backer , d.c . , 1996 , iau symp . 169 ( in press ) bastian , u. , rser , s. , yagudin , l.i . , nesterov , v.v . , polozhentsev , d.d . , potter , kh.i . , wielen , r. and yatskiv , ya.s . , 1993 , ppm star catalogue , vols . iii & iv , spektrum akademischer verlag , heidelberg blitz , l. and spergel , d.n . , 1991 , apj 370 , 205 brand , j. , 1986 , ph.d . thesis , univ . of leiden brand , j. and blitz , l. , 1993 , a&a 275 , 67 burton , w.b . and gordon , m.a . , 1978 , a&a 63 , 7 delhaye , j. , 1965 , in _ galactic structure _ , eds . a. blaauw and m. schmidt , univ . chicago press , 61 fich , m. , blitz , l. and stark , a.a . , 1989 , apj 342 , 272 fricke , w. , 1977 , basic material for the determination of precession and of galactic rotation and a review of methods and results , verffentlichungen astronomisches rechen - institut heidelberg no . 28 fricke , w. , schwan , h. , lederle , t. , bastian , u. , bien , r. , burkhardt , g. , du mont , b. , hering , r. , jhrling , r. , jahrei , h. , rser , s. , schwerdtfeger , h.m . and walter , h.g . , 1988 , fifth fundamental catalogue , verffentlichungen astronomisches rechen - institut heidelberg no . 32 georgelin , y. , 1975 , ph.d . thesis , univ . de provence hron , j. , 1987 , a&a 176 , 34 jahrei , h. and wielen , r. , 1983 , iau coll . 76 , 277 kholopov , p.n . 1985 - 1987 , general catalogue of variable stars , nauka publishing house , moscow lasker , b.m . , sturch , c.r . , mclean , b.j . , russell , j.l . , jenkner , h. and shara , m.m . , 1990 , aj 99 , 2019 lindegren , l. , rser , s. , schrijver , h. , lattanzi , m.g . , van leeuwen , f. , perryman , m.a.c . , bernacca , p.l . , falin , j.l . , froeschl , m. , kovalevsky , j. , lehnhardt , h. and mignard , f. , 1995 , a&a 304 , 44 merrifield , m.r . , 1992 , aj 103 , 1552 metzger , m.r . and schechter , p.l . , 1994 , apj 420 , 177 miyamoto , m. and sma , m. , 1993 , aj 105 , 691 miyamoto , m. , sma , m. and yoshizawa , m. , 1993 , aj 105 , 2138 pont , f. , mayor , m. and burki , g. , 1994 , a&a 285 , 415 rser , s. and bastian , u. , 1991 , ppm star catalogue , vols . i & ii , spektrum akademischer verlag , heidelberg rser , s. , bastian , u. and kuzmin , a. , 1994 , a&as 105 , 301 schneider , s.e . and terzian , y. , 1983 , apj 274 , l61 schwan , h. , 1988 , a&a 198 , 116 shu , f.h . , 1972 , apj 173 , 557 turon , c. et al . , 1992 , the hipparcos input catalogue , esa publications division , esa sp 1136 wielen , r. , 1979 , iau symp . no . 84 , 133 williams , j.g . , newhall , x x and dickey , j.o . , 1995 , highlights of astronomy , vol . 10 ( in press )	we present rotation curves of the galaxy based on the space - velocities of 197 ob stars and 144 classical cepheids , respectively , which range over a galactocentric distance interval of about 6 to 12kpc . no significant differences between these rotation curves and rotation curves based solely on radial velocities assuming circular rotation are found . we derive an angular velocity of the lsr of @xmath0mas / a ( ob stars ) and @xmath1mas / a ( cepheids ) , which is in agreement with the iau 1985 value of @xmath2mas / a . if we correct for probable rotations of the fk5 system , the corresponding angular velocities are @xmath3mas / a ( ob stars ) and @xmath4mas / a ( cepheids ) . these values agree better with the value of @xmath5mas / a derived from the vla measurement of the proper motion of sgra@xmath6 .
star forming galaxies in the early universe have been found in large numbers both by looking for strong lyman breaks , and by looking for line emission . galaxies found by these two methods differ greatly in their typical properties . this may indicate physically distinct galaxy populations , or selection effects inherent in the search methods , or a combination of the two . to help address these issues , we here examine the selection of lyman break galaxies in the hubble ultra deep field ( hudf ) . the depth of the hudf images means that emitting galaxies with fluxes typical of present surveys should all be detected ( down to flux @xmath7 ) , even if they have no continuum emission at all . we combine these deep hudf images with the deepest slitless spectra ever obtained , from the grapes and pears projects ( see below ) . these slitless spectra allow us to look for prominent line emission , with or without pre - selection for a lyman break . we study the continuum properties of a set of selected galaxies , to see what fraction pass our lyman break crieteria . conversely , we also study the emission line properties of a lyman break selected sample . moreover , we examine spectra for lyman break samples selected with two sets of photometric criteria , one `` traditional '' and the other more inclusive . the grism acs program for extragalactic science ( grapes ) project and probing evolution and reionization spectroscopically ( pears ) project are slitless spectroscopic surveys that exploit the potential of the g800l grism on the hubble space telescope s advanced camera for surveys ( acs ) to achieve the most sensitive unbiased spectroscopy yet at red optical wavelengths ( @xmath8 ) . two primary factors enhance our sensitivity relative to ground - based spectrographs . first , high redshift galaxies are typically compact ( e.g. , ferguson et al 2004 ; pirzkal et al 2007 ; hathi , malhotra , & rhoads 2008 ) , so that the sensitivity of ground - based observations is hampered by atmospheric blurring of the galaxy images ( and also usually by slit losses ) . second , the oh emission line forest in the night sky spectrum raises the background level for ground - based observations , introducing random noise ( and often systematic residuals as well ) in red - light spectra of high redshift galaxies from the ground . our hst spectra are free of both effects . moreover , because the acs grism is a slitless spectrograph , we obtain spectra of every source in our field of view , with no preselection required . grapes was targeted in the hudf region , to complement the hudf direct images , which are in turn the deepest optical imaging to date ( beckwith et al . 2006 ) . the grapes survey , and in particular our data analysis methods , are described in more detail by pirzkal et al . ( pears ( further described in malhotra et al 2008 ) included an additional forty orbits of g800l integration on the hudf , essentially doubling the integration time . a primary scientific goal of both surveys is to study the properties of lyman break galaxies using spectroscopically confirmed samples at unprecedented sensitivity . we are pursuing this effort through a targeted look at lyman break candidates identified in the hudf using both an @xmath9 dropout criterion ( malhotra et al 2005 ) and a @xmath10 dropout criterion ( this paper ) . we refer to acs filters by names of roughly corresponding ground - based filters : f435w @xmath11 b ; f606w @xmath11 v ; f775w @xmath11 i ; and f850lp @xmath11 z. throughout this paper we use the current concordance cosmology ( @xmath12 , @xmath13 , @xmath14 ; see spergel et al . 2003 , 2007 ) . we use ab magnitudes , so that magnitude zero corresponds to @xmath15 . we consider both lyman break and emission selected samples . for selection , we use the previously published sample from pirzkal et al ( 2007 ; hereafter p07 ) , which in turn is based on the hudf emission line catalog of xu et al ( 2007 ) . this sample contains 9 -emitting galaxies , spanning redshifts @xmath16 , line fluxes @xmath17 , and continuum magnitudes @xmath18 . for lyman break selection , we derive new samples based on hudf photometric pre - selection combined with grapes / pears grism spectroscopy . because we have spectra for all objects in our sample , we can examine a more inclusive set of lyman break galaxy candidates than is practical for photometric studies . we examined as candidate lbgs all objects having @xmath1 , and @xmath19 . the magnitude cut prevents the sample from being swamped by galaxies too faint for accurate photometry or spectroscopy . we inspected visually the grapes and pears spectra of all objects passing these selection cuts , and retained in our final sample those objects whose spectra support a lyman break identification . lllllrlll 119 & 53.1660037 & -27.8238734 & 27.51 & 2.14 & -0.62 & 28.97 & 4.88 & 2.5 + 201 & 53.1649949 & -27.8224088 & 27.59 & 1.23 & -0.24 & 32.92 & 4.60 & 2 + 478 & 53.1733579 & -27.8182711 & 27.06 & 1.11 & -0.06 & 30.65 & 4.52 & 2 + 546 & 53.158038 & -27.8179411 & 27.86 & 3.13 & 0.79 & 31.52 & 5.42 & 2.5 + 577 & 53.1614532 & -27.8174379 & 27.08 & 1.01 & 0.09 & 30.55 & 3.8 & 2.5 + 646 & 53.1660559 & -27.8167778 & 27.44 & 2.22 & 0.24 & 30.18 & 4.9 & 2 + 712 & 53.1783554 & -27.8162519 & 27.26 & 2.32 & 0.13 & 32.4 & 5.12 & 1.8 + 1115 & 53.1722726 & -27.8119732 & 26.40 & 1.85 & 0.17 & 28.89 & 4.72 & 1.8 + 1392 & 53.1563588 & -27.8095883 & 27.75 & 3.96 & -1.05 & 30.75 & 4.90 & 1 + 2408 & 53.1885357 & -27.8034637 & 26.76 & 1.77 & 0.14 & 31.82 & 4.86 & 2 + 2599 & 53.1626689 & -27.8022982 & 27.00 & 1.76 & -0.07 & 29.75 & 4.88 & 2 + 2736 & 53.1499325 & -27.8017358 & 26.94 & 1.39 & 0.11 & 32.27 & 4.56 & 2 + 2881 & 53.1415872 & -27.8005681 & 25.84 & 1.78 & 0.18 & 29.79 & 4.56 & 2 + 2894 & 53.1462482 & -27.8008077 & 27.58 & 2.57 & -0.10 & 29.85 & 5.3 & 2 + 2898 & 53.1798203 & -27.8008748 & 27.06 & 1.79 & -0.32 & 29.50 & 4.67 & 2.3 + 3094 & 53.1514271 & -27.7997637 & 25.72 & 1.18 & 0.02 & 29.57 & 4.62 & 1 + 3250 & 53.1326676 & -27.7989430 & 27.14 & 3.74 & -0.29 & 30.12 & 4.90 & 2.3 + 3968 & 53.1833331 & -27.7959556 & 27.56 & 3.35 & -0.82 & 31.55 & 4.7 & 2 + 5183 & 53.1437786 & -27.7908654 & 27.38 & 0.91 & -0.57 & 32.92 & 4.62 & 1 + 5225 & 53.1385743 & -27.7902115 & 25.83 & 1.52 & 0.04 & 27.87 & 5.42 & 2 + 5296 & 53.1907601 & -27.7903482 & 27.08 & 2.71 & 0.29 & 31.78 & 5.14 & 2.2 + 5307 & 53.1908538 & -27.7903656 & 27.15 & 2.53 & 0.38 & 30.39 & 5.14 & 2.2 + 5788 & 53.1456573 & -27.7882186 & 27.46 & 4.32 & 0.50 & 30.27 & 5.1 & 2 + 6066 & 53.1845519 & -27.7869713 & 26.11 & 0.93 & 0.01 & 29.12 & 4.42 & 1 + 6139 & 53.1581263 & -27.7863866 & 25.49 & 1.37 & -0.11 & 31.51 & 4.68 & 1.2 + 6515 & 53.1273697 & -27.7851656 & 27.15 & 1.25 & -0.24 & 29.75 & 4.75 & 1 + 6681 & 53.1926615 & -27.7841483 & 27.10 & 2.25 & 0.41 & 28.81 & 5.08 & 1.3 + 7050 & 53.1510392 & -27.7828658 & 27.36 & 4.13 & 0.50 & 30.44 & 5.45 & 2.2 + 7352 & 53.1376954 & -27.7812680 & 26.87 & 2.88 & -0.02 & 29.37 & 5.04 & 2.5 + 8301 & 53.1671692 & -27.7745246 & 27.18 & 2.91 & 0.22 & 32.47 & 5.0 & 2 + 8664 & 53.1890652 & -27.7770042 & 26.77 & 2.08 & 0.08 & 29.9 & 4.9 & 1.3 + 8682 & 53.1888007 & -27.7770931 & 25.62 & 1.94 & 0.06 & 28.70 & 5.08 & 2 + 8896 & 53.1900052 & -27.7790544 & 26.92 & 1.56 & 0.09 & 31.30 & 4.33 & 2 + 9040 & 53.1711852 & -27.7784585 & 25.97 & 3.42 & -0.49 & 28.3 & 4.55 & 2 + 9057 & 53.1829957 & -27.7804592 & 26.90 & 0.91 & -0.02 & 31.84 & 4.17 & 2 + 9275 & 53.1531485 & -27.766181 & 25.83 & 1.26 & 0.14 & 28.93 & 5.12 & 2 + 9777 & 53.1702387 & -27.7628552 & 26.17 & 1.95 & 0.85 & 30.20 & 5.41 & 2.5 + 9983 & 53.1671627 & -27.7598546 & 25.62 & 1.79 & 0.17 & 32.27 & 4.82 & 2 + 20191 & 53.1725558 & -27.8137100 & 25.73 & 1.41 & 0.09 & 31.48 & 4.62 & 1 + to determine redshifts , we fitted our spectra with a model consisting of a power law continuum attenuated by forest absorption , which we modeled using the formalism of madau ( 1995 ) . we thus fitted three parameters to each spectrum : the flux normalization , spectral slope , and redshift . in cases with strong line emission , the fitted slope is strongly biased towards blue values , since we do not expliclitly fit emission lines . we then assessed the best fitting model for each spectrum and assigned a grade to each . these grades were based on the @xmath20 of the fit , the signal to noise in the spectrum , and a visual inspection of the fit . grade 1 was given to the highest quality fits , where the lbg identification and redshift are very secure . grade 2 was given to reasonably secure lyman break objects . grade 3 was given to sources that remain possible lyman break objects but that can not be confirmed , whether because the signal - to - noise was insufficient , or because the spectra suffered from contamination . finally , grade 4 was assigned to photometric candidates that were refuted by the grism spectra . we assigned fractional grades to a modest number of galaxies where it seemed warranted . in total , we considered 216 candidates fulfilling the photometric selection criteria . a few additional objects were examined in the early stages of the project , resulting in the inclusion in the final sample of two v dropouts ( ids 546 and 1392 ) having @xmath21 . in total , we classified 39 objects as `` good '' ( 1 @xmath22 grade @xmath22 2.5 ) lyman break objects , 86 as refuted ( grade 4 ) objects , and 94 as grade 3 ( `` unidentified '' ) . representative spectra for grades from 1 to 2.5 are shown in figure [ fig : examples ] . the brightest confirmed lyman break galaxies have @xmath23 . the fraction of grade 3 ( `` unidentified '' ) sources rises steadily from 0 at @xmath24 to 100% at @xmath25 . a very rough linear fit is @xmath26 . among those sources adequately classified using the grism spectra , the fraction confirmed as lyman break galaxies _ rises _ weakly towards fainter magnitudes , from @xmath27 at @xmath28 to @xmath29 at @xmath30 . to understand this , consider the primary contaminants of the photometrically selected lyman break candidate lists . these are red stars in our galaxy , and early type galaxies at intermediate redshifts . while the number - magnitude relation for lyman break galaxies is steeply rising around @xmath31 , the number - magnitude relations for both contaminants are much flatter , and therefore the lbg fraction rises at the faint end . the v dropout sample includes a range of morphologies , from simple , compact galaxies to extended sources with tails and/or multiple peaks ( e.g. , hudf 5225 ; rhoads et al 2005 ) . detailed morphological analysis of these and other high redshift udf sources has been discussed by pirzkal et al ( 2005 , 2006 ) . of the 39 good lyman break galaxies , 25 meet the giavalisco et al selection criteria directly ( eq . [ giav_eqn ] ) . one of these lies within @xmath34 of the selection region boundary , as do five objects that would narrowly fail the giavalisco et al criteria . thus , the photometric v - dropout criteria have a completeness of @xmath35 compared to the simple @xmath1 cut of our spectroscopic sample . most of the confirmed v - dropout lbgs missed by eq . [ giav_eqn ] appear to be lbgs at a slightly lower redshift : our color - redshift calculations indicate that a star forming galaxy should lie at @xmath36 to meet eq . [ giav_eqn ] , and @xmath37 to meet @xmath1 . these redshift boundaries will of course be blurred by variations in the galaxies stellar populations , dust content , and line strength , all of which have some effect on galaxy colors . in practice , although only three of our spectroscopically measured redshifts actually fall at @xmath38 , there is a reasonable correlation between @xmath39 and redshift , and the inclusion of objects with @xmath40 does lower the mean redshift of the sample ( see figure [ viz ] ) . the reliability of the photometric criteria is broadly comparable . in total , we find 50 objects passing the criteria of eq . [ giav_eqn ] : the 25 confirmed lyman break galaxies discussed above , plus 27 other sources _ not _ confirmed as lyman break objects . among these , 21 had inconclusive ( grade 3 ) spectra , two are stars ( grade-5 ) , and only four are considered `` refuted '' with grade=4 . leaving aside the grade=3 sources , then , some @xmath41 of the objects that meet the photometric criteria are confirmed by inspection of their grapes / pears grism spectra . combining this with the @xmath2 completeness would naively imply that v drop galaxy counts derived directly from eq . [ giav_eqn ] are slightly underestimated , by a factor of @xmath42 . the @xmath43 color criterion in eq . [ giav_eqn ] seems unlikely to strongly affect which galaxies are included in the sample . the reddest galaxy actually selected still has @xmath44 . thus , if there are redshift @xmath0 lyman break objects with @xmath45 , they would have to form a disjoint population from the star - forming sample discussed in the present work . this cut does , however , help exclude l and t dwarfs from the sample . a final point about sample selection is that lyman break galaxies constitute only @xmath46 of those objects _ failing _ the giavalisco et al criteria while still having @xmath1 and @xmath47 . thus , for a purely photometric criterion , eq . [ giav_eqn ] is quite good , and in the absence of spectra we could not advocate a simple @xmath1 selection . expanding the selection region to include objects with @xmath48 and @xmath1 would be better than merely using @xmath1 alone ( since the @xmath49 cut eliminates many interlopers and no confirmed objects in our sample ) . this would increase the sample s completeness , but it would still reduce the overall reliability of the photometric selection when compared to the giavaliso et al selection . we show the redshift distribution of the v dropout sample in figure [ zhist ] . the distribution shows no convincing structure beyond a broad maximum . in particular , we do not see any large scale overdensity akin to the one we reported at @xmath50 based on the grapes @xmath9 dropout sample ( malhotra et al 2005 ) . the observed maximum spans @xmath51 , and can be understood by considering the redshift - dependence of the selection criteria . at @xmath52 , the @xmath39 color is not red enough for the selection . at @xmath53 , reduction of the @xmath9-band flux by the forest is becoming significant . we should expect the observed @xmath39 colors of the lyman break objects to grow redder with increasing redshift , as the forest absorption shifts through the v bandpass . this effect is apparent in figure [ viz ] , though there is considerable scatter at each redshift . with spectra of every lyman break galaxy in the hubble ultra deep field , we can examine the statistics of emission from galaxies in our sample . we have previously published lists of emission line objects in the field ( xu et al 2007 ; straughn et al 2008 ) and a more specific study of the galaxies , focussing on their ages , masses , and morphologies ( pirzkal et al 2007 ; hereafter p07 ) . starting from the galaxy sample of p07 and the v - dropout lyman break galaxy sample in this work , we find that five ( i d numbers 712 , 5183 , 5225 , 6139 , and 9040 ) have previously identified emission lines . there are an additional four emitting galaxies in the p07 sample . three of these ( i d numbers 631 , 9340 , and 9487 ) are too blue to be included in our @xmath1 sample , and the fourth ( i d number 4442 ) is too faint for our @xmath47 criterion . the p07 sample was based essentially on an emission line search combined with a stringent upper limit on @xmath55-band ( @xmath56 nm ) flux . the completeness of the emission line search is treated in xu et al ( 2007 ) . that sample is expected to be quite complete for galaxies that have @xmath57 , @xmath58 , @xmath59 , and @xmath60 . we conclude that @xmath5 of such emitting galaxies also pass our v dropout criteria . while the remaining galaxies are still well detected in the photometric catalog , many fail the dropout color criteria . in addition to these previously identified objects , we find good visual evidence for emission in objects 1392 , 2898 , 3250 , and 6515 . weaker evidence of a possible line is seen in 2408 and 8682 . object 119 also shows a clear emission line , which could be at @xmath61 . however , this line was interpreted in xu et al ( 2007 ) as . ( the ambiguity is linked to the object s faintness , @xmath62 , and we have assigned this object a grade of @xmath63 to reflect the uncertainty in its redshift . ) the detection of new emitters in the present paper is due to two factors . first , we have lower effective detection thresholds in line luminosity and/or equivalent width in the current work , because our line list includes objects identified by visual inspection of both 1-d and 2-d spectra in photometrically pre - selected high redshift galaxies . thus , an emission line does not need as high a statistical significance to enter the present sample as was required in xu et al ( 2007 ) . second , we are using not only the grapes data but also the pears - deep data , nearly doubling the available data . this provides better statistical signal - to - noise ratio on most objects , and improved robustness to contamination by overlapping spectra thanks to the additional four position angles of data . combining all the detected lines , we infer a emission fraction in the range @xmath64 to @xmath65 for the v dropout sample . there are two main ways in which hst grism searches can miss emission from well detected galaxies . first , the emission line is always located at the same wavelength as the continuum break introduced by forest absorption . this imposes a minimum observer - frame equivalent width that is comparable to the effective spectral resolution of the instrument . this spectral resolution , in turn , is set by the angular size of the target , given the slitless instrument . second , emission could come from a `` photosphere '' that is more extended than the star - forming regions that dominate the uv continuum light . there is a plausible physical mechanism for such extended emission , namely , resonant scattering of photons by neutral hydrogen in and around the emitting galaxy . moreover , observations of `` blobs '' up to @xmath66 in size ( e.g steidel et al 2000 ) provide direct evidence that can be scattered or emitted over wide spatial scales , at least in some rare objects . p07 report the effective radii in continuum emission for their sample of nine galaxies . among these , six have sizes corresponding to @xmath67 fwhm , and the remaining three have sizes near @xmath68 . if the typical photosphere is larger ( i.e. @xmath69 in extent ) , some lines would be missed in the slitless hst spectra , because their large spatial extent would translate to very broad line widths . @xmath70 translates to @xmath71 , which in turn corresponds to a 400 observed line width . xu et al ( 2007 ) report a 70% detection completeness for 400 lines , but this completeness drops ( approximately linearly ) to zero as the line width rises to 650 ( corresponding to @xmath72 size ) . to address this possibility , we stacked the 2-dimensional grism images of all 39 v - dropout galaxies , shifting them in wavelength to align the expected location of any emission in all spectra . all position angles were included in the average . the resulting average 2d spectrum is shown in figure [ fig : stack2dv1 ] . if spatially extended flux is prevalent in our sample , we should expect to see a region of diffuse emission at the wavelength . moreover , the profile of this emission would be a fair representation of the `` average '' emission morphology . in practice , we see no evidence for extended emission in this figure . the continuum emission , a break , and weak transmitted flux through the forest region are all clearly detected . we also made a stack of the ten best emitters , and a stack of all the remaining 29 objects . the `` stack '' shows a clear emission line . moreover , this line appears somewhat extended on visual inspection ( figures [ fig : stack2dv1 ] and [ fig : profcomp ] ) . the difference is not highly significant , but a direct measurement of the spatial width of the spectrum yields fwhm=0.19 ( or @xmath73 ) in the continuum region , and 0.26 ( or @xmath74 ) in the line region . as a quantitative test for extended , we have compared the spatial profile of the stacked spectrum ( perpendicular to the dispersion direction ) at both the location and in the adjacent continuum region . we find that the two profiles are indistinguishable ( figure [ fig : profcomp ] ) . thus , while we can not rule out extended emission in a minority of cases , it is at most an exception to the rule , rather than a generic phenomenon in lyman break galaxies . to place a quantitative upper limit on the flux in a spatially extended component , we first measure the @xmath34 noise level in the composite spectrum , which is about @xmath75 count / ksec / pixel . this corresponds to a flux level of @xmath76 in one @xmath77 acs wfc pixel . if we co - add the results for a larger solid angle @xmath78 , we expect the limit to scale as @xmath79 . we define two sets of pixels , each consisting of two @xmath80 rectangular regions and so totaling @xmath81 , and each excluding a strip @xmath82 wide along the trace of the stacked spectrum . the first set of pixels is centered at the expected location of emission , while the second is centered @xmath83 away , towards the red end of the spectrum . we take the total fluxes in these two sets of pixels , apply the wavelength - dependent sensitivity conversion factor , and compare the results . in the region where diffuse might be expected , we find a formal flux excess of @xmath84 . this is comparable to the typical fluxes of ground - based narrowband surveys . we have examined grapes and pears spectra of 216 photometrically selected candidate lyman break galaxies , and spectroscopically confirmed 39 of them . our pre - selection used a wider range of color space ( @xmath1 ) than more traditional `` v - dropout '' lyman break color criteria , in order to assess the completeness of those criteria , along with their reliability . we find that 64% of our confirmed objects meet the traditional criteria . among those galaxies passing the traditional criteria , and having adequate spectra for classification , we find that @xmath3 are confirmed as lyman break galaxies . our v drop sample includes 55% of emitting galaxies previously identified in this redshift range using grapes data ( xu et al 2007 , pirzkal et al 2007 ) . the `` missing '' galaxies are either too blue in @xmath39 ( in part due to the emission line in the v filter ) or too faint in the continuum for our v dropout selection criteria . we also detect 47 additional galaxies not included in the earlier samples . our overall detection fraction is comparable to that in spectroscopic followup of other lyman break surveys ( e.g. , steidel et al 2000 ) . we have examined our stacked 2d spectra for evidence of the diffuse emission that might result from resonant scattering of photons in neutral hydrogen near young galaxies . in the overall stack of 39 galaxies , we find no significant evidence for such emission down to a @xmath85 upper limit of @xmath86 over an @xmath87 region . on the other hand , when we examine the composite spectrum of just those galaxies with individually identified lines , we see a modestly broader spatial profile at the wavelength of than in the adjacent continuum . this is consistent with the possibility that scattering of photons results in a somewhat extended photosphere in these objects . this work has been supported under grant numbers hst - go-09793 and hst - go-10530 . c.g . acknowledges support from nsf - ast-0137927 . l.a.m . acknowledges support by nasa through contract number 1224666 issued by the jet propulsion laboratory , california institute of technology under nasa contract 1407 . we thank the stsci staff in general , and beth perriello in particular , for their assistance with the planning , scheduling , and execution of the grapes and pears programs . beckwith , s. v. w. , et al . 2006 , 132 , 1729 ferguson , h. c. , et al . 2004 , 600 , l107 giavalisco , m. , et al . 2004 , 600 , l103 hathi , n. p. , malhotra , s. , & rhoads , j. e. 2008 , 673 , 686 madau , p. 1995 , 441 , 18 malhotra , s. , et al . 2005 , 626 , 666 malhotra , s. , et al 2008 , in preparation pickles , a. j. 1998 , pasp , 110 , 863 pirzkal , n. et al . 2004 , 154 , 501 pirzkal , n. , malhotra , s. , rhoads , j. , & xu , c. 2005 , baas 207 , 22.16 pirzkal , n. , sahu , k. c. , burgasser , a. , moustakas , l. a. , xu , c. , malhotra , s. , rhoads , j. e. , koekemoer , a. m. , nelan , e. p. , windhorst , r. a. , panagia , n. , gronwall , c. , pasquali , a. , & walsh , j. r. 2005 , apj 622 , 319 . pirzkal , n. , xu , c. , ferreras , i. , malhotra , s. , mobasher , b. , rhoads , j. e. , pasquali , a. , panagia , n. , koekemoer , a. m. , ferguson , h. c. , & gronwall , c. 2006 , 636 , 582 . pirzkal , n. , malhotra , s. , rhoads , j. e. , & xu , c. 2007 , 667 , 49 rhoads , j. e. , et al . 2005 , 621 , 582 spergel , d. n. , et al . 2003 , , 148 , 175 spergel , d. n. , et al . 2007 , , 170 , 377 steidel , c. c. , adelberger , k. l. , shapley , a. e. , pettini , m. , dickinson , m. , & giavalisco , m. 2000 , 532 , 170 straughn , a. , et al 2008 , , submitted . xu , c. , et al 2007 , 134 , 169	we present the faintest spectroscopically confirmed sample of @xmath0 lyman break galaxies to date . the sample is based on slitless grism spectra of the hubble ultra deep field region from the grapes ( grism acs program for extragalactic science ) and pears ( probing evolution and reionization spectroscopically ) projects , using the g800l grism on the hst advanced camera for surveys . we report here confirmations of 39 galaxies , pre - selected as candidate lyman break galaxies using photometric selection criteria . we compare a `` traditional '' v - dropout selection to a more liberal one ( with @xmath1 ) , and find that the traditional criteria are about @xmath2 complete and @xmath3 reliable . we also study the emission properties of our sample . we find that emission is detected in @xmath4 of the sample , and that our broad - band color selected sample includes @xmath5 of previously published line - selected sources . finally , we examine our stacked 2d spectra . we demonstrate that strong , spatially extended ( @xmath6 ) emission is not a generic property of these lyman break galaxies , but that a modest extension of the photosphere ( compared to the starlight ) may be present in those galaxies with prominent emission .
in the next years , the first generation of large interferometric gravitational - wave detectors @xcite should reach a sensitivity good enough to expect the first direct detection of gw signals . in parallel of the experimental work consisting in operating the detectors at their working point with background noises as small as possible , the future data analysis methods are being prepared for the various expected sources of gw , each of them requesting specific tools . yet , the wiener filtering is used in most of these fields due to its optimal characteristics for signals whose time evolutions are known . indeed , it is not only the filter giving the highest signal - to - noise ratio ( snr ) among all the linear ones @xcite but it has also the property of having the lowest false dismissal rate for a given false alarm rate among all filters @xcite . conversely , its main drawback is its poor robustness : as soon as the physical signal and the filter do not match exactly , the snr can be dramatically reduced . even if the searched waveform is analytically computed with high precision , it always depends on a vector of parameters @xmath3 whose values are specific to a given source and thus unknown e.g. its mass , the main frequency of emission ... and whose accurate estimation is indeed a major aim of the data analysis . the set of physically possible values for the vector @xmath3 is the continuous parameter space @xmath4 . a given filter can only be efficient i.e. recovering a large fraction of the snr in a restricted region of this space , called the efficiency area . therefore , many different templates must be used in parallel for a matched filtering procedure , in order to cover the whole parameter space . the lattice must be dense enough to ensure a minimal loss in snr for any real signal whose parameters do not exactly match any of the available filters , while the number of templates @xmath0 must be kept small to limit the computing power needed for the analysis . given these two requirements , it clearly appears that the choice of the set of templates is very important as it has major consequences on both the ultimate filtering performances and its feasibility . on the other hand , tiling @xmath4 is most of the time very difficult as the efficiency area @xmath5 of a filter depends on @xmath3 and on the minimal match @xmath1 see section [ subsection : formalism ] . thus , in the general case , uniform coverage is not possible . in the one - dimensional case ( @xmath6 ) the tiling is easy as the efficiency areas are straight segments of various lengths which can be aligned one after the other one see e.g. @xcite where the matched filtering with gaussian peak templates ( only depending on their width ) is studied . as soon as @xmath7 , one has to cope with overlaps , holes and borders . the problem of computing a set of templates for matched filtering purpose has already been studied for the compact inspiralling binaries search . the gw signal is accurately estimated thanks to various development methods taylor @xcite or pade @xcite post - newtonian ( pn ) expansions and both the estimated snr and the event rate make such events good candidates for a first detection . if the spins are neglected and the orbit circular , the parameter space is two - dimensional ( the two masses of the stars ) and the number of templates can be very high @xcite if the low mass region is included in @xmath4 . using the formalism defined in ref.@xcite , owen and sathyaprakash present a method @xcite to cover this parameter space at the 2 pn order . to solve the question of placing new filters with respect to the previous ones , rectangles inscribed inside the efficiency areas are used instead of the real ellipses , bigger but more difficult to place accurately i.e. without holes . in this paper , a different tiling approach is studied for the case @xmath8 . the main idea is to construct the lattice of filters with an iterative method ensuring a ( rather optimal ) _ local _ template placement . the location of the first ellipse center is arbitrary and chosen by the user . then the procedure goes on and new ellipses increasing the coverage of @xmath4 are added one after the other . when it stops , the tiling of the full parameter space is normally complete @xmath4 fully covered but is certainly not optimal from the _ global _ point of view . as described in section [ section : tiling_details ] , the overlapping in the final configuration can be so important that a large fraction of ellipses are in fact completely covered by other ones and thus useless for the detection purposes , while wasting computing time . therefore , a second step is then started to clean the list of templates by selecting these ellipses and erasing the corresponding filters . in the end , the final set of templates all useful only depends on the position of the first filter in the parameter space from which the locations of all the other ones have been iteratively computed . due to border effects which can not be properly estimated , it is impossible to define the best initial position . so , a last trick to further decrease the number of templates is to merge some sets of templates corresponding to different starting points and to apply to the full list the reduction procedure previously mentioned . as our team is mostly involved in the search for gw bursts ( short gw signals usually lasting a few milliseconds for which waveforms can not be predicted as accurately as for coalescing binaries , emitted by e.g. supernovae or the merging phase of compact stars inspiralling one around the other ) , this algorithm was originally developed to search for damped sine - like signals . indeed , such time - behavior can be seen in a large fraction of supernova explosion gw signals computed numerically see e.g. the waveforms of ref.@xcite and is also expected to occur when excited black holes come back to equilibrium @xcite . the latter case will be used in the following as a benchmark of the tiling method performances ; details are presented in section [ section : damped_sine ] . another interesting feature of this example is that the number of templates remains small even when the allowed mismatch loss of snr @xmath9 is kept very low . thus , it is possible to check the quality of the tiling ( with respect to the prescription originating the tiling method ) and to see how the characteristics of the set of templates evolve when @xmath1 changes . this experience may be very useful for the inspiralling binaries search whose tilings can not be tested so easily yet , they must show similar behaviors . in particular , the most significant result we obtain is that one should be able to reduce the constraint on the template spacing while keeping small the mean false dismissal rate of events , ultimately the only important quantity in the search of rare signals occurring at random time . therefore , a much less numerous set of filters would still be efficient enough , but at a smaller computing cost . after the introduction of some hypothesis and useful notations , the main steps of the tiling method are presented . as a linear filtering consists in correlating the detector output @xmath10 with the corresponding filtering function @xmath11 , one defines the following scalar product to represent the filtering operation : @xmath12 where @xmath13 is the one - sided power spectrum density and the `` @xmath14 '' symbol means fourier transform . the normalization is chosen so that , in case of signal alone , @xmath15 is equal to the snr . if one now assumes that @xmath11 is almost monochromatic signal of frequency @xmath16 , one can show that if the noise spectrum is nearly flat around @xmath16 , eq . ( [ eq : inner_product_frequency ] ) becomes approximately equal to the following equation : @xmath17 as the definition of the ambiguity function involves normalized templates see section [ subsubsection : ambiguity ] , the frequency - depending factor @xmath18 vanishes . thus , representing the filtering operation by a scalar product in the time domain is in fact accurate also with a colored noise for signals with a narrow extension in the frequency domain . this will be approximately the case for the damped sine waveforms see ref.@xcite for a detailed discussion which will be used to test the tiling algorithm in the next sections . the ambiguity function @xmath19 between two templates @xmath20 et @xmath21 is defined by : @xmath22 the templates are normalized : @xmath23 . @xmath19 is thus a measurement of the closeness between two templates . it can also considered as a way to see how well the template @xmath20 can be used to detect the signal @xmath21 : the ambiguity function is the mean fraction of the optimal snr achieved at a given distance in parameter space . following ref.@xcite , if @xmath24 is small enough , the ambiguity function can be approximated by a second order power expansion the first order is null as the expansion is performed around the absolute maximum of @xmath19 . @xmath25 with @xmath26 defining a metric on the parameter space @xmath4 @xcite . finally , one defines the minimal match @xmath1 as the lower bound of the recovered snrs , which means that the loss of the snr due to the mismatch between the template and the signal must be kept below @xmath9 . following ref.@xcite , one can note that with this definition , a of course _ pessimistic _ estimator of the fraction of false dismissals @xmath27 is : @xmath28 for instance , with @xmath29 , one has @xmath30 . this value of @xmath1 is usually found in the literature as the correspondence with this particular value of @xmath27 is easy . this quantity @xmath1 is the only input parameter of the tiling procedure ; it allows one to define precisely the efficiency area @xmath31 of the template @xmath20 by the following equation : @xmath32 the area of @xmath4 including all the vectors @xmath21 which match this inequality is the inner part of an ellipsoid centered on @xmath3 whose proper volume scales as @xmath33 . the average fraction of recovered snr for physical signals with parameters belonging to @xmath31 is at least equal to @xmath1 . tiling the parameter space consists of finding a set of @xmath34 filters @xmath35 so that the union of the surfaces @xmath36 completely covers @xmath4 . the aim is to achieve this task by using as few templates as possible to keep the computing cost manageable . to do so , one has to solve two related questions : * which tiling algorithm to use ? * how to test its quality ? the main problems encountered by any tiling procedure are : overlapping , gaps , areas lost beyond the physical parameter space ... the difference between the ellipsoid and the real border of the efficiency area is not really taken into account : in practical cases for one - step searches , @xmath1 is close to 1 and thus the ellipsoid is assumed to be a good approximation of the ambiguity surface . moreover , the overlapping between close ellipses is an advantage to discard such a problem . in fact , the tests of the sets of templates constructed with the tiling method presented below show that the tilings have no significant holes , even with the choice @xmath37 , which validates a posteriori this hypothesis . clearly , the value of @xmath1 from which the ellipsoid approximation of the efficiency area is no longer valid depends on the precise tiling problem considered . it is well - known that the optimal tiling of an infinite plane by identical disks of radius @xmath38 consists in placing their centers on an hexagonal lattice , separated by a distance of @xmath39 see figure [ fig : optimal_circle_tiling ] . in this way , the overlapping is minimized and the ratio @xmath40 between the sum of the disk areas and the surface effectively covered is @xmath41 this value would increase in any real situation due to the border effect . this property of circular tiling is used by the algorithm on the following way . once the location of a template @xmath42 in @xmath4 has been chosen , one computes its efficiency area @xmath43 . through a simple plane transformation , @xmath43 becomes a circle @xmath44 of unit radius and the six neighbor centers are placed on the regular hexagonal lattice previously described . this choice is nearly optimal if the shape of the efficiency areas is slowly varying with respect to their characteristic dimensions . using the inverse transformation , the six new center positions are given in @xmath4 . then , each ellipse associated with a new center is tested in order to check if it covers a part of @xmath4 not already covered by the previously defined ellipses . if the ellipse is useful , its center is added to the center list and will be used later to place other centers . for example , we have found that with @xmath29 only about 20% of the centers are kept . this iterative algorithm stops either when the full surface of the parameter space is covered or when no new ellipse can be placed any longer . a successfully completed coverage computed with the former procedure can be usually redundant in the sense that a large fraction of ellipses are completely covered by others . to save computing time , useless templates need to be identified and deleted . like for the first step of the algorithm , one has to face the problem that erasing an ellipse is a local operation which can have global consequences : which template should be discarded first ? to answer this question , the following procedure has been set : * for each ellipse @xmath45 belonging to the tiling , one defines its _ utility _ @xmath46 which is the fraction of @xmath4 it covers alone . useless ellipses are thus characterized by @xmath47 . * among those ellipses , the one with the smallest area in the parameter space is dropped . this prescription is consistent with the idea that it is a priori better to keep big ellipses which have a higher potential of coverage . of course this general rule may not always be true but it looks reasonable . * utilities are then updated for the surrounding ellipses as they become more `` useful '' @xmath46 locally increases . then , this scheme is iterated until all the remaining ellipses have non zero utilities . this second step of the tiling algorithm is very important : for most of the simulations in the damped sine case , the number of templates is reduced by a factor of about two with respect to the first list . finally , the output of the whole procedure is a set of templates completely covering the parameter space and which depends on only one initial condition , that is the location of the first filter set at the beginning of the tiling generation . once more , this choice may have global consequences on the tiling quality for instance , shifting it horizontally/vertically may cause a column / line of templates to appear or vanish but can not be simply optimized . therefore , tilings starting at different points on the parameter space are computed in parallel ; then , all the lists of templates are merged in a single one . finally , the cleaning procedure is applied to this clearly redundant set of filters to obtain the final lattice of templates . thanks to this last step , its size is decreased by 10 or 15% with respect to the individual coverages in the damped sine case . in the damped sine case , five different tilings have been merged together to compute the final lattice of templates . there seems to be a good compromise between gain in the number of templates and computation time ( more than 80% of the ellipses of the full list are useless and so the cleaning procedure is much longer ) . this last merging step doubles the total computation time of the tiling , i.e. its duration is of the same order of magnitude as the sum of the computation times for the initial tilings . three variables can be used to control the quality of the tiling . * the number of templates @xmath0 . * the ratio @xmath48 between the sum of all the ellipse areas and the parameter space surface . * the ratio @xmath49 between the sum of all the ellipse areas inside @xmath4 and the parameter space surface . one has clearly @xmath50 . these two estimators allow to measure the overlapping between templates and the fraction of extra area outside the parameter space , while @xmath0 can be compared with its estimation computed by integrating the proper volume of the parameter space . due to the imperfections of the real tiling ( overlapping , border effects ... ) and to the fact that the analytical computation of the template numbers is only approximative by principle , it is interesting to check the consistency of the two numbers . an excited black hole , born e.g. after a supernova collapse or the merging of two compact objects , comes back to a stationary state by emitting gw . this emission can be described as a superposition of black hole quasi - normal modes @xcite . the dominant mode is expected to be quadrupolar with the longest damping time @xcite . in order to limit the number of free parameters defining the matched filtering parameter space , one assumes that after some transitory phase the waveform becomes ( with a proper choice of time origin ) : @xmath51 it was observed @xcite that a given couple @xmath52 is connected to one single set of physical parameters @xmath53 , the mass and the reduced angular momentum of the black hole . moreover , the corresponding relation can be expressed analytically with a 10% precision @xcite . introducing the quality factor @xmath54 , one has in geometrical units ( @xmath55 ) : @xmath56\end{aligned}\ ] ] the normal mode frequency is , as expected , a decreasing function of the black hole mass , increasing with the rotation parameter as the quality factor , which is also independent of @xmath57 . therefore , detecting such gw signal would give direct physical information on its source . following ref.@xcite , two different methods estimating the optimal snr give identical results : * a calculation in the direct space assuming that the noise is white at the oscillation frequency ; * computing the fourier spectrum and approximating it by a dirac at the oscillation frequency . one has finally @xmath58 where @xmath13 is the one - sided power spectrum density of the noise . the proportionality constant depends on the physical process at the origin of the black hole formation . for high mass inspiralling binaries , it can be very large @xcite and leads to detections at cosmological distances , while for supernova collapses , it can be high enough to be detected in the local group @xcite . using this framework , it is sufficient to study the two quadratures and to compute the corresponding sets of templates covering the two dimensional parameter space @xmath59 . as in the black hole oscillation case , no correlation is assumed between the oscillation frequency @xmath60 and the quality factor @xmath61 . therefore , @xmath4 is rectangular : @xmath62 \ ; \times \ ; \left [ f_{\text{min } } \ , ; \ , f_{\text{max } } \right ] \ ] ] its borders are chosen in the following way : * for q , by using the black hole normal mode range + @xmath63 * for f , by using the interferometric detector main characteristics + @xmath64 the mismatch between two templates of parameters @xmath65 and @xmath66 ) can be written ) there is a factor 2 between the ellipse coefficients and the metric ones : @xmath67 , @xmath68 and @xmath69 . ] @xmath70 with @xmath9 being the maximal expected loss in snr . the dimensionless coefficients @xmath71 , @xmath72 and @xmath73 only depend on the quality factor @xmath61 . their expressions are the following : * for pure damped cosines : @xmath74 * for pure damped sines : @xmath75 in the former case , assuming that the quality factor is much greater than 1 , the first order expansion of these coefficients in power of @xmath61 gives the results computed in ref.@xcite . the hypothesis @xmath76 is not valid in the total range @xmath77 $ ] we have considered in this paper and the difference between the real coefficients and their first order approximations can reach 10% in the region of low quality factor . the number of templates needed to cover the parameter space can be roughly estimated by the usual formula : @xmath78 where @xmath79 is the template proper volume and @xmath80 is the metric determinant . as @xmath81 , we can express this number with the help of @xmath1 : @xmath82 from equation ( [ eq : nb_templates ] ) , it is possible to estimate the numbers of templates @xmath0 needed to cover the parameter space . one gets at the first order in q : @xmath83 for the two families of signals studied . the number of templates is logarithmic in the frequency band and scales linearly in the quality factor range . with the numerical border values previously defined , one gets @xmath84 . the expressions giving the coefficients @xmath71 , @xmath72 and @xmath73 clearly show a dependence on the ellipse location in the parameter space . using the vocabulary of differential geometry , this is a strong indication that @xmath4 is curved . following ref.@xcite , it is possible to verify this by calculating the curvature of the search manifold . as the waveform is a function of the parameters @xmath85 only , we have a two dimensional problem and can thus calculate the gaussian curvature , @xmath86 , directly . following the petrov classification , the riemann tensor @xmath87 associated with the metric defined in eq . ( [ eq : metric ] ) has only one independent component , which we can take to be @xmath88 , in its fully covariant form . we can thus define the gaussian curvature on the search manifold by @xmath89 where @xmath80 is the metric determinant . for the damped sine case for instance , the gaussian curvature is constant over @xmath90 : @xmath91 . therefore , the search manifold has the same topology as a 2-sphere . as the template tilings appear to be very similar for the damped cosine and sine signals compare for instance the expressions giving the coefficients @xmath71 , @xmath72 and @xmath73 , we mainly focus on the latter in the following . [ cols="^,^,^,^,^",options="header " , ] table ix : template numbers and fraction of events lost for tilings with different values of the minimal match parameter @xmath1 table ix shows the results achieved for the case of sine filters with @xmath1 values between 85% and 97% . the evolution of the tiling characteristics with respect to @xmath1 can also be seen on figure [ fig : tiling_characteristics ] . for each value of the minimal match , a procedure similar to the @xmath29 case has been followed : computation of 5 different tilings corresponding to different starting points and then merging of the template lists to reduce the number of filters needed . the first thing to note is that @xmath0 clearly decreases with @xmath1 and follows quite perfectly the expected behavior ( [ eq : nb_templates ] ) . with respect to the theoretical number of templates , @xmath0 is slightly larger , due to border effects . yet , they affect in the same relative way the tilings corresponding to different values of the minimal match and so the relation @xmath92 is preserved . the top graph of figure [ fig : tiling_characteristics ] shows this relation ; the line corresponds to the best linear fit computed with the available data . one gets : @xmath93 the slope of the function is of the same order of magnitude as the estimated value given by eq.[eq : n_owen ] : @xmath94 . the second result and perhaps the most important is that even with @xmath95 the estimated fraction of lost events is smaller than the expected 10% , as shown on the bottom graph of figure [ fig : tiling_characteristics ] . with a template number reduced by a factor of 3 , the parameter space coverage remains well inside the initial specification : @xmath96 . it is only with the choice @xmath37 that @xmath97 slightly exceeds 10% , but with a reduction of a factor 4 in the number of templates . numerically , one obtains @xmath98 this feature is of small importance here as the total number of templates @xmath0 is only a few hundred but would be essential for the compact inspiralling binaries searches . in this case , for a total mass of two solar masses , the number of templates is expected to be close to @xmath99 or even more depending on detector bandwidth see e.g. ref.@xcite . any tiling method based on the minimal match criterion like the one developed in this paper should provide a set of templates able to recover on average a fraction of snr much better than @xmath1 the _ minimal _ match indeed , and thus miss less events than assumed during the building phase of the lattice . therefore , by using a monte - carlo study of the generated tiling performances , it should be possible to decrease by a large fraction the number of templates really needed to guarantee that for instance 90% of the events potentially detectable will be seen . such a refinement appears to be a way to save a large amount of computing time and would simplify any data analysis strategy based on matched filtering methods . as known since many years , mean losses in snr recovery are well below @xmath9 , which indeed implies that the mean fraction of lost events is much smaller than @xmath2 . in the previous section , the evolution of @xmath100 was shown to be also well - fitted by a power law @xmath101 , with the exponent @xmath102 . it would be interesting to see if such dependences could be analytically foreseen , without any simulation and without any assumption on the tiling problem . let us consider first the simpler case of a single parameter @xmath103 ( @xmath6 ) . one then assumes that the match is well approximated by its quadratic expansion around a given template located at @xmath104 . thus , the efficiency area @xmath5 is bounded by the equation : @xmath105 with @xmath1 the minimal match and @xmath106 the metric coefficient . @xmath5 is a segment @xmath107 $ ] with @xmath108 . assuming a uniform distribution of physical signals in this area , it is easy to compute the mean fraction of recovered snr : @xmath109 with @xmath29 , one gets @xmath110 which is very close to the value of @xmath111 achieved in our simulations . it is worth noting that the result is independant of @xmath106 ( and so of the particular efficiency area considered ) , which validates the assumption of computing the mean value of the fraction of snr recovered . one can estimate @xmath112 in the same way : @xmath113 the choice of @xmath29 gives @xmath114 , while with @xmath115 , one has @xmath116 . these values exceed the results of the numerical simulations ( respectively 2.6% and 10.7% ) but are quite comparable . how do these results change with a two - dimensional parameter space @xmath117 which is indeed the situation considered in this article ? by a proper choice of coordinates , one can like for the 1d - case assume that the chosen template is located at @xmath118 and that the efficiency area is defined by the equation of an ellipsoid written in its simplest form : @xmath119 straightforward calculations allow one to compute both the mean fraction of recovered snr @xmath120 and the mean loss of events @xmath112 : @xmath121 like for the one - dimensional case , these expressions do not depend on the particular template and are thus estimators of the mean values of these quantities . with @xmath29 , one gets values higher than those computed numerically : @xmath122 and @xmath123 respectively . the discrepancy between these numbers and those given in the case @xmath6 is higher , even if the tiling studied here is 2-dimensional . the effect of the overlap between close templates for @xmath124 has been checked with the fully simplified model presented in figure [ fig : optimal_circle_tiling ] : efficiency areas are identical disks of a given radius . in this case , a more accurate expression of @xmath125 can be computed analytically by selecting for each physical signal the closest template , i.e. the one allowing to recover the highest fraction of snr . but the gain is very small : the fraction of false dismissals decreases only by about 1 - 2% and remains in all cases higher than for the 1d - model . in fact , one could not expect much more from this refinement . figure [ fig : optimal_circle_tiling ] shows that only a fraction ( @xmath126 ) of the physical signals are affected by this improvement and that these particular signals have the worst matches . finally , it appears that the good performances in detection efficiency of the template bank are the consequence of two aspects : firstly , the parameter space considered in this study is much more complex than the ideal case of a parameter space where one could use a uniform lattice of templates , and secondly the generated tilings remain redundant despite all the steps used to reduce it as much as possible . there is certainly still room for future improvements for the geometrical tiling algorithm . the comparison of the numerical data with the analytical parameterization computed for the model @xmath124 ( 2d ) is also shown on figure [ fig : tiling_characteristics ] . on this plot , it is clear that it overestimates the snr losses with respect to the simulated tiling performances ; yet , it gives values closer to real data than the @xmath2 curve . thus , the fractions of false dismissals are also higher . this paper presents a method to tile any two dimensional parameter space @xmath4 with a given maximal loss in snr . the template list is build in two steps : the first one is an iterative algorithm which provides a complete coverage of @xmath4 without any area left uncovered . as the computed set of filters can be redundant , a second step including two cleaning steps allows one to significantly decrease the number of templates by dropping those which are useless . this algorithm has been used for the lattice of templates needed to look for damped sine signals with matched filtering . even if the number of filters involved is by some orders of magnitude below those computed for the inspiralling binaries search , this choice of waveform allowed us to strongly test the capability of the algorithm as the ellipse characteristics show large variations in @xmath4 or , equivalently that @xmath4 is curved . moreover , the algorithm using these templates can be directly implemented in an on - line filtering . as it is not possible to choose the optimal tiling of a given parameter space , the final configuration of filters is computed by mixing various sets of templates computed with different initial conditions ( the location of the initial filter used to develop the iterative algorithm ) and by applying to the global list the reduction procedure . this kind of average decreases the number of filters by about 15% with respect to an unique procedure . the number of templates finally computed is comparable to the analytical estimations and monte - carlo simulations show that the set of filters fulfills the initial requirement : minimizing the loss in snr in the whole parameter space . in the damped sine case , the study of the sets of templates shows that the effective loss of events @xmath97 ( as a function of @xmath1 ) is much less than its usual estimation : @xmath97 approximately behaves as @xmath127 instead of @xmath128 . with the usual prescription of 10% for the loss , the number of templates can be decreased by a factor close to 4 . seen in a reverse way , for a given @xmath1 prescription , the snr fraction recovery is on average much higher than @xmath1 : 99.1% for @xmath29 as an example . even if the exponent value depends on the specific waveform details , such a behavior should be expected for the inspiralling binary case . this feature can allow us either to decrease the number of templates ( and then the computing power ) or perhaps obtain a more robust ( with respect to noise for example ) tiling of the parameter space . finally , simple parameterizations allowing one to predict the dependence in @xmath1 of the mean fraction of recovered snr @xmath120 and of the false dismissal fraction @xmath97 are presented . they are derived independently of the particular tiling method studied in this paper and should thus be generic . they do not give the power - law scaling of @xmath97 inferred from the results of our numerical simulations and they both overestimate the losses due to the finite lattice of templates w.r.t . the tilings we generated . it would be interesting to compare this situation with different tilings generated with other algorithms . t. zwerger & e. mller , _ astron . _ * 320 * 209 ( 1997 ) . + h. dimmelmeier , j.a . font , and e. mller _ astron . _ * 388 * , 917 , ( 2002 ) . + h. dimmelmeier , j.a . font , and e. mller _ astron . _ * 393 * 523 ( 2002 ) .	searching for a signal depending on unknown parameters in a noisy background with matched filtering techniques always requires an analysis of the data with several templates in parallel in order to ensure a proper match between the filter and the real waveform . the key feature of such an implementation is the design of the filter bank which must be small to limit the computational cost while keeping the detection efficiency as high as possible . this paper presents a geometrical method which allows one to cover the corresponding physical parameter space by a set of ellipses , each of them being associated to a given template . after the description of the main characteristics of the algorithm , the method is applied in the field of gravitational wave ( gw ) data analysis , for the search of damped sine signals . such waveforms are expected to be produced during the de - excitation phase of black holes the so - called ringdown signals and are also encountered in some numerically computed supernova signals . first , the number of templates @xmath0 computed by the method is similar to its analytical estimation , despite the overlaps between neighbor templates and the border effects . moreover , @xmath0 is small enough to test for the first time the performances of the set of templates for different choices of the minimal match @xmath1 , the parameter used to define the maximal allowed loss of signal - to - noise ratio ( snr ) due to the mismatch between real signals and templates . the main result of this analysis is that the fraction of snr recovered is in average much higher than @xmath1 , which dramatically decreases the mean percentage of false dismissals . indeed , it goes well below its estimated value of @xmath2 used as input of the algorithm . thus , as this feature should be common to any tiling algorithm , it seems possible to reduce the constraint on the value of @xmath1 and indeed the number of templates and the computing power without loosing as much events as expected in average . this should be of great interest for the inspiralling binaries case where the number of templates can reach some hundreds of thousands for the whole parameter space . = 2
the heavy elements ( z@xmath130 ) are mainly produced by neutron - capture processes @xcite . the slow neutron - capture process ( s - process ) contains two categories . the weak s - process mainly produces the lighter neutron - capture elements and takes place in the massive stars @xcite . the main s - process occurs in the agb stars and is directly responsible for the heavier elements @xcite . the rapid neutron - capture process ( r - process ) also contains two categories . the main r - precess could occur in sne ii with 8 - 10 @xmath2 progenitor and mainly produces the heavier elements @xcite . however the weak r - process might take place in the sne ii with 11 - 25 @xmath2 progenitors . the sne ii primarily produce the weak r - process elements and eject the light and fe - peak elements @xcite . globular clusters are the living fossils to study the galactic evolution history since their extremely ancient age . historically palomar 1 ( pal 1 ) was thought as a globular cluster , since its location is higher than the galactic plane @xcite . however , the red giants in pal 1 , whose averaged metallicity [ fe / h ] is about -0.6 , show so metal - rich for a globular cluster of the outer halo @xcite . based on the study of the color - magnitude diagram of pal 1 , @xcite suggested its age lies the range of 6.3 - 8 gyr . in this case , pal 1 was regarded as the youngest globular cluster . recently , using high - resolution spectra , @xcite derived chemical abundances for four red giants in pal 1 . they found that element abundances of pal 1 are unusual comparing to those of the disk stars . the abundances of mg , si , ca , ti , co , v and y of pal 1 are lower than those of the disk stars with similar metalliciy . however the abundances of ba and la are higher than those of the disk stars . @xcite suggested that pal 1 is not the traditional gc and should be accreted from a dwarf galaxy by our galaxy . the unusual abundances should imply that pal 1 should have an unusual formation environment and history . nearby galaxies should have complex formation histories and chemical evolution process @xcite . they should have a close relation to our galaxy . for example , some stars and clusters in the galaxy should originate from the dwarf galaxies . in this case , these stars retain the chemical signatures of its progenitor system . the chemical signatures , such as lower [ @xmath0/fe ] ratios , and star formation histories of nearby dwarf galaxies have been studied for many years @xcite . historically , the explanation of low [ @xmath0/fe ] ratios in dwarf galaxies is that the light elements are deficient , since a low star formation rate ( sfr ) and extra fe from sne ia @xcite . based on the detailed abundance analysis , @xcite concluded that the @xmath0-element deficiencies in the sagittarius ( sgr ) dwarf spheroidal galaxy result from an top - light initial mass function ( imf ) , which is relatively deficient in the highest mass stars . @xcite studied the effects of galaxy mass on stellar imf . they found that the integrated galaxy imf varies from top - light to top - heavy in dependence of galaxy type . the top - light imf should be associated to the dwarf galaxies because the galaxies have not sufficient massive clouds to form the highest - mass stars @xcite . on the other hand , the imfs seem to vary among more massive early type galaxies @xcite . @xcite found that more massive early type galaxies are associated to bottom - heavy imf , because many early supernovae drive up the supersonic turbulence in the giant clouds , which leads to formation of more low mass stars . it is important to note that the abundances of individual stars of pal 1 present the accumulated effects from the time of the progenitor - system formed to the time of pal 1 formed . for purpose of investigating the complex formation mechanism and history of pal 1 , it is necessary to study the elemental abundances completely , containing light elements , fe - peak elements and heavy elements . in this paper , using the abundance approach given by @xcite , we investigated the astrophysical origins of chemical abundances of pal 1 and the disk stars . the results and discussions are given in section 2 . our conclusions are presented in section 3 . it is indicated that the stellar abundances could not be explained by single astrophysical reason @xcite . for exploring the origins of light and heavy elements in pal 1 , we adopt the abundance approach from @xcite . the abundance of element i should be express as : @xmath3},\ ] ] where @xmath4 , @xmath5 , @xmath6 , @xmath7 and @xmath8 are the abundances of the main r- , the main s- , the primary , the secondary process and sne ia , respectively , which have been scaled to the corresponding abundances of solar system . @xmath9 , @xmath10 , @xmath11 , @xmath12 and @xmath13 are the corresponding component coefficients . the abundance @xmath14 is the number of atoms of element i which is scaled to @xmath15 si atoms of the solar abundances . the primary light elements and fe - peak elements are produced in the massive stars ( @xmath16 ) during hydrostatic burning and the weak r - process elements were produced in the sne ii with progenitor mass @xmath16 . because the primary light elements , fe - peak elements and the weak r - process elements are ejected from the massive stars , these elements could be combined as one component , which is called the primary component " . on the other hand , the secondary light elements and fe - peak elements are produced in the massive stars ( @xmath16 ) during hydrostatic burning and the weak s - process elements are produced in the massive stars during core he burning and shell c burning . the yields of weak s - process also have the secondary nature . because the secondary light elements , fe - peak elements and the weak s - process elements come from the massive stars , these elements could be combined as secondary component " @xcite . note that the abundances of secondary elements are observable only for higher metallicity , since their yields decrease with decreasing metallicity . in this case , the primary component @xmath6 contains the abundances of primary elements and the weak r - elements . the secondary component @xmath7 contains the abundances of the secondary elements and the weak s - elements . both primary and secondary components are produced in the massive stars . the abundances of the secondary elements and the weak s - elements are adopted from @xcite and @xcite respectively . the abundances @xmath4 and @xmath6 are adopted from @xcite . the abundances @xmath8 are adopted from @xcite and updated the fe , cu , and zn values from @xcite . @xcite reported that the sne ia events could not occur for the progenitors with low metallicity [ fe / h]@xmath17 . this implies that the sne ia events only occur on the higher metalllicy . in this work , we suppose that the abundance pattern produced by sne ia do not vary with metallicity , because the yields of sne ia depend on metallicity weakly @xcite . in this case , the adopted abundance pattern produced by sne ia can be thought as the average abundance pattern produced by sne ia . from the observations , @xcite reported that compared to the disk stars , the [ ba / y ] values are apparently high , which indicates that the pal 1 stars were contaminated by the low - metallicity and low - mass agb stars . in these agb stars , the s - process prefers heavier s - nuclei rather than lighter s - nuclei , because an iron - seed nucleus can capture more neutrons @xcite . obviously , the different abundance patterns between pal 1 and the disk stars attribute to the different chemical characteristics between pal 1 and the disk stars . the abundances @xmath5 in equation ( 1 ) are adopted from the results with [ fe / h]=-0.6 presented by @xcite for pal 1 . by comparison , for the disk stars , the abundance pattern @xmath5 in equation ( 1 ) are adopted from the results with [ fe / h]=-0.6 calculated by @xcite ( see their figures 6@xmath1812 ) . equation ( 1 ) contains five components . the corresponding component coefficients can be derived by looking for the minimum @xmath19 . adopting the average abundances in pal 1 stars @xcite and the average abundances of the disk stars with [ fe / h]=-0.6 @xcite , we can obtain the best - fit coefficients . for the abundances of the solar system , @xmath20 . furthermore , we can compare the component coefficients of pal 1 with the corresponding coefficients of the disk stars to investigate the astrophysical reasons of unusual abundances in pal 1 stars . the calculated results for pal 1 and the disk stars are shown in figure 1 ( a ) and ( b ) . in top panels , the filled circles indicate the observed abundances and the solid lines represent the fitted results . in bottom panels , the individual relative offsets ( @xmath21 ) and standard calculated errors for pal 1 and the disk stars are shown , respectively . it is confirmed from the figure 1 ( a ) and ( b ) that the calculated results is validity . in figure 2 , the values of the component coefficients @xmath9 , @xmath11 , @xmath10 , @xmath12 , @xmath13 and the associated errors are displayed for pal 1 and the disk stars , respectively . from the figure we can see that the values of @xmath10 , and @xmath13 of pal 1 are higher than those of the disk stars , which implies that the contributions from the two components to the abundances of pal 1 are larger than those of the disk stars . on the other hand , the value of @xmath11 of pal 1 is lower than that of the disk stars , which implies that the contribution of the component to pal 1 are smaller than that to the disk stars . furthermore , the values of @xmath12 and @xmath9,of pal 1 are close to those of the disk stars . the abundances of each component can be derived quantitatively using corresponding component coefficient . since the various categories of elements are produced in different astrophysical sites , the element abundances of pal 1 should be directly effected by imf of its progenitors system . we can study the characters of the imf through the contributions of different processes to the element abundance of pal 1 . the primary component and secondary component are produced by the massive stars with m@xmath22 . on the other hand , the main r - process should take place in the sne ii with 8@xmath23 progenitors @xcite and the main s - process mainly occurs in the 1.5@xmath24 agb stars @xcite . based on the calculated results , the coefficients ratios of pal 1 to the disk stars for primary process , secondary process , main r - process , sne ia and main s - process are about 0.6 , 1.0 , 1.2 , 3.0 and 4.5 respectively . the ratios have a trend that increase with decreasing progenitor mass , which should be an evidence that the imf of pal 1 s progenitor - system ( e.g. , dwarf galaxy in which cluster pal 1 formed ) is top - lighter than that of the galaxy . note that the coefficient ratios of pal 1 to the disk stars are determined by their imfs dominantly and the coefficient ratios should depend on their sfrs weakly . if the imf of pal 1 is similar to that of the galaxy , the coefficient ratios of pal 1 to the disk stars should be close to a constant . so the unusual abundances of pal 1 are mainly due to the top - light imf of its progenitor system . recently , the signatures of the top - light imf in the elemental abundances of the fornax dwarf galaxy and the sagittarius dwarf galaxy have been reported @xcite . the results imply that the imf of pal 1 s progenitor - system have the similar characters with that of the dwarf spheroidal galaxies mentioned above . our approach is based upon the measured abundances of the sample stars , thus the associated uncertainties would be included in the calculations . taking the component coefficient of main s - process of pal 1 as an example , we discuss the uncertainties of the coefficients . adopting @xmath9=3.5 , @xmath11=1.2 , @xmath12=1.1 and @xmath13=0.9 , the top panel in figure 3 shows the calculated ratios [ ba / h ] as a function of the component coefficients @xmath10 . there is only a range of the component coefficients in figure 3 , @xmath25 , in which the calculated ratios fall into observed limits of [ ba / h ] . the bottom panel in figure 3 illustrates the reduced @xmath26 values are the function of the component coefficients @xmath10 . there is a minimum of @xmath27 at @xmath10=1.3 with a @xmath28 error bar , which indicates that the component coefficient is constrained well . adopting this approach , we derived the errors of other component coefficients : @xmath29 , @xmath30 , @xmath31 and @xmath32for pal 1 . @xmath33 @xmath34 , @xmath35 , @xmath36 and @xmath37 for the disk stars . the calculated errors of component coefficients are shown in figure 2 . in this case , the calculated abundance errors should be close to the observational errors , since the component coefficients are restrained by the observational errors . the average observed errors of all elements for the pal 1 stars and the disk stars are about 0.17 and 0.07 dex , respectively @xcite . in this work , for simplicity , we take the average observed errors as the standard calculated errors of elemental abundances for the two category stars . because different elements in pal 1 came from different astrophysical sites on different epoches , the abundances of individual stars present the accumulated effects from the time of the progenitor - system formed to the time of pal 1 formed . in order to reveal the astrophysical reasons of unusual abundances of pal 1 , we show the component abundance ratios of fe , mg , co , y and ba in pal 1 and compare them with the disk stars in figures 4 - 8 . in the figures , the filled pentagons represent the calculated abundance ratios and the associated error bars represent the calculated errors . from figure 4 , we know that for the disk stars , the fe element mainly produced by primary process . although the observed fe abundances of the disk stars are similar to those of pal 1 , the abundances of the primary and sne ia components of the disk stars are obviously different from those of pal 1 . for the pal 1 , the fe abundance mainly originate from the sne ia . the calculated results imply that the primary fe abundances of pal 1 are lower than those of the disk stars , which agree in a top - lighter imf of pal 1 s progenitor - system . @xcite reported that the observed mg abundances of pal 1 are lower than those of the disk stars by about 0.4 dex . from figure 5 we know that for the disk stars and pal 1 , although the mg abundances dominantly produced by the primary process , the primary mg abundances of pal 1 are smaller than those of the disk stars . this should mean that the proportion of the massive stars in pal 1 s progenitor - system is smaller than that of the massive stars in the galaxy . the observed unusual abundances of other @xmath0 elements in pal 1 could be also explained by the top - light imf of pal 1 s progenitor - system . @xcite showed that abundances of some fe - peak elements , such as v and co of pal 1 are smaller than those of the disk stars . taken co as an example , figure 6 shows that , for the disk stars and pal 1 , although the co abundances mainly originate from the primary and secondary components of the massive stars , the contributions of the massive stars to pal 1 is lower than those of the massive stars to the disk stars , which also agree in the top - light imf of pal 1 s progenitor - system . the lower abundance of v in pal 1 could be explained by the similar reason . it is indicated that the y abundances mainly originate from the weak r - component for the disk stars in figure 6 . however , for pal 1 , the y abundance originate from three components : main r- , weak r- and main - s components in the same figure . @xcite have found that the y abundances of pal 1 are lower than those of the disk stars about 0.5 dex . from the figure we know that the main reason of the lower y abundance is that the abundance of weak r - component of pal 1 is smaller than those of the disk stars . the abundances of s - process component for pal 1 should be mainly contaminated by the low metallicity agb stars with 1.5 - 3 @xmath2 . in these agb stars , the s - process prefers heavier s - nuclei ( e.g. , ba , la ) rather than lighter s - nuclei ( e.g. , sr , y ) , because an iron - seed nucleus can capture more neutrons @xcite . this should be another reason of the lower y abundances in pal 1 . as figure 7 shows , for element y , the observed abundance including error bar falls into the calculation error limits . from figure 8 , we can see that the ba abundances mainly originate from the main r - component for the disk stars . on the other hand , for pal 1 , the ba abundance mainly originate from the main s - component . @xcite reported that the ba abundances of pal 1 are larger than those of the disk stars . from the figure we know that the main reason of higher ba abundance is that the ba abundance of main s - component of pal 1 is higher than those of the disk stars , which should attribute to the top - light imf for the pal 1 s progenitor - system . the element abundances of pal 1 contained a great deal of information of element nucleosynthesis and evolution history . in this paper , using the abundance - decomposed approach , we explore the astrophysical reasons of the unusual observed abundances in pal 1 stars . we found that the component coefficients of main s - process and sne ia of pal 1 are higher than those of the disk stars . however , the component coefficient of primary process of pal 1 is lower than that of the disk stars . the results should imply that the imf of pal 1 s progenitor - system is top - lighter than that of the galaxy . the fe abundances mainly produced by primary process for the disk stars and the fe abundances mainly originate from the sne ia for pal 1 stars . although the @xmath0-element abundances dominantly produced by the primary process for the disk stars and pal 1 , the contribution of primary component of the massive stars to pal 1 is lower than those of the massive stars to the disk stars , which should be the astrophysical reason of the observed low-@xmath0 abundances in pal 1 . the fe - peak elements v and co mainly produced in the massive stars for the disk stars and pal 1 , but the contributions of the massive stars to pal 1 is lower than those of the massive stars to the disk stars , which might be the astrophysical reason of the observed low abundances of v and co in pal 1 . that the y abundances of weak r - component for pal 1 are smaller than that of weak r - component for the disk stars should be the main reason of the low y abundances in pal 1 stars furthermore , that the ba abundances of the main s - component for pal 1 are higher than that of main s - component for the disk stars is the astrophysical reason of the high ba abundance in pal 1 stars . as a whole , the observed unusual abundances in pal 1stars could be explained by the top - light imf for the pal 1 s progenitor - system . our results obtained from the abundance - decomposed approach can provide more constraints on the astrophysical environment of pal 1 . overall , we hope our work can give a beneficial guidance for judging chemical origin of pal 1 . of cause , more observed elemental abundances about the stars of pal 1 are needed to investigate the relation between pal 1 and a dwarf galaxy . we thank the anonymous referees for insightful comments which improved this paper greatly . we also thank dr jianrong shi for the careful review of the manuscript and for fruitful discussion . this work has been supported by the national natural science foundation of china under 11273011 , u1231119 , 10973006 and 11003002 , the natural science foundation of hebei provincial education department under grant z2010168 , xjpt002 of shijiazhuang university , the natural science foundation of hebei province under grant a2011205102 , a2011210017 , and the program for excellent innovative talents in university of hebei province under grant cprc034 . allen , d.m . , barbuy , b. 2006 , a&a , 454 , 917 arlandini , c. , k@xmath38ppeler , f. , wisshak , k. , gallino , r. , lugaro , m. , busso , m. , & straniero , o. 1999 , , 525 , 886 burbidge , e. m. , burbidge , g. r. , fowler , w. a. , & hoyle , f. 1957 , rev . phys , 29 , 547 busso , m. , gallino , r. , wasserburg g. j. 1999 , , 37 , 239 busso , m. , gallino , r. , lambert , d.l . , travaglio , c. , smith , v.v . 2001 , astrophys . j. 557 , 802 cowan , j. j. , thielemann , f .- k . , truran , j. w. 1991 , physrep , 208 , 267 conroy , c. , & van dokkum , p. g. 2012 , , 760 , 71 harris , w.e . 1996 , , 112 , 1487 ( 2010 edition ) iwamoto , k. , brachwitz , f. , nomoto , k. , et al . 1999 , , 125 , 439 kobayashi , chiaki . , tsujimoto , takuji . , nomoto , kenichi . , hachisu , izumi . , & kato , mariko . 1998 , , 503 , l155 kroupa , pavel . , weidner , carsten . , pflamm - altenburg , jan . , thies , ingo . , dabringhausen , jorg . , marks , michael . , maschberger , thomas . 2013 , pss5.book , 115 lanfranchi , g. a. , matteucci , f. , & cescutti , g. 2008 , a&a , 481 , 635 lamb , s. , howard , w. m. , truran , j. w. , iben , i. 1977 , , 217,213 li , h. j. , shen , x. j. , liang , s. , cui , w. y. , zhang , b. 2013a , , 125 , 143 li , h. j. , cui , w. y. , zhang , b. 2013b , , 775 , 12 mcwilliam , andrew . , wallerstein , george . , mottini , marta . 2013 , , 778 , 149 mishenina , t. v.,et al . , kovtyukh , v. v. , soubiran , c. , travaglio , c. , busso , m. 2002 , a&a , 396 , 189 oey , m. s. 2011 , , 739 , l46 reddy , b.e . , tomkin , j. , lambert , d.l . , allende prieto , c. 2003 , , 340 , 304 reddy , b.e . , lambert , d.l . , & prieto , c.a . 2006 , , 367 , 1329 raiteri , c. m. , busso , m. , gallino , r. , picchio , g. , pulone , l.1991 , , 371 , 665 raiteri , c. m. , gallino , r. , busso , m. , neuberger , d. , kappeler , f. 1993 , , 419 , 207 rosenberg , a. , saviane , i. , piotto , g. , aparicio , a. , & zaggia , s.r . 1998a , , 115 , 648 rosenberg , a. , piotto , g. , saviane , i. , aparicio , a. , & gratton , r. 1998b , , 115 , 658 sneden , c. , cowan , j. j. , gallino , r. 2008 , , 46 , 241 sakari , c. m. , venn , k. a. , irwin , m. , aoki , w. , arimoto , n. , dotter , a. 2011 , , 740 , 106 tinsley , b. m. 1979 , , 229 , 1046 tsujimoto , t. 2011 , , 736 , 113 tolstoy , e. , venn , k. a. , shetrone , m. , et al . 2003 , , 125 , 707 tolstoy , e. , hill , v. , tosi , m. , 2009 , , 47 , 371 the , l .- s . , el eid , m. f. , meyer , b. s. 2000 , , 533 , 998 travaglio , c. , galli , d. , gallino , r. , busso , m. , ferrini , f. , & straniero , o. 1999 , , 521 , 691 travaglio , c. , gallino , r. , arnone , e. , cowan , j. , jordan , f. , & sneden , c. 2004 , , 601 , 864 timmes , f. x. , woosley , s. e. , & weaver , t. a. 1995 , , 98 , 617 van dokkum , & pieter g. , conroy , charlie . 2012 , , 760 , 70 weidner , c. , kroupa , p. , pflamm - altenburg , j. , vazdekis , a. 2013 , , 436 , 3309 ( top panel ) and the reduced @xmath26 ( bottom panel ) as a function of the component coefficients @xmath10 , in a model with @xmath40 , @xmath41 , @xmath42 and @xmath43 . solid curves represents the calculated results , and dashed horizontal lines refer to the observed values , with errors plotted by dotted lines . the shaded areas illustrate the allowed regions for the theoretical model.,scaledwidth=100.0% ] = -0.6 for element fe in pal 1 and the disk stars . for pal 1 , the filled circles with error bars represent observed abundance ratios and the associated error ; the filled pentagons with the error bars represent the calculated abundance ratios and the associated error ; the filled up triangles , filled left triangles , filled stars , filled diamonds , and filled down triangles represent the ratios of primary , secondary , sne ia , main r- and main s - component , respectively . for the disk stars , the open circles with error bars represent observed abundance ratios and the associated error ; the open pentagons with the error bars represent the calculated abundance ratios and the associated error ; the open up triangles , open left triangles , open stars , open diamonds and open down triangles represent the ratios of primary , secondary , sne ia , main r- and main s - component , respectively.,scaledwidth=100.0% ]	observed unusual elemental abundances in globular cluster palomar 1 ( pal 1 ) could provide important information for us to study the relation between the globular cluster and our galaxy . in this work , we study the abundances of @xmath0 elements , fe - peak elements and neutron - capture elements in pal 1 . we found that the abundances of the sne ia and main s - process components of pal 1 are larger than those of the disk stars and the abundances of the primary component of pal 1 are smaller than those of the disk stars with similar metallicity . the fe abundances of pal 1 and the disk stars mainly originate from the sne ia and the primary component , respectively . although the @xmath0 abundances dominantly produced by the primary process for the disk stars and pal 1 , the contributions of the primary component to pal 1 are smaller than the corresponding contributions to the disk stars . the fe - peak elements v and co mainly originate from the primary and secondary components for the disk stars and pal 1 , but the contributions of the massive stars to pal 1 are lower than those of the massive stars to the disk stars . the y abundances mainly originate from the weak r - component for the disk stars . however the contributions of the main s- and main r - components to y are close to those of the weak r - component for pal 1 . the ba abundances of pal 1 and the disk stars mainly originate from the main s - component and the main r - component , respectively . our calculated results imply that the unusual abundances of pal could be explained by the top - light imf for pal 1 s progenitor - system .
the density of dark matter in the present universe is observed via its gravitational effects on galaxies and clusters of galaxies to constitute at least about 30% of the critical density of the universe , but the nature of this dark matter is still unknown . primordial nucleosynthesis constrains the density of baryonic matter to be less than about 5% of the critical density , thus most of the dark matter is non - baryonic . the leading candidate for the dark matter is the lightest supersymmetric particle in supersymmetric extensions of the standard model that is stable by conservation of r - parity . in most scenarios this weakly interacting massive particle ( wimp ) is the neutralino , @xmath0 ( for a review , see @xcite ) . neutralinos may be detected directly as they traverse the earth or indirectly by the observation of their annihilation products . direct neutralino searches are now underway in a number of low temperature experiments with no consensus detection as of yet . indirect searches have been proposed both for gamma rays and synchrotron emission from the annihilation of wimps in the galactic center @xcite where the wimp density and magnetic field around a central massive black hole may enhance the emission significantly @xcite . the rate of annihilation is proportional to the neutralino density squared ( @xmath1 ) , therefore the strongest flux is expected to come from the galactic center where the dark matter halo density peaks . n - body cold dark matter ( cdm ) simulations have shown that the dark matter halos have a density profile with a cusp at the center ( within a core radius @xmath2 kpc ) and a steeper profile in the outer regions . the slope of the inner cusp is still a matter of debate ranging from @xmath3 to @xmath4 in different simulations . superimposed on the smooth component , the high - resolution simulations find a large degree of substructure formed due to the constant merging of smaller halos to form the present dark matter halo ( see , for example , @xcite ) . the large number of clumps generated through the hierarchical clustering of dark matter comprise about @xmath5 of the total mass and can enhance significantly the emission of gamma rays and neutrinos from neutralino annihilation in higher density clumps @xcite . we show here ( and in more detail in @xcite ) that the synchrotron radiation of electrons and positrons generated as decay products of wimp annihilation in the galactic magnetic field can provide a crucial test of wimp models , since the predicted fluxes are in the microwave region and exceed the signal of cmb anisotropies at some frequencies . the detection of this excess radiation from small angular size regions in the sky may provide the first signal from wimp annihilations . the annihilation of neutralinos produces high energy particles through several processes , depending on the mixture of supersymmetric fields that form the neutralino . in addition to the gamma ray line generated in the channel @xmath6 , the annihilation of two neutralinos also results in a continuum of particles ( gamma rays , neutrinos , electrons , positrons , muons , etc . ) that have energy spectra well represented by an @xmath7 power law . this is typical of the process of fragmentation and hadronization of quarks into hadrons ( mainly pions ) and their decay into secondary products . here we concentrate on the process @xmath8 hadrons and the decay of the resulting pions . ( we neglect the small contribution from kaons and other mesons . ) in particular , we are interested in the @xmath9 pairs generated by the decay of the charged pions . when the annihilation of neutralinos occurs in the galaxy , the secondary products are injected in the ambient magnetic field . clumps can be brightened by the synchrotron emission of @xmath9 pairs in the galactic field , depending on their position in the halo . we initially assume that the electrons emit in the same region in which they are generated , which is the case when @xmath10 are magnetically constrained in the higher magnetic field regions . as the field decays , the diffusion of the emitting @xmath10 needs to be included ( see @xcite ) . the spectrum of the generated @xmath10 naturally cuts off at about the neutralino mass , @xmath11 . therefore , the relevant frequency range for @xmath10 synchrotron emission lies below the maximum frequency , @xmath12 where @xmath13 g . since the galactic magnetic field is around @xmath14 g , for @xmath15 gev , the radiation will extend up to microwave frequencies . note that the electron - positron spectra are flatter than @xmath16 which implies that most of the energy is carried by the most energetic @xmath10 particles , and most of the synchrotron emission occurs at frequencies approaching the cutoff . in our calculations , we assumed that the galactic magnetic field has an exponential scale height as in @xcite . to model the smooth halo component , we considered the nfw halo profile @xcite @xmath17^{-2 } \ , \ ] ] where @xmath18 is the core radius and @xmath19 is the number density at @xmath18 the two parameters , @xmath18 and @xmath19 , can be set by requiring that the halo contains a given total mass ( @xmath20 ) and that the velocity dispersion at some distance from the center is known ( in the case of the galaxy , the velocity dispersion is @xmath21 km / s in the vicinity of our solar system ) . we modeled the clumpy halo following @xcite where they fit the simulations to a joint distribution of clump mass , @xmath22 , and position , @xmath23 , by @xmath24^{-3/2},\ ] ] where @xmath25 is a normalization constant , @xmath26 is the core of the clumps distribution , and @xmath27 fits well the simulations in @xcite . in @xcite , a halo with @xmath28 contains about @xmath29 clumps with mass larger than @xmath30 . we present here the results for the case of dark matter halos following a nfw profile while other cases are considered in detail in @xcite . the density of each clump is taken to be as in eq . ( 2 ) , where the normalization constant is calculated from the total mass , the core radius of the clump is assumed to be 0.1 of the clump radius and the latter is taken from the condition that the clump density equals the local density of dark matter in the galaxy at the clump position . we simulated several realizations of a clumpy halo each with about 4000 clumps of masses @xmath31 . figure 1 shows one realization with 3972 clumps in galactic coordinates . here we adopt a cross section for neutralino annihilation @xmath32 . the results can be rescaled for the present choice of dark matter density profile ( but this rescaling is not generic , see @xcite ) . in figure 1 , 754 clumps have synchrotron fluxes in the range 1 to 1000 ghz above @xmath33 times the cmb flux . these clumps are potentially detectable by anisotropy experiments where the isotropic cmb emission is subtracted and only anisotropies at the level of @xmath33 times the cmb flux remain . of these , 132 are above 30 degrees galactic latitude ( or below -30 ) . the clumps range in angular sizes from the size of a pixel ( 10x10 ) to about 10 degrees , occupying 2% of the solid angle of the sky . the crosses shown in the figure represent the solid angle ( amplified by a factor of 5 for clarity ) inside which 90% of the radiation for each clump is located . figure 2 shows the location and solid angles of clumps that are observable between 10 and 400 ghz . this range is expected to be relatively quiet of cmb foregrounds ( see , e.g. , @xcite ) , therefore the most sensitive cmb anisotropy experiments are planned for this range in frequency . the histogram in fig . 3 shows the number of observable clumps of different solid angles . there are over 100 clumps that can be observed by experiments with angular resolution around 0.2 degrees . a few very large objects can be seen even by experiments with poor angular resolution . these correspond to clumps located close to earth . if a large object is identified in this frequency range , a spectral study would verify its nature as a dark matter clump . in addition , the radial dependence of the flux within the clump could further constrain cdm clustering behavior . figure 4 shows a histogram of the number of observable clumps with total flux in or above a given energy bin . the histograms in figs . 3 and 4 were generated for the same realization in fig . 1 . for different realizations see @xcite . the behavior shown in these figures is generic and can act as a guide to determine if microwave sources are annihilating dark matter clumps or some other foreground . as an example of the spectral dependence of the synchrotron emission from wimp annihilation , we show in figure 5 the spectrum of a dark matter clump ( thick line ) compared with the cmb anisotropies ( thin line ) in the same solid angle occupied by the clump . this clump is chosen to lie at galactocentric coordinates [ -4,0,0 ] kpc ( where the sun is located at [ -8.5,0,0 ] ) , with @xmath34 , occupying a half - angle of 1 degree on the sky . the neutralino mass was chosen to be 100 gev . the dashed line shows the case of a neutralino with 10 tev for comparison . the cutoff is moved to higher frequencies as expected from eq . if an annihilating dark matter clump were to be observed , the cutoff would give us the neutralino mass directly . the dotted line shows the flux for the same clump with @xmath35 gev but with the interaction strength , given by the wimp annihilation cross section times velocity , @xmath36 , increased by a factor of 100 . the clumpy nature of cdm halos can be used to detect and constrain wimp candidates for the non - baryonic dark matter . neutralino annihilation in the higher density clumps can be observed via the synchrotron radiation of electrons and positrons as these annihilation products radiate in the galactic magnetic field . the spatial structure of the galactic magnetic field implies that the synchrotron emission from annihilation gets stronger as clumps get closer to the galactic plane . this behavior gives a different angular distribution than the distribution from the gamma ray signature of the same clumps . this unique combination will help distinguish wimp clumps from other extragalactic gamma ray sources . the frequency range where these clumps are better observed overlaps with highly sensitive experiments planned for cmb anisotropy measurements . the possibility of detecting these clumps will be soon within reach . depending on the density profile of dark matter clumps , hundreds of clumps have detectable fluxes and angular sizes . even more clumps may be present if the lower limit on clump masses is lowered from @xmath37 . the spectral shape , spatial distribution , and angular size of annihilating neutralino clumps discussed above represent some particular choices of @xmath38 and @xmath36 which are hard to constrain a priori . the neutralino mass sets the cutoff of the spectrum and changes the overall flux while the cross section mostly influences the flux amplitude . finally , the galactic magnetic field structure above and below the plane of the galaxy is poorly known and will also influence the exact observable clump distribution . the best strategy is to search for varying sizes of cmb foregrounds at a number of frequencies and select for those with the spectral dependence given in fig . 5 . once some extended synchrotron sources have been selected , the particular radial distribution and flux shape will help determine if these sources are dark matter clumps . the combination of these synchrotron measurements with the direct gamma ray and neutrino signature will make these sources unique . in addition , the synchrotron signature will help determine the structure of the magnetic field above the galactic disk . future cmb experiments such as map and planck can be used in conjunction with future gamma ray and neutrino experiments . full sky coverage helps this determination since the galactic plane is usually avoided by small area experiments . map will observe above about 20 ghz while planck should start at 30 ghz , with a large increase in angular resolution that makes these objects easier to detect .	weakly interacting massive particles ( wimps ) remain the strongest candidates for the dark matter in the universe . if wimps are the dark matter , they will form galactic halos according to the hierarchical clustering observed in n - body simulations . cold dark matter ( cdm ) simulations show that large dark matter structures such as galactic and cluster halos are formed by the merging of many smaller clumps of dark matter . each clump or halo is characterized by a centrally cusped density profile that can enhance the rate of wimp annihilation and make the annihilation products more easily detectable . electrons and positrons generated as decay products of wimp annihilation emit synchrotron radiation in the galactic magnetic field . we study the synchrotron signature from the clumps of dark matter in our galactic halo . we find that the emission in the radio and microwave region of the electromagnetic spectrum can be above the cmb anisotropy level and should be detectable by cmb anisotropy experiments . depending on the density profile of dark matter clumps , hundreds of clumps can have detectable fluxes and angular sizes .
in this paper we consider the gibbsian properties of homogeneous low - temperature ising gibbs measures on trees , subjected to an infinite - temperature glauber evolution . this problem has been considered before on regular lattices see e.g. @xcite , for ising spins , @xmath0-vector and unbounded spins , and for various finite- or infinite - temperature dynamics , of glauber , kawasaki , or diffusion type , and even for non - markovian evolutions . at high initial temperatures , or for suffiently short times , standard methods can be used to prove gibbsianness , also in our situation . thus the interesting case is to find out what happens for low initial temperatures . as usual ( but see the mean - field analysis of @xcite ) low - temperature dynamics are beyond reach so far . for simplicity we will consider infinite - temperature dynamics , but high - temperature evolutions are expected to behave qualitatively similarly . in contrast to what happens on regular lattices such as @xmath1 , the gibbsian properties of evolved gibbs measures for models on trees turn out to depend on which of the different gibbs measures ( plus or minus , versus intermediate ) one considers . in all cases there are two transition times : for the intermediate measure after the first transition time it becomes non - gibbsian in the familiar sense that some , but not all , configurations are `` bad '' ( that is , they are points of discontinuity ) , while it turns out that after a certain later time the evolved intermediate gibbs measure becomes `` totally bad '' ; thereafter it has the surprising property that _ all _ spin configurations are discontinuity points . this last property is something which will not happen for the other two extremal invariant gibbs measures . for those measures , although after a first transition time they also become non - gibbsian , after the second transition time they become gibbsian again . we will provide proofs for these results on the cayley tree , and do nt aim for the greatest generality here , but we will indicate why these results should be expected to hold more generally . we present our results both in zero and non - zero external fields . our analysis illustrates ( again ) how different models on trees are as compared to models on regular ( amenable ) lattices . non - gibbsian properties of some other measures of statistical mechanical origin on trees have been considered before , @xcite . for fk measures as well as fuzzy potts models on trees , the possibility of having a positive - measure set of `` bad '' discontinuity points was found , and for a particular renormalised ising measure the set of bad points was shown to have measure zero , while having a fractal dimension . let @xmath2 be a cayley tree for some @xmath3 , that is the unique connected tree with @xmath4 for all @xmath5 . let @xmath6 , endowed with the product topology . elements in @xmath7 are denoted by @xmath8 . a configuration @xmath8 assigns to each vertex @xmath9 a spin value @xmath10 . denote by @xmath11 the set of all finite subtrees of @xmath2 . for @xmath12 and @xmath13 we denote by @xmath14 the restriction of @xmath8 to @xmath15 , while @xmath16 denotes the set of all such restrictions . let @xmath17 be any set , finite or infinite . we denote by @xmath18 its set of edges and by @xmath19 its set of vertices . let now @xmath20 , hence finite . we will consider the nearest - neighbour ising model on the tree . the finite - volume gibbs measure on any finite subtree @xmath15 for an ising model in an inhomogeneous external field , given by fields @xmath21 at sites @xmath22 , boundary condition @xmath23 , at inverse temperature @xmath24 , is defined by the following boltzmann - gibbs distribution @xmath25 infinite - volume gibbs measures are defined by having their conditional probabilities of finite - volume configurations , conditioned on the configurations outside the volume , of this gibbsian form , see e.g. @xcite and @xcite . in equation form we require that for all volumes @xmath15 and configurations @xmath26 @xmath27 satisfies @xmath28 the infinite - volume gibbs measures are parametrized by the external magnetic fields ( in most of what follows we will consider a homogeneous field @xmath29 ) , and by the inverse temperature @xmath30 . this will lead us to consider finite - volume gibbs measures with this same homogeneous field plus a possibly different boundary field . we put @xmath31 and , for @xmath32 , @xmath33 \mathbb{i}_{\b>\b(d)},\crcr \end{split}\ ] ] where @xmath34 . it is known ( see @xcite ) , that if @xmath35 and @xmath36 , then the system exibits a phase transition . throughout the paper we will assume @xmath37 , @xmath35 , and @xmath32 , whenever the opposite is not indicated . this condition ensures the existence of three homogeneous phases @xmath38 , @xmath39 , @xmath40 . these phases are extremal in the set of invariant infinite - volume gibbs measures ; @xmath40 and @xmath38 are also extremal in the set of all infinite - volume gibbs measures , whereas @xmath39 becomes non - extremal in this set below a certain temperature strictly smaller than the phase transition temperature @xcite ; however , this second transition will not concern us here . let @xmath41 be the cayley tree with @xmath0 generations and @xmath42 the ( sub- ) cayley tree with @xmath43 generations , where @xmath44 stands for the inner boundary of @xmath41 . it is a known result for the ising model on trees that the marginal on @xmath45 of the finite - volume gibbs measure on @xmath41 is a finite - volume gibbs measure on @xmath45 , with a possibly different external magnetic field at the boundary . see appendix for how this works out in marginalizing infinite - volume gibbs measures by using boundary laws . marginalizing on @xmath45 , that is to a tree of one generation less , leaves us with a finite - volume gibbs measure on @xmath45 , parametrized by the following external fields @xmath46 where @xmath47 . thus , summarising , taking the marginal of an ising model gibbs measure on a tree with @xmath0 generations with homogeneous boundary field @xmath48 results in an ising model on an @xmath49-generation tree with a homogeneous boundary field @xmath50 . the map from @xmath48 to @xmath50 , , has three fixed points @xmath51 and @xmath52 . ( equivalently , one could consider the map between the magnetisation at generation @xmath0 to the magnetisation at generation @xmath43 , which again has the corresponding three fixed points @xmath53 and @xmath54 . ) whereas @xmath55 and @xmath52 are stable , @xmath56 is an unstable fixed point which implies that weak positive boundary conditions will result in a plus state , once one is far enough from the boundary . in other words , the phase transition is _ robust _ @xcite . these three fixed points determine the three homogeneous extremal invariant infinite - tree gibbs measures mentioned above . let @xmath57 be the set of all probability measures on @xmath7 and @xmath58 be the set of all gibbs measures of the ising model with an inverse temperature @xmath59 and external field @xmath29 . let @xmath60 denote the set of all @xmath61 which are invariant under all the graph automorphisms ( translations , rotations , reflections etc ) . let @xmath62 , where @xmath63 . we aim to study here the time - dependence of the gibbsian property of the tree gibbs measure @xmath64 , for @xmath65 , under an infinite - temperature glauber dynamics . this is the stochastic evolution @xmath66 which is obtained by having independent spin flips at each vertex at a certain given rate . in other words , we want to investigate whether or not @xmath67 is a gibbs measure at a given time @xmath68 . by assumption the initial measure @xmath27 is a gibbs measure . this immediately guarantees the non - nullness of the measure @xmath69 for all @xmath70 ( including @xmath71 ) . it will thus suffice to study whether the transformed measure is quasi - local or not . define @xmath72 to be a realization of the corresponding conditional expectation for bounded @xmath73 , finite @xmath74 , @xmath75 . we also use the notation @xmath76 when we condition only on configurations on a finite subset of sites @xmath77 . with this notation we have e.g@xmath78 , for volumes @xmath79 where @xmath80 denotes integration over the variables outside of @xmath81 . the measure @xmath82 is not quasilocal , if it is not consistent with any quasilocal specification . to prove this , it is enough to find a single , nonremovable , point of discontinuity ( in the product topology ) for a single @xmath83 for a single ( quasi)local function @xmath73 @xcite . the definition of the non - quasilocality for the transformed measure can be refined , see in particular @xcite . the relevant definitions read as follows : the measure @xmath82 is not quasilocal at @xmath84 if there exists @xmath85 and @xmath73 local ( given that @xmath86 is finite it suffices to look for @xmath73 local , with support @xmath87 ) such that no realization of @xmath88 is quasilocal at @xmath89 . in other words , any realization of @xmath88 must exhibit an essential discontinuity at @xmath89 ; one that survives zero - measure modifications . ( remember that conditional probabilities are only defined up to measure - zero sets . ) for a local function @xmath73 as above , @xmath90 is @xmath82-essentially discontinuous at @xmath89 , if there exists an @xmath91 such that @xmath92 if @xmath90 is @xmath82-essentially discontinuous at @xmath89 , informally it means that there exists an @xmath91 such that for every @xmath12 there exists @xmath93 and configurations @xmath94 , such that @xmath95 for @xmath96 , where @xmath97 is of positive @xmath82-measure . @xmath90 is strongly discontinuous at @xmath89 , iff there exists an @xmath91 such that @xmath98 intuitively the difference is that whereas for @xmath82-essential discontinuity one needs to estimate a difference on two measurable sets of positive measure , for a strong discontinuity one needs an estimate of a difference on open sets ; however , because of the impossibility of conditioning on individual configurations , we get the somewhat unwieldy definitions above . a useful tool to study whether @xmath82 stays gibbs is to consider the joint two - time distribution @xmath99 on @xmath100 , where the initial spins @xmath8 are distributed according to @xmath27 , and the evolved spins @xmath101 according to @xmath82 . this joint distribution will be denoted by either @xmath99 or @xmath102 . it can be viewed as a gibbs measure on @xmath103 with @xmath104 consisting of two `` layers '' of @xmath2 . formally , the hamiltonian of @xmath102 is @xmath105 where @xmath106 is the transition kernel of the dynamics . we consider independent spin - flip dynamics , so @xmath107 let us denote @xmath108 this approach to study the evolved measure as the marginal of a two - layer gibbs measure was introduced in @xcite , and has been applied repeatedly since . here we will find for @xmath109 , by making the choices @xmath110 , + @xmath111 , that in any open neighborhood of @xmath89 two positive - measure sets exist , on which the limits differ , however , in contrast to amenable graphs , these sets are not open ( which allows different behaviour between different evolved gibbs measures @xmath39 and @xmath40 as regards their gibbsianness , something which is excluded on amenable graphs such as @xmath1 ) . in other words we will show a @xmath82-essential , although non - strong , discontinuity . as explained in the appendix we have the representation of the conditional probabilities of the time - evolved measure @xmath112 of the form @xmath113(d\s_0)p_t ( \s_0,\eta_0 ) \end{split}\ ] ] with the perturbed @xmath101-dependent measure on spin configurations + @xmath114(d\s)\equiv\mu[\eta_{\l\ba 0},\eta_0=0](d\s)$ ] whose finite - volume marginals look like @xmath115(\s_{\l'})=c \exp \left\ { \b \sum_{(i , j)\in \l'}\s_i\s_j + \sum_{i\in \l'\setminus \del\l ' } h_i \s_i + \sum_{i\in \del\l ' } \tilde h_i \s_i\right\ } , \end{split}\ ] ] where @xmath116 where the external fields at the boundaries are given in terms of @xmath117 . this value represents the fixed point of the recursion relation with homogeneous field @xmath29 , , and is bijectively related with the starting measure @xmath64 . more generally , such a representation is always valid if the initial measure is a markov chain on the tree . markov chains can be described by boundary laws , and conditional probabilities of infinite - temperature time evolutions , are , for finite - volume conditionings , described by boundary laws obeying recursions which are local perturbations of those of the initial measure , see the appendix and @xcite . in what follows we choose @xmath118 and @xmath119 . with this notation , for non - gibbsianness it is enough to prove that , at @xmath89 , there exists an @xmath120 such that , for all @xmath15 there exists @xmath79 such that @xmath121(\s_0 ) -\mu[\bar\eta_{\l\setminus0},\xi^2_{\l'\setminus\l}](\s_0 ) \right\|>\e\ ] ] to prove the non - gibbsianness of @xmath82 , we will have to consider the phase transition behaviour of the gibbs measures on the first layer in various external fields . these external fields are determined by the various conditionings , as well as by the choice of the initial gibbs measure . let @xmath122 be integers with @xmath123 , let us denote @xmath124 and @xmath125 . consider first the case @xmath126 . marginalizing on @xmath127 leaves us with a finite - volume gibbs measure on @xmath127 denoted by @xmath128 and parametrized by the following external fields @xmath129 in order to apply the ( marginalisation ) procedure to the @xmath101-dependent finite - volume gibbs measure @xmath128 on @xmath127 we need to identify the role played by @xmath101 . it can be shown that taking the marginal on @xmath130 of the finite - volume gibbs measure on @xmath127 ( summing out the spin @xmath131 ) gives us a finite - volume gibbs measure on @xmath130 with an external field at the boundary equal to @xmath132 here the sum is over the nearest neighbors @xmath133 . the equation tells us how the configurations @xmath134 will affect the field acting on @xmath135 after having taken one - generation marginal . the configuration @xmath136 will govern the value of the fields at @xmath137 , when the marginal on @xmath138 is taken . let us see how : * @xmath139 @xmath140 * @xmath141 @xmath142 note that the above chosen @xmath101-conditioning on the annulus makes the recursion homogeneous . choosing @xmath143 big enough guarantees that the recursions , approach their time - dependent fixed points ; we denote them respectively by @xmath144 , @xmath145 and @xmath146 , @xmath147 , see figure . assume that we start at time @xmath71 with the measure @xmath39 , then @xmath148 . it ensures that the recursions , will approach , respectively , @xmath149 and @xmath150 . @xmath151 represents the biggest stable fixed point for the @xmath152 recursion , and @xmath153 the smallest stable fixed point for the @xmath154 recursion . the fact that both recursions have as a starting point the unstable fixed point @xmath155 guarantees that the plus conditioning will drag the field towards @xmath151 and the minus one towards @xmath153 . this will not be the case for @xmath156 and @xmath157 as we will see later . the @xmath158-dependent marginals on @xmath138 , of the measure on @xmath127 , are finite - volume gibbs measures parametrized by the following fields : + for the case @xmath159 @xmath160 and in the case @xmath161 @xmath162 notice that only the fields at @xmath137 depend on @xmath136 and not the ones acting on the interior . we emphasize that the broadcasting is absorbed by the boundary and has no direct influence on the interior . now we investigate how the recursion relation @xmath163 , obtaining by summing out generations in @xmath138 , will depend on the fixed configurations @xmath164 , namely on the fields @xmath165 acting on the generation @xmath166 . we emphasize that the annulus configurations determine the starting point of the recursion . we will also show how the aforementioned recursion relation can be bounded from below if we are coming from @xmath167 , and from above for @xmath168 . furthermore these bounds will turn out to be uniform with respect to @xmath169 and with respect to @xmath170 ( number of iterations ) . [ recur ] given the recursion relation @xmath163 we have : + @xmath171 , for all @xmath22 and @xmath170 , if @xmath172 ; and @xmath173 , for all @xmath22 and @xmath170 , if @xmath174 . here @xmath175 is the fixed point for the homogeneous recursion @xmath176 with @xmath177 . * proof * : fixed points of the discussed recursion relation are given in the picture . the proof follows by induction . take first the case @xmath178 . naturally @xmath179 , so @xmath180 for all @xmath22 . if we now assume @xmath181 for all @xmath22 , then @xmath182 . the case @xmath183 follows by symmetry ; the corresponding recursion relation will be bounded from above by @xmath184 . let @xmath186 be defined by @xmath187 [ mu - hash ] if @xmath8 is distributed according to @xmath39 , then after time @xmath186 all configurations @xmath101 are bad configurations ( points of essential discontinuity ) for the transformed measure @xmath109 . the main idea is as follows : if the plus configuration is bad ( and by symmetry the same is true for the minus configuration ) , then all configurations @xmath89 will be bad . this is because if minus boundary conditions give a minus magnetisation for the conditioned @xmath188-spin at the origin , and plus boundary conditions a positive one , the same holds for all @xmath89 ( due to fkg e.g. ) . so take @xmath89 to be plus . choosing @xmath189 to be plus in a large enough annulus @xmath190 and integrating the outside with @xmath39 will lead to an effective plus boundary condition at @xmath191 . the reason is that the positive magnetisation @xmath192is an attractive fixed point for the recursive relation , and any positively magnetised field in @xmath193 will lead into its domain of attraction . the same is true for the negative magnetisation . as there are different magnetisations with plus and minus boundary conditions , even in the presence of a weak plus field ( the field is plus due the @xmath89 being plus ) , the choice of plus or minus in the annulus influences the expected magnetisation at the origin , however big @xmath191 is . the definition of @xmath186 , , will assure we are in the phase - transition regime for the transformed system ( for @xmath194 ) . making use of lemma , the value of @xmath195 we are after , in order to prove the essential discontinuity , is given by @xmath196 . this value corresponds to taking , for the measure coming from @xmath167 , the smallest positive field along all the @xmath197 iterations , namely @xmath175 . the field at the origin is given by @xmath198 and could be roughly bounded from below @xmath199 thus the corresponding single - site measure is given by @xmath200 , so @xmath201(\s_0)=\tanh(h^+_t)\ ] ] analogously for the measure coming from @xmath168 , we take the biggest negative value along all the @xmath197 iterations , that is @xmath202 , therefore @xmath203 and @xmath204(\s_0)=\tanh(-h^+_t)\ ] ] for @xmath196 the inequality holds . let us notice that @xmath195 is chosen uniformly with respect to @xmath101 , thanks to the uniform bounds appearing in lemma . this ensures the @xmath82-essential discontinuity in any point . @xmath185 as we mentioned before , the previous argument does not hold for @xmath40 and @xmath157 . we treat here only the @xmath40 case , the @xmath38 case is completely symmetrical . so , in case we start with the plus measure , even conditioning on a minus configuration in the annulus , due to the plus influence from the boundary will lead to a measure on @xmath138 that looks like the plus measure in a negative field . [ no_bad ] given the starting measure @xmath40 , the fields acting on @xmath205 for the marginal measure on @xmath127 , which are given by @xmath206 , @xmath207 , satisfy the following inequality @xmath208 for all @xmath32 , @xmath35 and for all @xmath209 . * proof * : let @xmath186 be as in . we need to show that @xmath210 in the aforementioned region of parameters . first of all we note that the expression on the right - hand side is zero in the limit @xmath211 , and it is a decreasing function of @xmath70 . so in order to prove the lemma it is enough to show @xmath212 using that @xmath213 is a fixed point for the @xmath214 recursion at @xmath215 , we arrive at @xmath216 note that @xmath217 , where @xmath218 is a tangent point to @xmath219 such that @xmath220 . we show that @xmath221 . in fact we know that @xmath222 . using the mean - value theorem together with the fact that @xmath223 , we write @xmath224 . it implies that @xmath189 is such that @xmath225 . using then that @xmath226 is a decreasing function it follows that the domain of @xmath189 , namely @xmath227 has to contain @xmath218 ; so @xmath221 . using then the monotonicity of the functions @xmath228 the claim is proved . @xmath185 [ + allgood ] if @xmath8 is distributed according to @xmath40 , then after time @xmath186 all configurations @xmath101 are good configurations for the transformed measure @xmath229 . * proof. based on lemma , choosing @xmath127 big enough we make sure that the recursion relation coming from the fixed `` + ' ' -annulus @xmath230 will approach its fixed value @xmath151 , so do we for the fixed `` @xmath231''-annulus to approach its fixed value @xmath175 . then the magnetic fields for the finite - volume gibbs measure on @xmath138 are respectively given by @xmath232 and @xmath233 define @xmath234 . this maximum is always positive , as an inductive argument shows . we are about to prove that @xmath235 such that @xmath236 ; this is equivalent to say that @xmath237 . @xmath238 \cr & = \max_i\left[\sum_{l\sim i}\frac{d\phi'(c_l)}{d}\left(h_l^{+,j } - h_l^{-,j}\right)\right ] \leq ( 1-\delta ) \max_i\left[\frac{1}{d}\sum_{l\sim i}\left(h_l^{+,j } - h_l^{-,j}\right)\right ] \leq \cr & = ( 1-\delta)\max_i \max_{l\sim i}\left((h_l^{+,j } - h_l^{-,j})\sum_{l\sim i}\frac{1}{d}\right ) = ( 1-\delta)\max_i \max_{l\sim i}\left(h_l^{+,j } - h_l^{-,j}\right)\crcr & = ( 1-\delta)\d^j \end{split}\ ] ] we used the mean - value theorem together with the fact that @xmath239 for @xmath240 . @xmath185 for @xmath8 distributed according to @xmath39 , we will show the existence of an intermediate time interval , where some , but not all , configurations are bad for @xmath82 . theorem will express this . we will show that the all plus and all minus configurations are good for @xmath241 at all times in @xmath242 . moreover we will impose a condition on the field @xmath243 ( therefore on @xmath70 itself ) , such that it guarantees the existence of at least one bad configuration for @xmath109 . we will find a @xmath244 , which is larger than the minimal value of time for which this condition is satisfied . this value @xmath244 will turn out to be strictly less than @xmath186 . this will guarantee that @xmath244 is small enough so that the transformed measure , conditioned on an all plus or all minus @xmath101 will not exibit a phase transition . note that this implies that at the same time @xmath186 the intermediate state has a transition to a totally non - gibbsian regime , where all spin configurations are discontinuity points , whereas the plus and minus state have a transition to a gibbsian regime , without discontinuity points . [ good : t_2 ] if @xmath8 is distributed according to @xmath39 then for all @xmath245 the @xmath152 and @xmath154 configurations are good configurations for the transformed measure @xmath109 . proof. as was shown before , the recursions , ( related to the annuli ) give us respectively @xmath151 and @xmath153 . let first @xmath101 be the plus configuration . in this case @xmath246 for all @xmath22 and @xmath170 . in other words the field will stick to the fixed point value along the iterations . using an inductive argument we show that @xmath247 ; that s to say that it does not depend on @xmath22 . based on that , it is straightforward to get a monotonicity property for @xmath248 , namely that @xmath249 for all @xmath170 . indeed the last inequality follows from the fact that @xmath251 for all @xmath252 , due to the chosen range of @xmath70 . recalling that for @xmath245 the recursion relation @xmath253 has only one fixed point , namely @xmath151 , the lemma is proven for @xmath152 . the @xmath154 case follows by symmetry . @xmath185 the chosen range of times enables the existence of a unique fixed point for each of the recursions , , independently of @xmath117 . this means that the fields we obtain at @xmath137 depend on the annuli , but they do not depend on the exterior @xmath254 . for this reason lemma applies to @xmath8 s distributed according to @xmath40 and @xmath38 too . for the sake of clarity , let us recall that @xmath255 indicates the positive stable fixed point for the recursion with @xmath126 . [ lemma : t_1 ] let @xmath244 be given by @xmath256 then @xmath257 proof. recalling equation , the fact that @xmath244 lies in the interval @xmath242 is guaranteed by the truth of the inequality @xmath258 , for @xmath35 and @xmath32 . indeed @xmath259 knowing that @xmath260 , the monotonicity of the function @xmath261 concludes the proof . @xmath185 define the `` alternating '' configuration @xmath262 to be @xmath263 for @xmath264 and @xmath265 , i.e. all vertices at each generation have the same sign different from the sign of the previous and the next generations . naturally the configuration for which @xmath266 is also an `` alternating '' one . let us call @xmath267 the field at the vertex @xmath268 after @xmath269 applications of the recursion formula , starting respectively at @xmath270 or @xmath271 . the particular structure of the `` alternating '' configuration makes the fields homogeneous at each generation ; i.e. , @xmath272 , for all @xmath268 . [ thm : somebad ] if @xmath8 is distributed according to @xmath39 , and @xmath244 is given by , then for all @xmath273 some , but not all , configurations @xmath101 are bad for the transformed measure @xmath109 . proof. making use of lemma , lemma , to prove the theorem it is enough to find a particular configuration @xmath101 that will be bad for all @xmath273 . the `` alternating '' configurations will be shown to be bad for all @xmath274 , in other words they transmit the influence of the annulus to the origin , no matter how `` distant '' the annulus and the origin are . as remarked before , @xmath275 associated to the @xmath262 configurations depend only on @xmath170 , and we call the corresponding values @xmath276 . without loss of generality let us assume @xmath277 , for @xmath278 . by an inductive argument , based on the hypothesis @xmath279 ( which in terms of fields means @xmath280 ) , and on the particular structure of the configuration @xmath262 , we show that @xmath281 and @xmath282 , for all @xmath170 even , namely for those @xmath170 which relate to generations at which @xmath262 is set to be @xmath283 , and that @xmath284 and @xmath285 for @xmath170 odd . this will imply @xmath286 for all @xmath170 . consider the case @xmath170 even . + for @xmath287 we have : @xmath288 , @xmath289 + both inequalities hold , because @xmath290 is a decreasing function of @xmath70 whose lower bound is given by @xmath255 . assuming the statement is true for @xmath170 , let us see that it holds for @xmath291 . we focus first on @xmath292 . @xmath293 where the second equality is justified by the particular structure of the alternating configuration . using the assumption @xmath281 and the monotonicity of @xmath261 we arrive at @xmath294 the fact that @xmath295 ensures that @xmath296 . this concludes the proof for @xmath297 . + for @xmath298 we have : @xmath299 using always the assumption @xmath282 , the monotonicity of @xmath261 , and the assumption @xmath300 , which guarantees @xmath301 , we obtain @xmath302 the case @xmath170 odd is analogous . @xmath185 the above result also applies to the evolved plus and minus measures . indeed the alternating configuration displays a strong discontinuity here , whereas the above analysis shows that for large times all configurations display a @xmath109-essential but nonstrong discontinuity . whether the @xmath244 used above is optimal in any sense is not known . we conjecture that it may be for the intermediate state , but not for the plus or minus states . recall that @xmath303 , @xmath35 and @xmath32 ; these conditions guarantee existence of three homogeneous phases for the original measure ; we denote them , even if not fully consistent with the notation we have been using so far , @xmath304 , @xmath305 , and @xmath306 , just to emphasize their dependence on @xmath29 . we show that the previous results found for @xmath126 will also apply to the case @xmath307 but for different time values . let @xmath308 be given by the following equations : @xmath309 call @xmath310 depending on the sign of the initial field , @xmath311 might be either bigger or smaller then @xmath312 , as follows from . nevertheless the definitions of @xmath313 , and @xmath314 will always assure @xmath315 ( e.g. for @xmath316 the order is @xmath317 ) . the time @xmath313 indicates the time value for which the dynamic field @xmath243 , taken in the opposite direction to @xmath29 , will first reach a value which guarantees phase transition for the conditioned transformed measure . the time @xmath314 refers to the analogous value , but for @xmath243 taken with the same sign to @xmath29 . suppose w.l.o.g . that @xmath316 . note that for @xmath29 negative the magnetization corresponding to @xmath318 is positive , see @xcite , chapter 12 . for @xmath319 there exist three fixed points for the @xmath214-recursion @xmath320 , namely two stable ones @xmath321 , @xmath322 , and an unstable @xmath323 . the existence of several fixed points makes the convergence to them be dependent on the starting point . in particular the recursion will take us to @xmath322 iff the starting point , @xmath324 , lies to the right of the unstable one , that is when @xmath325 ; it will take us to @xmath321 iff @xmath326 , and will stick to @xmath323 iff @xmath327 . given that @xmath328 , the assumption @xmath319 ensures the existence of three fixed points also for the @xmath329-recursion @xmath330 ; they are denoted by @xmath331 , and @xmath332 . assume that we start at time @xmath71 with the measure @xmath306 , then the starting point for the @xmath333-recursions is @xmath334 . however , for the chosen range of time , @xmath319 , it can be shown that @xmath335 will always lie to the right of @xmath332 and always to the left of @xmath323 . so the next theorem reads : if @xmath8 is distributed according to @xmath306 , then after time @xmath314 all configurations @xmath101 are bad configurations for the transformed measure @xmath336 . @xmath185 analogously to the analysis for @xmath126 , the former result will not hold for @xmath8 distributed according to @xmath337 . two other results , obtained in the previous section , have equivalents for non - zero external field . [ good : t_2_h0 ] if @xmath8 is distributed according to @xmath306 , then for all @xmath338 the @xmath152 and @xmath154 configurations are good configurations for the transformed measure @xmath339 . @xmath185 [ + allgood_h0 ] if @xmath8 is distributed according to @xmath337 , then after time @xmath314 all configurations @xmath101 are good configurations for the transformed measure @xmath340 . @xmath185 [ goodness - shifted ] it is worth remarking that the strict inequality @xmath315 , always holding for @xmath341 , implies the non - emptiness of the interval of times @xmath342 . a similar result to the one given in theorem holds in the case @xmath341 , namely that for @xmath343 some , but not all , configurations are bad . in fact , it can be shown , for example in case @xmath316 , that the time @xmath313 corresponds to the time for which the plus configuration becomes bad , while for all times @xmath344 the minus configuration will remain good . in case @xmath345 , as symmetry may suggest , the time @xmath313 will be the threshold for the minus configuration to become bad , while the plus configuration will be good till @xmath346 . encouraged by the many analogies between the @xmath126 case and the @xmath341 case , one might ask what one can say about the @xmath347-equivalent of the time @xmath244 , . pursuing the former , let us define the values of times @xmath348 , @xmath349 by the following equalities @xmath350 and define further @xmath351 the picture helps to understand the role played by the different times so far defined . it can be shown that @xmath352 for @xmath353 . nonetheless the relation between time @xmath354 and @xmath313 is not so trivial as we will show . the next lemma formalizes that for all time @xmath355 the `` alternating '' configurations are bad for @xmath8 distributed according to @xmath306 . [ good : somebad_h0 ] if @xmath8 is distributed according to @xmath356 , and @xmath354 is given by , then for all @xmath357 `` alternating '' configurations are bad for the transformed measure @xmath336 . * the proof follows the same route taken in the proof of theorem with some modifications on the bounds . nontheless we reckon it is instructive to sketch the main points at least for @xmath316 . for @xmath357 an inductive argument leads to the following bounds : @xmath358 @xmath359 therefore @xmath360 for all @xmath170 . @xmath185 the previous lemma together with remark shows that if @xmath8 is distributed according to @xmath306 , then for all @xmath361 some , but not all , configurations are bad . there are then two different time intervals where some , but not all , configurations are bad . we will not leave the reader wondering how these two intervals relate . we will show the existence of a critical value @xmath362 such that for @xmath363 we have @xmath364 , for @xmath365 the inclusion is reversed , namely @xmath366 , and for @xmath367 the two intervals coincide . in the small - field regime @xmath365 we have that the `` alternating '' configuration becomes bad before the all plus and the all minus configurations . in that case , the dominant effect is that the alternating character of the conditioning provides some cancellations , just as in the zero - field case . in the other regimes we can just say what follows from @xmath352 , i.e. that the `` alternating '' configurations become bad before the homogeneous configuration with all @xmath101 s aligned with @xmath29 , that is @xmath368 . the impossibility to state something more in the other regimes is due to the fact that @xmath354 is not a `` sharp '' threshold for the `` alternating '' configurations to become bad . however , in this case having a `` bad '' configuration , one may need to counteract the effect of the field , thus in a positive external field , the minus configuration becomes bad at an earlier time than the alternating one . to explore the latter inclusions we need to compare the values @xmath354 and @xmath313 , or equivalently @xmath369 and @xmath370 . consider the difference between the fields @xmath371 based on the definitions of the times , , it turns out that the function @xmath73 is even . so we might focus on its behaviour only for negative values of the initial field @xmath29 . for such values of the field the function has the following form @xmath372 first of all the limit values of @xmath73 in the interval @xmath373 are given by @xmath374 @xmath375 note that the second limit value is positive , as has been explained in the proof of lemma , while the first one is negative by the definition of @xmath376 , and by . taking now the derivative of @xmath73 with respect to @xmath29 we obtain @xmath377 using the only thing we know about @xmath378 , @xmath335 , namely that they are fixed points for the recursion @xmath379 , the following equalities turn out to hold @xmath380 @xmath381 because @xmath382 and @xmath383 , the monotonicity of @xmath226 assures that @xmath384 for all @xmath385 . therefore the existence and uniqueness of @xmath362 is guaranteed by an application of the intermediate - value theorem . we point out that the function @xmath73 is not differentiable in @xmath126 . indeed , being @xmath73 an even function and @xmath386 clarify the discontinuity . we would like to remark that the case @xmath126 might be obtained from the previous analysis by taking the limit @xmath387 . indeed @xmath388 , @xmath389 . we have shown that the gibbs - non - gibbs transition on trees has a number of different aspects , as compared to that on regular lattices . in particular , we have shown that different evolved gibbs measures can have different gibbsian properties . for the evolved intermediate state there are two transitions , one from being gibbsian to being `` standard non - gibbsian '' ( having some , but not all configurations bad ) and a second transition to a `` totally non - gibbsian '' regime where _ all _ configurations are bad . both these properties do not occur in the more familiar lattice and mean - field situations . for the plus and minus measure there are also two transitions , namely one after which the evolved measure becomes non - gibbsian , and some , but not all , configurations become discontinuity points and a second one after which the measure becomes gibbsian again ; this is the behaviour which on the lattice occurs for an initial gibbs measure in an external field . although we have worked out the case of cayley trees , we expect our results to hold for a much wider class of trees . the instability of the fixed point @xmath56 for example corresponds with the phase transition being robust , which is true in general for ising models on trees @xcite . also , the property of plus boundary conditions in a not too strong minus field inducing a positively magnetised state , which was used in the proof that the plus configuration was good for the plus state holds quite generally . the choice of bad configuration in the intermediate regime may be somewhat tree - dependent . moreover , it seems problematical to identify a unique measure @xmath39 in a field ( on random galton - watson trees for example ) . it is the purpose of this appendix to explain the relation between the notion of a boundary law as it is used in the book by georgii @xcite and the one - sided simple recursions which are used in the paper . the notion of a boundary law is necessary to describe all the extremal phases ( or more generally , all markov chains on trees ) . to follow the notation used in georgii , let us denote , for @xmath390 , by @xmath391 the transition matrix of the random field ising model on the tree with hamiltonian @xmath392 , where @xmath393 , so the local field at each site has been symmetrically distributed among the edges to its neighbors . every extremal gibbs measure @xmath27 for the random field ising model on the cayley tree is a markov chain on the tree ( theorem 12.12 of georgii ) . to define what it means to be a markov chain on the tree , consider an oriented bond @xmath394 , draw this bond horizontally such that @xmath22 lies to the left of @xmath170 , and draw the tree embedded into the plane in such a way that there is no intersection between the tree and the axis crossing the oriented bond @xmath394 in a perpendicular way . a measure @xmath27 is a markov chain on the tree if conditioning on the semi - infinite spin configurations extending from @xmath22 to the left ( the past ) is the same as conditioning on the spin configuration at the site @xmath22 alone , and this holds for all oriented bonds @xmath394 . not all markov chains are extremal gibbs measures however , as the example of the free boundary condition gibbs measure of the ising model in zero field at sufficiently low temperatures shows . the meaning and importance of a boundary law lies in the following fact . a markov chain on the tree always has a representation in terms of a boundary law @xmath395 , @xmath396 , that is for the finite - volume marginals it holds @xmath397(\s_{\l \cup \del_+ \l } ) = \frac{1}{z_{\l}(\b , h ) } \prod_{k\in \partial_+ \l}l_{k k_\l}(\s_k ) \prod_{\{ij\ } \cap \l \neq \emptyset}q_{ij}(\s_i , \s_j)\ ] ] where @xmath398 denotes the outer boundary of @xmath15 and @xmath399 is the unique nearest neighbor of @xmath400 in @xmath15 . a boundary law is a function on oriented edges @xmath394 which depends on the possible spin values . from its appearance in the last formula we see that , at any @xmath394 , it is defined only up to a multiplicative constant , not depending on the spin configuration @xmath401 . define therefore @xmath402 in the ising case . this quantity has the character of a local field at the site @xmath22 and contains the full information about the boundary law in the ising case . more precisely @xmath403 has the meaning of a local field acting on the spin @xmath404 which has to be added to the hamiltonian with free boundary conditions in the volume @xmath405 if the site @xmath400 is attached at the site @xmath406 . assuming the validity of the last formula for the finite - volume marginals one arrives at a @xmath407-dependent consistency ( or recursion ) relation that a boundary law has to satisfy . this recursion is formulated as ( 12.10 ) in georgii ; in the case of the ising model with site - dependent fields it translates equivalently into the recursion @xmath408 conversely , a function @xmath409 on all oriented bonds which is consistent in the sense of defines a markov chain by formula with the corresponding boundary law @xmath410 . note that is a one - sided recursion which has no beginning and no end . it is interesting in the first step to look at homogeneous solutions , i.e. solutions not depending on the bond @xmath394 , but there may be also many other solutions , even in the case when the local magnetic field in the initial hamiltonian is site - independent . in that case there can be non - homogeneous solutions when there are more than one fixed points for the homogeneous recursion . indeed , to construct a non - homogeneous solution one picks a site @xmath170 and looks to all oriented bonds @xmath394 pointing to it , and picks values of @xmath409 not at the fixed point . then one defines a boundary law by preimages for @xmath411 s for the oriented bonds @xmath412 going up to @xmath394 . in order to make sure that there are such preimages under all orders of iterations , the value has to be chosen such that it lies between a stable and an unstable fixed point . to see the meaning of the boundary law in a more intuitive or physical way let us make explicit the difference to the field which is already present in the original hamiltonian . we look at the asymmetric quantity which is centered at the local field for the first spin , namely @xmath413 and note that it satisfies the equation @xmath414 with @xmath415 . with this variable we have @xmath416(\s_{\l \cup \del_+ \l } ) = \frac{1}{z_{\l}(\b , h ) } e^{\sum_{\{ij\ } \cap \l \neq \emptyset}\b \s_i \s_j + \sum_{i\in \l\cup \partial_+ \l}h_i \s_i+ \sum_{k\in \partial_+ \l } f_{k k_\l}\s_k}\ ] ] so the @xmath417 has the meaning of an additional boundary field at the site @xmath22 acting on top of the local fields which are present already in the hamiltonian , when one computes the finite - volume marginals in a volume with a boundary site @xmath22 when @xmath22 is attached via the site @xmath170 to the inside of the volume . now , let us enter in more detail the discussion on the dependence of boundary laws on a variation of local fields entering in the hamiltonian . suppose that a boundary law @xmath418 $ ] , not necessarily homogeneous , is given for the ( not necessarily but possibly homogeneous ) hamiltonian with a field @xmath419 . recall that , as we just explained , homogeneous fields @xmath419 may have very well inhomogeneous boundary laws . let us consider the system now in the presence of a local perturbation of the field @xmath420 , possibly site - dependent , but bounded , i.e. @xmath421 . any gibbs measure @xmath422 $ ] gives rise to a gibbs measure @xmath423 $ ] which is related by the formula involving the local perturbation of the hamiltonian of the form @xmath424(\phi(\tilde \s ) ) & = \frac{\mu[h](\phi(\tilde \s ) e^{\sum_{i}\d h_i \tilde \s_i } ) } { \mu[h](e^{\sum_{i}\d h_i \tilde \s_i } ) } \end{split}\ ] ] where it is understood that integration is over @xmath425 . if the original gibbs measure is actually a markov chain described by the boundary law @xmath426 $ ] , the perturbed measure is described by the boundary law @xmath427 $ ] which is obtained by putting @xmath427:=l_{ij}[h]$ ] for oriented bonds @xmath394 in the outside of the region of the perturbation of the fields which are pointing towards the perturbation region . when passing with the recursion through the perturbation region of the local fields the @xmath410 s obtain a dependence on the size of the perturbations . then the forward iteration is used to obtain an assignment of @xmath428 s to all oriented bonds . [ 616161 ] suppose that @xmath419 is an arbitrary external - field configuration , @xmath429 is an arbitrary finite - volume perturbation of the external fields , and @xmath430 $ ] is the measure which results from a local perturbation from a markov chain @xmath422 $ ] which is described by a boundary law @xmath418 $ ] . then @xmath430 $ ] behaves in a quasilocal way ( i.e. all expected values @xmath430(\phi)$ ] on local spin functions @xmath261 are quasilocal functions of @xmath429 ) if and only if the boundary laws @xmath431 $ ] , depending on field perturbations @xmath432 s for @xmath400 in the past of the oriented bond @xmath394 , behave in a quasilocal way , and this holds for all oriented bonds @xmath394 . here a vertex @xmath400 is said to be in the past of @xmath394 if the path from @xmath400 to @xmath170 passes through @xmath22 . quasilocality is meant in the same way as it has been introduced in the context of finite - volume variations of spins , i.e. we say that @xmath428 depends quasilocally on a variation of fields iff @xmath433 where the supremum is taken over perturbations @xmath434 in the finite volume @xmath81 which look the same on @xmath15 . we note again that there is a one - to - one correspondence between simple directed field recursions with @xmath435 neighbors , as used in the paper , and boundary laws . so we obtain the following corollary , which is used extensively in the paper . * corollary . * suppose that @xmath419 is a homogeneous external field , @xmath429 is an arbitrary finite - volume perturbation of external fields , and @xmath430 $ ] is the measure which results from a local perturbation from either one of the homogeneous measures @xmath422 $ ] , corresponding to the plus , the minus or the unstable fixed points . then the measures @xmath430 $ ] behave in a non - quasilocal way on the field perturbations @xmath429 iff the corresponding solutions of the one - sided simple recursions for the effective fields behave in a non - quasilocal way . * some non - homogeneous gibbs measures . * the discussion just given has consequences also for those gibbs measures @xmath436 which are obtained by pasting boundary laws @xmath437 for oriented bonds @xmath412 of the form @xmath438 for some fixed subtree @xmath15 , so that is true for the particular volume @xmath15 . then extend the boundary laws to have a prescription in the whole volume . then the parameter region for non - quasilocal behavior of the resulting measure will be the union of the parameter regions of non - gibbsianness of the original measures taken over the @xmath412 s . * connection to gibbs vs. non - gibbs under time evolution . * since the gibbs properties of time - evolved ising measures in infinite - temperature evolution can be expressed via quasilocality properties of @xmath439 $ ] , for _ finite - volume _ @xmath429 , we are left with the investigation of the locality properties of the boundary law iteration . a local variation of the image spins amounts to a local perturbation @xmath429 of the local fields . indeed , denoting the time - evolved measure by @xmath440 , starting from the measure @xmath441 , we have for finite @xmath442 the formula @xmath443(d\s_0)p_t ( \s_0,\eta_0 ) \end{split}\ ] ] with a measure @xmath114(d\s)$ ] of the form @xmath430 $ ] with a perturbation in the finite volume @xmath444 . finite - volume marginals of this measure have a representation , according to lemma [ 616161 ] , of the form with an @xmath101-dependent transition matrix @xmath445(\s_i,\s_j)= e^{\frac{h_t \eta_i 1_{i\in \l\ba 0}}{d+1 } + \frac{h_t \eta_j 1_{j\in \l\ba 0}}{d+1 } } q_{i j}(\s_i,\s_j)$ ] where @xmath446 is the transition matrix for the initial measure @xmath27 , and an @xmath447-dependent boundary law @xmath448 $ ] which obeys the locally modified iterations for the boundary law described below . hence , non - gibbsianness of time - evolved measures is detected by non - quasilocality of the perturbed boundary laws @xmath448 $ ] . a consequence of these remarks is that a time - evolved measure resulting from an initial gibbs measure which is constructed by pasting finitely many boundary laws @xmath437 as described above , will be non - gibbsian at a parameter regime which is the union of the non - gibbsian parameter regimes of the time - evolved markov chains corresponding to @xmath437 , over @xmath412 .	in this paper we study homogeneous gibbs measures on a cayley tree , subjected to an infinite - temperature glauber evolution , and consider their ( non-)gibbsian properties . we show that the intermediate gibbs state ( which in zero field is the free - boundary - condition gibbs state ) behaves different from the plus and the minus state . e.g. at large times , all configurations are bad for the intermediate state , whereas the plus configuration never is bad for the plus state . moreover , we show that for each state there are two transitions . for the intermediate state there is a transition from a gibbsian regime to a non - gibbsian regime where some , but not all configurations are bad , and a second one to a regime where all configurations are bad . for the plus and minus state , the two transitions are from a gibbsian regime to a non - gibbsian one and then back to a gibbsian regime again . * ams 2000 subject classification : * . _ keywords : _ non - gibbsianness , ising models , tree graphs , glauber dynamics .
neutron stars are spotted as pulsars by radio telescopes and x - ray satellites . they are more massive ( i.e. @xmath1 ) than our sun but are typically only about @xmath2 kilometers across so that the matter in their centers is compressed to densities that are up to an order of magnitude higher than the density of atomic nuclei @xcite . at such densities it is quite plausible that numerous subatomic particle processes will compete with each other and novel phases of matter like the quark - gluon plasma being sought at the most powerful terrestrial particle colliders could exist in the center of neutron stars . figure [ fig : cross ] summarizes the possible scenarios graphically , with the associated equations of state shown in figure [ fig : eos ] . the strangeness - carrying @xmath3 quark is likely to play a vital role inside neutron stars , for several of the building blocks of matter may contain an @xmath3 quark as one of their constituents . examples are the hyperons @xmath4 , the @xmath5 meson , and the h - dibaryon . more than that , strangeness may also exist in the form of unconfined @xmath3 quarks , which could populate , in chemical equilibrium with @xmath6 and @xmath7 quarks and/or hadrons , extended regions inside neutron stars . this paper aims at giving a brief overview of the possible manifestations of strangeness inside neutron stars . it is most noteworthy that these objects can nowadays be observed with state - of - the - art radio telescopes ( e.g. arecibo , jodrell bank , parkes , vla ) and x - ray satellites ( e.g. chandra , hst , rxte , xmm ) with such a remarkable clarity that previously was only imagined , which enables us for the first time ever to seriously constrain the properties of superdense matter from astrophysical observations . only in the most primitive conception , a neutron star is constituted from neutrons . at a more accurate representation , a such stars will contain neutrons ( @xmath8 ) and a small number of protons ( @xmath9 ) whose charge is balanced by leptons ( @xmath10 , @xmath11 ) . at the densities in the interiors of neutron stars , the neutron chemical potential , @xmath12 , easily exceeds the mass ( modified by interactions ) of the @xmath13 which then allows neutrons to transform into @xmath13 s . mathematically , this is expressed as @xmath14 , which indicates that for neutron fermi momenta greater than @xmath15 , @xmath13 particles could make their appearance . such fermi momenta correspond to densities of just @xmath16 , with @xmath17 the baryon number density of ordinary nuclear matter . so , in addition to nucleons and electrons , neutron stars may be expected to hide considerable populations of strangeness - carrying @xmath13 hyperons , eventually accompanied with @xmath18 and @xmath19 hyperon populations too @xcite . the total hyperon population may be as large as 20% @xcite . once the reaction @xmath20 becomes possible in a neutron star , it is energetically advantageous for the star to replace the fermionic electrons with the bosonic @xmath5 mesons . the associated chemical potentials must obey @xmath21 . whether or not this is actually fulfilled in neutron stars depends on the mass of the @xmath5 in dense matter . information about this is provided by the @xmath5 kinetic energy spectra extracted from ni+ni collisions at sis energies , as measured by the kaos collaboration at gsi @xcite . an analysis of the kaos data shows that the attraction from nuclear matter may bring the @xmath5 mass down to @xmath22 at @xmath23 . for neutron - rich matter , the relation @xmath24 was established @xcite , with @xmath25 mev the @xmath5 vacuum mass . values around @xmath22 lie in the vicinity of the electron chemical potential , @xmath26 , in neutron star matter @xcite so that the threshold condition for the onset of @xmath5 condensation , @xmath27 , could be fulfilled in the cores of neutron stars . the situation is graphically illustrated in figure [ fig : kmass ] , with the uncertainties inherent in the calculations indicted by the shaded regions . similar to the @xmath28 condensate , a @xmath5 condensate would also soften the equation of state and enhance the star s neutrino luminosity @xcite . the softening of the equation of state can be quite substantial , as pointed in @xcite , reducing the maximum neutron star mass from @xmath29 to about @xmath30 . based on this , the existence of a large number of low - mass black holes in the galaxy was proposed @xcite . another boson that could make its appearance in the center of a neutron star is the so - called h - dibaryon , a doubly strange six - quark composite with spin and isospin zero , and baryon number two @xcite . since its first prediction in 1977 , the h - dibaryon has been the subject of many theoretical and experimental studies as a possible candidate for a strongly bound exotic state . in neutron stars , which may contain a significant fraction of @xmath13 hyperons , the @xmath13 s could combine to form h - dibaryons , which could give way to the formation of h - matter at densities somewhere between @xmath31 @xcite and @xmath32 @xcite , depending on the in - medium properties of the h - dibaryon . h - matter could thus exist in the cores of moderately dense neutron stars . in @xcite it was pointed out that h - dibaryons with a vacuum mass of about 2.2 gev and a moderately attractive potential in the medium of about @xmath33 mev could go into a bose condensate in the cores of neutron stars if the limiting star mass is about that of the hulse taylor pulsar psr 1913 + 16 , @xmath34 . conversely , if the medium potential were moderately repulsive , say around @xmath35 mev , the formation of h - dibaryons may only take place in neutron stars heavier than @xmath36 . if formed , however , h - matter may not remain dormant in neutron stars but , because of its instability against compression could trigger the conversion of neutron stars into hypothetical strange stars @xcite . it has been suggested already back in the 1970 s by a number of researchers @xcite that neutrons , protons plus the heavier constitutes ( @xmath37 ) can melt , creating the quark - gluon plasma state being sought at the most powerful terrestrial heavy - ion colliders at cern and rhic . ( evidence for the creation of such matter in the framework of cern s lead beam program has been claimed earlier this year @xcite . ) at present one does not know from experiment at what density the expected phase transition to quark matter occurs , and one has no conclusive guide yet from lattice qcd simulations . from simple geometrical considerations it follows that nuclei begin to touch each other at densities of @xmath38 , which , for a characteristic nucleon radius of @xmath39 fm , is less than twice @xmath40 . above this density , therefore , one may expect that the nuclear boundaries of hadrons begin to dissolve and the formerly confined quarks populate free states outside of the hadrons . depending on rotational frequency and stellar mass , densities as large as two to three times @xmath40 are easily surpassed in the cores of neutron stars , so that the neutrons and protons in the centers of neutron stars may have been broken up into their constituent quarks by gravity @xcite . more than that , since the mass of the strange quark is only @xmath41 mev , high - energetic up and down quarks will readily transform to strange quarks at about the same density at which up and down quark deconfinement sets in @xcite . it is therefore three - flavor quark matter that could exist as a permanent component of matter in the centers of neutron stars @xcite . it is most fascinating that observational astrophysicists may be able to spot evidence for the existence of this novel phase of matter in the neutron star data , as we shall see next . whether or not quark deconfinement occurs in neutron stars makes only very little difference to their static properties , such as the range of possible masses and radii , which renders the detection of quark matter in such objects extremely complicated . this turns out to be strikingly different for rotating neutron stars ( i.e. pulsars ) which develop quark matter cores in the course of spin - down . the reason being that as such stars spin down , because of the emission of magnetic dipole radiation and a wind of electron - positron pairs , they become more and more compressed . for some rotating neutron stars the 0.50 cm mass and initial rotational frequency may be just such that the central density rises from below to above the critical density for dissolution of baryons into their quark constituents . this is accompanied by a sudden shrinkage of the neutron star , which , as calculated in @xcite , occurs at frequencies around 200 hz ( angular velocities @xmath42 ) . this effects the star s moment of inertia dramatically , as shown in figure [ fig : ioab ] . depending on the ratio at which quark and normal matter change with frequency , the moment of inertia can decrease very anomalously , and could even introduce an era of stellar spin - up lasting for @xmath43 years @xcite . since the dipole age of millisecond pulsars is about @xmath44 years , we may roughly estimate that about 10% of the @xmath45 solitary millisecond pulsars presently known could be in the quark transition epoch and thus could be signaling the ongoing process of quark deconfinement ! changes in the moment of inertia reflect themselves in the braking index , @xmath8 , of a rotating neutron star , as can be seen from @xmath46 @xmath47 the right - hand - side of this expression reduces to the well - known canonical constant @xmath48 only if @xmath49 is independent of frequency . evidently , this is not the case for rapidly rotating neutron stars , and it fails completely for stars that experience pronounced internal changes which alter the moment of inertia significantly . figure [ fig : nvso ] illustrates this for the neutron star of figure [ fig : ioab ] . because of the changes in @xmath49 , driven by the transition into quark matter , the braking index deviates dramatically from 3 at the transition frequency , when pure quark matter is generated . such dramatic anomalies in @xmath50 are not known for conventional neutron stars , because their moments of inertia appear to vary smoothly with @xmath51 @xcite . the future astrophysical observation of such anomalies in the braking behavior of pulsars may thus be interpreted as a signal for quark deconfinement in neutron stars ( `` quark astronomy '' ) . accreting x - ray neutron stars provide a very interesting contrast to the spin - down of isolated neutron stars . these x - ray neutron stars are being spun up by the accretion of matter from a lower - mass ( @xmath52 ) , less - dense companion . if the critical deconfinement density falls within that of the canonical pulsars , quark matter will already exist in them but will be `` spun out '' of x - ray stars as their frequency increases during accretion . this scenario has been modeled in @xcite , where it was found that quark matter remains relatively dormant in the core of a neutron star until the star has been spun up to frequencies at which the central density is about to drop below the threshold density at which quark matter exists . as known from above , this manifests itself in a significant increase of the star s moment of inertia . the angular momentum added to a neutron star during this phase of evolution is therefore consumed by the star s expansion , inhibiting a further spin - up until the quark matter has been converted into a mixed phase of matter made up of hadrons and quarks . such accreters , therefore , tend to spend a greater length of time in the critical frequencies than otherwise . there will be an anomalous number of accreters that appear at or near the same frequency . this is what was found recently with the rossi x - ray timing explorer ( rxte ) . quark deconfinement constitutes a most natural explanation for this phenomenon , though alternative explanations are possible too @xcite . recently it was discovered that instantons may cause strong correlations between up and down quarks , which could give way to the existence of colored diquark pairs in superdense matter @xcite . these pairs would form a bose condensate in cold and dense quark matter . carrying color charges , the condensate ought to exhibit color superconductivity @xcite . both the magnitude of the gap and the critical temperature associated with the color superconductive phase were estimated to be on the order of @xmath53 mev @xcite . the implications of such tremendous gaps for the magnetic fields of pulsars and their thermal evolution were explored in @xcite and @xcite , respectively . color superconductivity may also be linked to the sudden irregularities observed in the rotational frequencies of some pulsars , known as glitches , as shown in the very recent study @xcite . so far we have assumed that quark matter forms a state of matter higher in energy than atomic nuclei . this most plausible assumption , however , may be quite deceiving @xcite because for a collection of more than a few hundred @xmath54 quarks , the energy per baryon ( @xmath55 ) of quark matter can be just as well below the energy of the most stable atomic nucleus , @xmath56 , whose energy per baryon number is @xmath57 mev , with @xmath58 the mass of the @xmath59fe atom . a simple estimate indicates that for strange quark matter @xmath60 , so that bag constants of @xmath61 ( i.e. @xmath62 mev ) and @xmath63 ( @xmath64 mev ) place the energy per baryon of such matter at @xmath65 mev and 915 mev , respectively @xcite . obviously , these values correspond to quark matter which is absolutely bound with respect to @xmath59fe . in this event the ground state of the strong interaction would be strange quark matter ( strange matter ) , made up of @xmath54 quarks , instead of nuclear matter . this finding is one of the most startling predictions of modern physics with far - reaching implications for neutron stars , for all hadronic stellar configurations in figure [ fig : cross ] would then be only metastable with respect to stars made up of absolutely stable 3-flavor strange quark matter @xcite . if this is indeed the case , and if it is possible for neutron matter to tunnel to quark matter in at least some neutron stars , then in appears likely that all neutron stars would in fact be strange stars @xcite . in sharp contrast to the other stars in figure [ fig : cross ] , made up of hadronic matter , possibly in phase equilibrium with quarks , strange stars consist nearly entirely of pure 3-flavor quark matter , eventually enveloped in a thin nuclear crust whose density is always less than neutron drip density ( @xmath66 ) @xcite . the hypothetical , absolute stability of strange matter gives way to a variety of novel objects which stretch from strangelets at the small baryon number end , @xmath67 , to strange machos and strange dwarfs , to the compact strange stars at the high end , @xmath68 , where strange matter becomes unstable against gravitational collapse @xcite . the strange counterparts of ordinary atomic nuclei are the strange nuggets , searched for in high - energy collisions at brookhaven ( e.g. e858 , e864 , e878 , e882-b , e886 , e896-a ) , cern ( newmass experiment na52 ) , balloon - borne experiments ( crash ) , and terrestrial experiments ( e.g. hadron ) @xcite . ( for a recent review , see @xcite as well as this volume . ) one of the most distinguishing features of the compact strange stars are their anomalously small radii which enables the whole branch of compact strange stars to withstand very rapid rotation , down to periods in the half - millisecond regime @xcite . in contrast to this , only neutron stars very close to the limiting - mass value may be able to withstand such rapid rotation . an interesting candidate for a small - radius `` neutron '' star could be sax j1808.4 - 3658 whose radius , as argued in @xcite , may be significantly smaller than the canonical neutron star radius of @xmath69 km ( see however @xcite ) . other distinguishing features originate from the fact that the surfaces of bare strange stars are made up of quark matter which allows for stellar luminosities as high as several @xmath70 @xcite . the predominant cooling mechanism of hot ( temperatures of several @xmath71 k ) newly formed neutron stars immediately after formation is neutrino emission , with an initial cooling time scale of seconds . already a few minutes after birth , the internal neutron star temperature drops to @xmath72 k @xcite . photon emission overtakes neutrinos only when the internal temperature has fallen to @xmath43 k , with a corresponding surface temperature roughly two orders of magnitude smaller . neutrino cooling dominates for at least the first @xmath73 years , and typically for much longer in standard cooling ( modified urca ) calculations . being sensitive to the adopted nuclear equation of state , the neutron star mass , the assumed magnetic field strength , the possible existence of superfluidity , meson condensates and quark matter , theoretical cooling calculations , as summarized in figure [ fig : cool ] , provide most valuable information about the interior matter and neutron star structure . the stellar cooling tracks in figure [ fig : cool ] are computed for a broad collection of equations of state @xcite , including those shown in figure [ fig : eos ] . knowing the thermal evolution of a neutron star also yields information about such temperature - sensitive properties as transport coefficients , transition to superfluid states , crust solidification , and internal pulsar heating mechanisms such as frictional dissipation at the crust - superfluid interfaces @xcite ) . this work was supported by the deutsche forschungsgemeinschaft ( dfg ) . 10 madsen j 1994 _ physics and astrophysics of strange quark matter , in : proc . of 2nd int . conf . on phys . and astrophys . of quark - gluon plasma , calcutta _ ed b sinha y p viyogi and s raha ( singapore : world scientific ) p 186	it is generally agreed on that the tremendous densities reached in the centers of neutron stars provide a high - pressure environment in which numerous novel particles processes are likely to compete with each other . these processes range from the generation of hyperons to quark deconfinement to the formation of kaon condensates and h - matter . there are theoretical suggestions of even more exotic processes inside neutron stars , such as the formation of absolutely stable strange quark matter , a configuration of matter even more stable than the most stable atomic nucleus , iron . in the latter event , neutron stars would be largely composed of pure quark matter , eventually enveloped in a thin nuclear crust . no matter which physical processes are actually realized inside neutron stars , each one leads to fingerprints , some more pronounced than others though , in the observable stellar quantities . this feature combined with the unprecedented progress in observational astronomy , which allows us to see vistas with remarkable clarity that previously were only imagined , renders neutron stars to nearly ideal probes for a wide range of physical studies , including the role of strangeness in dense matter . [ sctn ] fw - undap-08 - 00 + * strangeness in neutron stars * + fridolin weber + university of notre dame + department of physics + 225 nieuwland science hall + notre dame , in 46556 - 5670 , usa + http://nta0.lbl.gov/@xmath0fweber + + _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ presented at the 5th international conference + on + strangeness in quark matter ( strangeness 2000 ) + berkeley , california , usa + july 2025 , 2000 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * strangeness in neutron stars * + fridolin weber + university of notre dame + department of physics + 225 nieuwland science hall + notre dame , in 46556 - 5670 , usa + http://nta0.lbl.gov/@xmath0fweber + 1.0truecm * strangeness in neutron stars *
in the framework of discrete flavor symmetries one of the most used ansatz for the lepton mixing matrix @xmath0 is the tribimaximal mixing ( tbm ) matrix @xcite , which despite the non - zero value for @xmath2 recently confirmed by the t2k collaboration @xcite , double chooz @xcite , daya bay @xcite and reno @xcite experiments , can still be used as a good approximation . contributions from the charged lepton sector and/or from renormalization effects , can generate a non - zero value for @xmath2 in agreement with the experimental data . for a classification of models predicting tbm mixing see @xcite and references therein , and for an overview of flavor symmetry models for neutrino mixing , based on non - abelian discrete symmetries that give the tbm pattern , see ref . @xcite . other ansatzs for the neutrino mass matrix have been explored . one such case consists on the so - called @xmath3-zero textures where , for example , taking the case of majorana neutrinos , it has been found that only two independent zero entries are allowed by current data @xcite . for dirac mass matrices the situation is different ( since the mass matrices are not symmetric ) as shown in @xcite , where a study of the minimal mass matrices with as many zeros as possible determined that up to five zero entries are allowed . in both of these cases the neutrino mass matrices were expressed in the charged lepton diagonal basis . fritzsch - like @xcite textures have also been considered for both the charged lepton and neutrino mass matrices in @xcite , where dirac neutrinos are considered , and in @xcite , where all possible hermitian six - zero fritzsch - like as well as non fritzsch - like textures for lepton mass matrices have been investigated , for both majorana and dirac matrices . for a recent overview of all possible cases of fritzsch - like as well as non fritzsch - like matrices as possible textures of the fermion mass matrices see @xcite . in this paper we take the following approach : assuming that the neutrino mass matrix is diagonalized by the tbm matrix , we determine the textures for the charged lepton mass matrix that leads to a lepton mixing matrix in agreement with the experimental data . we find both the number and location of zeros in the charged lepton mass matrix as well as the relative sizes of the non - zero entries in the corresponding yukawa matrix . the results obtained can be of use to model builders interested in obtaining flavor models with tbm mixing in the neutrino sector . in section [ sec : setup ] we present the general idea and setup of our analysis . the @xmath3-zero textures for @xmath4 are obtained in section [ sec : zeros ] . section [ sec : analysis ] contains a description of the numerical analysis performed , as well as our results and discussion . finally , section [ sec : conclusions ] contains our conclusions . [ sec : setup ] the lepton mixing matrix @xmath0 is given by @xmath5 where @xmath6 relates left - handed leptons in the weak basis @xmath7 to the mass basis leptons @xmath8 : @xmath9 ; and @xmath10 corresponds to the same basis transformation for left - handed neutrinos @xmath11 the @xmath0 matrix can be parametrized as follows @xcite : @xmath12 where @xmath13 , @xmath14 and @xmath15 is the dirac cp violating phase . since the @xmath16 matrices in are unitary , we can express @xmath6 as @xmath17 in our approach , we demand that @xmath10 be the tbm matrix , namely @xmath18 and so , if @xmath19 is the matrix that diagonalizes the neutrino mass matrix @xmath20 and @xmath6 the one that diagonalizes the _ squared _ charged lepton mass matrix @xmath21 , we want to determine the form of @xmath22 such that @xmath23 has values in the allowed experimental range . as described above the matrix @xmath6 diagonalizes @xmath24 and thus satisfies @xmath25 where @xmath26 and @xmath27 . at this point @xmath28 is an arbitrary mass matrix with unknown values and we parametrize it as @xmath29 in general the mass matrix @xmath28 has complex entries , however , in this first approach we consider the particular case of a real mass matrix and thus assume all parameters in @xmath22 to be real . in this particular case , eq . becomes a set of equations for the entries of @xmath30 : @xmath31 there are nine variables and only six equations encoded in eq . . thus we have an incomplete set of equations . to proceed we can start by making some assumptions on the _ form _ of the lepton mass matrix . just as an example take @xmath28 to be a lower triangular matrix ( by making @xmath32 ) . this gives @xmath33 the number of variables has now been reduced to six variables , and with the six independent equations ( as we show later in the paper ) it is easy to find solutions for the parameters in eq . ( [ muem ] ) in agreement with the experimental data . we are interested in determining the textures with the maximum number of zeros . the main motivation for this is that having such textures can be useful to model builders interested in obtaining them using an underlying flavor symmetry . with this in mind , we explore the textures with the maximum number of zeros such that eq . has real solutions in the allowed experimental range for the mixing angles . note that our assumption of real mass matrices implies that in our case the cp violating phase satisfies @xmath34 or @xmath35 . we also assume that the majorana cp violating phases are zero but note however , that since they can be factorized from @xmath0 as a diagonal matrix , their inclusion would not affect the magnitudes of the @xmath0 entries . we are concerned with the possible textures for the non - diagonal mass matrix for charged leptons ( as we said before , assuming that the neutrino mass matrix is diagonalized by the tbm matrix ) . we assume that the entries are real and so @xmath36 . zeros in the entries of our matrices may be due to some underlying symmetry(ies ) and because of that we will look for the textures with the highest possible number of zeros . as discussed above , we have a system of six coupled , quadratic equations and ( in general ) nine variables . thus there is a need for some extra constraints on the variables . for matrices with no zeros at all we need three of this extra constraints , for matrices with one zero we need only two . we begin our analysis with three - zero mass matrices requiring no extra constraints . a combinatorial analysis shows that , for the three - zero matrices , we have @xmath37 ways to place the three zeros and thus @xmath38 different matrix forms . nevertheless , it is well known that in forming the real product @xmath39 , any permutation of the columns of the matrix involved does not change the resulting product matrix . thus , this reduces the number of matrices to @xmath40 . from these textures we discard those with a row or column made of zeros because they are singular . there are @xmath41 cases of this and we are left with only @xmath42 different textures . another condition that has to be satisfied , in view of the non - zero values for the charged lepton masses , is @xmath43 . denoting the possible mass matrices by @xmath44 , where indexes stand for the @xmath45 texture of the non - diagonalized charged lepton matrix with @xmath3 zeros , for @xmath4 we obtain : @xmath46 as none of the matrices shown above is obtained from another by means of permuting columns , these are all the possible three - zero textures to be analyzed . we are interested in identifying the textures which provide solutions to eq . . in order to find which textures provide solutions we performed ( for each texture ) a scan over the whole allowed experimental range for the mixing angles taking the values @xmath47 and @xmath48 for the cp - violating phase . to carry out the numerical analysis we used the experimental data at 3@xmath49 from the global neutrino data analysis @xcite @xmath50 with normal ( inverted ) hierarchy , and the recently daya bay results @xcite ( confirmed at 5@xmath49 ) @xmath51 which can be rewritten as @xmath52 . for the charged leptons masses we use the values given in @xcite @xmath53 for instance by taking the central values of the mixing angles and those of the charged lepton masses ( setting @xmath47 ) we obtain @xmath54 and from eq . we get for @xmath6 @xmath55 if we use these values for the example of the lower triangular matrix presented in section [ sec : setup ] , equating eq . to the explicit value of the right hand side of eq . , and using eq . and @xmath56 evaluated at the central value for the lepton masses , we find : @xmath57 + where @xmath58 is in mev and the yukawa matrix @xmath59 shown above is obtained from : @xmath60 being @xmath61 gev the vacuum expectation value ( vev ) of the higgs field @xcite ( here we are assuming only one higgs doublet contributes to the mass matrix . in general in this paper we assume either that , or if there are more su(2 ) doublets involved , that they all acquire vev s of @xmath62 gev ) . to perform the complete scan we used a grid of dimensions @xmath63 in the experimental range of the three mixing angles . for each combination of the mixing angles that provides solutions to eq . , we select only the real solutions . once those are determined , we then also determine the order of magnitude of the yukawa mass entries by finding their minimum and maximum values . from all the possible three - zero textures ( shown in eq . ) only the following provide ( real ) solutions : @xmath64 , @xmath65 , @xmath66 , @xmath67 , @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72 , and @xmath73 . for the textures @xmath74 , @xmath75 , @xmath76 , @xmath77 , @xmath78 , and @xmath79 we did not find any solution . the reason is that all these textures lead to an extra constriction because the matrix product @xmath80 has a zero entry . those entries are @xmath81 , @xmath82 , @xmath83 , @xmath84 , @xmath85 , and @xmath86 . as mentioned above , once we have performed all calculations ( including the normalization of the mass matrix by the vev of the higgs field ) we find the maximum and minimum orders of magnitude of the entries for each texture of the yukawa matrix . they are shown here ( up to @xmath87 coefficients ) : @xmath88 as part of the analysis we performed , we searched for the solution volume of each three - zero texture , that is , the set of points given by the three mixing angles that make possible to find real solutions to the entries of the @xmath28 matrix . in the case of textures @xmath65 @xmath73 the solution volume fills the complete space given by the experimental intervals of the three mixing angles ( the experimentally allowed parameter space ) . this happens for both @xmath47 and @xmath48 . the most interesting case is the texture @xmath64 , which has the zeros on the major diagonal ( and its permutations ) . the relevance of this case is that the angle @xmath89 is now very restricted . thus , we took a closer look within a smaller interval for @xmath89 using a finer grid of @xmath90 , and analyzing a @xmath89 interval of @xmath91 $ ] , we find that the solution volume of @xmath64 , for both @xmath47 and @xmath48 , is a thin curved surface ( presented in figure [ fig1 ] for @xmath48 ) ; the only difference between these two cases is a displacement of the interval allowed for @xmath89 , which is @xmath92 $ ] for @xmath47 , and @xmath93 $ ] for @xmath48 . these intervals are near but exclude the central value of @xmath89 . for @xmath48 , found with a grid of @xmath90 . similar results are obtained for @xmath47.,width=340 ] the yukawa matrices in can explicitly be written as @xmath94 then , the diagonal entries @xmath95 are just the sum of the square of the elements of the i - th row of @xmath96 : @xmath97 the textures obtained in subsection [ subsec : results ] present some interesting features : first note that in all textures the largest upper bound ( @xmath98 ) always appears in the second row . thus , from eq . , @xmath99 . this is a little higher than @xmath100 ) . note also that ( except for the @xmath101 ) the third row has always an upper bound of @xmath102 and then , from eq . , we can deduce that @xmath103 . this is quite close to the value of the ratio @xmath104 . more precisely and using the explicit values of the yukawa matrix of eq . we find @xmath105 we thus find that there are extra conditions on the entries of @xmath96 in terms of the charged leptons masses . because of equation ( [ equ ] ) , we can expect that these conditions may be due to an internal structure of the mixing matrix @xmath6 for the charged leptons . in what follows we show that there is indeed a parametrization for this matrix that leads to specific relations between the entries in the yukawa matrices and the charged lepton masses that reproduce the above results , including the small deviations in the orders of magnitude . motivated by the fact that the charged leptons are analogous to the down type quarks in terms of representations of the standard model gauge group , we consider a ckm - like parametrization for the @xmath6 mixing matrix in terms of the three angles @xmath106 , @xmath107 and @xmath108 , taking values between @xmath109 and @xmath110 , and all phases set to zero in order to use real values for the @xmath6 matrix , as we have done in the whole analysis . then , we can write : @xmath111 using @xmath112 instead of @xmath0 ( to avoid agglomeration of indexes ) one can easily find that : @xmath113 @xmath114 @xmath115 + here the @xmath116 are , as before , the parameters of @xmath0 ; different from @xmath117 , which parametrize @xmath1 . we evaluated equations ( [ e1])-([e3 ] ) with @xmath118 $ ] , and took all the points that matched the experimental intervals given by ( [ experim1 ] ) and ( [ experim2 ] ) . the points ( @xmath119 ) are presented in figure [ fig2 ] . in figure [ fig2 ] the blue region is made of those values of the angles that solve equation ( [ e1 ] ) within the experimental values , the purple points are those that solve equation ( [ e1 ] ) and ( [ e2 ] ) within the experimental ranges , and the dark yellow region is the set of points that satisfies the three equations ( [ e1 ] ) , ( [ e2 ] ) and ( [ e3 ] ) within all the experimental values . we then obtain a small region given by two small angles and a large one : @xmath120\\ \theta^{l}_{13}&=&[0.07 - 0.16]\nonumber\\ \theta^{l}_{23}&=&[1.43 - 1.57\approx(\pi/2)]\nonumber\end{aligned}\ ] ] now we use the parametrization of the @xmath1 matrix given by equation ( [ para ] ) and the values of its angles to make an analysis of the entries in a general charged lepton mass matrix , of the form given by ( [ gen ] ) . let us remember that @xmath1 appears in ( [ equ ] ) . using the values given in ( [ values ] ) , one can built a simpler form for @xmath1 . indeed , according to equation ( [ para ] ) and using the fact that @xmath121 and @xmath122 are small , we can use a first - order approximation for their respective rotation matrices . the size of the angle @xmath123 allows us to use a first order approximation too , but around @xmath124 . then , @xmath125 and @xmath126 . thus , we end up with : @xmath127 + using ( [ apro0 ] ) in ( [ equ ] ) , we obtain the matrix @xmath128 in terms of @xmath106 , @xmath107 and @xmath129 . comparing the result with ( [ mue ] ) , we can see that the variables from the first row of the mass matrix must satisfy : @xmath130 the variables of the second row of ( [ gen ] ) satisfy : @xmath131 and the third row of ( [ gen ] ) must have a contribution of the form @xmath132 these expressions explain the observations considered at the beginning of the present subsection , including the corrections to the order of magnitude estimates , and represent the main result of this analysis . taking our specific example above , dividing equations ( [ row1])-([row3 ] ) by @xmath133 as in ( [ overhiggs ] ) , and using the values @xmath134 , @xmath135 , @xmath136 as well as the central values for the charged lepton masses we have : @xmath137 these are the values from ( [ yuk ] ) , which are relations that exhibit the dependence of the entries of @xmath96 on the masses and the mixing angles of @xmath1 . it is easily observed that for the first row , although the leading term ( zeroth order on the angles ) is @xmath138 , the other contributions play a more important role than this one , because the masses of the @xmath139 and @xmath140 leptons are much greater than that of the electron and even overcome the small size of their coefficients ( which are powers of the angles ) . finally , we note that these expressions are independent of the texture chosen for the yukawa matrix and also independent of the two values of the dirac phase used in the pmns matrix . all the yukawa textures found exhibit the relations given by equations ( [ row1])-([row3 ] ) , so an important feature that we want to point out is that the matrices we have found show a very large mixing ( nearly @xmath110 ) between the second and third generations of the charged leptons . we have already explored the textures with three zeros and found that it is possible to have solutions in the allowed ranges for many of them . now we are interested in checking if this is also possible for textures with four zeros . to start , we list the possible textures with four zeros and their equivalent ( column ) permutations , discarding those with a row or column completely filled with zeros ( thus avoiding singular matrices ) : @xmath141 we have also calculated the product @xmath80 for each texture and we can observe that most of them render one or two zero entries . as in the previous case , this increases the difficulty in finding solutions to the system of equations in eq . . moreover , since we are assuming only real values in the mass matrix entries , for the four - zero textures we are left with @xmath142 equations and only @xmath143 variables to satisfy them and the system is over - constrained . the zero entries in the product @xmath80 for each texture are the following : @xmath144 the only textures which give non - zero entries in @xmath80 are @xmath145 , and @xmath146 . those would be the more interesting cases to study if we allow complex parameters ( this would imply also to allow the @xmath15 cp violating phase to take all its possible values ) . @xmath147 , @xmath148 and @xmath149 also render non - zero entries in @xmath150 but they are singular . we performed a scan over all the allowed experimental range for the mixing angles trying to solve the system of equations for each particular texture , in the case of real parameters , without success . the only possibility is to consider cases where one of the zero entries in the four - zero texture gets slightly turned on . this case could correspond then to one of the three - zero texture above where one of the entries is much smaller than all others . analyzing the three - zero textures and the relative size of their entries ( presented above ) , we observe that except for @xmath64 , all of them contain entries in the range @xmath151 @xmath152 mev , and that the smallest one is at least two orders of magnitude smaller than the next one . it is then possible , from a model building perspective , to consider a four - zero texture where one of the zeros is _ lifted _ ( perhaps due to radiative corrections , or higher order operator mixings ) such that at the end one ends up with one of the successful three - zero textures above . here we show which four - zero texture could be used in such a scenario @xmath153 as a final remark we comment that even though at this point we considered only real matrices for the charged lepton mass matrices , it is possible to extend some of our results to the case of complex matrices with factorizable phases , i.e. mass matrices that can be written as @xmath154 , where @xmath155 are matrices with real entries and @xmath156 a diagonal phase matrix , @xmath157 . then , if @xmath158 is the orthogonal matrix which diagonalizes @xmath159 we can write , under our assumption @xmath160 , @xmath161 . nevertheless , from the property of orthogonality and @xmath162 we can infer that the only possible values for the factorizable phases are integer multiples of @xmath35 . we checked exactly the same three - zero textures as before , but allowing complex entries and a cp - violating phase @xmath163 and found solutions for the same textures as in the real case . the difference with our previous results is that in these cases solutions are found in all the allowed volume of angles , even for @xmath101 , and the order of magnitude of the entries are generally larger . in the four - zero texture we again find no solutions and due to the fact that the entries in the three - zero textures are larger in this case , it is not possible to easily start from a four - zero texture and lift one zero ( here we are thinking of a naturally small correction to the zero entry in question ) . we analyze several textures for the charged lepton mass matrix under the assumption that the neutrino mass matrix is diagonalized by the tbm matrix , and with the intention of maximizing the number of zeros in them . requiring the @xmath0 mixing matrix to be consistent with the allowed experimental range for mixing angles ( fixing the cp violating phase ) , and with ( the central values of ) the charged lepton masses , we explore the maximum number of zeros possible in a texture under the assumption of real charged lepton mass matrix entries . we find that there are ten three - zero textures which provide @xmath0 values in agreement with and also determine the size range for their entries . among the successful textures , the one with zeros in the diagonal shows an interesting behaviour in the sense that in order to work , it requires the mixing angle @xmath89 to lie in a very restricted range which , albeit consistent with the experimental range , it excludes the central experimental values . a general analysis of the successful textures showed that there are relations between their entries and the charged lepton masses . through a ckm - like parametrization of the @xmath1 mixing matrix we are able to obtain the texture - independent specific relations in terms of the three rotation angles in @xmath1 . we find no solutions for four - zero textures but observe that it is possible to consider cases where one the zeros is lifted in such a way that the remaining mass matrix corresponds to some of the successful three - zero textures . finally , we explore a possible extension to consider the cases with complex factorizable textures . we find similar results , i.e. solutions available for the three - zero textures and no solutions for the four - zero - texture , with the difference that the typical entry size of the charged lepton mass matrix is generally larger in this case . while preparing this paper , a new preprint appeared in the arxiv @xcite , where an analysis of the charged lepton mixing matrix was performed using the tbm matrix form as well as the bimaximal mixing matrix ( bm ) for the neutrino mixing matrix . they consider the charged lepton mixing matrix to be nearly the identity but deviated from it by one and two parameters for the bm and the tbm case respectively . the parameters are the cabbibo angle @xmath164 and @xmath165 for the tbm case , assuming that the @xmath6 has its own ckm - like parametrization . in our case the assumption of @xmath6 as a matrix near the identity is not used . instead , we determine this matrix by allowing the three rotation angles to vary over the whole experimental range and find that in fact , since @xmath166 , the matrix is not close to the identity also , in our context , the goal included to find the textures that are diagonalized with eq . . this is not explored in @xcite and we feel it is a useful result that can be used by model builders interested in flavor symmetries . k. abe _ et al . _ [ t2k collaboration ] , `` indication of electron neutrino appearance from an accelerator - produced off - axis muon neutrino beam , '' phys . lett . * 107 * , 041801 ( 2011 ) [ arxiv:1106.2822 [ hep - ex ] ] . l. dorame , d. meloni , s. morisi , e. peinado and j. w. f. valle , `` constraining neutrinoless double beta decay , '' nucl . b * 861 * , 259 ( 2012 ) [ arxiv:1111.5614 [ hep - ph ] ] . s. morisi and j. w. f. valle , `` neutrino masses and mixing : a flavour symmetry roadmap , '' arxiv:1206.6678 [ hep - ph ] . p. h. frampton , s. l. glashow and d. marfatia , `` zeroes of the neutrino mass matrix , '' phys . b * 536 * , 79 ( 2002 ) [ hep - ph/0201008 ] . h. fritzsch , z. -z . xing and s. zhou , `` two - zero textures of the majorana neutrino mass matrix and current experimental tests , '' jhep * 1109 * , 083 ( 2011 ) [ arxiv:1108.4534 [ hep - ph ] ] . c. hagedorn and w. rodejohann , `` minimal mass matrices for dirac neutrinos , '' jhep * 0507 * , 034 ( 2005 ) [ hep - ph/0503143 ] . h. fritzsch , phys . b * 70 * , 436 ( 1977 ) ; ibid . `` weak interaction mixing in the six - quark theory , '' phys . lett . b * 73 * , 317 ( 1978 ) . g. ahuja , m. gupta , m. randhawa and r. verma , `` texture specific mass matrices with dirac neutrinos and their implications , '' phys . d * 79 * , 093006 ( 2009 ) [ arxiv:0904.4534 [ hep - ph ] ] . n. mahajan , m. randhawa , m. gupta and p. s. gill , `` investigating texture six zero lepton mass matrices , '' arxiv:1010.5640 [ hep - ph ] . m. gupta and g. ahuja , `` possible textures of the fermion mass matrices , '' int . j. mod . a * 26 * , 2973 ( 2011 ) [ arxiv:1206.3844 [ hep - ph ] ] . k. nakamura _ et al . _ [ particle data group collaboration ] , `` review of particle physics , '' j. phys . g * g37 * , 075021 ( 2010 ) and 2011 partial update for the 2012 edition ( url : http://pdg.lbl.gov ) . t. schwetz , m. tortola , j. w. f. valle , `` where we are on @xmath2 : addendum to global neutrino data and recent reactor fluxes : status of three - flavour oscillation parameters , '' new j. phys . * 13 * , 109401 ( 2011 ) . [ arxiv:1108.1376 [ hep - ph ] ] .	assuming that the neutrino mass matrix is diagonalized by the tribimaximal mixing matrix , we explore the textures for the charged lepton mass matrix that render an @xmath0 lepton mixing matrix consistent with data . in particular we are interested in finding the textures with the maximum number of zeros . we explore the cases of real matrices with three and four zeros and find that only ten matrices with three zeros provide solutions in agreement with data . we present the successful yukawa textures including the relative sizes of their non - zero entries as well as some new and interesting relations among the entries of these textures in terms of the charged lepton masses . we also show that these relations can be obtained directly from a parametrization of the charged lepton mixing matrix @xmath1 .
large extra dimensions have recently attracted a lot of interest . they provide many new theoretical ideas to explore questions in particle physics . most excitingly , they also predict signals which will be tested at the upcoming collider experiments . in this talk we focus on the scenario of universal extra dimensions ( ueds ) @xcite . in ueds , all standard model ( sm ) fields propagates in extra dimensions of size @xmath1 . there are many theoretical motivations for this scenario , such as electroweak symmetry breaking @xcite , proton decay @xcite , the number of generations @xcite , neutrino masses @xcite , etc . the current experimental bounds allow kaluza - klein ( kk ) states in ueds to be as light as a few hundred gev @xcite . such light kk states can be copiously produced at the current or future colliders with center of mass energies @xmath2 tev . it is therefore interesting to study their collider phenomenology , and one might expect that it would be easy to discover these kk states . however , although their production cross section can be very large , their subsequent detection is non - trivial because they decay nearly invisibly . the phenomenology of ueds shows interesting parallels to supersymmetry . every standard model field has kk partners . the lowest level kk partners carry a conserved quantum number , kk parity , which guarantees that the lightest kk particle ( lkp ) is stable . heavier kk modes cascade decay to the lkp by emitting soft standard model particles . the lkp escapes detection , resulting in missing energy signals . given that the signals are very similar to supersymmetry ( susy ) , it is important to distinguish them by studying the properties of the new states produced in colliders in details . linear colliders often provide a clean environment for such kind of studies . however , in @xmath3 collisions , the signals will be plagued by the huge two photon backgrounds due to the softness of the observable decay products . on the other hand , the @xmath0 mode , as we will see , stands as a unique opportunity for a detailed study of the kk excitations of the electrons . this talk was based on work in collaboration with konstantin t. matchev and martin schmaltz @xcite . let us first describe the model . the simplest ued scenario has all of the standard model fields ( no supersymmetry ) propagating in a single extra dimension . in 4 + 1 dimensions , the fermions [ @xmath4 , where upper ( lower ) case letters represent @xmath5 doublets ( singlets ) ] are four - component and contain both chiralities when reduced to 3 + 1 dimensions . to produce a chiral 4d spectrum , we compactify the extra dimension on an @xmath6 orbifold . fields which are odd under the @xmath7 orbifold symmetry do not have zero modes , hence the unwanted fields ( zero modes of fermions with the wrong chiralities and the 5th component of the gauge fields ) can be projected out . the remaining zero modes are just the standard model particles in 3 + 1 dimensions . the full lagrangian of the theory comprises both bulk and boundary interactions . gauge and yukawa couplings and the higgs potential are contained in the bulk lagrangian in one - to - one correspondence with the couplings of the standard model . the boundary lagrangian interactions are localized at the orbifold fixed points and do not respect five dimensional lorentz invariance . ignoring the localized terms for the moment , the tree - level mass of the @xmath8-th kk mode is @xmath9 where @xmath10 is the radius of the compact dimension , and @xmath11 is the zero mode mass . the spectrum at each kk level is highly degenerate except for particles with large zero mode masses ( @xmath12 ) . the bulk interactions preserve the 5th dimensional momentum ( kk number ) . the corresponding coupling constants among kk modes are simply equal to the sm couplings ( up to normalization factors such as @xmath13 ) . the feynman rules for the kk modes can easily be derived ( e.g. , see ref . @xcite ) . in contrast , the coefficients of the boundary terms are not fixed by standard model couplings and correspond to new free parameters . in fact , they are renormalized by the bulk interactions and hence are scale dependent @xcite . one might worry that this implies that all predictive power is lost . however , since the wave functions of standard model fields and kk modes are spread out over the extra dimension and the new couplings only exist on the boundaries , their effects are volume suppressed . we can get an estimate for the size of these volume suppressed corrections with naive dimensional analysis by assuming strong coupling at the cut - off . the result is that the mass shifts to kk modes from boundary terms are numerically equal to corrections from loops @xmath14 . we will assume a symmetry which exchanges the two orbifold fixed points . the boundary terms on the two fixed points are equal under this symmetry . it is anomaly free and preserved by the radiative corrections of the bulk interactions . most relevant to the phenomenology are localized kinetic terms for the sm fields , such as @xmath15,\ ] ] where the dimensionless coefficients @xmath16 and @xmath17 are arbitrary and not universal for the different standard model fields . these terms are important phenomenologically for several reasons : ( _ i _ ) they split the near - degeneracy of kk modes at each level , ( _ ii _ ) they break kk number conservation down to a kk parity ( due to the symmetry between the two fixed points , ) under which modes with odd kk numbers are charged , ( _ iii _ ) they may introduce new flavor violations . for simplicity and definiteness , we will concentrate on the minimal universal extra dimensions ( mueds ) @xcite , where all boundary terms are assumed to vanish at some cutoff scale @xmath18 . at a lower scale , they will be generated by radiative corrections from the bulk interactions which are calculable . this is analogous to the case of the minimal supersymmetric standard model ( mssm ) where one has to choose a set of soft supersymmetry breaking parameters at some high scale in order to study its phenomenology . in some sense , our choice of the boundary terms may be viewed as analogous to the simplest minimal supergravity boundary condition universal scalar and gaugino masses . given this choice of the boundary condition , we can calculate the corrections to the kk spectrum , which depends on the boundary terms at low scales . since the corrections are small one can use the one - loop leading log approximation . in addition to the boundary terms , the kk spectrum also receives corrections from pure bulk effects . loops wrapping around the compact extra dimension with nonzero net winding numbers break the 5-dimensional lorentz invariance and therefore also correct the kk mass formula [ tree - spectrum ] . the bulk corrections are finite and exactly calculable due to their non - local nature . all these corrections , including both bulk and boundary contributions , were computed at one - loop in ref . @xcite . an example of the spectrum for the first level kk modes is shown in fig . [ fig : spectrum ] . typically , the corrections for kk modes with strong interactions are @xmath19 while those for states with only electroweak interactions are a few percent . we find that the corrections to the masses are such that @xmath20 . the lightest kk particle @xmath21 , is a mixture of the first kk mode @xmath22 of the @xmath23 gauge boson @xmath24 and the first kk mode @xmath25 of the @xmath26 @xmath27 gauge boson . we will usually denote this state by @xmath21 . however , note that the corresponding `` weinberg '' angle @xmath28 is much smaller than the weinberg angle @xmath29 of the standard model @xcite , so that the @xmath21 lkp is mostly @xmath22 and @xmath30 is mostly @xmath25 . the mass splittings among the level 1 kk modes are large enough for the prompt decay of a heavier level 1 kk mode to a lighter level 1 kk mode . but since the spectrum is still quite degenerate , the ordinary sm particles emitted from these decays will be soft , posing a challenge for collider searches . the terms localized at the orbifold fixed points also violate the kk number by even units . however , assuming that no explicit kk - parity violating effects are put in by hand , kk parity remains an exact symmetry . the boundary terms allow higher ( @xmath31 ) kk modes to decay to lower kk modes , and even level states can be singly produced ( with smaller cross sections because the boundary couplings are volume suppressed ) . thus kk number violating boundary terms are important for higher kk mode searches . once the radiative corrections are included , the kk mass degeneracy at each level is lifted and the kk modes decay promptly . the collider phenomenology of the first kk level is therefore very similar to a supersymmetric scenario in which the superpartners are relatively close in mass all squeezed within a mass window of 100 - 200 gev ( depending on the exact value of @xmath10 ) . each level 1 kk particle has an exact analogue in supersymmetry : @xmath32 bino , @xmath33 gluino , @xmath34 left - handed ( right - handed ) squark , etc . the decay cascades of the level 1 kk modes will terminate in the @xmath21 lkp ( fig . [ fig : transitions ] ) . just like the neutralino lsp is stable in @xmath10-parity conserving supersymmetry , the @xmath21 lkp in mueds is stable due to kk parity conservation and its production at colliders results in generic missing energy signals . at hadron colliders , the largest cross sections for kk states come from strongly interacting kk quarks and gluons because of the approximately degenerate kk spectrum . the level 1 kk states have to be pair produced due to kk parity conservation . their cross sections at the tevatron run ii and the lhc can be found in ref . @xcite and @xcite . in mueds , the heaviest kk state at level 1 is the kk gluon @xmath35 . it decays to a kk quark and an ordinary quark . the @xmath5 singlet kk quarks will decay to @xmath21 directly . the signature is jets@xmath36 , similar to the traditional squark and gluino searches @xcite . using the existing studies of the analogous susy studies , one might expect that run ii can probe @xmath37 gev while the lhc reach for @xmath38 can be @xmath39 tev @xcite . however , the jets will be relatively soft in the case of mueds due to the approximate degeneracy of the kk initial and final states , and the measured missing energy is correlated with the energy of the soft recoiling jets hence also small , even though the total missing mass is large . they may cause more difficulties in triggering and separation of backgrounds . further studies in the context of mueds are needed to obtain a better estimate of the reaches with the jetty signatures . the @xmath5 doublet kk quarks will mostly decay to kk @xmath40 and @xmath41 bosons , then subsequently decay to kk leptons and @xmath21 . they give rise to the much cleaner signatures with multi - leptons . using the very clean 4 lepton channel in ref . @xcite we estimate that run iib can go slightly beyond the current indirect bounds ( @xmath42 gev ) from precision data , and lhc can extend the reach to @xmath43 tev . although lhc will be able to extend the reach much beyond the current bounds , to identify the extra - dimensional nature of the new physics once some signals are found will be rather challenging . all signals from level 1 kk states look very much like supersymmetry all sm particles have `` partners '' with similar couplings . of course , the spins of these sm particle partners are different in susy and ueds , but this difference will most likely escape detection at a hadron collider . in addition , all the observed jets , leptons , and @xmath44 are relatively soft . it would be difficult to know exactly the mass scale of these kk states . additional experimental information will be needed to identify which scenario of new physics is realized and the mass scale associated with it . linear colliders provide a clean environment for precise studies of particle properties . taking the slepton production in supersymmetry as an example , one can get a precise measurement of the slepton mass and the lsp mass from the end points of the final state lepton energy distribution @xcite . the flat energy distribution between the end points will also tell the scalar nature of the sleptons produced . therefore , a linear collider can be a great tool to distinguish ueds from susy . given the bounds on the masses of the level 1 kk modes , the required energy for pair producing them is probably too high for the first stage of the next generation linear colliders , but may be reachable with energy upgrades or at clic . however , in @xmath3 collisions there are huge backgrounds from the two photon processes for very soft final state leptons or jet ( fig . [ fig:2photon ] ) . soft leptons or jets are produced by the collisions of the two soft photons from the initial state radiation , while the initial @xmath45 and @xmath46 go down the beam pipe , resulting in large missing energies . the two photon background for slepton search at a 500 gev linear collider is studied in details in ref . it was shown that for very soft final state leptons ( @xmath47 ) , the two photon background is still several order magnitude larger than the signal even after the cuts . in the case of mueds , the mass splittings between the kk leptons and @xmath21 are only about 13% . therefore the leptons coming from decays of the kk leptons will be completely buried in the two photon background . kk quarks receive larger mass corrections and the jets coming from their decay would be harder . their detection in the @xmath3 collisions is more promising . on the other hand , only kk electrons can be produced in the @xmath0 collisions ( if there is no flavor violation from the boundary terms ) . the signal will be two soft electrons of the _ same _ charge plus large missing energy , while the two photon processes produce fermions of the opposite charges . by charge identification one can remove the problematic two photon background . therefore , the @xmath0 mode of a linear collider stands as a unique opportunity for kk electron studies . all particle properties of the kk electrons should be able to be measured unambiguously , providing valuable information for identifying the corresponding new physics . it would also be useful to study possible lepton flavor violations if they are present as in the case of supersymmetry @xcite . universal extra dimensions with compactification radius near the tev scale promise exciting phenomenology for future colliders . all standard model particles have kk partners which can be produced with enormous cross sections at the lhc . however , the subsequent decays produce very soft standard model particles which can be difficult to see above the backgrounds , hence their detection and identification are rather non - trivial . the @xmath3 collider has one more piece of information as the total missing energy can be measured , but also suffers from the huge two photon background if the sm particles from the decays of the kk states are too soft . the @xmath0 collider provide a unique clean environment for the detailed studies of the kk electrons . as different possible new physics scenarios often give rise to complicated but sometimes similar collider phenomenologies , it is essential to have as many tools as possible to provide enough experimental information for our understanding of any new physics . i would like to thank k. t. matchev and m. schmaltz for collaboration on this work . i also thank n. arkani - hamed , m. chertok , a. cohen , b. dobrescu and b. schumm for useful discussions . this work is supported by the department of energy grant de - fg02 - 90er-40560 . t. appelquist , h .- c . cheng and b. a. dobrescu , phys . d * 64 * , 035002 ( 2001 ) , hep - ph/0012100 . n. arkani - hamed , h .- c . cheng , b. a. dobrescu and l. j. hall , phys . d * 62 * , 096006 ( 2000 ) , hep - ph/0006238 . t. appelquist , b. a. dobrescu , e. ponton and h. u. yee , phys . rev . lett . * 87 * , 181802 ( 2001 ) , hep - ph/0107056 . b. a. dobrescu and e. poppitz , phys . lett . * 87 * , 031801 ( 2001 ) , hep - ph/0102010 . t. appelquist , b. a. dobrescu , e. ponton and h. u. yee , hep - ph/0201131 . k. agashe , n. g. deshpande and g. h. wu , phys . lett . b * 514 * , 309 ( 2001 ) , hep - ph/0105084 . t. appelquist and b. a. dobrescu , phys . lett . b * 516 * , 85 ( 2001 ) , hep - ph/0106140 . cheng , k. t. matchev and m. schmaltz , hep - ph/0204342 . cheng , k. t. matchev and m. schmaltz , hep - ph/0205314 . c. macesanu , c. d. mcmullen and s. nandi , hep - ph/0201300 . d. a. dicus , c. d. mcmullen and s. nandi , phys . d * 65 * , 076007 ( 2002 ) , hep - ph/0012259 . h. georgi , a. k. grant and g. hailu , phys . b * 506 * , 207 ( 2001 ) , hep - ph/0012379 . t. g. rizzo , phys . d * 64 * , 095010 ( 2001 ) , hep - ph/0106336 . b. abbott _ et al . _ [ d0 collaboration ] , phys . lett . * 83 * , 4937 ( 1999 ) , hep - ex/9902013 ; t. affolder _ et al . _ [ cdf collaboration ] , phys . rev . lett . * 88 * , 041801 ( 2002 ) , hep - ex/0106001 . s. abel _ et al . _ [ sugra working group collaboration ] , hep - ph/0003154 . s. i. bityukov and n. v. krasnikov , phys . b * 469 * , 149 ( 1999 ) , hep - ph/9907257 . t. tsukamoto , k. fujii , h. murayama , m. yamaguchi and y. okada , phys . rev . d * 51 * , 3153 ( 1995 ) . t. abe _ et al . _ [ american linear collider working group collaboration ] , in _ proc . of the aps / dpf / dpb summer study on the future of particle physics ( snowmass 2001 ) _ ed . r. davidson and c. quigg , hep - ex/0106056 . n. danielson and l. goodman , `` point 2 selections and the standard model background , '' university of colorado , colo - hep-422 ( 1998 ) ; n. danielson , `` search for supersymmetry and the two photon background , '' university of colorado , colo - hep-423 ( 1998 ) , http://hep-www.colorado.edu/susy .	universal extra dimensions ( ueds ) with compactification radius near the tev scale provide interesting phenomenology at future colliders . the collider signals of the first kaluza - klein ( kk ) level are very similar to those of a supersymmetric model with a nearly degenerate superpartner spectrum . the heavier first level kk states cascade decay to the lightest kk particles ( lkp ) , which is neutral and stable because of kk - parity . the signatures involve missing energy and relatively soft jets and leptons which can be difficult for detection . the kk electron signal in @xmath0 collisions is free from the problematic two photon background therefore provides a unique opportunity for a detailed studies of the kk electrons in the universal extra dimension scenario . june 3 , 2002 efi-2002 - 86 + * universal extra dimensions + at the @xmath0 colliders * .5 in * hsin - chia cheng * .5 in _ enrico fermi institute , the university of chicago , chicago , il 60637 , usa _ .5 in
following the proposal of paczyski ( 1986 ) , experiments to search for lensing - induced light variations of stars ( microlensing events ) located in the galactic bulge and the magellanic clouds have been or are being conducted by several groups ( macho : alcock et al . 1993 ; eros : aubourg et al . 1993 ; ogle : udalski et al . 1993 ; moa : bond et al . 2001 ; duo : alard & guibert 1997 ) . these experiments have successfully detected a large number of events ( @xmath12 ) , most of which are detected towards the galactic bulge . despite a large number of event detections , the nature of the lenses is still poorly known . this is because the einstein ring radius crossing time @xmath13 ( einstein timescale ) , which is the only observable providing information about the physical parameters of the lens ( lens parameters ) , results from a combination of the lens parameters , i.e.@xmath14^{1/2},\ ] ] where @xmath15 is the einstein ring radius , @xmath16 is the lens mass , @xmath17 is the lens - source transverse speed , and @xmath18 and @xmath19 are the distances to the lens and the source from the observer , respectively . under this circumstance , the only approach one could pursue would be identifying the major lens population by statistically determining the lens mass function based on the observed timescale distribution . however , this approach requires _ a prior _ knowledge about the geometrical distribution of the lens , the lens kinematics , and the functional form of the lens mass function , which are all poorly known . in addition , even if all lensing objects were of the same mass , they would give rise to a broad range of timescale . as a result , it is difficult to identify the major lens population from this approach ( mao & paczyski 1996 ; gould 2001 ) . recently , from the _ hubble space telescope _ ( hst ) images of one of the large magellanic cloud ( lmc ) events ( macho lmc-5 ) taken 6.3 years after the original lens measurement , alcock et al . ( 2001 ) were able to resolve the lens from the lensed source star . by directly imaging the lens , they could identify that the event was caused by a nearby low mass star located in the galactic disk . besides the identification of the lens as a normal star , direct lens imaging is of scientific importance due to following reasons . first , by directly and accurately measuring the lens proper motion with respect to the source , @xmath20 , one can better constrain the physical parameters of the individual lenses . the previous method to determine @xmath20 is based on the analysis of the lensing light curves of events affected by the finite source effect , such as source - transit single lens events and caustic - crossing binary lens events ( gould 1994 ; witt & mao 1994 ; nemiroff & wickramasinghe 1994 ) . by analyzing the part of the light curves near the source transit or the caustic crossing of these events , one can measure the source star angular radius normalized by the angular einstein ring radius @xmath21 , i.e. @xmath22 , where @xmath23 is the angular source star radius . then , the lens proper motion is determined by @xmath24 for the proper motion determination by using this method , however , one should know the source star angular radius , which can only be deduced from an uncertain color - surface brightness relation . as a result , the proper motions determined in this way suffer from large uncertainties . by contrast , if the lens is resolved , the proper motion can be directly and thus accurately measured from the observed image . measuring the proper motion is equivalent to measuring the angular einstein ring radius because @xmath25 , where the event timescale is determined from the light curve . while @xmath13 depends on three lens parameters of @xmath16 , @xmath18 , and @xmath17 , @xmath21 does not depend on @xmath17 , and thus the lens mass can be better constrained . second , if the lens is resolved for an event where the lens - source relative parallax , @xmath26 , was previously measured during the lensing magnification , one can completely break the lens parameter degeneracy and the lens mass is uniquely determined by @xmath27 where @xmath28 ( gould 2001 ) . third , if the source of an event was resolved via either a source transit or a caustic crossing and thus @xmath29 was precisely measured , one can determine the angular source star radius by reversing the process of the classical method of the proper motion determination , i.e. @xmath30 . by measuring @xmath23 , one can determine the effective temperature of the source star , which is important for the accurate construction of stellar atmosphere models ( e.g. , alonso et al . 2000 ) . although the first directly imaged lens was identified for an lmc event , much more numerous direct lens identifications are expected if high resolution followup observations are performed for events detected towards the bulge . there are several reasons for this expectation . first , compared to the total number of lmc events , which is @xmath31 , there are an overwhelmingly large number of bulge events . second , while the majority of lmc events are suspected to be caused by dark ( or very faint ) objects , most bulge events are supposed to be caused by normal stars , for which imaging is possible . third , an important fraction of lenses responsible for bulge events are believed to be located in the galactic disk with moderate distances , and thus more likely to be imaged due to their tendency of being bright and having large proper motions . the goal of this work is to estimate the fraction of galactic bulge events whose lenses can be directly imaged . for this estimation , we first compute the expected distribution of the lens - source proper motions of the currently detected galactic bulge events based on standard models of the geometrical and kinematical distributions of lenses and their mass function ( 2 ) . we then apply realistic detection criteria for lens resolution and the result is presented as a function of the time elapsed after the original lensing measurement , @xmath0 ( 3 ) . based on the result in 3 , we discuss some of the observational aspects of the future high resolution followup lensing observations aimed for direct lens imaging ( 4 ) . we summarize the result and conclude in 5 . the first requirement for direct lens imaging is that the lens should have a large relative proper motion with respect to the source so that it can be widely separated from the lensed source star within a reasonable amount of @xmath0 . in this section , we , therefore , estimate the expected distribution of proper motions of galactic bulge events . with the models of the lens mass function , @xmath32 , the matter density distributions of the lens and source stars along the line of sight towards the galactic bulge field , @xmath33 and @xmath34 , and the kinematical distribution of lens - source transverse velocities , @xmath35 , the distribution of lens proper motions of galactic bulge events is computed by @xmath36 where @xmath37 and @xmath38 are the components of the transverse velocity which are respectively parallel with and normal to the galactic disk plane , @xmath39 is the detection efficiency of events as a function of @xmath13 , and the notation @xmath40 represents the dirac delta function . we note that the factors @xmath17 and @xmath15 are included in equation ( 4 ) to weight the event rate by the transverse speed and the lensing cross section . for the galactic bulge and disk matter density distributions , we adopt the models of dwek et al . ( 1995 ) and bahcall ( 1986 ) , respectively . in the bulge model , the bulge has a triaxial shape and the matter density is represented by an analytic form of @xmath41 ^ 2+(z'/z_0)^4,\ ] ] where @xmath42 kpc , and the coordinates @xmath43 represent the axes of the triaxial bar from the longest to the shortest , and the longest axis is misaligned with the line of sight toward the galactic center by an angle @xmath44 . the bahcall disk model is expressed by a double exponential form of @xmath45,\ ] ] where the radial scale length and the vertical scale height are @xmath46 pc and @xmath47 kpc , respectively . for the transverse velocity distribution , we adopt the model of han & gould ( 1995 ) . in this model , the velocity distributions for both disk and bulge components have a gaussian form of @xmath48 , \qquad i \in y,\ z.\ ] ] the means and the standard deviations of the individual velocity components for events with disk lenses and bulge source stars ( disk - bulge events ) are @xmath49 where @xmath50 corresponds to the rotation speed of the galactic disk and @xmath51 are the adopted velocity dispersions of stars in the solar neighborhood . for events with bulge source stars lensed by another foreground bulge stars ( bulge self - lensing events ) , the means and standard deviations of the transverse velocity distributions are @xmath52 for the barred bulge , the velocity dispersions along the axes of the bar are deduced from the tensor virial theorem ( binney & tremaine 1987 ) , resulting in @xmath53 @xmath54 . due to the projection effect caused by the bar misalignment , the projected velocity dispersions are computed by @xmath55^{1/2 } , \sigma_{z'})= ( 82.5 , 66.3)\ { \rm km}\ { \rm s}^{-1}$ ] , which correspond to the standard deviations in ( 9 ) . we note that the adopted velocity model is a rough approximation in the sense that it does not include factors such as the systematic mean motion of bulge stars discussed by evans & belokurov ( 2002 ) , the figure rotation of the bulge discussed by blum ( 1995 ) , and the possibility of non - gaussian nature of disk star velocity distribution discussed by evans & collett ( 1993 ) , which may actually be important in the resulting distribution of proper motions . since the inclusion of the mean motion or the figure rotation of bulge stars result in high proper motions , we note that the fraction of resolvable lenses predicted by the adopted bulge velocity model is the lower limit . the einstein timescale distribution of events observed by the macho group is claimed to be consistent with the distribution from normal stars ( alcock et al . 2000b ) . we , therefore , model the mass function of lenses based on the present day main - sequence stars determined by kroupa , tout & gilmore ( 1993 ) . the adopted mass function has a three power - law functional form of @xmath56 for the detection efficiency , we adopt the latest determination of the macho experiment ( alcock et al . 2000b ) , whose data were analyzed by using the ` difference image analysis ' method . = 12.0 cm -0.5 cm -0.2 cm = 12.0 cm -0.2 cm in figure 1 , we present the obtained distributions of relative lens - source proper motions for disk - bulge ( solid curve ) and bulge self - lensing ( dashed curve ) events , respectively . due to the large uncertainties in the geometrical and kinematical distributions of lenses , the relative contribution of the disk and bulge lenses to the total event rate is very uncertain . we , therefore , leave the proper motion distributions of the disk - bulge and bulge self - lensing events separately , instead of estimating the distribution of total events by arbitrarily normalizing the ratio between the two populations of events . from the figure , one finds that , as expected , the average proper motion of events caused by disk lenses is larger than that of events caused by bulge lenses . the lens detectability is additionally restricted by the lens brightness . this is not only simply because the lens should be brighter than a detection limit , but also because the threshold lens - source separation for lens detection , @xmath57 , varies depending on the apparent lens / source flux ratio . therefore , we compute the angular threshold as a function of the magnitude difference between the lens and the source , @xmath58 . for the computation of @xmath59 , we first model the point - spread - function ( psf ) of a stellar image as a gaussian , where the standard deviation @xmath60 of the psf characterizes the resolving power of the instrument to be used for followup lensing observations . we then generate the combined image of the lens and the source by normalizing such that the volume under each psf is proportional to the flux of each star . once the combined image is constructed , we then judge whether the lens and the source can be resolved each other . for this judgment , we assume that the lens can be resolved if the sign of the derivative of the combined image s one - dimensional psf profile changes more than two times in the overlapping region between the centers of the lens and the source images ( see figure 2 for illustration ) . then , the angular threshold is defined as the lens - source separation at which the lens is just to be resolved from the source star . in figure 3 , we present the computed angular threshold as a function of @xmath61 . the two curves correspond to the distributions expected when the followup observations are carried out by using two different instruments , whose resolving powers are characterized by @xmath62 ( dashed curve ) and @xmath63 ( solid curve ) , respectively . we note that the advanced camera for surveys ( acs ) recently installed on hst can achieve a resolution of @xmath64 . we also note that the near infrared camera ( nircam ) of _ next generation space telescope _ ( ngst ) , which will have an aperture of 6 7 m , will be sensitive in the wavelength range of 0.6 to 5 microns , and thus can achieve @xmath65 . from the figure , one finds that as the lens becomes fainter , it becomes difficult to resolve the lens from the source star . = 12.0 cm -0.5 cm -0.2 cm with the computed distribution of @xmath59 , we recalculate the distribution of proper motions of events by using equation ( 4 ) , but in this time only for events with detectable lenses . to be detected , the lens should meet the condition of @xmath66 . in addition , we also restrict that detectable lenses should be brighter than a threshold magnitude of @xmath67 . for this computation , we assume that the source star has a fixed brightness of @xmath68 , which corresponds to that of a bulge clump giant , while the lens brightness is deduced from its mass by using the mass - luminosity relation provided by kroupa , tout & gilmore ( 1993 ) . once the absolute magnitudes of the source and the lens are set , their apparent magnitudes are computed considering their distances from the observer , i.e. @xmath19 for the source and @xmath18 for the lens . = 12.0 cm -0.5 cm -0.2 cm in figure 4 , we present the finally determined fractions of events with detectable lenses as a function of @xmath0 . in the figure , the thick and thin curves represent the expected fractions when followup observations are conducted by using instruments with @xmath7 and @xmath69 , respectively . if the instrument has a resolution of @xmath1 , we estimate that lenses can be resolved for @xmath2 and @xmath3 of disk - bulge events and for @xmath4 and @xmath5 of bulge self - lensing events after @xmath6 and 20 years , respectively . the fraction increases substantially with the increase of the resolving power . if followup observations are performed by using an instrument with @xmath7 , we estimate that lenses can be resolved for @xmath8 and @xmath9 of disk - bulge events and for @xmath10 and @xmath11 of bulge self - lensing events after @xmath6 and 20 years , respectively . in the previous section , we estimated the fraction of galactic bulge events for which one can directly image lenses from followup observations by using high resolution imaging instruments and presented the result as a function of a time after original lensing measurements . in this section , based on the result in the previous section we discuss some of the observational aspects of the future followup lensing observations aimed for direct lens imaging . first , from the distributions in fig . 4 , we find that the proper choice of the instrument for the future high resolution followup lensing observations will be ngst . bulge events has been reported since 1993 ( udalski et al . 1994 ; alcock et al . 1995a ) . under a rough assumption that disk - bulge and bulge self - lensing events equally contribute to the total galactic bulge event rate and considering the life expectancy of hst , the fraction of events with resolvable lenses from hst observations will be just @xmath70 even if the followup observations are performed at the end stage of hst for the first generation of lensing events . however , by using ngst , which is scheduled to be launched in 2009 , it will be possible to resolve lenses for a significant fraction of events . second , given that all lensing events will be unable to be followed up due to the limited observation time of ngst , priority of targets should be given to events from which one can obtain extra information about the lens or the source star other than the identification of the lens as a normal star and the measurement of the lens - source proper motion . as mentioned in 1 , one can measure the angular radius of the source star of an event , for which the source was previously resolved , and uniquely determine the lens mass of an event , for which the lens parallax was measured . therefore , these events should be at the top of the target list . in the published literature , we find 10 events , for which the source star was well resolved ( alcock et al . 1997 , 2000a ; albrow et al . 1999 , 2000 , 2001a , b ; afonso et al . 2000 ; an et al . 2002 ) , and 11 events , for which parallax effect was measured ( alcock et al . 1995b ; mao 1999 ; soszyski et al . 2001 ; smith , mao , & woniak 2002 ) . another possible high priority targets will be the longest events with @xmath71 days . these events can not be well explained by the standard models of the geometrical and kinematical distributions of lenses and their mass function ( han & gould 1996 ) . although these events comprise a small fraction ( @xmath72 ) of the total number of events , they are important because their contribution to the total optical depth is important . detection of lenses and the measured proper motions ( or non - detection of lenses ) will allow one to better constrain the nature of these mysterious events . we have estimated the fraction of galactic bulge microlensing events for which the lenses can be directly imaged from future high resolution followup observations by computing the distribution of proper motions of the currently detected bulge events and imposing realistic criteria for lens resolution . from this computation , we find that lens identification will be possible for a significant fraction of bulge events from followup observations using ngst under the assumption that most bulge events are caused by normal stars . besides identifying lenses as stars , direct lens imaging will allow one to accurately determine the lens proper motion , from which the physical parameters of the individual lenses can be better constrained . if lenses are imaged for events where lens parallaxes were measured , the lens parameter degeneracy can be completely broken and the lens mass can be uniquely determined . in addition , high resolution followup observations will provide a valuable chance to measure the angular radii of remote bulge source stars involved with events for which the source was previously resolved via either a source transit or a caustic crossing . afonso , c. et al . , 2000 , , 532 , 340 alard c. , guibert j. , 1997 , , 326 , 1 albrow m. d. et al . , 1999 , , 522 , 1011 albrow m. d. et al . , 2000 , , 534 , 894 albrow m. d. et al . , 2001a , , 549 , 759 albrow m. d. et al . , 2001b , , 550 , l173 alcock c. et al . , 1993 , nature , 365 , 621 alcock c. et al . , 1995a , , 445 , 133 alcock c. et al . , 1995b , , 454 , l125 alcock c. et al . , 1997 , , 491 , 436 alcock c. et al . , 2000a , , 541 , 270 alcock c. et al . , 2000b , , 541 , 734 alcock c. et al . , 2001 , nature , 414 , 617 alonso a. , salaris m. , arribas s. , martinez - roger c. , asensio ramos a. , 2000 , , 355 , 1060 an j. a. et al . , 2002 , , 572 , 521 aubourg e. et al . , 1993 , nature , 365 , 623 bahcall j. n. , 1986 , , 24 , 577 binney j. , tremaine s. , 1987 , galactic dynamics ( princeton : princeton univeristy press ) , 67 blum r. d. , 1995 , , 444 , l89 bond i. et al . , 2001 , , 327 , 868 dwek e. et al . , 1995 , , 445 , 716 evans n. w. , belokurov v. , 2002 , , 567 , l119 evans n. w. , collett j. l. , 1993 , , 264 , 353 gould a. , 1994 , , 421 , l71 gould a. , 2001 , , 113 , 903 han c. , gould a. , 1995 , , 447 , 53 han c. , gould a. , 1996 , , 467 , 540 kroupa p. , tout c. a. , gilmore g. , 1993 , , 262 , 545 mao s. , paczyski b. , 1996 , , 473 , 57 mao s. 1999 , , 350 , l19 nemiroff r. j. , wickramasinghe w. a. d. t. , 1994 , , 424 , l21 paczyski b. , 1986 , , 304 , 1 smith m. c. , mao s. , woniak p. , 2002 , , 332 , 962 soszyski i. et al . , 2001 , , 552 , 731 udalski a. , szymanski m. , kaluzny j. , kubiak m. , krzeminski w. , mateo m. , preston g. w. , paczynski , b. 1993 , acta astron . , 43 , 289 udalski a. , szymanski m. , kaluzny j. , kubiak m. , 1994 , acta astron . , 44 , 1 witt h. j. , mao s. , 1994 , , 430 , 505	recently , from the _ hubble space telescope _ ( hst ) images of one of the large magellanic cloud ( lmc ) events taken 6.3 years after the original lensing measurement , alcock et al . were able to directly image the lens . although the first resolved lens was identified for an lmc event , much more numerous lenses are expected to be resolved for galactic bulge events . in this paper , we estimate the fraction of galactic bulge events whose lenses can be directly imaged under the assumption that all bulge events are caused by normal stars . for this determination , we compute the distribution of lens proper motions of the currently detected galactic bulge events based on standard models of the geometrical and kinematical distributions of lenses and their mass function . we then apply realistic criteria for lens resolution , and the result is presented as a function of the time elapsed after an original lensing measurement , @xmath0 . if followup observations are performed by using an instrument with a resolving power of @xmath1 , which corresponds to that of hst equipped with the new advanced camera for surveys , we estimate that lenses can be resolved for @xmath2 and @xmath3 of disk - bulge events and for @xmath4 and @xmath5 of bulge self - lensing events after @xmath6 and 20 years , respectively . the fraction increases substantially with the increase of the resolving power . if the instrument has a resolution of @xmath7 , which can be achieved by the _ next generation space telescope _ , we estimate that lenses can be resolved for @xmath8 and @xmath9 of disk - bulge events and for @xmath10 and @xmath11 of bulge self - lensing events after @xmath6 and 20 years , respectively . epsf # 1 [ firstpage ] gravitational lensing stars : fundamental parameters
in the recent years effective field theories constructed from quantum chromodynamics ( qcd ) have been enormously successful to describe phenomena where masses and momenta follow certain limits . among them is non - relativistic qcd ( nrqcd ) @xcite which is applicable to a system of two heavy quarks moving with small relative velocity . next to properties of the @xmath1 and @xmath2 families also the threshold production of top quark pairs is among the prominent examples ( see , e.g. , ref . @xcite for a review ) . the common method to construct an effective theory is based on the so - called matching procedure : appropriately chosen green s functions are computed in the full and effective theory and equality is required up to power - suppressed terms . in this way the couplings of the effective operators ( i.e. the matching coefficients ) are determined which completely specifies the effective theory . a crucial operator both in qcd and nrqcd is the vector current of a heavy quark - antiquark pair . the corresponding matching coefficient enters as a building block in a variety of physical observables , for example the bottom - quark mass from @xmath2 sum rules ( see , e.g. , refs . @xcite for recent analyses ) and top - quark threshold production at a future electron positron linear collider @xcite . the latter process allows for an extraction of the top - quark mass with an accuracy below 100 mev @xcite an improvement of about an order of magnitude as compared to the current results from the fermilab tevatron or the the cern large hadron collider @xcite . several quantities are needed in order to perform a third - order analysis of a heavy quark - antiquark system at threshold . ultrasoft effects have been considered in refs . @xcite , the three - loop static potential has been computed in refs . @xcite and in ref . @xcite a preliminary analysis of the top - quark threshold production cross section has been presented including also third - order potential effects . details on the potential contributions can be found in refs . @xcite . in this paper we compute the three - loop matching coefficient between the vector current in qcd and nrqcd . thus all ingredients are available to obtain the complete next - to - next - to - next - to leading order qcd prediction of the cross section @xmath3 close to threshold or the decay width of the @xmath4 meson to leptons . the results for the latter are presented in an accompanying paper @xcite where all building blocks are combined to a phenomenological analysis . the vector current in the full theory ( qcd ) is given by @xmath5 where @xmath6 denotes a generic heavy quark with mass @xmath7 . on the other hand in the effective theory ( nrqcd ) the current is represented by an expansion in @xmath8 where at each order effective operators have to be considered which are multiplied by coefficient functions . the leading contribution involves one operator given by @xmath9 where @xmath10 and @xmath11 are two - component pauli spinors for quark and anti - quark , respectively , and @xmath12 ( @xmath13 ) are the pauli matrices . hence , the matching coefficient of the vector current is defined through @xmath14 note that the 0-component of @xmath15 is only relevant for the power - suppressed contributions . the purpose of this paper is the evaluation of the purely gluonic three - loop corrections to @xmath16 . the fermionic contributions have already been considered in refs . @xcite . in order to compute @xmath16 it is convenient to consider on - shell vertex corrections involving the currents @xmath17 and @xmath18 . after taking into account the wave function renormalization one obtains ( see also ref . @xcite ) @xmath19 where the quantities with a tilde are defined in the effective theory and the ellipsis represents terms suppressed by the heavy - quark mass . @xmath20 is the renormalization constant of the current @xmath18 which is used to subtract the remaining poles after renormalization . these poles are due to the separation of long and short distance contributions in the effective theory . in order to evaluate physical quantities it is important that the same subtraction scheme is also adopted in the contributions originating from the effective theory @xcite . it is well - known that in the full theory the renormalization constant of the vector current is equal to one . in eq . ( [ eq::def_cv ] ) @xmath21 is the on - shell wave function renormalization constant which has been computed up to three - loop accuracy in refs . @xmath22 denotes the one - particle irreducible vertex diagrams with on - shell quarks carrying momenta @xmath23 and @xmath24 . it incorporates all one - particle irreducible vertex graphs and the corresponding counterterms for @xmath7 and @xmath25 . sample feynman diagrams are shown in fig . [ fig::sample ] . feynman diagrams contributing to @xmath22 . straight and curly lines denote heavy quarks with mass @xmath7 and gluons , respectively . ] the counterparts of @xmath22 and @xmath21 in the effective theory can be found on the right - hand side of eq . ( [ eq::def_cv ] ) . it is convenient to apply the threshold expansion @xcite to eq . ( [ eq::def_cv ] ) . this requires the identification of the hard , soft , potential and ultrasoft momentum regions in the integrals contributing to @xmath22 and @xmath26 . since nrqcd is obtained from qcd by integrating out the hard modes one has by construction that the soft , potential and ultrasoft modes agree in @xmath22 and @xmath26 and thus drop out from eq . ( [ eq::def_cv ] ) . as a consequence @xmath22 is evaluated for @xmath27 , which corresponds to the leading term of the hard integration region , and @xmath28 . furthermore , we have @xmath29 . there are several technical difficulties which one has to overcome in order to compute the vertex corrections . among them are the large number of diagrams which leads to several thousand feynman integrals to be evaluated in the first place , their reduction to a small set of about 100 basis integrals , so - called master integrals , and the evaluation of the latter in an expansion in @xmath30 , where @xmath31 is the space - time dimension . the last two tasks become more complicated by the additional condition @xmath32 on the external momentum . an automated setup for the calculation has been described in ref . @xcite and applied to the fermionic contributions . its core parts are a powerful implementation of laporta s algorithm in the program crusher @xcite , and fiesta @xcite which is based on sector decomposition and is used for the numerical integration of the master integrals . the main differences to the gluonic part considered in this paper are the larger number of diagrams and the increased complexity of the integrals which have to be reduced to master integrals . furthermore , the master integrals are more numerous and more involved . before discussing the matching coefficient it is instructive to consider the renormalization constant @xmath34 . the analytical results can be extracted from refs . @xmath35 \nonumber \\ & & \mbox { } + c_fc_a \left[\frac{1}{864\epsilon^2 } + \left ( \frac{113}{324 } + \frac{1}{4 } \ln2 + \frac{5}{32 } { l_\mu}\right ) \frac{1}{\epsilon } \right ] \nonumber \\ & & \mbox { } + c_a^2 \left[-\frac{1}{16\epsilon^2 } + \left ( \frac{2}{27 } + \frac{1}{4 } \ln2 + \frac{1}{24 } { l_\mu}\right ) \frac{1}{\epsilon } \right ] \nonumber \\ & & \mbox { } + t n_l \left [ c_f\left ( \frac{1}{54\epsilon^2 } -\frac{25}{324\epsilon } \right ) + c_a \left ( \frac{1}{36\epsilon^2 } - \frac{37}{432\epsilon } \right ) \right ] \nonumber \\ & & \mbox { } + c_f t n_h \frac{1}{60\epsilon } \bigg\ } + { \cal o}(\alpha_s^4)\ , , \label{eq::zv}\end{aligned}\ ] ] where @xmath36 , @xmath37 and @xmath38 for a @xmath39 gauge group and @xmath40 . note that the strong coupling is defined in the effective theory with @xmath41 active quarks where @xmath42 is the total number of quark flavors . in our case we have @xmath43 , however , we keep @xmath44 in the formulae for convenience . from our calculation we can extract the renormalization constant @xmath34 and compare with eq . ( [ eq::zv ] ) . the central values of our numerical coefficients agree at the per cent level with the analytical result of eq . ( [ eq::zv ] ) which constitutes a first non - trivial check and provides quite some confidence in the overall set - up of our calculation . we write the perturbative expansion of the matching coefficient in the form @xmath45 and decompose @xmath46 according to the color structures as @xmath47 \nonumber\\&&\mbox { } + \mbox{singlet terms } \label{eq::cv3ldef } \,.\end{aligned}\ ] ] note that all color factors of the non - singlet part can be expressed in terms of @xmath48 , @xmath49 , and @xmath50 . in this paper we do not consider the singlet contribution where the external current does not couple to the fermion line in the final state . at two - loop order such contributions have been computed for axial - vector , scalar and pseudo - scalar currents in ref . their numerical effect in those cases is below 3% as compared to the non - singlet contributions and thus quite small . whereas the one- @xcite and two - loop @xcite terms have been known since long the fermionic three - loop corrections became available only a few years ago @xcite . the so - called renormalon contribution , consisting of the one - loop diagram with arbitrary many massless quark loop insertions in the gluon propagator , has been computed in ref . supersymmetric one - loop corrections to @xmath16 have been computed in ref . @xcite . in the following we present the results for the individual coefficients in eq . ( [ eq::cv3ldef ] ) parameterized in terms of @xmath51 . the reconstruction of the full dependence on the renormalization scale is straightforward ; the corresponding expressions can be obtained from @xcite . our results read . this explains the difference in the logarithmic part of the coefficient @xmath52 . ] @xmath53 all uncertainties originating from the individual master integrals are added quadratically . in order to obtain a conservative error estimate we interpret the uncertainty of the numerical integration as one standard deviation from a gaussian distribution and multiply it by a factor of five which is accounted for in eq . ( [ eq::cv3 ] ) in eq . ( [ eq::cv3 ] ) . ] . the coefficients of @xmath54 could be obtained in analytic form since all renormalization constants are known analytically . in most applications it is sufficient to know the result for the matching coefficient with numerically evaluated color factors . setting @xmath55 , @xmath56 , @xmath38 and @xmath43 before inserting the master integrals and combining the numerical uncertainties we get @xmath57 \nonumber\\&&\mbox { } + \left(\frac{\alpha_s^{(n_l)}}{\pi}\right)^3\left [ -2091(2 ) + 120.66(0.01)\ , n_l \right.\nonumber\\&&\left.\mbox { } -0.823\ , n_l^2 \right ] + \mbox{singlet terms } \ , , \label{eq::cv3num}\end{aligned}\ ] ] where @xmath58 has been chosen . note that the @xmath41-independent three - loop term contains the contribution with a closed massive quark loop which amounts to @xmath59 @xcite . one observes that for @xmath60 and @xmath61 all coefficients in eq . ( [ eq::cv3num ] ) have the same sign and that they grow quite rapidly when going from nlo to nnnlo . at nnlo and nnnlo the fermionic corrections screen the non - fermionic ones , but even for @xmath62 only a reduction of at most 30% is obtained . a first glance at eq . ( [ eq::cv3num ] ) would suggest that for the quantity @xmath16 perturbation theory breaks down even though the momentum scale involved in the problem , @xmath7 , is quite large . however , as already mentioned above , @xmath16 itself does not represent a physical quantity . it has to be combined with contributions originating from soft , potential and ultrasoft momentum regions which can compensate the large coefficients in eq . ( [ eq::cv3num ] ) . further discussions on this topic can be found in ref . it might very well be that the @xmath0 scheme adopted in our calculation is not well suited for separating the divergences occuring in the different regimes . in fact , also the ultrasoft contribution studied in ref . @xcite shows large numerical effects . we have performed several checks on the correctness of our result which we want to mention in the following . in our calculation we allowed for a general gauge parameter @xmath63 which manifests as a polynomial dependence of the individual diagrams . after summing the three - loop results for @xmath21 and @xmath22 ( taking into account the corresponding quark mass counterterm contribution ) we concentrated on the coefficient of the linear @xmath63-dependence and have verified that it vanishes . as a further check we recomputed the @xmath41 contribution @xcite using our automated setup . in this context we want to mention that in ref . @xcite all occuring master integrals have been computed either analytically or using a numerical method different from the one used in the present paper . as already mentioned above , with our calculation we could also reproduce the renormalization constant in eq . ( [ eq::zv ] ) with high accuracy which checks all but the highest @xmath64 coefficients of the master integrals . we note in passing that we have a similar accuracy for the cancellation of the spurious poles up to seventh order occuring due to our reduction procedure . typical master integral appearing in our calculation . the solid lines and dashed lines represent massive and massless lines , respectively . for the external momenta we have the conditions @xmath65 and @xmath66 . , title="fig : " ] + @xmath23 @xmath24 + at this point it is instructive to show a result for a typical master integral contributing to @xmath16 . for the feynman diagram in fig . [ fig::mi ] , which we need up to order @xmath64 , we obtain with the help of fiesta @xcite @xmath67 a very powerful check on the correctness of our result is provided by the change of basis for the master integrals . we employ the integral tables generated during the reduction procedure in order to re - express the master integrals , which are not known analytically , through different , in general more complicated ones . this transformation is done analytically for general space - time dimension @xmath31 . in a next step the new master integrals are again evaluated with fiesta and inserted in the new expression for @xmath16 . in tab . [ tab::basis ] we compare the results for the purely gluonic coefficients and the complete result for @xmath68 obtained in the two bases . we observe an excellent agreement within the uncertainties . in the case of the `` alternative basis '' one has to keep in mind that the integrals to be evaluated numerically are significantly more complicated which explains the larger uncertainties for the coefficients in tab . [ tab::basis ] . .[tab::basis]comparison of the purely gluonic coefficients of eq . ( [ eq::cv3 ] ) and @xmath46 with @xmath69 and @xmath62 for two different choices of the master integral basis . for convenience @xmath58 has been adopted . the given uncertainties are obtained by combining the numerical uncertainties of each master integral contribution in quadrature . in contrast to eqs . ( [ eq::cv3 ] ) and ( [ eq::cv3num ] ) no factor five has been introduced for this comparison . [ cols="^,^,^",options="header " , ] as a further check on the numerical evaluation of the master integrals we have used a different momentum assignment in the input for fiesta . as a consequence different expressions are generated in intermediate steps leading to different numerical integrations . the final results are in complete agreement with eq . ( [ eq::cv3num ] ) . residue for the top quark system normalized to @xmath70 as a function of the renormalization scale @xmath71 . dotted , dash - dotted , short - dashed and solid lines correspond to lo , nlo , nnlo and nnnlo prediction . in the long - dashed curve only the fermion contributions to @xmath46 are taken into account ( nnnlo ( ferm . ) ) . ] we are now in the position to have a first look to the phenomenological consequences of our result for @xmath16 . we consider the residue of the two - point function of the vector currents @xmath72 which is obtained by considering @xmath73 close to the @xmath74 threshold . in this limit @xmath73 is dominated by pole contributions originating from bound - state effects @xmath75 where the ellipsis denotes contributions from the continuum . @xmath76 and @xmath77 are the residue and energy of the @xmath78 resonance which determine the height and position of the threshold cross section , respectively . in the following we consider the residue of the @xmath79 ( pseudo ) bound state of top quarks and extend the considerations of ref . @xcite should read @xmath80 . ] by including the non - fermionic contribution of @xmath46 and the @xmath81 term of the @xmath8 potential @xcite . choosing the potential subtracted scheme @xcite with @xmath82 gev to define the top quark mass we obtain @xmath83 gev which leads to @xmath84 \nonumber\\ & = & \frac{(c_f m_t^{\rm ps } \alpha_s)^3}{8\pi } \left[1 - 2.13\alpha_s + 23.66\alpha_s^2 \right . \nonumber \\&&\mbox { } \left . - 113.0(0.1)\alpha_s^3 + { \cal o}(\alpha_s^4 ) \right ] \,,\end{aligned}\ ] ] where @xmath85 , @xmath86 , and @xmath87 . we have used @xmath88 to compute @xmath89 where the soft scale @xmath90 gev has been adopted after the second equality sign . in order to get an impression about the importance of the individual contributions we mark the @xmath71-independent coefficients from the three - loop static potential ( @xmath91 ) , from the two - loop @xmath81 term of the @xmath92 potential ( @xmath93 ) , and from the three - loop fermion ( @xmath94 ) and purely gluonic ( @xmath95 ) contribution to @xmath46 separately . for this choice of @xmath71 one observes quite big nnnlo contributions which are dominated by @xmath95 . thus , it is instructive to investigate the @xmath71-dependence of @xmath96 which is shown in fig . [ fig::zmu ] . around the soft scale no convergence is observed . allowing , however , for higher scales one finds a quite flat behavior of the nnnlo curve . furthermore , the nnnlo corrections become quite small . e.g. , considering the top quark system for @xmath97 gev , the nlo terms amount to about @xmath98% and the nnlo to roughly @xmath99% . the third - order contribution is practically zero . similar observations also hold for the bottom quark case , see ref . @xcite . the third - order contribution to the matching coefficient of the vector current between qcd and nrqcd has been computed . an automated setup has been developed where even the occuring master integrals are identified automatically , processed with the help of the computer program fiesta , and prepared for the insertion into the analytic reduction of @xmath46 . in the @xmath0 scheme the numerical impact of @xmath46 is quite big as can be seen from eq . ( [ eq::cv3num ] ) which constitutes the main result of this paper . in a dedicated analysis one has to investigate the consequences for the bottom - quark mass extracted from @xmath2 sum rules and the top - quark threshold production cross section at a future linear collider . we would like to thank martin beneke and alexander penin for many useful discussions and communications and alexander smirnov for continuous support on fiesta . this work is supported by dfg through sfb / tr 9 `` computational particle physics '' . the feynman diagrams were drawn with the help of axodraw @xcite and jaxodraw @xcite . w. e. caswell and g. p. lepage , phys . b * 167 * ( 1986 ) 437 . g. t. bodwin , e. braaten and g. p. lepage , phys . d * 51 * ( 1995 ) 1125 [ erratum - ibid . d * 55 * ( 1997 ) 5853 ] [ arxiv : hep - ph/9407339 ] . n. brambilla _ et al . _ , arxiv : hep - ph/0412158 . a. pineda and a. signer , phys . d * 73 * ( 2006 ) 111501 [ arxiv : hep - ph/0601185 ] . a. hoang , p. ruiz - femenia and m. stahlhofen , jhep * 1210 * ( 2012 ) 188 [ arxiv:1209.0450 [ hep - ph ] ] . a. h. hoang _ et al . _ , eur . j. directc * 2 * ( 2000 ) 1 [ arxiv : hep - ph/0001286 ] . m. martinez and r. miquel , eur . phys . j. c * 27 * ( 2003 ) 49 [ arxiv : hep - ph/0207315 ] . k. seidel , f. simon , m. tesar and s. poss , eur . j. c * 73 * ( 2013 ) 2530 [ arxiv:1303.3758 [ hep - ex ] ] . t. horiguchi , a. ishikawa , t. suehara , k. fujii , y. sumino , y. kiyo and h. yamamoto , arxiv:1310.0563 [ hep - ex ] . j. beringer _ [ particle data group collaboration ] , phys . d * 86 * ( 2012 ) 010001 . m. beneke , y. kiyo and a. a. penin , phys . b * 653 * ( 2007 ) 53 [ arxiv:0706.2733 [ hep - ph ] ] . m. beneke and y. kiyo , phys . b * 668 * ( 2008 ) 143 [ arxiv:0804.4004 [ hep - ph ] ] . a. v. smirnov , v. a. smirnov and m. steinhauser , phys . b * 668 * ( 2008 ) 293 [ arxiv:0809.1927 [ hep - ph ] ] . a. v. smirnov , v. a. smirnov and m. steinhauser , phys . * 104 * ( 2010 ) 112002 [ arxiv:0911.4742 [ hep - ph ] ] . c. anzai , y. kiyo and y. sumino , phys . rev . * 104 * ( 2010 ) 112003 [ arxiv:0911.4335 [ hep - ph ] ] . m. beneke , y. kiyo , a. penin and k. schuller , econf c * 0705302 * ( 2007 ) top01 [ arxiv:0710.4236 [ hep - ph ] ] . m. beneke , y. kiyo and k. schuller , pos radcor * 2007 * ( 2007 ) 051 [ arxiv:0801.3464 [ hep - ph ] ] . m. beneke , y. kiyo and k. schuller , arxiv:1312.4791 [ hep - ph ] . m. beneke , y. kiyo and k. schuller , to be published . p. marquard , j. h. piclum , d. seidel and m. steinhauser , nucl . b * 758 * ( 2006 ) 144 [ arxiv : hep - ph/0607168 ] . p. marquard , j. h. piclum , d. seidel and m. steinhauser , phys . b * 678 * ( 2009 ) 269 [ arxiv:0904.0920 [ hep - ph ] ] . d. j. broadhurst , n. gray and k. schilcher , z. phys . c * 52 * ( 1991 ) 111 . k. melnikov and t. van ritbergen , nucl . b * 591 * ( 2000 ) 515 [ arxiv : hep - ph/0005131 ] . p. marquard , l. mihaila , j. h. piclum and m. steinhauser , nucl . b * 773 * ( 2007 ) 1 [ arxiv : hep - ph/0702185 ] . m. beneke and v. a. smirnov , nucl . b * 522 * ( 1998 ) 321 [ arxiv : hep - ph/9711391 ] . v. a. smirnov , _ analytic tools for feynman integrals _ , springer tracts mod . phys . * 250 * ( 2012 ) 1 . p. marquard and d. seidel , unpublished . a. v. smirnov and m. n. tentyukov , comput . commun . * 180 * ( 2009 ) 735 arxiv:0807.4129 [ hep - ph ] . a. v. smirnov , v. a. smirnov and m. tentyukov , comput . commun . * 182 * ( 2011 ) 790 [ arxiv:0912.0158 [ hep - ph ] ] . a. v. smirnov , arxiv:1312.3186 [ hep - ph ] . m. beneke , a. signer and v. a. smirnov , phys . lett . * 80 * ( 1998 ) 2535 [ arxiv : hep - ph/9712302 ] . b. a. kniehl , a. a. penin , m. steinhauser and v. a. smirnov , phys . * 90 * ( 2003 ) 212001 [ erratum - ibid . * 91 * ( 2003 ) 139903(e ) ] [ arxiv : hep - ph/0210161 ] . b. a. kniehl , a. onishchenko , j. h. piclum and m. steinhauser , phys . b * 638 * ( 2006 ) 209 [ arxiv : hep - ph/0604072 ] . g. kllen and a. sarby , k. dan . vidensk . mat .- fis . 29 , n17 ( 1955 ) 1 . a. czarnecki and k. melnikov , phys . * 80 * ( 1998 ) 2531 [ arxiv : hep - ph/9712222 ] . e. braaten and y. -q . chen , phys . d * 57 * ( 1998 ) 4236 [ erratum - ibid . d * 59 * ( 1999 ) 079901 ] [ hep - ph/9710357 ] . y. kiyo , m. steinhauser and n. zerf , phys . d * 80 * ( 2009 ) 075005 [ arxiv:0907.1146 [ hep - ph ] ] . https://www.ttp.kit.edu / progdata / ttp14/ttp14 - 002/. m. beneke , phys . lett . b * 434 * ( 1998 ) 115 [ hep - ph/9804241 ] . j. a. m. vermaseren , comput . commun . * 83 * ( 1994 ) 45 . d. binosi and l. theussl , comput . commun . * 161 * ( 2004 ) 76 [ arxiv : hep - ph/0309015 ] .	we evaluate the three - loop corrections to the matching coefficient of the vector current between quantum chromodynamics ( qcd ) and non - relativistic qcd . the result is presented in the @xmath0 scheme where large perturbative corrections are observed . the implications on the threshold production of top quark pairs are briefly discussed .
the high precision of current collider data requires comparable precision in the phenomenological predictions . the state of the art in high - energy calculations is at next - to - next - to - leading order ( nnlo ) in quantum chromodynamics ( qcd ) . consequently , major efforts have been undertaken to produce nnlo parton distribution functions ( pdfs ) from a global analysis of the available data . these include the ct14nnlo pdfs @xcite as well as others @xcite , all of which include lhc data in the determination of the pdfs . in this paper we describe the introduction of qed evolution at leading order ( lo ) with the next - to - leading order ( nlo ) qcd evolution in the same cteq global analysis package that was used to produce the ct14 pdfs @xcite . past studies of qed effects in global analysis have been done by the mrst @xcite and the nnpdf @xcite groups . we have checked our code against other qed@xmath5qcd evolution codes @xcite and find good agreement . the mrst and nnpdf analyses used different approaches for modeling the photon pdf . the mrst group used a parametrization for the photon pdf based on radiation off of `` primordial '' up and down quarks , with the photon radiation cut off at low scales by constituent or current quark masses @xcite . the nnpdf group used a more general photon parametrization , which was then constrained by high - energy @xmath6 , @xmath7 and drell - yan data at the lhc @xcite . they found constraints on the size of the photon pdf , which was still consistent with zero at the initial scale of @xmath8 gev . as discussed by martin and ryskin @xcite , the photon pdf has a large _ elastic _ contribution in which the proton remains intact , in addition to the _ inelastic _ contribution in which the proton breaks into a multihadron final state . neither mrst nor nnpdf addresses these separate contributions to the photon pdf , although we can assume that the nnpdf photon is _ inclusive _ , containing both inelastic and elastic components , since it was constrained using inclusive drell - yan and vector boson data . given the limited amount of data to constrain the shape of the photon pdf , we will use a generalization of the mrst approach . we parametrize the inelastic contribution to the photon at the initial scale gev is the same as that used in the standard ct14 pdf sets , and was chosen to be just below the input charm pole mass of @xmath9 gev . ] @xmath1 gev by @xmath10 where @xmath11 is the convolution of the quark - to - photon splitting function @xmath12 with the `` primordial '' quark distribution @xmath13 , which we take to be the initial ct14 nlo up and down valence distributions . we then set @xmath14 to obtain a single parameter family of photon distributions , which we can label by their initial inelastic momentum fraction @xmath15 . for comparison , in analogy with the mrst approach , we will also show results for a `` current mass '' ( cm ) photon distribution , given by defining @xmath16 , and setting the @xmath17 to the quark current masses ; _ i.e. , _ @xmath18 mev and @xmath19 mev . we will constrain the inelastic photon pdf using data on deep inelastic scattering ( dis ) with isolated photons from the zeus collaboration @xcite . the advantage of using this process is that the initial - state photon contributions are at leading order in the perturbation expansion . in contrast , the initial - state photon contribution to drell - yan or @xmath6 and @xmath7 production is suppressed by factors of @xmath20 relative to the leading quark - antiquark production . however , to use the dis - plus - photon data we will first need to address some technical issues relating to the combination of different subprocess contributions to the observed final state . the organization of this paper is as follows : in sec . [ sec : qedevol ] we describe the inclusion of qed evolution in the cteq global analysis code and give more details about our initial pdf parametrizations . in sec . [ sec : constraints ] we discuss constraints on the photon pdf coming from the ct14 global analysis data set and from the zeus dis with isolated photon data . in this section we present a new calculation for the dis - plus - isolated - photon process , which consistently combines the photon - initiated contribution with the quark - initiated contribution . we show that this data gives significant constraints on the initial photon pdf . in sec . [ sec : conclude ] we discuss our findings and give conclusions . we also include an appendix where we show comparisons between our qcd@xmath5qed evolution code and other publicly - available codes . in this section we discuss the implementation of the qed evolution and the initial photon pdf in the context of the cteq - tea global analysis program . the evolution of the pdfs , @xmath21 , including qed contributions at leading order ( lo ) and qcd contributions at higher orders , is described by the equations : @xmath22 where @xmath23 , the indices @xmath24 and @xmath25 run over active quark flavors , and the convolution is defined by @xmath26 the qcd splitting functions , given by @xmath27 are known up to @xmath28 , next - to - next - to - leading order ( nnlo ) @xcite . the lo qed splitting functions can be extracted from the lo qcd splitting functions , giving @xmath29\nonumber\\ \tilde{p}^{(0)}_{\gamma q}(z)&=&\frac{1+(1-z)^2}{z}\\ \tilde{p}^{(0)}_{\gamma \gamma}(z)&=&-\frac{2}{3}n_c\sum_i e_i^2\,\delta(1-z)\ , \end{aligned}\ ] ] where @xmath30 is the number of colors . we have modified the fortran nlo evolution code evolve , which was used for previous cteq - tea pdfs ( cteq6 - 6.6 @xcite-@xcite and ct09 @xcite ) , to include the lo qed contributions . this code solves the evolution equations directly in @xmath31-space , so that the only new technical issue introduced by the qed corrections is the separation of the quark singlet distributions into separate up and down contributions , based on the quark charges . we have checked our evolution code against the public qcd@xmath5qed codes , partonevolution @xcite and apfel @xcite , and we find good agreement . details of this comparison are given in the appendix . the initial photon pdf at the scale @xmath4 is a nonperturbative input that must be obtained by a fit to data . even a choice of zero initial photon pdf is ambiguous , since it depends on the arbitrary scale @xmath4 . so far there have been two different approaches to the initial photon pdf . in the mrst analysis @xcite the initial photon pdf was given by an ansatz , obtained from radiation off primordial valence up and down quark distributions , cut off at low scales given by the current quark masses , @xmath32 mev and @xmath33 mev , or by constituent quark masses @xmath34 mev . alternatively , the nnpdf approach @xcite was to use a general parametrization for the initial photon pdf to be constrained by high - energy @xmath6 , @xmath7 , and drell - yan production at the lhc . in this work we will use a generalization of the mrst ansatz , but we must first address a subtlety of the photon pdf . unlike the case for colored partons , the photon pdf has a large elastic component , in which the proton remains intact @xcite . this is in addition to the inelastic component , in which the proton dissociates . the elastic component can be parametrized by the equivalent photon approximation ( epa ) @xcite , which involves an integration over the proton electromagnetic form factors . for this work we focus on the inelastic component , which we parametrize by a radiative ansatz , but with free parameters to be fit by data . given the weak constraints from data on the photon pdf , we find it useful to limit the number of parameters to one or two for the time being . we shall see that the zeus dis - plus - isolated - photon data @xcite constrains the inelastic photon pdf roughly in the range @xmath35 for @xmath36 gev@xmath37 . we parametrize the inelastic contribution to the initial photon pdfs in the proton and neutron by @xmath38 where @xmath39 and @xmath40 are `` primordial '' valence - type distributions in the proton , and the initial photon pdf in the neutron is obtained by an approximate isospin symmetry . defining @xmath16 , we can trade the parameters @xmath41 for mass scales @xmath17 , and we see that the nonperturbative inputs @xmath42 are modeled by the radiation of a single photon off the `` primordial '' quarks , with a collinear cutoff given by the scales @xmath17 . the mrst initial photon pdfs can be obtained from this parametrization by setting @xmath43 gev , using the functions for @xmath44 and @xmath45 given in ref . @xcite , and letting @xmath46 and @xmath47 be either the constituent or current quark masses . for our analysis , we use @xmath1 gev as in ct14 , and we set @xmath48 the initial up and down ct14 nlo valence distributions in the proton . the presence of the photon pdf violates isospin between the neutron and proton . continuing with the radiative ansatz and working to first order in @xmath49 , we can neglect the isospin violation in the gluon and sea - quark pdfs @xcite and use @xmath50 for the valence quarks at first order in @xmath49 , the radiative ansatz plus approximate isospin symmetry implies @xmath51 this suggests a consistent set of pdfs for the valence quarks in the neutron given by @xmath52 note that eqs . ( [ eq : photonpdf ] ) and ( [ eq : valencen ] ) together ensure that if the number and momentum sum rules ( including the photon contribution ) are satisfied for the pdfs in the proton , they are automatically satisfied for the pdfs in the neutron , regardless of the choices for @xmath39 and @xmath40 . again , for our analysis , we choose @xmath44 and @xmath45 to equal the initial up and down ct14 nlo valence distributions in the proton . thus , from eqs . ( [ eq : photonpdf])-([eq : sea ] ) and ( [ eq : valencen ] ) we can obtain the quark , gluon , and photon pdfs in both the proton and neutron in terms of the parametrization of the quark and gluon pdfs in the proton , plus the two additional parameters @xmath53 and @xmath54 . for this paper , we shall make the further simplification of @xmath14 , which corresponds to cutting off the radiation from both valence quarks at the same scale . with this choice , everything is specified by one additional parameter , which can be taken to be the cutoff scale @xmath55 , defined by @xmath56 . alternatively , we can trade this parameter for the initial inelastic photon momentum fraction in the proton : @xmath57 for the remainder of this paper , unless otherwise specified , the photon pdfs will be in this one - parameter radiation ansatz labeled by @xmath0 . for comparison purposes , we will make one exception to this by defining a `` current mass '' ( cm ) photon pdf , analogous to the mrst current mass pdf , and given by @xmath16 with @xmath18 mev and @xmath19 mev . for this choice the initial inelastic photon momentum fraction is determined to be @xmath58% . for all other partons in our analysis we use the ct14 nlo initial distributions , except that to maintain a total momentum fraction of 1 , we re - normalize the initial up , down , and strange sea - quark distributions , to account for the additional photon momentum fraction . given that the relevant photon momentum fractions are very small , we find that this reduces the sea - quark distributions by typically less than 1% , and it is inconsequential in our analysis . ( the sea - quark distributions are reduced by 0.9% for @xmath59% , and they are reduced by 1.6% for the cm photon pdf . ) in fig . [ fig : xfq ] we plot the quantity @xmath60 for three representative photon pdfs , relative to the quark and gluon pdfs , at the scales @xmath61 gev and @xmath62 gev . we note that the effect of the initial photon pdf and the qed evolution on the quark and gluon pdfs is imperceptible in these plots . for the photon pdf , we plot for @xmath63% ( @xmath64 gev ) and for @xmath59% ( @xmath65 mev ) , and for the cm photon pdf . for @xmath61 gev ( left ) and @xmath62 gev ( right ) . three representative photon pdfs are plotted : the `` current mass '' photon pdf ( @xmath66 , red dotted ) , and photon pdfs with initial inelastic photon momenta fractions of @xmath63 and 0.14% ( @xmath67 , blue dashed , and @xmath68 , green dot - dashed , respectively ) . the effects of the different initial photon pdfs on the quark and gluon pdfs are imperceptible in these plots . [ fig : xfq],title="fig:",scaledwidth=47.0% ] for @xmath61 gev ( left ) and @xmath62 gev ( right ) . three representative photon pdfs are plotted : the `` current mass '' photon pdf ( @xmath66 , red dotted ) , and photon pdfs with initial inelastic photon momenta fractions of @xmath63 and 0.14% ( @xmath67 , blue dashed , and @xmath68 , green dot - dashed , respectively ) . the effects of the different initial photon pdfs on the quark and gluon pdfs are imperceptible in these plots . [ fig : xfq],title="fig:",scaledwidth=47.0% ] and @xmath69 for @xmath70 gev ( black dots ) , 3.2 gev ( red dashes ) , and 85 gev ( blue solid ) . the left plot is for zero initial photon momentum and the right plot is for the cm photon pdf . [ fig : isospin],title="fig:",scaledwidth=47.0% ] and @xmath69 for @xmath70 gev ( black dots ) , 3.2 gev ( red dashes ) , and 85 gev ( blue solid ) . the left plot is for zero initial photon momentum and the right plot is for the cm photon pdf . [ fig : isospin],title="fig:",scaledwidth=47.0% ] the constraints on the photon pdf from the dis and tevatron data , used in the ct14 analysis , are relatively weak . these come from two main sources : isospin violation effects in nuclear scattering and constraints from the momentum sum rule . in general , isospin violation will arise through qed evolution , as well as from the initial conditions given by eq . ( [ eq : valencen ] ) . this isospin violation can be seen in fig . [ fig : isospin ] , where we plot @xmath71 and @xmath69 for several values of @xmath72 for the case where the initial photon pdf is zero , and for the case where the initial photon is the cm choice . note that the isospin violation is small and most important at large @xmath31 . given that cuts of @xmath73 gev@xmath37 , applied to enforce perturbativity in the calculations , typically require @xmath74 , we expect constraints from isospin violation to be small in the present data , as observed in the mrst analysis @xcite . constraints from the momentum sum rule arise because any momentum carried by the photon implies less momentum available for the quark and gluon pdfs . in this way , constraints from data on the colored parton pdfs indirectly impact the photon pdf . we have performed a preliminary analysis using the data sets included for ct10 @xcite . for a fixed initial photon momentum fraction , with the photon pdf parametrized as discussed in sec . ii , we minimized the global @xmath75 by varying the quark and gluon pdfs . using the usual cteq - tea choice of @xmath76 tolerance , we obtain a limit on the photon momentum fraction of @xmath77% at the 90% confidence level , which is similar in magnitude to the results found by the mrst and nnpdf analyses . the best fit for the initial photon momentum fraction from this global analysis is @xmath78% , but with only a small change of @xmath79 , relative to the fit with @xmath63% . for comparison , we find the elastic contribution to the initial photon momentum fraction , as calculated in the equivalent photon approximation , to be @xmath80% . unfortunately , this limit on the initial photon momentum fraction is much larger than one would expect for a photon pdf . in the analysis of the nnpdf group , additional constraints were made on the initial photon pdf by including lhc data on high - energy @xmath6 , @xmath7 , and drell - yan production , and comparing with theoretical predictions that included photon initial - state contributions . although the photon - induced contribution to these processes is small compared to the dominant quark - antiquark annihilation subprocess , the precision of these measurements was enough to substantially increase the constraints on the photon pdf @xcite . however , the small relative contribution of the photon - photon subprocess puts a stringent requirement on the precision needed for both experimental and theoretical analyses . any small misjudgment of systematic errors on the experimental side , or uncalculated higher - order corrections on the theoretical side could have a significant effect on the extraction of the photon pdf . in particular , given that the initial photon pdf is nominally of order @xmath49 , one might expect that the uncalculated @xmath81 quark - initiated contributions to drell - yan production would contribute at the same level as the photon - initiated contributions . for this reason , we consider a different experimental process , isolated photon production in dis , to constrain the photon pdf . at the partonic level , the process of dis with isolated photon occurs at lo through compton scattering of a photon coming from the proton off the lepton , as shown in fig . [ fig : feyndiags](a ) . thus , this process probes the photon pdf at lo , having no large backgrounds with which to compete . however , the quark - initiated subprocess shown in figs . [ fig : feyndiags](b ) and ( c ) , while formally suppressed by @xmath82 , is just as large because of the small size of @xmath83 relative to @xmath84 . in fact , if we consider the photon pdf to be @xmath82 , then the photon - initiated subprocess and the quark - initiated subprocess are actually the same order in @xmath49 . thus , the correct way to calculate the cross section for dis with isolated photons is to include both subprocesses consistently without double - counting . in the literature there have been two approaches to calculations of the process @xmath2 . the calculation of mrst @xcite , which was preceded by studies of blmlein _ et al . _ @xcite , included just the photon - initiated contribution of fig . [ fig : feyndiags](a ) . the calculation of gehrmann - de ridder , gehrmann , and poulsen ( ggp ) @xcite included just the quark - initiated contributions of figs . [ fig : feyndiags](b , c ) . in the ggp analysis it was found convenient to make the lorentz - invariant separation of the cross section into three components , depending on the fermion line off which the final - state photon is emitted : @xmath85 for emission off the lepton line , given by the square of the partonic amplitude in fig . [ fig : feyndiags](b ) ; @xmath86 for emission off the quark line , given by the square of the partonic amplitude in fig . [ fig : feyndiags](c ) ; and @xmath87 for the interference between the two sets of diagrams . actually represents two feynman diagrams , where the final - state photon is emitted off the initial - state lepton or quark as well as off the final - state lepton or quark . ] in the ggp calculation a cut on the outgoing quark was necessary to remove the divergence in the amplitude as the photon off - shellness went to zero in the @xmath85 amplitude . a hybrid calculation was also considered by the zeus collaboration in their analysis of the dis - plus - isolated photon data @xcite , where the @xmath85 component of the quark - initiated subprocess of ggp was replaced by the photon - initiated subprocess of mrst . . for each diagram shown there is an additional diagram where the photon is emitted off the initial - state lepton or quark . [ fig : feyndiags],scaledwidth=80.0% ] in this section we introduce a consistent and systematic method of combining the photon- and quark - initiated subprocesses , which also reduces the factorization scale dependence of either calculation . first , consider the calculation of the differential cross section as a power series in @xmath49 without consideration of the relative sizes of @xmath83 and @xmath84 . it can be written as a convolution over partonic cross sections @xmath88 where each of the partonic cross sections can be expanded in a power series in @xmath49 , @xmath89 the diagram in fig . [ fig : feyndiags](a ) corresponds to a lo contribution ( @xmath90 ) , while the diagrams in figs . [ fig : feyndiags](b , c ) correspond to nlo contributions ( @xmath91 ) . through nlo in @xmath49 the cross section can still be written as a sum of @xmath85 , @xmath86 , and @xmath87 components , @xmath92 where the @xmath85 component also includes the photon - initiated contribution , @xmath93\ , \ ] ] using the modified minimal subtraction ( @xmath94 ) scheme , we can factorize the initial - state singularity from the nlo quark - initiated subprocess into the definition of the photon pdf , leaving a nlo subprocess cross section , @xmath95 where the first term on the right - hand side , @xmath96 , is the hard partonic quark - induced subprocess , and the second term is the collinear subtraction counterterm . here , we have distinguished the renormalization scale @xmath97 from the factorization scale @xmath72 , and we explicitly note that the initial - state collinear singularity cancels within the @xmath85 component . calculating everything in dimensional regularization with @xmath98 dimensions , the combined @xmath85 component of the nlo quark - initiated subprocess cross section , @xmath99 is finite as @xmath100 . in principle , there are additional virtual and real contributions at nlo in @xmath49 , besides the quark - initiated contributions . however , all other nlo terms are proportional to @xmath83 , which is in fact suppressed by an amount of order @xmath49 relative to @xmath84 , as seen in the previous section . thus , by keeping the photon - initiated contribution at lo and the quark - initiated contributions at nlo , and including the collinear - subtraction counterterm of eq . ( [ eq : ap ] ) we have a well - defined calculation in the @xmath94 scheme , while maintaining all contributions of the same size in @xmath49 . note that this is reminiscent of the acot scheme @xcite for including charm and bottom quark pdfs , although for the photon , there is no equivalent of the fixed - flavor scheme , due to its zero mass and the consequently nonperturbative photon pdf . the calculations of the @xmath86 and @xmath87 components are identical to those in the ggp calculation . for the kinematics of interest to us , the @xmath87 component is negligible , but it is included for completeness . the @xmath86 contribution has a final - state singularity , when the photon and final - state quark become collinear , which can be handled by including a fragmentation contribution and an associated collinear subtraction counterterm in the @xmath94 scheme . thus , we obtain @xmath101 where @xmath102 is the hard partonic subprocess , @xmath103 is the lo subprocess cross section for @xmath104 , and @xmath105 is the fragmentation function for finding the photon ( with momentum @xmath106 ) in the quark @xmath107 ( with momentum @xmath107 ) , with momentum fraction @xmath108 at the fragmentation scale @xmath109 . the singularities in the first two terms on the right - hand side of this equation cancel as @xmath100 . we will discuss the choice for the fragmentation function in the next subsection . we have calculated the @xmath85 and the @xmath86 components of the cross section , using the subtraction method to handle the collinear divergences @xcite in the hard cross - section term . following this method , we subtract a term with a two - particle final state mapped onto a 3-particle final - state , with the third particle phase space unintegrated , and then add back the exact same term with the third particle phase space integrated out . the subtraction term is designed to have the same collinear - singular limit as the hard term in the same region of phase space of the third particle , so that the hard cross section term minus the subtraction is integrable in @xmath110 dimensions , while the @xmath111 singularities in the remaining terms cancel . using this method we obtain for the @xmath85 quark - initiated contribution : @xmath112 where @xmath113\ , \,d\hat\sigma_\gamma^{(0)}(z\xi ) \ .\label{eq : ap3}\ ] ] the hard term minus the subtraction term can be written as an integral over the phase space of the additional quark : @xmath114 where @xmath115 with @xmath116 and @xmath117 the scattering angles of the final - state quark in the initial parton - parton center - of - momentum frame . note that the hard term on the right - hand side is treated with three - body final - state phase space , while the subtraction term is treated with two - body final - state phase space . similarly , we obtain for the @xmath86 contribution : @xmath118 where @xmath119\bigg\}\,d\hat\sigma_{eq\rightarrow e^\prime q^\prime}\ .\nonumber\end{aligned}\ ] ] the hard term minus the subtraction term can be written @xmath120 where in this case we have found it convenient to use a different parametrization of the final - state quark phase space . letting @xmath121 be the hard partonic subprocess , we use @xmath122 , where @xmath123 is the angle between @xmath124 and @xmath125 in the @xmath126 center - of - momentum frame and @xmath127 is the azimuthal angle between the @xmath128 plane and the @xmath129 plane . as before , the hard term on the right - hand side is treated with three - body final - state phase space , while in this case the subtraction term and the fragmentation term are treated in the limit where the final - state photon and quark are collinear , with momenta satisfying @xmath130 and @xmath131 . the zeus experiment @xcite used proton and lepton beam energies of @xmath132 gev and @xmath133 gev , respectively , corresponding to a center - of - mass energy and rapidity of @xmath134 respectively . ( we neglect the proton mass , @xmath135 , in all calculations here . ) for the process @xmath2 , with momentum satisfying , @xmath136 , one can define the standard dis variables that describe the kinematics of the scattered lepton , @xmath137 , @xmath138 , and @xmath139 . the zeus collaboration measured distributions for two leptonic variables @xmath140 and @xmath31 , and for two photonic variables , @xmath141 and @xmath142 , the transverse energy and pseudorapidity of the photon , respectively . the collaboration combined data that were 59.1% @xmath143 and 40.9% @xmath144 scattering . note that the sign of the charged lepton has no effect on the @xmath85 or @xmath86 components of the calculation , but the combination of the two charged lepton contributions produces a significant cancellation of the already small @xmath87 component , so that it is negligible in the analysis . the kinematic region defined by the experiment was @xmath145 the cut on the final - state lepton angle , @xmath146 , can be written in terms of its rapidity as @xmath147 there are two additional cuts that require discussion . the experimentalists reported a cut of @xmath148 gev , where @xmath149 . naively , this cut looks problematic because it would remove the photon - initiated contribution , which occurs at exactly @xmath150 . however , upon closer investigation it appears that this cut was only applied to the theoretical and monte carlo calculations . to quote from ref . @xcite , `` the keen reader will note that no such cut was applied at the detector level . this proved impossible due to the poor @xmath151 resolution at detector level and poor description of the data by mc ... '' . the relevant detector - level cut was the requirement of at least one reconstructed track , well separated from the lepton , which was used to ensure some hadronic activity and to remove deeply virtual compton scattering events . after this cut , it was found that the number of events in the monte carlo calculations with @xmath152 gev was negligible . for our purposes , we interpret the forward track cut as a requirement to tag inelastic events , and we include no explicit @xmath151 cut . note that the forward track cut should equally remove elastic isolated photon events , and so remove the contribution from the elastic part of the photon pdf . in this way , the zeus data probe only the inelastic part of the photon pdf , and therefore , we only include this inelastic contribution in comparison with the experimental data . in doing so we have made the approximation @xmath153 ; _ i.e. , _ the elastic and inelastic components of the photon pdf evolve separately . this approximation is good because @xmath154 changes very little from @xmath4 to @xmath140 due to the rapid falloff of the proton electromagnetic form factor , while the inelastic contribution evolves additively , @xmath155 up to corrections suppressed by extra factors of @xmath49 . we have verified by explicit calculation that this additive approximation replicates the consistently evolved inclusive photon pdf , with errors that are far smaller than other theoretical uncertainties that we will discuss below . the second additional important cut is the isolation cut on the photon , enforcing that 90% of the energy in the jet containing the photon belongs to the photon , where jets are formed with the @xmath156 cluster algorithm with parameter @xmath157 . we will model this isolation cut in our calculation in two different ways . first , we can model the experimental cut at the parton level , requiring @xmath158 if the photon - quark separation satisfies @xmath159 . for later reference , we call this the `` sharp '' isolation cut . since this does not completely remove the quark - photon collinear singularity , the theoretical calculation of the @xmath86 component will depend on the choice of the quark - to - photon fragmentation function @xmath160 . for this , we use the lo fragmentation function determined by the aleph collaboration @xcite , parametrized by @xmath161\ , \nonumber\end{aligned}\ ] ] where @xmath162 gev and @xmath163 . note that this parametrization of @xmath160 is an exact solution to the evolution equation at @xmath82 , so that the dependence on the fragmentation scale @xmath109 cancels exactly in our calculation . in the ggp analysis @xcite , other parametrizations of @xmath160 with different assumptions were compared , with only a small effect on the calculated cross sections . one of the disadvantages of having a dependence on the quark - to - photon fragmentation function , in addition to the uncertainties due to the phenomenological fit to the function , is that it assumes that the cross section is inclusive in the fragmentation remnant . in our calculation , the combination of the experimental constraints on the photon and on the lepton indirectly imposes constraints on the remnant quark in the process . therefore , we also consider an alternative model of the experimental isolation cut , replacing it with a `` smooth '' isolation cut @xcite , so as to avoid the necessity of the fragmentation contribution . the smooth isolation cut is given by requiring that the hadronic energy @xmath164 inside all cones of radius @xmath165 around the photon direction satisfy @xmath166 where we take @xmath167 and @xmath168 . these values of @xmath169 and @xmath170 are chosen to ensure that the photon will contain at least 90% of the energy inside a cone of @xmath171 centered on the photon , just as for the experimental isolation cut . however , the smooth cut does not translate exactly to the experimental isolation cut , because it requires the photon to carry a greater fraction of the energy as the cone size @xmath172 becomes smaller . in practice , because of this , the theoretical calculation with the smooth cut is better behaved than with the sharp cut . in addition , the smooth isolation cut is more restrictive than the sharp isolation ; for a strictly positive - definite differential cross section , the smooth isolation prescription must always give a smaller predicted cross section . in this way , a comparison of the two calculations can give some indication of the theoretical uncertainty due to the isolation cut . finally , we note that the jet - clustering algorithm also includes the electron , so that the isolation cut effectively imposes @xmath173 . and correspond to the total prediction ( light gray solid ) , the @xmath86 component ( blue hashed ) , the @xmath85 component ( dark red solid ) , and the photon - initiated contribution only ( green hashed ) . also shown are the zeus data points with combined statistical and systematic errors . [ fig : scaledep],title="fig:",scaledwidth=47.0% ] and correspond to the total prediction ( light gray solid ) , the @xmath86 component ( blue hashed ) , the @xmath85 component ( dark red solid ) , and the photon - initiated contribution only ( green hashed ) . also shown are the zeus data points with combined statistical and systematic errors . [ fig : scaledep],title="fig:",scaledwidth=47.0% ] before discussing the comparison of the theory with the data , it is useful to understand the theoretical uncertainties of the calculation by studying the factorization scale dependence and the dependence on the isolation prescription . in fig . [ fig : scaledep ] we plot the differential cross sections for @xmath174 and @xmath175 as functions of @xmath141 and @xmath142 , respectively , while varying the factorization scale within @xmath176 . here we have used the sharp isolation cut and calculated with zero initial inelastic photon pdf at @xmath1 gev . the four bands on each of the two plots correspond to the photon - initiated contribution only ( green hashed band ) , @xmath85 component ( dark red solid band ) , @xmath86 component ( blue hashed band ) , and the total calculation ( light gray solid band ) . the @xmath87 contribution is imperceptible on the scale of these plots . from these plots we learn several important facts . first , the scale dependence of the @xmath85 component is reduced dramatically compared to the photon - initiated contribution alone . this large scale dependence of the photon - initiated contribution cancels greatly with that of the collinear subtraction counterterm in the combined @xmath85 component . second , the @xmath85 and @xmath86 components dominate in different regions of phase space . for instance , the cross sections are most sensitive to the @xmath85 component , and consequently to the photon pdf , at large @xmath141 and small @xmath142 . thus , the shapes of these distributions can give information about the nonperturbative contribution to the photon pdf . finally , we note that the scale dependence of the @xmath86 component is still large , being only lo in @xmath177 , and it dominates the overall scale uncertainty of the theoretical calculation . in fig . [ fig : isodep ] we plot the total predictions of the same two distributions , again for zero initial inelastic photon pdf , but now comparing the two different isolation prescriptions in the theoretical calculation . in these plots we show the predictions with the smooth isolation prescription ( blue hashed band ) and the sharp isolation prescription ( red hashed band ) , again varying the factorization scale within @xmath176 . the first thing to note here is that the difference between the predictions is about the same size as the scale uncertainty , with similar dependence on the kinematic variables . another striking feature is that the prediction using the smooth isolation prescription is uniformly larger than that using the sharp isolation , in contrast to expectations . this is probably due to incomplete cancellations in the sharp isolation calculation between the large negative collinear fragmentation contribution and the positive real contribution , due to indirect constraints on the emitted final - state quark in the real emission contribution . presumably higher - order qcd corrections will affect the predictions for both isolation predictions , to resolve this issue . as noted previously , our calculation is only lo in @xmath177 ; we expect both the factorization scale uncertainty and the isolation prescription discrepancy to be reduced at nlo in @xmath177 . in any event , we will use the two isolation prescriptions , as well as the factorization scale dependence , as a measure of the theoretical uncertainty of our calculation . . the blue hashed band is calculated using the smooth isolation prescription and the red hashed band is calculated using the sharp isolation prescription . also shown are the zeus data points with combined statistical and systematic errors . [ fig : isodep],title="fig:",scaledwidth=47.0% ] . the blue hashed band is calculated using the smooth isolation prescription and the red hashed band is calculated using the sharp isolation prescription . also shown are the zeus data points with combined statistical and systematic errors . [ fig : isodep],title="fig:",scaledwidth=47.0% ] with this understanding of the theoretical uncertainties of the calculation we can now compare the zeus data against predictions for the differential distributions , while varying the initial inelastic photon momentum fraction @xmath0 of the photon pdf , described in sec . [ sec : photonpdf ] . for this analysis , the initial quark and gluon pdfs are just the ct14nlo pdfs , except that the sea - quark normalizations are rescaled in order to maintain a total momentum fraction of 1 . this rescaling has little effect in our analysis , because the photon momentum fractions considered here are small . for instance , a photon momentum fraction of @xmath178 induces a reduction of the sea - quark momentum by only 0.9% , while the cm photon pdf induces a reduction of the sea - quark momentum by 1.6% . at this stage of the analysis we have not refit the quark and gluon pdfs , since the @xmath179 process is dominantly sensitive to the photon pdf directly , whereas the indirect sensitivity through changes in the quark and gluon pdfs is negligible . and @xmath142 , with the smooth isolation prescription , with factorization scale @xmath180 . the curves , from bottom to top are with initial inelastic photon momentum fractions of @xmath181 ( black dashed ) , 0.1% ( blue dotted ) , 0.2% ( red dot - dashed ) , and for the cm photon ( green solid ) . also shown are the zeus data points with combined statistical and systematic errors . [ fig : eteta],title="fig:",scaledwidth=47.0% ] and @xmath142 , with the smooth isolation prescription , with factorization scale @xmath180 . the curves , from bottom to top are with initial inelastic photon momentum fractions of @xmath181 ( black dashed ) , 0.1% ( blue dotted ) , 0.2% ( red dot - dashed ) , and for the cm photon ( green solid ) . also shown are the zeus data points with combined statistical and systematic errors . [ fig : eteta],title="fig:",scaledwidth=47.0% ] in fig . [ fig : eteta ] we plot the differential cross sections for @xmath174 and @xmath175 as a function of the photon variables , @xmath141 and @xmath142 , using the smooth isolation prescription , with a factorization scale of @xmath180 . the curves , from bottom to top are with initial inelastic photon momentum fractions of @xmath181 ( black dashed ) , 0.1% ( blue dotted ) , 0.2% ( red dot - dashed ) , and for the cm initial photon ( green solid ) , which has initial momentum fraction 0.26% . also shown are the zeus data points with combined statistical and systematic errors . with these choices of @xmath72 and the isolation prescription , we see that the theory can fit the data well for @xmath182% . on the other hand , the theory fits poorly for the cm initial photon , overshooting the data at large @xmath141 and small @xmath142 . of course , the best fit for @xmath183 is correlated with the choice of @xmath72 and the isolation prescription . however , since these choices tend to move the curves up or down uniformly , it is still possible to constrain the initial photon pdf by the shape of the distributions . in particular , it is impossible to get a good fit to the prediction using the cm initial photon pdf , regardless of the choices of @xmath72 and the isolation prescription . in fig . [ fig : q2x ] we plot the differential cross sections for @xmath184 and @xmath185 as a function of the lepton variables , @xmath186 and @xmath31 against the zeus data , using the exact same theoretical choices and initial inelastic photon pdfs as for the previous plot . in this case we see that it is impossible to fit the data , regardless of the initial photon pdf or the choices of factorization scale and isolation prescription . in particular the theory fits the data very poorly at small @xmath31 and @xmath186 . in fact , we note that prediction for the smallest bin in @xmath31 is far from the zeus data point , and is essentially independent of the initial photon pdf . we expect that the predictions in these bins are highly sensitive to higher - order qcd radiation , so that it is difficult to fit the full lepton distributions with a fixed - order calculation . and @xmath31 , with the smooth isolation prescription , with factorization scale @xmath187 . the curves , from bottom to top are with initial inelastic photon momentum fractions of @xmath181 ( black dashed ) , 0.1% ( blue dotted ) , 0.2% ( red dot - dashed ) , and for the cm photon ( green solid ) . also shown are the zeus data points with combined statistical and systematic errors . [ fig : q2x],title="fig:",scaledwidth=47.0% ] and @xmath31 , with the smooth isolation prescription , with factorization scale @xmath187 . the curves , from bottom to top are with initial inelastic photon momentum fractions of @xmath181 ( black dashed ) , 0.1% ( blue dotted ) , 0.2% ( red dot - dashed ) , and for the cm photon ( green solid ) . also shown are the zeus data points with combined statistical and systematic errors . [ fig : q2x],title="fig:",scaledwidth=47.0% ] the fact that these fixed - order calculations are more reliable for the photon distributions than for the lepton distributions can be seen further by looking at the phase - space constraints for the two sets of variables . in fig . [ fig : limits ] we show plots of the constraints on the photon variables @xmath141 and @xmath142 and on the lepton variables @xmath186 and @xmath31 . in these figures , the dashed lines indicate the bins that are plotted by the zeus data . the combined dark red@xmath5light blue regions indicate the regions of phase space allowed by the zeus kinematic constraints of eq . ( [ eq : zeuscuts ] ) , for the fully inclusive event , whereas the dark red region only is allowed for the lo photon - initiated subprocess . for the photon distributions , the constraints are dominated by the photon cuts on @xmath141 and @xmath142 , with only a small cut on the photon - initiated contribution in the upper - left corner due to the requirement of @xmath188 . thus , all of the bins in @xmath141 and @xmath142 have a large photon - initiated contribution and can be considered very inclusive . in contrast , for the lepton distributions the additional photon constraints have a large effect in many of the bins . for instance , the photon - initiated contribution to the smallest @xmath186 bin is largely removed by the requirement of @xmath189 gev , and the photon - initiated contribution to the smallest @xmath31 bin is completely removed by the requirement of @xmath190 . these bins are dominated by events with additional particles in order to satisfy the kinematics , so we would not expect our fixed - order calculation to do well at predicting the lepton distributions . and @xmath142 ( left ) and @xmath186 and @xmath31 ( right ) . the dashed lines indicate the bins that are plotted by the zeus experiment . the dark red region only is kinematically accessible by the photon - initiated contribution , while the dark red@xmath5light blue regions are kinematically allowed in general . [ fig : limits],title="fig:",scaledwidth=46.0% ] and @xmath142 ( left ) and @xmath186 and @xmath31 ( right ) . the dashed lines indicate the bins that are plotted by the zeus experiment . the dark red region only is kinematically accessible by the photon - initiated contribution , while the dark red@xmath5light blue regions are kinematically allowed in general . [ fig : limits],title="fig:",scaledwidth=47.2% ] based on these arguments we will use only the distributions in the photon variables @xmath141 and @xmath142 to constrain the initial inelastic photon pdf , for a total of eight data points . we also reiterate that the constraints due to the remaining ct14 experimental data set are much weaker than these data , and are neglected in the present analysis . we define the chi - squared function for these data points by @xmath191 where @xmath192 , @xmath193 , and @xmath194 are the theory prediction , the experimental measurement , and its combined statistical and systematic error for the data point @xmath24 . in fig . [ fig : chi2 ] we plot @xmath75 versus the initial inelastic photon momentum fraction @xmath0 for both the smooth and sharp isolation prescriptions and for several values of the factorization scale @xmath72 . note that the value of @xmath0 determined by the minimum of @xmath75 depends significantly on the isolation prescription and on the factorization scale , giving best fits for the initial momentum fraction that can vary from less than 0 to above 0.1% . in addition , due to this theoretical uncertainty in the current calculation , it is not possible to unambiguously determine an error band on @xmath0 , using the standard ct approach of applying some tolerance criterion on the rise in the @xmath75 around the best fit . however , from fig . [ fig : chi2 ] we do see that not all choices of theoretical parameters are able to fit the shape of the data points equally well . therefore , we can determine a conservative limit on the value of @xmath0 by requiring that the data and theory not disagree beyond some level . a @xmath75 distribution with eight data points will have @xmath195 at the 90% confidence level . for the remaining 2947 data points used in the ct14 analysis is @xmath196 in going from @xmath63% to @xmath59% . ] therefore , we define that any theoretical prediction with @xmath197 is ruled out as a bad fit to the data at the 90% confidence level . it is impossible to satisfy this criterion for @xmath198% for either choice of the isolation prescription and for any value of @xmath72 . furthermore we find that the cm choice of the photon pdf has @xmath199 for any choice of isolation and factorization scale and so is ruled out by this data . versus initial inelastic photon momentum fraction @xmath0 using the smooth isolation prescription ( left ) and the sharp isolation prescription ( right ) for factorization scales @xmath200 , @xmath141 , @xmath201 , and @xmath202 . the horizontal line at @xmath203 is the 90% confidence level limit for eight data points . [ fig : chi2],title="fig:",scaledwidth=47.2% ] versus initial inelastic photon momentum fraction @xmath0 using the smooth isolation prescription ( left ) and the sharp isolation prescription ( right ) for factorization scales @xmath200 , @xmath141 , @xmath201 , and @xmath202 . the horizontal line at @xmath203 is the 90% confidence level limit for eight data points . [ fig : chi2],title="fig:",scaledwidth=47.2% ] thus , we find our maximal initial inelastic photon pdf to have @xmath59% at the 90% confidence level . of course , the exact value of the momentum fraction is correlated with the shape of the initial photon pdf . from fig . [ fig : limits ] we see that the zeus dis - plus - isolated - photon data constrains the photon pdf in the kinematic region given roughly by @xmath35 for @xmath36 gev@xmath37 . outside of this region the photon pdf is very weakly constrained , but we believe that the radiative ansatz gives a reasonable expectation for its overall shape . as for the minimal possible value of the initial inelastic photon momentum fraction , it could , in principle , be negative , which is not ruled out by the analysis of this section . for instance , one could begin the evolution with zero initial photon pdf at a lower value of the scale @xmath4 . however , we take the reasonable assumption that it should be nonzero at the low scale of @xmath1 gev . we thus propose the initial pdfs with @xmath63% and @xmath59% as our 90% c.l . photon pdfs . a similar analysis gives @xmath204% at the 68% confidence level , but the data are still consistent with @xmath63% at the 68% c.l . in fig . [ fig : comp1 ] we compare , at the scale @xmath205 gev , the ct14qed photon pdfs with @xmath63% and @xmath59% against the nlo mrst2004qed photon pdfs , using the current quark masses ( labeled mrst0 ) and using the constituent quark masses ( labeled mrst1 ) , and against the nlo nnpdf2.3qed average photon pdf with @xmath206 . we should emphasize that the ct14qed photon pdfs only contain the inelastic contribution in these plots . the nnpdf2.3 average photon pdf has more structure in its shape at large and small values of @xmath31 than do the other pdfs , but it is still consistent with the zeus data in the @xmath31 range that is probed by the experiment . gev : ct14qed with @xmath181 ( solid green ) , ct14qed with @xmath178 ( short - dashed black ) , mrst2004qed0 using current quark masses ( dotted orange ) , mrst2004qed1 using constituent quark masses ( dot - dashed brown ) , and nnpdf2.3qed with @xmath206 and average photon ( long - dashed blue ) . [ fig : comp1],scaledwidth=46.0% ] in fig . [ fig : comp2 ] we compare the same set of photon pdfs at the higher scale of 85 gev and the very high scale of 1 tev . here we can make some very interesting observations . the most obvious thing in these figures is that the ct and mrst photon pdfs become very similar at large @xmath186 , whereas the nnpdf photon pdf is distinctly different and much smaller at small values of @xmath31 . this difference is due to the different approaches to the evolution of the pdfs taken by the different groups . whereas in the mrst and cteq - tea approaches , the qcd and qed scales are chosen to be identical and evolved together , in the nnpdf2.3qed pdfs that are included in the lhapdf library @xcite , the qcd and qed scales are separate and the two scales are evolved successively ; first the qed scale is evolved from @xmath4 to @xmath140 and then the qcd scale is evolved from @xmath4 to @xmath140 . as discussed in ref . @xcite , the successive evolution of qed and qcd differs from the combined evolution by terms that are subleading by @xmath207 and can induce large unresummed logarithms between the two scales . this difference in the evolution at small @xmath31 is also seen to be consistent with the behavior seen in the right panel of fig . 2 in ref . @xcite , where the nnpdf photon pdf also is smaller at small @xmath31 and large @xmath186 than when it is evolved using the code partonevolution @xcite . we expect that the difference between the nnpdf2.3qed photon pdf and the other photon pdfs at high @xmath140 would be less significant if the nnpdf2.3qed pdfs were evolved from the low scale simultaneously in qed and qcd . another observation from fig . [ fig : comp2 ] , concerning the ct14qed and mrst2004qed photon pdfs is that the impact of the initial photon distribution becomes less significant as @xmath186 increases and more photons are produced through radiation off the quarks . from these plots we see that the fractional deviation between the different photon pdfs decreases with increasing @xmath186 . in fact at very small @xmath31 and large @xmath186 the differences in the sea - quark distributions of the pdfs presumably have more impact on the photon pdf than does the initial photon distribution . gev ( left ) and @xmath208 tev ( right ) : ct14qed with @xmath181 ( solid green ) , ct14qed with @xmath178 ( short - dashed black ) , mrst2004qed0 using current quark masses ( dotted orange ) , mrst2004qed1 using constituent quark masses ( dot - dashed brown ) , and nnpdf2.3qed with @xmath206 and average photon ( long - dashed blue ) . [ fig : comp2],title="fig:",scaledwidth=46.0% ] gev ( left ) and @xmath208 tev ( right ) : ct14qed with @xmath181 ( solid green ) , ct14qed with @xmath178 ( short - dashed black ) , mrst2004qed0 using current quark masses ( dotted orange ) , mrst2004qed1 using constituent quark masses ( dot - dashed brown ) , and nnpdf2.3qed with @xmath206 and average photon ( long - dashed blue ) . [ fig : comp2],title="fig:",scaledwidth=46.0% ] in this paper , we have presented ct14qed , which is the first set of ct14 parton distribution functions obtained by including qed evolution at leading order ( lo ) with next - to - leading - order ( nlo ) qcd evolution in the global analysis by the cteq - tea group . this development will provide better theory predictions to compare with the precision data , such as drell - yan pair production , measured at the lhc . the ct14qed pdfs are based on the ct14 nlo initial distributions with the addition of an initial photon pdf . ( there is also an inconsequential rescaling of the quark sea pdfs , in order to maintain the momentum sum rule . ) the inelastic contribution to the photon pdf is parametrized at the initial scale @xmath4 using a generalization of the radiative ansatz introduced by the mrst group in their previous study . the initial photon pdf then depends on two independent parameters [ cf . ( [ eq : photonpdfintro ] ) ] , which are related to the scales at which the radiation off the up and down valence quarks is cut off . however , given the weak constraints on the initial photon pdf we find it convenient at this time to set the scales equal , so that the initial photon pdf is parametrized by a single parameter , which we take to be the momentum fraction carried by the inelastic photon at the initial scale @xmath4 . for comparison purposes we have also defined a `` current mass '' ( cm ) photon pdf , comparable to the mrst current mass pdf , for which the initial photon momentum fraction is @xmath209 . a set of neutron pdfs can also be obtained with a small amount of isospin breaking , suggested by the radiative ansatz applied to first order in @xmath49 , and which automatically ensures that the number and momentum sum rules are satisfied . however , as previously seen by both the mrst and nnpdf groups , we find that the constraints from isospin violation effects ( generally small and most important at large @xmath31 ) in nuclear scattering and from the requirement of the momentum sum rule , imposed by the dis and tevatron data in the ct14 , are relatively weak . thus , in order to constrain the photon pdf , we focused on the scattering process @xmath210 , which was measured by the zeus experiment at hera . this process is dominantly sensitive to the inelastic photon pdf directly , with negligible indirect sensitivity through the modification of the quark and gluon pdfs by qed effects . it also has the advantage that the initial - state photon subprocess contribution occurs at leading order , so that it does not compete with other much larger contributions . in this paper we have produced for the first time a consistent and systematic calculation for this cross section that combines both the photon- and quark - initiated subprocesses , and simultaneously reduces the factorization scale dependence of either calculation . details of this calculation were presented in sec . [ sec : theory ] . the photon isolation cut , which required that the final - state photon must contain at least 90% of the energy in the jet to which it belongs ( where jets are formed with the @xmath156 cluster algorithm with parameter @xmath157 ) , was modeled using two different models of photon isolation . we used the two different isolation prescriptions , as well as the factorization scale dependence as a measure of the theoretical uncertainty of our calculation . by comparing the zeus data for the distributions of transverse energy and pseudorapidity of the final - state photon against our calculation of the differential distributions , we were able to constrain the initial inelastic photon momentum fraction inside the proton to be @xmath211 at the 90% confidence level hence , the cm choice of photon pdf has been ruled out by this data . for completeness , we also compared the ct14qed pdfs to some of the nlo ( in @xmath177 ) photon pdfs published by the mrst and the nnpdf groups . phenomenological applications of the ct14qed pdfs will be discussed in future publications . as shown in fig . [ fig : isodep ] , the theoretical uncertainties due to the factorization scale dependence and the isolation prescription are currently larger than the experimental uncertainties of the zeus data . thus , extending our calculation to nlo in @xmath177 should be able to further constrain the initial photon pdf . this is a project that we are currently undertaking . parametrizations for the ( inelastic photon ) ct14qed pdf sets ( both proton and neutron versions ) will be distributed in a standalone form via the cteq - tea web site @xcite , or as a part of the lhapdf6 library @xcite . for backward compatibility with version 5.9.x of lhapdf , our web site also provides ct14 grids in the lhapdf5 format , as well as an update for the cteq - tea module of the lhapdf5 library , which must be included during compilation to support calls of all eigenvector sets included with ct14 @xcite . we will also distribute sets with the inclusive photon pdfs , ct14qedinc . for the proton , ct14qedinc at the initial scale @xmath4 is the sum of the ( inelastic ) ct14qed and the elastic component of the photon pdf , given by the equivalent photon approximation . the proton ct14qedinc pdfs are then evolved from @xmath4 to @xmath140 as discussed in sec . [ sec : qcdplusqed ] . for the neutron , ct14qedinc is equal to ct14qed , since the neutron has zero electric charge , and therefore it has no elastic component of the photon pdf . this work was supported by the national science foundation under grant no . we also thank sayipjamal dulat , tie - jiun hou , joey huston , pavel nadolsky and jianwei qiu for helpful discussions . gev to @xmath212 gev . [ fig : apfelcomp],scaledwidth=60.0% ] we have checked our code against the public evolution code partonevolution @xcite , which also solves the evolution equations at lo in qed and nlo in qcd . the main difference ( other than technicalities of implementation ) in the partonevolution code is that it also includes the charged leptons as partons in the proton . using the toy model of ref . @xcite with zero initial photon pdf , ensuring all input parameters agree , using the same formulation for the running of @xmath177 and @xmath49 , and removing the lepton pdfs and their contribution to @xmath213 from partonevolution , we find excellent agreement between the two programs . evolving from @xmath214 gev to @xmath212 gev , we find differences of less than @xmath215% over most of the range of @xmath31 for all of the pdfs , including the photon . reinstating the lepton pdf contribution to the evolution equations in partonevolution , which in principle should be included for consistency , we find their effects on the quark and gluon pdfs to be negligible , with changes of less than @xmath216% . the only noticeable effect is the reduction of the photon pdf by about @xmath217% with mild @xmath31 dependence , due to splitting of the photons into lepton - antilepton pairs . this is presumably much less than other uncertainties in our analysis , so it is reasonable to leave out the lepton pdf contribution in our code . the percent difference in our prediction for the photon pdf , relative to the partonevolution prediction with and without the inclusion of the lepton pdfs , is shown in fig . [ fig : pecomp ] . we have also checked our code against the program apfel @xcite , which includes qed at lo and qcd at up to nnlo . the main difference in the apfel program is that the qcd and qed factorization scales can be taken to be independent , and the evolution with respect to each scale can be done successively . however , using the setting `` qavdp '' in apfel averages the two possible orderings for performing the evolutions , which should agree with our approach to @xmath81 . in addition , the apfel code has been recently updated to allow the simultaneous evolution of the qed and qcd scales , using the `` qunid '' setting . we have compared our code with apfel 2.6.0 , starting with the ct10nlo pdfs @xcite with zero initial photon pdf at @xmath214 gev and evolving in qed at lo and qcd at nlo to @xmath212 gev . we have done the comparison using both the `` qavdp '' and the `` qunid '' settings for apfel . we obtain excellent agreement for the quark and gluon pdfs , with differences of less than @xmath215% over most of the range of @xmath31 for both apfel settings . in fig . [ fig : apfelcomp ] we show the results for the photon pdf . we obtain pretty good agreement with apfel with the `` qavdp '' setting , with differences of less than @xmath218% except at large @xmath219 . this is consistent with the @xmath81 differences expected in the different evolution procedures . we obtain excellent agreement with apfel with the `` qunid '' setting , with differences of less than @xmath220% over the full range of @xmath221 shown . we note that in fig . [ fig : apfelcomp ] we replace the evolution subroutine for @xmath49 in the apfel program with the code used in the ct global analysis code ; however , using the original @xmath49 subroutine in apfel still gives differences of less than 1% for the evolved photon pdf over the full range of @xmath31 when using the `` qunid '' setting . this is certainly much smaller than the uncertainties in the initial photon pdf itself . we have not checked our code directly against the mrst evolution code or the recently developed qcdnum+qed evolution code @xcite , but we do note that the comparison between qcdnum+qed and apfel `` qavdp '' for the evolution of the photon pdf in ref . @xcite looks qualitatively similar to the results that we have found in fig . [ fig : apfelcomp ] . in addition benchmarking studies between apfel and these two evolution codes in ref . @xcite show agreement at a similar level to that which we have found with our code here . s. dulat _ et al . _ , phys . d * 93 * , no . 3 , 033006 ( 2016 ) doi:10.1103/physrevd.93.033006 [ arxiv:1506.07443 [ hep - ph ] ] . r. d. ball _ et al . _ [ nnpdf collaboration ] , jhep * 1504 * , 040 ( 2015 ) [ arxiv:1410.8849 [ hep - ph ] ] . l. a. harland - lang , a. d. martin , p. motylinski and r. s. thorne , eur . j. c * 75 * , no . 5 , 204 ( 2015 ) [ arxiv:1412.3989 [ hep - ph ] ] . s. alekhin , j. blumlein and s. moch , phys . rev . d * 89 * , no . 5 , 054028 ( 2014 ) [ arxiv:1310.3059 [ hep - ph ] ] . a. d. martin , r. g. roberts , w. j. stirling and r. s. thorne , eur . j. c * 39 * , 155 ( 2005 ) [ hep - ph/0411040 ] . a. d. martin and m. g. ryskin , eur . j. c * 74 * , 3040 ( 2014 ) doi:10.1140/epjc / s10052 - 014 - 3040-y [ arxiv:1406.2118 [ hep - ph ] ] . s. chekanov _ et al . _ [ zeus collaboration ] , phys . b * 687 * , 16 ( 2010 ) [ arxiv:0909.4223 [ hep - ex ] ] . s. moch , j. a. m. vermaseren and a. vogt , nucl . b * 688 * , 101 ( 2004 ) [ hep - ph/0403192 ] . a. vogt , s. moch and j. a. m. vermaseren , nucl . b * 691 * , 129 ( 2004 ) [ hep - ph/0404111 ] . j. pumplin , d. r. stump , j. huston , h. l. lai , p. m. nadolsky and w. k. tung , jhep * 0207 * , 012 ( 2002 ) . [ hep - ph/0201195 ] . v. m. budnev , i. f. ginzburg , g. v. meledin and v. g. serbo , phys . rept . * 15 * , 181 ( 1975 ) . doi:10.1016/0370 - 1573(75)90009 - 5 j. gao _ et al . _ , d * 89 * , no . 3 , 033009 ( 2014 ) doi:10.1103/physrevd.89.033009 [ arxiv:1302.6246 [ hep - ph ] ] . j. blumlein , z. phys . c * 47 * ( 1990 ) 89 . j. blumlein , g. levman and h. spiesberger , j. phys . g * 19 * ( 1993 ) 1695 . a. arbuzov , d. y. bardin , j. blumlein , l. kalinovskaya and t. riemann , comput . commun . * 94 * ( 1996 ) 128 [ hep - ph/9511434 ] . a. gehrmann - de ridder , t. gehrmann and e. poulsen , phys . lett . * 96 * , 132002 ( 2006 ) [ hep - ph/0601073 ] .	we describe the implementation of quantum electrodynamic ( qed ) evolution at leading order ( lo ) along with quantum chromodynamic ( qcd ) evolution at next - to - leading order ( nlo ) in the cteq - tea global analysis package . the inelastic contribution to the photon parton distribution function ( pdf ) is described by a two - parameter ansatz , coming from radiation off the valence quarks , and based on the ct14 nlo pdfs . setting the two parameters to be equal allows us to completely specify the inelastic photon pdf in terms of the inelastic momentum fraction carried by the photon , @xmath0 , at the initial scale @xmath1 gev . we obtain constraints on the photon pdf by comparing with zeus data @xcite on the production of isolated photons in deep inelastic scattering , @xmath2 . for this comparison we present a new perturbative calculation of the process that consistently combines the photon - initiated contribution with the quark - initiated contribution . comparison with the data allows us to put a constraint at the 90% confidence level of @xmath3 for the inelastic photon pdf at the initial scale of @xmath1 gev in the one - parameter radiative ansatz . the resulting inelastic ct14qed pdfs will be made available to the public . in addition , we also provide ct14qedinc pdfs , in which the _ inclusive _ photon pdf at the scale @xmath4 is defined by the sum of the inelastic photon pdf and the elastic photon distribution obtained from the equivalent photon approximation .
neutrino interactions have been studied in the resonance region in terms of form factors and have been compared with available data @xcite with considerable success . after the early experiments with larger errors , there is a new generation of experiments that provide more accurate data . in addition , the oscillation phenomena and the finite mixing angle @xmath3 @xcite require more accurate theoretical predictions in order to decipher properties of oscillations including cp asymmetries . all these require on the theoretical side estimates of amplitudes that are reliable . one property of production cross sections is the fact that for @xmath4 and @xmath5 the axial contribution is given by the partially conserved axialvector current ( pcac ) . we worked out this framework for resonance production @xcite and for coherent scattering on nuclei it has been reported by two groups @xcite . in this article we wish to present our results for many reactions , calculate explicitly the energy spectra of the produced pions and point out special properties . for example , in the production of the delta resonance in the energy region @xmath6 to @xmath7{gev}$ ] the vector squared and the interference contributions are almost equal @xcite . they add up for neutrinos and cancel each other for antineutrinos leaving the axial current squared as the dominant contribution in the latter reaction . this will be tested in the experiments . it will also be useful for measuring the flux by using this and additional channels in the nearby detector and to predict the yield in the far away detector in order to establish deviations from normal oscillations , like the presence of a cp violating phase . the above property for antineutrino reactions is especially useful because the flux of the antineutrino beam will be smaller . for antineutrino interactions we find the cross section on a proton target to be smaller than on a neutron . early qualitative evidence for this ordering is already available for integrated cross sections @xcite . in this article we adopt the neutrino reaction @xmath8 as benchmark for comparing other reactions of charged and neutral currents . for @xmath9{gev}$ ] the delta resonance dominates and it has been shown that there is only a small background of 10% for other reactions . we shall assume that the @xmath10 amplitude dominates and thus obtain the antineutrino and neutral current reactions . later on , we may revisit the topic in order to study the changes brought about when we introduce a small nonresonant background . in section [ sec : energyspectrumandangulardependenceofpions ] we calculate the energy and angular spectrum of the pions for the small region of @xmath11 . for this we introduce the @xmath12 cross section in the rest frame of the resonance and then transform it to the laboratory frame of the neutrino interaction . details of the calculation are given in section [ sec : energyspectrumandangulardependenceofpions ] and in the appendix . one worrisome aspect concerns the modifications brought about by the subsequent rescattering of pions in nuclei . here we make the observation that charge symmetry predicts similar corrections for @xmath13 and @xmath14 interactions on isoscalar nuclei . we shall use this property . for nuclear corrections we adopt a model for rescatterings which is both simple and transparent @xcite . in fact , since some of the neutrino - nucleon reactions are predicted , we can use them to test the accuracy of nuclear corrections . for isoscalar targets many structure functions are related by charge symmetry at both steps of the reaction : the initial neutrino - nucleon scattering and the subsequent pion nucleus interactions . consequently , several relations follow to be tested experimentally . finally , the results for @xmath15{gev^2}$ ] are useful for matching them to the form factors at higher values of @xmath11 . the contents of the article are arranged as follows . in section [ sec : chargedandneutralcurrentreactions ] we define the general method and present charged and neutral current cross sections . we point out several regularities that are inherent in the cross sections . results for pion spectra in energy @xmath16 from the axial current are presented in section [ sec : energyspectrumandangulardependenceofpions ] . for these calculations we introduce the energy and angular dependence of the reaction @xmath17 in the rest frame of the resonance and transform it to the laboratory frame . the details of the lorentz transformation are given in the appendix . in section [ sec : nuclearcorrections ] we compute nuclear corrections for isoscalar targets and pay special attention to find charge symmetric reactions . the results will be useful for long baseline experiments in which we point out that in our kinematic region products of cross sections @xmath18 nuclear corrections will be very similar for the regenerated @xmath19 and @xmath20 beams . since the mass @xmath21 of the electron is very small we repeated the calculations with @xmath21 and the results are shown in figure [ fig : electron ] . the article closes with a general discussion . the method can be extended to higher resonances in order to fill the transition region between resonances and deep inelastic scattering . in fact , another group @xcite uses a dynamic hadron model for this purpose . we consider the reactions listed in table [ tab : isoreactions ] . we take reaction ( [ eq : vccp ] ) as a standard where for @xmath22 there is only the @xmath23 amplitude . the amplitude for @xmath24 is given by the clebsch - gordan coefficient ( cgc ) shown in the fourth column of table [ tab : isoreactions ] . the cgc is the same for vector and axial - vector form factors and reduces the cross section by a factor of @xmath25 . when the subsequent decays to specific pion - nucleus channels are considered , there are additional cgcs . the results for vector , axial - vector and interference terms were computed in an earlier article @xcite and the results are shown again in figure [ fig : pcacold ] . considering next the reaction @xmath26 we use the same structure functions as for ( [ eq : vccp ] ) , since they are related by charge symmetry , but the sign of the interference term in the cross section changes . we note in figure [ fig : pcacold ] that the vector and interference terms are approximately equal and cancel each other for antineutrinos , leaving the axial contribution alone . .input quantities and isospin factors for various reactions . [ cols="^,^,^,^,^,^",options="header " , ] the results of this study point to general properties which may be useful to other investigations . for the determination of oscillation parameters and the observation of cp asymmetries we need the yields of single pions from nuclear targets which have the same energy dependence for neutrino and antineutrino reactions . we point out that for isoscalar targets the nuclear corrections for @xmath13 and @xmath14 are equal provided that their energy spectrum at the production vertex are the same . this requires the production by neutrinos and antineutrinos to be the same . we pointed out that the structure functions for reactions ( [ eq : vccp ] ) and ( [ eq : avcc ] ) are the same by charge symmetry . the difference in the cross sections comes from the sign in front of @xmath27 . we remark now that the contribution of @xmath27 to the cross sections diminishes as @xmath28 increases . this is evident in equation ( 2.10 ) of reference @xcite and is apparent in figures 6 8 of reference @xcite . the merging together of neutrino and antineutrino cross sections also occurs as @xmath29 . we pointed this out for neutral currents in figure [ fig : isospinrelationsnc ] . it also happens in charged current reactions and we demonstrate it for @xmath30 and @xmath31 by repeating the calculation and replacing the muon mass by the electron mass . in figure [ fig : electron ] we show the two cross sections together and for various energies . for the subsequent interaction the nuclear corrections on isoscalar targets will be the same for @xmath13 and @xmath14 . these remarks suggest to use several reactions in the nearby detector in order to determine the fluxes and at the same time check the shapes and normalizations of pion production . then measure in the far away detector charge - symmetric reactions . ratios that follow from charge symmetry should provide checks at every step . for instance , in the charged current reactions the @xmath13 yield is dominant , but for antineutrinos the @xmath14 yield dominates . for neutral currents the @xmath32 yield is dominant and the sum @xmath33 is , within 5 % , equal to the @xmath32 yield . then one can use the determined fluxes of @xmath34 and @xmath35 in order to compute the regenerated , through oscillations , @xmath19 and @xmath20 fluxes far away . for charge symmetric reactions for which the products cross sections @xmath18 nuclear corrections are the same , the change in the yields at the far away detector should be proportional to the change in the fluxes , coming from the oscillations with or without cp violation . we carried out an extensive program of calculations in the kinematic region @xmath36 and @xmath37{gev^2}$ ] , where pcac gives the dominant contribution for the axial current . we give results for charged and neutral current reactions of neutrinos and antineutrinos . one interesting result , in the energy range @xmath38{gev}$ ] , is that the axial contribution dominates the @xmath39 reaction . in the calculations we kept the mass @xmath40 which produces the turn over in the differential cross sections @xmath1 as @xmath29 . the long baseline oscillation experiments will be searching for @xmath30 type reactions where @xmath41 and we repeated the results in figure [ fig : electron ] . the pcac limiting point at @xmath42 is now finite . the energy spectra for the pions are also interesting and measurable . in section [ sec : energyspectrumandangulardependenceofpions ] we outlined a method and carried out the calculations obtaining the energy and angular spectra of the pions with the help of transformations summarized in the appendix . since our estimates for the axial current are valid for @xmath43{gev^2}$ ] we limited the figures to that kinematic region . for nuclear corrections we adopted the anp model which is both simple and transparent . we presented corrections for the incoherent sum of a proton and neutron target . they can be scaled up for other isoscalar nuclei ( c@xmath44 , o@xmath45 , ... ) by multiplying with the appropriate number of particles in the target and using the appropriate transfer matrix . then we pointed out that some general properties follow from charge symmetry . these relations follow from isospin rotations and should hold beyond the kinematic domains introduced by the validity of pcac . we hope that the method and results can be adopted in oscillation experiments searching for cp asymmetry in the leptonic sector . one of us ( eap ) wishes to thank drs . w. bardeen and s. parke for the hospitality at fermi laboratory and the humboldt foundation for a traveling grant . numerous discussions with dr . j. morfin and members of the minerva group helped us formulate the physical issues of this work . we define as center of mass system the frame where the resonance is at rest . in the laboratory the four - momentum of the excited resonance after the collision is @xmath46 we can bring the resonance to rest by the transformation parameter @xmath47 and the corresponding @xmath48 in the rest frame of the resonance with its invariant mass within its width by @xmath49 , the energy of the pion is @xmath50 with this information we can relate various quantities in the two frames . the scattering angles of the pion satisfy @xmath51 the differential cross sections are related by @xmath52 and @xmath53 \label{eq : apb}\end{aligned}\ ] ] with the jacobian @xmath54 derived from the above equation @xmath55 finally , we obtain @xmath56	we present the results of extensive calculations for charged and neutral current reactions of neutrinos and antineutrinos in the small @xmath0{gev^2}$ ] region . the results include single @xmath1 and double @xmath2 differential cross sections at energies relevant for oscillation experiments . we include nuclear corrections in the adler - nussinov - paschos model and point out that on isoscalar nuclear targets there are charge symmetry relations that hold in extended kinematic regions . we discuss how the results can be used in long baseline experiments in order to study oscillation parameters and search for cp asymmetries .
the identification of scalar mesons and particularly the determination of the lightest scalar @xmath13 nonet is of extreme importance both for quark systematics and searches for non-@xmath13 objects like glueballs and multiquark states ( see the discussion of this problem in the minireview on scalar mesons by s.spanier and n.trnqvist , p.390 of ref . it is generally accepted that the probabilities of the electric dipole radiative transitions of the @xmath0 meson are crucial for the clarification of the nature of @xmath14 and @xmath15 mesons @xcite . a search for these decays has been earlier performed by the nd @xcite and cmd-2 @xcite groups . a number of new results has been recently published by snd @xcite and cmd-2 @xcite collaborations at the @xmath16 collider vepp-2 m @xcite where the radiative decays @xmath17 , @xmath18 and @xmath12 have been observed for the first time . the mode with two charged pions has very large background because of the radiative processes @xmath19 where a photon comes from initial electrons or from final pions . therefore a signal from the @xmath3 final state can be seen as an interference structure at the energy @xmath20 40 mev in the photon spectrum . the shape of this structure depends on the @xmath21 mass and width . in our first paper @xcite a search for the @xmath12 decay led to the upper limit of @xmath22 that looked at first puzzlingly low compared to the decay @xmath17 observed with the branching ratio of about @xmath23 @xcite . as was shown in @xcite , in case of the @xmath24 intermediate state it could be explained by the destructive interference between bremsstrahlung processes and the @xmath0 decay . in this paper results of the study of the @xmath25 process with the cmd-2 detector @xcite are presented . in total , the 14.2 @xmath26 of data have been collected since 1993 at 14 energy points around the @xmath0 mass . for the analysis of the @xmath27 decay mode 13.1 @xmath26 corresponding to @xmath28 @xmath0 decays were used . seven scans of this energy region were performed allowing control of systematic errors caused by possible detector instability . the results obtained from individual scans were found to be consistent . event candidates were selected by requiring only two minimum ionizing tracks in the drift chamber ( dc ) and one or two photons with energy greater than 20 mev in the csi calorimeter . the following selection criteria were used : 1 . the average momentum of two charged particles is higher than 240 mev / c to remove the background from @xmath29 decays . detected tracks have a polar angle between 1.05 and 2.1 radians so that they enter the inner muon system . the sum of the energy depositions of two clusters associated with two tracks is less than 450 mev to remove bhabha events . the radial distance of the closest approach of each track to the beam axis is less than 0.3 cm . the z - coordinate of the vertex ( along the beam ) is within 10 cm from the detector center . this cut reduces cosmic ray background by a factor of two . detected photons have a polar angle between 0.85 and 2.25 radians so that they enter the `` good '' region in the csi barrel calorimeter . this requirement suppressed the background from the photons emitted by initial electrons . events with an invariant mass of two photons close to the @xmath30 mass ( @xmath31 mev ) were removed . after the above cuts the main background for the @xmath32 final state comes from : a ) the radiative process @xmath33 , b ) the decay @xmath34 when one of the photons from the @xmath30 escapes detection and c ) collinear events @xmath35 in which secondary decays and interactions of muons or pions with the detector material produce a background cluster mimicking a photon . the inner muon system was used to separate muons from pions . the muon system uses streamer tubes grouped in two layers ( inner and outer ) with a 15 cm magnet yoke serving as an absorber and has 1 - 3 cm spatial resolution . the requirement of hits in the inner muon system for both charged particles selects muon events , together with some pion events in which both pions pass the calorimeter without nuclear interaction . separation of pion and muon events in the csi calorimeter is illustrated in fig . [ ec1-ec2 ] , where scatter plots of the energy deposition of one track vs. that of the other one are presented for events with one or no hits in the muon system ( a ) and selected as muons ( b ) . energy depositions are corrected for the incident angle . pions can have nuclear interactions and in some cases leave more energy , while muons mostly exhibit de / dx losses only . [ tbh ] to select a cleaner sample of muon events , in addition to the information from the muon system both tracks were required to show only minimum ionizing energy deposition in the calorimeter ( 60 - 130 mev ) . all the remaining events were considered as pion candidates . in this raw separation the pion sample @xmath37 contains muons because of some inefficiency of the muon system while the muon sample @xmath38 contains pions since a pion can reach the muon system without nuclear interaction . the observed numbers of events @xmath37 and @xmath38 are related to the true numbers @xmath39 and @xmath40 as : + @xmath41 , + @xmath42 , + where @xmath43 and @xmath44 are the muon system efficiency and probability for a pion pair to be selected as muons respectively . the magnitudes of @xmath43 and @xmath44 were determined by studying correlations between the energy deposition in the csi calorimeter and response of the muon system to collinear @xmath45 and @xmath46 events . the results of this study are shown in fig . [ pimu_eff ] for one of the experimental scans . figure [ pimu_eff]a presents the muon system efficiency for different energy points . only statistical errors are shown . the probability for two pions to be selected as muons is presented in fig . [ pimu_eff]c and is independent of the possible instability of the detection system . the number of pions and muons has been corrected for the dc reconstruction efficiency shown in fig . [ pimu_eff]b . this efficiency was determined from collinear bhabha events as described in @xcite . our study of the efficiencies gives a correlated 3% systematic uncertainty in the final number of @xmath45 and @xmath46 events . for events with photons @xmath43 and @xmath44 depend on photon energy . this dependence was studied using simulated @xmath27 events . the probability of two pions to be selected as muons varies with photon energy as presented in fig . [ pi_loss ] and was used to correct the final number of events with photons . at low photon energy this probability is consistent with that obtained from collinear events ( see fig . [ pimu_eff]c ) . to reduce the background from collinear events as well as that from the three pion @xmath0 decay a constrained fit was used requiring total energy - momentum conservation for a three body decay . about 20% of the selected events had an additional photon . in this case the constrained fit was applied to both possible combinations and that with a minimum @xmath47 was chosen . an additional cut was applied to the photon direction : an azimuthal angle should be more than 0.25 radians from the charged track direction . this cut removed the remaining collinear events with a background photon which survived after the constrained fit . the @xmath47/d.f . distribution for events selected as muons had very small background and was found to be in good agreement with simulation as shown in fig . [ xi2_pmg]a . motivated by simulation , a cut that @xmath47/d.f . be less than 3 was imposed for pion and muon events selecting 95% of signal events . after the above cut the pion sample still contained some background mainly from three pion @xmath0 decays . such three pion background appears when one of the photons from the @xmath48 has energy below 20 mev and is not detected so that the event looks like a three body decay with the remaining photon energy higher than 150 mev . the @xmath47/d.f distribution for these events was flat and events with 3 @xmath49/d.f.@xmath50 6 , shown in fig . [ xi2_pmg]b by open points , were used to estimate the background spectrum . at each energy the background spectrum was subtracted from that for the signal candidates . from a constrained fit one can obtain an improved estimate for the photon energy . this effect was studied by using simulation and results are presented in fig . [ gam_eff ] . simulation shows that after the constrained fit photons have energy resolution about 5 mev in the whole energy range instead of @xmath51 csi resolution as shown in figs . [ gam_eff]a , b . figure [ gam_eff]c demonstrates the simulated photon detection efficiency in the csi calorimeter vs. photon energy . the overall detector efficiency vs. photon energy is shown in fig . [ gam_eff]d for @xmath27 events . [ tbh ] to extract the resonant contribution associated with the @xmath0 , two data sets were used . the first set ( `` @xmath0 '' region ) for e@xmath52 from 1016.0 to 1023.2 mev had the integrated luminosity of 9.24 @xmath26 while the second one ( `` off-@xmath0 '' region ) taken at e@xmath52 = 996 - 1013 and 1026 - 1060 mev with the integrated luminosity of 3.89 @xmath26 containing less than 3% @xmath0 decays was used for a background estimate . figure [ gam_sum ] presents photon spectra obtained after background subtraction and corrections for the detector efficiency at the `` @xmath0 '' ( a , c ) and `` off-@xmath0 '' ( b , d ) regions for pions and muons respectively . the solid line corresponds to the theoretical calculation @xcite taking into account the integrated luminosity at each energy point and @xmath53 mixing for the bremsstrahlung process . a peak at 220 mev is due to the radiation by initial electrons resulting in the process @xmath54 , @xmath55 . [ tbh ] the photon energy range 20 to 120 mev has minimum background and some excess of events over the expected bremsstrahlung spectrum ( the solid line in fig . [ gam_sum]a ) can be seen in the @xmath56 sample at the `` @xmath0 '' region . in total , 30175 @xmath56 events and 27188 @xmath36 events have been selected in this energy range . the cross section for each energy point was calculated as @xmath57 = @xmath58/(l@xmath59 ) . @xmath58 is the number of selected events with photons in the energy range from 20 to 120 mev . the integrated luminosity l for each energy point was determined from bhabha events with about 2% systematic accuracy @xcite . the detection efficiency @xmath60 was obtained by simulation and the approximation shown by the solid line in fig . [ gam_eff]d was used to correct photon spectra . the resulting cross sections of the processes @xmath61 and @xmath6 versus energy are presented in fig . [ fnot - gamma]a , b . only statistical errors are shown . the systematic error in the experimental cross sections was estimated to be about 5% dominated by the uncertainty of the pion - muon separation efficiencies . [ tbh ] the cross sections can be described according to the calculations performed in @xcite . it has been shown there that the amplitudes describing initial bremsstrahlung differ from those for the final bremsstrahlung and the @xmath0 meson , so that the differential cross section for @xmath27 events is : + @xmath62 . + here @xmath63 and @xmath63 are the amplitudes describing initial and final bremsstrahlung processes , and the amplitude @xmath64 introduces the influence of the @xmath0 upon the photon propagator ( a vacuum polarization term ) . this contribution gives rise to an interference pattern in the cross section at the @xmath0 mass and can be referred to as an `` electromagnetic '' decay @xmath65 or @xmath66 . the amplitude of the interference is determined by the @xmath0 meson leptonic coupling constant . the amplitude @xmath67 represents the possible @xmath0 decay into the @xmath27 final state via @xmath68 . for @xmath36 events @xmath67 = 0 . the model considers the @xmath69 meson as a two- or four - quark state or @xmath70 molecule depending on the values of the coupling constants g@xmath71 and g@xmath72 . it also includes the interference of the @xmath0 amplitude with the final bremsstrahlung amplitude @xmath73 as well as the correction for the @xmath74 scattering in the final state @xcite . the latter effect gives a contribution of about 15% to the branching ratio , but within our accuracy can be introduced as an additional phase shift @xmath75 between the amplitudes of the @xmath0 and bremsstrahlung process . this shift was predicted to be 1.2 - 1.4 radians , but the sign of the term was unknown and was determined from the fit . in general , all amplitudes have different dependence on s and @xmath76 . the initial bremsstrahlung process presented by the amplitude @xmath63 is suppressed by selecting photons transverse to the beam direction , but it still accounts for about 2/3 @xcite of the observed @xmath19 and one half of the @xmath77 cross section . integration over @xmath76 from 20 to 120 mev gives @xmath78 which can be used for the cross section aproximation in fig . [ fnot - gamma ] . assuming no hadronic @xmath2 decay ( @xmath67=0 ) a simple formula has been used where the final radiation amplitude of the bremsstrahlung process interferes with the breit - wigner amplitude from the @xmath0 vacuum polarization : + @xmath79 ; + @xmath80 . + here @xmath81 , @xmath82 and @xmath83 are @xmath0 meson parameters , @xmath84 is the @xmath0 decay amplitude in the peak and @xmath85 is the relative phase between the bremsstrahlung and @xmath0 decay amplitudes . this formula is valid under the natural assumption of the same photon spectrum for the final radiation amplitude of the bremsstrahlung process and the breit - wigner amplitude of the `` electromagnetic '' @xmath0 decay . the initial @xmath86 and final @xmath87 bremsstrahlung cross sections have different energy dependence for 20 - 120 mev photons because of the pion formfactor energy behaviour . according to @xcite the power functions @xmath88 and @xmath89 can describe the bremsstrahlung process @xmath90 in our energy range . the parameter @xmath91 represents the sum of initial and final bremsstrahlung cross section at the @xmath0 mass . the function @xmath92 was used for muons . the fit of experimental data with @xmath91 , peak amplitude @xmath84 and @xmath85 as free parameters shows good agreement of obtained cross sections with the theoretical calculations @xcite : + @xmath93 , + @xmath94 + for pions and muons respectively . the above values also indicate that the @xmath95 separation uncertainty does not exceed the estimated systematic error . using the obtained amplitude one can calculate the peak cross section @xmath96 , where @xmath97 is the total peak cross section of the @xmath0 resonance obtained from the leptonic width @xcite . the following results have been obtained after inserting corrections for angular acceptance equal to 0.68 and 0.42 for pions and muons respectively : @xmath98 , @xmath99 , @xmath100 radians for pions and @xmath101 radians for muons . the obtained values should be compared to the theoretical calculations based on `` electromagnetic '' @xmath0 decays only @xcite : @xmath102)=0.047@xmath103 , @xmath104)=1.15@xmath105 . + for muons the experimental value is in good agreement with theory , but for pions the value of the measured branching ratio is 9 times larger than the theoretical expectation and points to the presence of the hadronic decay of @xmath0 to @xmath27 . the obtained branching ratios supersede our results in @xcite based on the part of the total data sample . the corresponding fit curves are shown in fig . [ fnot - gamma ] . to search for the @xmath24 decay contribution to the observed @xmath12 decay , photon energy spectra were studied . the photon spectra from the `` @xmath0 '' region are shown in fig . [ f0-spectr ] for six c.m.energy points and photons in the 20 - 160 mev energy range . the spectra were corrected for all experimental inefficiencies and normalized to the integrated luminosity . the signal from the decay of the @xmath24 is seen as a structure in the photon spectra at 40 - 60 mev . also shown by the dotted lines are the theoretical predictions for the expected bremsstrahlung spectra @xcite including vacuum polarization . because of the presence of the resonance in the photon spectra the branching ratio obtained in the previous chapter for the @xmath12 decay can be used only as some indication to the existence of the @xmath0 hadronic decay into this mode . it should be mentioned that the excess of events over the bremsstrahlung spectra can not be used for the branching ratio calculation . as it was shown in @xcite , the destructive interference with the bremsstrahlung process can reduce a visible signal in the selected energy range for photons . to demonstrate the effect of interference the photon spectra in fig . [ f0-spectr ] were fit as a group with the spectra calculated for each energy point using the differential cross section from @xcite ( see previous section ) . the number of events was not sufficient to keep all model parameters free and therefore the model parameters g@xmath106 and g@xmath107 were varied to keep the @xmath21 width at about 40 mev @xcite . the fit had the following free parameters : the branching ratio , @xmath21 mass and phase shift @xmath75 . the data can be described by the model only for the destructive interference between the @xmath69 amplitude and the bremsstrahlung process . the following results have been obtained : @xmath108 ; @xmath109 mev ; @xmath110 radians . + the results above include a systematic error from the weak dependence on the @xmath21 width . the relatively large value of the branching ratio obtained in the fit can not be accounted for by the models assuming the normal @xmath13 structure of @xmath21 @xcite and is predicted only in case of its four quark structure . note also that it significantly differs from the branching ratio obtained from the simple fit assuming the same photon spectra for all processes . the selected @xmath27 events with photon energies from 100 to 300 mev ( see fig . [ gam_sum ] ) can be used to search for the c - violating decay @xmath112 , @xmath55 . the cross section vs. energy for the events with photons in the 100 - 300 mev range is presented in fig . [ rho - sech ] . the final state in the c - violating @xmath112 decay has the same quantum numbers as in the initial bremsstrahlung process which dominates in the cross section and to extract a possible signal the interference of these two processes is assumed : + @xmath113 , + where @xmath114 . the comparison of the obtained cross section with the theoretical calculation @xcite gives the ratio @[email protected] . as a result of the fit , @xmath117 and @xmath118 radians have been obtained . taking into account the 80% detection efficiency , the corresponding upper limit is : @xmath119 at 90% c.l . this result should be compared to the previous measurements which gave upper limits @xmath120 @xcite and @xmath121 @xcite . the selected @xmath27 events can be used to search for the p- and cp- violating decay @xmath123 , where the @xmath10 comes from the radiative @xmath124 decay . from 19.7 millions of @xmath0 decays used for the @xmath27 channel analysis one could expect about 248,000 events which decayed via the @xmath125 channel . these p- , cp - violating decays should be observed as peaks in the invariant mass of two pions at @xmath126 . figure [ etapipi ] represents experimental distributions of @xmath45 masses from selected @xmath27 events . the line corresponds to a fit with a linear function and gaussian distribution representing a possible signal . it was found that the signal does not exceed 10 events at 90% cl . the histogram in the box shows a simulated signal from @xmath123 decays at 90% cl . the detection efficiency found by simulation was 0.124 . the following result has been obtained : br(@xmath127 . + which should be compared to the best previous limit of @xmath128 @xcite . using 13.1 pb@xmath129 of data collected around the @xmath0 meson ( about 20 millions of the @xmath0 decays ) @xmath19 and @xmath77 events were selected . for the first time the decay @xmath12 has been observed in the 20 - 120 mev photon energy range . the fit assuming that the @xmath0 contributes only to the photon propagator ( no direct @xmath0 decay ) gave the branching ratio : @xmath130 . + this value is nine times higher than the expected and points to the presence of the hadronic decay of the @xmath0 into this final state . the analysis of the photon spectra shows the presence of a resonance in the @xmath45 system with a mass of about 980 mev . the obtained branching ratio can be affected by the complicated interference of the hadronic @xmath0 decay with the bremsstrahlung process . the effect of interference was demonstrated by fitting photon spectra assuming the @xmath24 decay and the following results have been obtained :	radiative decays of the @xmath0 meson have been studied using a data sample of about 20 million @xmath0 decays collected by the cmd-2 detector at vepp-2 m collider in novosibirsk . from selected @xmath1 events the @xmath2 decay has been observed for the first time . under the assumption that the intermediate @xmath3 state dominates in the @xmath4 decay , the corresponding branching ratio is @xmath5 . selected @xmath6 events were used to obtain @xmath7 for @xmath8 mev . using the same data sample , upper limits at 90% cl have been obtained for the c - violating decay of the @xmath0 : @xmath9 ; and for the p- and cp - violating decay of the @xmath10 : @xmath11 . * first observation of the @xmath12 decay * r.r.akhmetshin , e.v.anashkin , m.arpagaus , v.m.aulchenko , v.s.banzarov , l.m.barkov , n.s.bashtovoy , a.e.bondar , d.v.bondarev , a.v.bragin , d.v.chernyak , a.s.dvoretsky , s.i.eidelman , g.v.fedotovich , n.i.gabyshev , a.a.grebeniuk , d.n.grigoriev , p.m.ivanov , s.v.karpov , v.f.kazanin , b.i.khazin , i.a.koop , p.p.krokovny , l.m.kurdadze , a.s.kuzmin , i.b.logashenko , p.a.lukin , a.p.lysenko , k.yu.mikhailov , i.n.nesterenko , v.s.okhapkin , e.a.perevedentsev , e.a.panich , a.s.popov , t.a.purlatz , n.i.root , a.a.ruban , n.m.ryskulov , a.g.shamov , yu.m.shatunov , b.a.shwartz , a.l.sibidanov , v.a.sidorov , a.n.skrinsky , v.p.smakhtin , i.g.snopkov , e.p.solodov , p.yu.stepanov , a.i.sukhanov , v.m.titov , yu.v.yudin , s.g.zverev + budker institute of nuclear physics , novosibirsk , 630090 , russia j.a.thompson + university of pittsburgh , pittsburgh , pa 15260 , usa
1.0truecm _ institut fr kernphysik , technische universitt wien , _ + + 1.0truecm 1.0truecm 1.0truecm pacs:11.10.ef , 13.75 . cs , 14.20 . + keywords : relativistic field theory , deuteron , neutron , proton , radiative capture , weak interaction , w boson , fusion 1.0truecm in a recent publication [ 1 ] we have suggested a relativistic field theory model of the deuteron . this model , being some kind of a @xmath4model [ 2 ] , is based on the assumption that the physical deuteron state should be produced due to integration over low energy proton neutron fluctuations at energies restricted by the scale @xmath5 . for simplification we have suggested a one nucleon loop approximation for the integration over proton neutron fluctuations . unfortunately , this approximation does not have a perturbative parameter and should be considered as effective one , like that suggested by nambu and jona lasinio [ 3 ] . in this case the scale @xmath5 has the meaning of the cut off . we have defined @xmath5 in terms of the effective radius of the deuteron @xmath6 , i.e. @xmath7 . for the estimate of @xmath5 we have applied the non relativistic formula : @xmath8 [ 4,5 ] , where @xmath9 is the binding energy of the physical deuteron [ 5 ] and @xmath10 is the mass of the nucleon . we used equal masses for the proton and neutron , i.e. @xmath11 . this corresponds to the chiral limit when masses of current u and d quarks vanish , i.e. @xmath12 . our estimate of @xmath5 gives : @xmath13 . the interactions of the deuteron with proton and neutron were described in terms of two coupling constants @xmath14 and @xmath15 , defining the interactions via vector and tensor nucleon currents , respectively . in the one nucleon loop approximation we calculated the anomalous magnetic dipole moment of the deuteron @xmath16 in units of the nucleon magneton @xmath17 , where @xmath18 is the proton charge , and @xmath19 the electric quadrupole moment . imposing the constraint @xmath20 , being valid in the lowest approximation , we found the correlation between coupling constants @xmath14 and @xmath15 , i.e. @xmath21 . then the coupling constant @xmath14 has been fixed in terms of the electric quadrupole moment @xmath19 . eventually we calculated in one nucleon loop approximation the binding energy of the physical deuteron in terms of @xmath14 and @xmath5 . the theoretical result obtained was in good agreement with the experimental value : @xmath22 [ 5 ] . the physical nature of the coupling constants @xmath14 and @xmath15 is connected with one meson exchange . of course , the @xmath23meson exchange should give the main contribution . this assumption has been confirmed by the magnitude of @xmath14 , i.e. @xmath24 , where @xmath25 is the coupling constant of the @xmath26interaction . the application of the model to the calculation of processes of low energy interactions of the deuteron encounters the problem of the unambiguous computation of one nucleon loop digrams that should describe the interactions of the deuteron with other particles in the suggested model . indeed , since these are fermion loops , most of them , contributing to processes like radiative neutron proton capture n + p @xmath0 d + @xmath1 , low energy proton proton reaction p + p @xmath0 d + @xmath2 + @xmath3 and so on , depend explicitely on the shift of virtual momenta of nucleons in the loop . this introduces substantial ambiguities interfering with the direct application of the model . however , free parameters appearing due to shifts of virtual momenta in nucleon loops must not be considered as free parameters of the model introduced especially for the computation of every nucleon diagram . in this paper we revise our model [ 1 ] and apply it to the calculation of the cross sections of the radiative neutron proton capture n + p @xmath0 d + @xmath1 and two proton fusion p + p @xmath0 d + @xmath2 + @xmath3 . in order to compute unambigously the effective anomalous magnetic and electric quadrupole moments of the deuteron we impose constraints on the computation of one nucleon loops , caused by the requirement of electromagnetic gauge invariance of contributions for individual loops . this distinguishes our effective model from a field theory like qed , where the electromagnetic gauge invariance should be required for the complete set of diagrams in fixed order of perturbation theory . the requirement of electromagnetic gauge invariance applied to the individual nucleon loops allows one to fix ambiguities and use one nucleon loops as well defined quantum theory objects . in order to remove ambigutities of the computation of the one nucleon loop diagrams describing the effective lagrangian of the radiative neutron proton capture we turn to the selection rules . in the case of the computation of the effective lagrangian responsible for the low energy two proton fusion p + p @xmath0 d + @xmath2 + @xmath3 we can remove ambigutities of the computation of the one nucleon loop diagrams by using the gauge invariance under gauge transformations of the deuteron field . the present paper is organized as follows . in sect . 1 we adduce the starting assumptions that we have put in the foundation of the model , and phenomenological parameters in terms of the parameters characterizing the physical deuteron . in sects . 2 and 3 we give a detailed derivation of the effective corben schwinger and aronson lagrangians , respectively , describing electromagnetic interactions of the deuteron in terms the magnetic and electric quadrupole moments . in sect . 4 and sect . 5 we apply the relativistic field theory model of the deuteron to the computation of the amplitudes and cross sections of the radiative neutron proton capture n + p @xmath0 d + @xmath1 and the low energy proton proton fusion p + p @xmath0 d + @xmath2 + @xmath3 , respectively . in the conclusion we discuss the obtained results . in ref . [ 1 ] we have suggested a relativistic field theory model for the deuteron as a bound state of proton and neutron . the main idea , that has been put into the foundation of the approach , has been adopted from the nambu jona lasinio ( njl ) model [ 3 ] . as has been suggested by nambu and jona lasinio [ 3 ] , the physical deuteron should appear as the proton neutron collective excitation due to a local four nucleon interaction and vacuum fluctuations of nucleons , calculated in one nucleon loop approaximation , producing both a kinetic term of the physical deuteron and its interactions with other particles . unfortunately , the direct realization of the nambu jona lasinio program , i.e. the use of the local four nucleon interaction , would lead to a strongly bound localized proton neutron state , whereas a physical deuteron is weakly bound and rather `` extended '' object . therefore , we have borrowed from the njl model only the admission of the applicability of the one loop approximation . this is used for the computation of observed parameters for physical states in leading order in the long wavelength expansion [ 69 ] . in our description of the physical deuteron , being some kind of the @xmath4model , we use the following scheme : we start from the lagrangian of an unphysical deuteron field @xmath27 , considered as a bound proton neutron state at zero binding energy and with a mass equal to the sum of the proton and neutron masses . then in the one nucleon loop approximation and leading order in the long wavelength expansion we obtain an effective lagrangian of a physical deuteron field describing a physical deuteron with observable binding energy @xmath28 [ 5 ] , anomalous magnetic dipole moment @xmath29 , determined as @xmath30 , where @xmath31 , @xmath32 and @xmath33 are the magnetic dipole moments of the physical deuteron , proton and neutron , respectively , and the electric quadrupole moment @xmath34 [ 5 ] . the magnetic dipole moment of the deuteron is measured in nuclear magnetons @xmath35 , where @xmath18 and @xmath36 are the electric charge of the proton and the mass of the proton and neutron , then @xmath37 and @xmath38 are the anomalous magnetic dipole moments of the proton and neutron . in the case of the neutron the total magnetic dipole moment coincides with the anomalous one . below , to simplify matter , we neglect the proton neutron mass difference and use equal masses of the proton and neutron , i.e. @xmath39 . this should correspond to the chiral limit approximation with zero masses of current @xmath40 and @xmath41quarks , i.e. @xmath42 . in this connection we would like to refer to the paper written by sakita and goebel [ 10 ] , where , for the aim of the investigation of the low energy limit of the photodisintegration , the deuteron has been described in terms of a local field operator . the lagrangian of the unphysical deuteron field @xmath43 , which interacts strongly with the proton @xmath44 and neutron @xmath45 fields , reads @xmath46 d^{(0)}_\mu ( x ) - \nonumber\\ & & - ig_{\rm v } [ \bar{p^c}(x ) \gamma^\mu n(x ) - \bar{n^c}(x ) \gamma^\mu p(x ) ] d^{\dagger(0)}_\mu ( x ) + \\ & & + \frac{g_{\rm t}}{m_0 } [ \bar{p}(x ) \sigma^{\mu\nu } n^c(x ) - \bar{n}(x ) \sigma^{\mu\nu } p^c(x ) ] d^{(0)}_{\mu\nu}(x ) + \nonumber\\ & & + \frac{g_{\rm t}}{m_0 } [ \bar{p^c}(x ) \sigma^{\mu\nu } n(x)- \bar{n^c}(x ) \sigma^{\mu\nu } p(x ) ] d^{\dagger(0)}_{\mu\nu}(x ) + \nonumber\\ & & + \bar{p}(x ) ( i\gamma^\mu \partial_\mu - m_{\rm n } ) p(x ) + \bar{n}(x ) ( i\gamma^\mu \partial_\mu - m_{\rm n } ) n(x).\nonumber\end{aligned}\ ] ] here @xmath47 , @xmath48 is the mass of the unphysical deuteron , then @xmath49 and @xmath50 , and @xmath51\,$ ] . the operator @xmath52 denotes charge conjugation , @xmath53 is a transposition , @xmath14 and @xmath15 are the phenomenological constants which will be fixed below . we assume that the coupling constants @xmath14 and @xmath15 are caused by one meson exchanges . the @xmath23meson exchange should give the main contribution . the guiding principle for the construction of the lagrangian ( [ label1 ] ) has been only the fact that the deuteron is a bound state of the proton and neutron with spin one [ 2 ] . in order to compute the electric quadrupole and anomalous magnetic dipole moments of the physical deuteron we have to include the interaction the unphysical deuteron field @xmath54 with the electromagnetic field . having performed this inclusion by a minimal coupling we obtain the lagrangian @xmath55 where @xmath56 a_\mu(x ) d^{(0)}_{\nu}(x ) -\nonumber\\ & & - ie\frac{2g_{\rm t}}{m_0 } [ \bar{p^c}(x ) \sigma^{\mu\nu } n(x)- \bar{n^c}(x ) \sigma^{\mu\nu } p(x ) ] a_\mu(x ) d^{\dagger(0)}_{\nu}(x ) - \\ & & -i\,e\,\frac{\lambda}{m^2_{0}}\,d^{\dagger ( 0)}_{\mu\nu}(x)\,d^{(0)\nu \alpha}(x)\,{f_{\alpha}}^{\mu}(x)\,- e\,\bar{p}(x ) \gamma^\mu p(x ) a_\mu(x)\,+\nonumber\\ & & + ie\,\frac{\kappa_{\rm p}}{4m_{\rm n}}\,\bar{p}(x)\,\sigma^{\mu\nu}p(x)\,f_{\mu\nu}(x)\,+ \,ie\,\frac{\kappa_{\rm n}}{4m_{\rm n } } \bar{n}(x)\,\sigma^{\mu\nu}n(x)\,f_{\mu\nu}(x)+o(e^2).\nonumber\end{aligned}\ ] ] here @xmath57 and @xmath58 are the electromagnetic field strength tensor and the electromagnetic potential , respectively . also we have added the aronson interaction [ 11 ] describing the electric quadrupole and anomalous magnetic moments of the unphysical deuteron : @xmath59 and @xmath60 [ 11 ] . now let us give some arguments for the justification of the application of the aronson interaction . recall , that the electric quadrupole and the anomalous magnetic dipole moments characterise the inhomogeneity of the structure of the unphysical deuteron represented in terms of the field @xmath54 . the availability of such an inhomogeneity is not obvious beforehand and does not follow from the lagrangian ( [ label1 ] ) via a minimal coupling inclusion of the interaction of the unphysical deuteron with an external electromagnetic field . therefore , the inclusion of nonzero @xmath61 and @xmath62 is an additional assumption concerning the internal structure of the unphysical deuteron field . since , the minimal coupling leads to the appearance of electromagnetic interactions linear in @xmath63 , the required interaction of the unphysical deuteron field with an external electromagnetic field , caused by electric quadrupole and anomalous magnetic dipole moments , should be of order @xmath64 . this does not contradict to the main principles of classical electrodynamics [ 12 ] and gives the required interaction in the form , being irreducible with respect to those interactions induced by a minimal coupling . the aronson effective lagrangian provides a simplest form of an interaction of order @xmath65 [ 11 ] . the inclusion of the aronson interaction can be also justified by the assumption that the electromagnetic interactions induced by electric quadrupole and anomalous magnetic dipole moments of the unphysical deuteron field with an external electromagnetic field become available only in next to leading order in long wavelength expansion , i.e. in powers of gradients of the deuteron field @xmath63 . in leading order in the long wavelength expansion , linear in @xmath63 , the interaction of the unphysical deuteron with an external electromagnetic field , caused by electric quadrupole and anomalous magnetic dipole moments , vanishes . we have taken into account the anomalous magnetic dipole moments of the proton and neutron that play an important role for the radiative neutron proton capture n + p @xmath0 d + @xmath1 . the inclusion of @xmath66 and the anomalous magnetic dipole moments of the proton and neutron distinguishes the present model from that given in [ 1 ] . some of the results and assertions adduced in this section repeat those given in [ 1 ] . this concerns mainly the computation of the binding energy @xmath67 and the discussion of the validity of the approximation having been applied to the computation of one nucleon loop diagrams . this repetition is caused by the consideration of the completeness of the exposition , i.e. in order not to relegate a reader to [ 1 ] for consultation , and the change of the magnitudes of phenomenological parameters of the model . the former is due to different procedures of the removal of ambiguities of one nucleon loop diagrams , induced by shifts of virtual nucleon momenta , we use here and we have used in ref . [ 1 ] . in the present version all shift ambiguities are fully fixed by requirements of gauge invariance . following [ 1 ] we compute the effective lagrangian of the physical deuteron field in the one nucleon loop approximation ( see also [ 69 ] ) . the nucleon diagrams describing the contributions to the kinetic term of the deuteron field and leading to the appearance of the non zero value of the binding energy are depicted in fig . 1 . by calculating these diagrams in leading order in the long wavelength expansion we get where @xmath68 and @xmath69 are the following divergent integrals the ultra violet cut off @xmath70 restricts the 3momenta of fluctuations of virtual nucleons taking part in the formation of the physical deuteron field . one should expect that @xmath70 is connected with the region of localization of a wave packet procreated by fluctuations of virtual nucleons , i.e. @xmath71 . we shall specify the value of @xmath72 below . now we discuss the problem of the applicability of the long wavelength expansion for the calculation of the one nucleon loop diagrams in fig . it is well known that a two body @xmath73wave bound state being denoted as @xmath74 with a reduced mass @xmath75 and a binding energy @xmath67 is localized in the region restricted by @xmath76 [ 4 ] . this quantity can be considered as the effective radius of the bound state [ 5 ] . in the case of the physical deuteron we have @xmath77 [ 5 ] . this value exceeds three times the effective radius of nuclear forces @xmath78 , where @xmath79 is the pion mass [ 13 ] . hence the physical deuteron looks like a rather `` extended '' bound state . it should be obvious that such an `` extended '' bound state can be formed at the expense of the main contributions of long wavelength fluctuations of the bound proton and neutron . it implies that we can expect a cut off @xmath70 satisfying the inequality @xmath80 . it should be obvious that @xmath70 is to be identified with @xmath81 , i.e. @xmath82 . the infinitesimality of the derivatives @xmath83 necessary for the validity of the long wavelength expansion can be justified as follows . we assume that the deuteron field @xmath27 is a bound state of a proton and a neutron at zero binding energy . as has been mentioned above , this means that the field @xmath27 is localized in the region whose upper boundary @xmath84 goes to infinity . obviously it results in a smooth variation of the field @xmath27 at the scale of the effective radius of the physical deuteron . thus we have adduced some arguments on behalf of the validity of the long wavelength expansion which has been applied for the calculation of the effective lagrangian ( [ label4 ] ) . now we proceed to the computation of the binding energy of the physical deuteron . first , following [ 1 ] , we introduce the field of the physical deuteron @xmath85 where @xmath86 is the wave function normalization constant . after the renormalization ( 7 ) we get the effective lagrangian of the free physical deuteron field @xmath87 [ 1 ] @xmath88 where @xmath89 is the mass of the physical deuteron and @xmath67 is the binding energy . in the one nucleon loop approximation and in leading order of the long wavelength expansion the binding energy @xmath67 is determined by the expression @xmath90 for the derivation of eq . ( [ label9 ] ) we have used the inequality @xmath91 being consistent with the inequality @xmath92 discussed above . for @xmath92 eq . ( [ label9 ] ) reads @xmath93\,.\end{aligned}\ ] ] it is seen that the binding energy of the physical deuteron is expressed in terms of three phenomenological parameters of our model : @xmath94 and @xmath15 . then we have to fix the coupling constants @xmath14 and @xmath15 to obtain the theoretical value of the binding energy . ( [ label9 ] ) and ( [ label11 ] ) have been calculated in [ 1 ] . we need to reproduce them here for two reasons . first , we give below the new relation between the coupling constant @xmath14 and the electric quadrupole moment @xmath19 . second , in the conclusion we shall discuss eq . ( [ label11 ] ) setting @xmath95 . for these reasons it is convenient to have eqs . ( [ label9 ] ) and ( [ label11 ] ) at hand and not to relegate a reader to ref . [ 1 ] . in order to fix the coupling constants @xmath14 and @xmath15 we suggest to calculate the anomalous magnetic dipole @xmath16 and electric quadrupole @xmath19 moments in terms of @xmath14 and @xmath15 . we admit that @xmath16 and @xmath19 appear in the one nucleon loop approximation , and only at the expense of interactions taken into account in the lagrangian ( [ label2 ] ) [ 1 ] . the complete set of one nucleon loop diagrams is depicted in fig . 2 . the non trivial contributions come from the diagrams in figs . the detailed computation of the diagrams depicted in figs . 2c , d is given in sects . 2 and 3 , respectively . here we only adduce the results . the effective lagrangian , defined by the one nucleon loop diagrams in fig . 2 and describing the electromagnetic interactions of the physical deuteron field via the anomalous magnetic and electric quadrupole moments , is given by @xmath96-\lambda\bigg\}\frac{1 } { m^2_{\rm d}}\,d^{\dagger}_{\mu\nu}(x ) d^{\nu\alpha}(x ) { f_{\alpha}}^{\mu}(x).\nonumber\end{aligned}\ ] ] the lagrangian ( [ label12 ] ) contains only physical deuteron fields . the first term in @xmath97 is the corben schwinger interaction [ 14 ] , while the second term presents the interaction which has first been introduced by aronson [ 11 ] . these interactions describe the anomalous magnetic dipole and electric quadrupole moments of the charged vector field . it should be emphasized that we have neglected the divergent contributions which are small in comparison with the convergent ones due to the restriction ( [ label10 ] ) . note that the effective lagrangian ( [ label12 ] ) differs from that obtained in ref . [ 1 ] by the additional contributions caused by anomalous magnatic moments of the proton and neutron , and the phenomenological coupling @xmath66 induced by the aronson interaction . if one would set @xmath98 , one would find a discrepancy in the main terms.the former is due to different procedures of the removal of ambiguities of one nucleon loop diagrams , induced by shifts of virtual nucleon momenta , we use here and we have used in ref . [ 1 ] . in the present version all shift ambiguities are fully fixed by requirements of gauge invariance . the anomalous magnetic dipole moment @xmath16 measured in units of the nuclear magneton @xmath99 and the electric quadrupole moment @xmath19 are given by [ 10,13 ] @xmath100\,-\,\lambda\bigg\}\,,\end{aligned}\ ] ] @xmath101\,-\,\lambda\bigg\}\bigg\}\frac{1}{m^2_{\rm d}}\,.\end{aligned}\ ] ] the experimental value @xmath29 of the anomalous magnetic dipole moment of the deuteron is small compared with the magnetic dipole moment of the deuteron @xmath102 . therefore , we suggest to set @xmath103 giving @xmath104\,.\end{aligned}\ ] ] in this case the observed magnitude @xmath29 can be the matter of next to leading order corrections to the accepted approximation , for example , chiral corrections , etc . by applying eq . ( [ label15 ] ) we can express the electric quadrupole moment @xmath19 in terms of the coupling constant @xmath14 only , i.e. @xmath105 the relation between @xmath14 and @xmath19 , represented by eq . ( [ label16 ] ) , differs by a factor 2/3 from that obtained in ref . such a discrepancy has the following explanation . the relation eq . ( [ label16 ] ) is caused by effective couplings of the deuteron to electromagnetic field , i.e. the corben schwinger and aronson interactions . these interactions are defined by one nucleon loop diagrams the computation of which depends substantially on the shift of virual nucleon momenta ( see sects . 3 and 4 ) . the mathematical procedure of the removal of such ambiguities which has been used in ref.[1 ] , unfortunately , has left room for residual ambiguities . in the present version ambiguities induced by shifts of virtual nucleon momenta are fully fixed by requirements of gauge invariance under gauge transformations of the electromagnetic and the deuteron fields . the former has led to the factor 2/3 in the relation between @xmath19 and @xmath106 . by using the experimental value @xmath107 and @xmath108 we can estimate the value of @xmath14 . this gives : @xmath109 . without loss of generality we can use the positive sign , i.e. @xmath110 the coupling constant @xmath14 satisfies the relation @xmath111 , where @xmath112 [ 4 ] is the coupling constant of the @xmath26interactions . thus the magnitude of @xmath14 corroborates our assumption that the phenomenological interactions of the deuteron and the proton and neuteron , given in the lagrangian ( [ label1 ] ) , are caused by one meson exchange and the @xmath23meson exchange gives the main contribution . this admission can be justified by comparing the effective radii of one meson exchanges . the radii of pion , @xmath113meson [ 6,9 ] and @xmath114meson exchanges are defined in terms of their masses , i.e. @xmath115 at @xmath116 [ 13 ] , @xmath117 at @xmath118 [ 5,9 ] and @xmath119 at @xmath120 , respectively . the radii of the @xmath113 and @xmath114meson exchanges are much smaller than the radius of the @xmath23meson exchange . therefore , these interactions of the proton and neutron with @xmath113 and @xmath114 mesons should contribute perturbatively to the deuteron problem . we have put @xmath20 as the experimental value @xmath121 is small enough in comparison with the magnetic dipole moment of the deuteron @xmath102 . we assume that the nonzero value of @xmath16 can be obtained perturbatively by taking into account nonzero values of current quark masses , for example , within chiral perturbation theory at the quark level ( chpt)@xmath122 [ 9 ] , based on the extended nambu jona lasinio model with linear realization of chiral @xmath123 symmetry , and interactions of the proton and neutron with @xmath113 and @xmath114mesons within the one meson exchange approximation . from eqs . ( [ label11 ] ) and ( [ label16 ] ) we can express the binding energy @xmath67 in terms of the electric quadrupole moment @xmath19 and the ratio @xmath124 @xmath125\,.\end{aligned}\ ] ] this formula agrees with the statement that the main contribution to the binding energy of the deuteron comes from the tensor forces producing the non zero value of the quadrupole moment @xmath19 [ 3 ] . for the computation of the magnitude of the binding energy we have to fix the ratio @xmath126 that is still free . putting @xmath21 [ 1 ] we obtain the best fit of the binding energy @xmath127 the accuracy of the fit makes up 2@xmath128 . thus , we have fixed all phenomenological parameters and described all quantities , characterizing the physical deuteron . the total effective lagrangian of the physical deuteron describing strong and electromagnetic interactions of the deuteron reads @xmath129 d^{\dagger}_\mu(x ) d_\nu(x ) f^{\mu\nu}(x ) -\nonumber\\ & & -ie\bigg[\frac{g^2_{\rm v}}{2\pi^2}\bigg ] \,\frac{1 } { m^2_{\rm d}}\,d^{\dagger}_{\mu\nu}(x ) d^{\nu\alpha}(x ) { f_{\alpha}}^{\mu}(x ) -\nonumber\\ & & -ig_{\rm v } [ \bar{p}(x ) \gamma^\mu n^c(x)-\bar{n}(x ) \gamma^\mu p^c(x ) ] d_\mu ( x ) - \nonumber\\ & & -ig_{\rm v } [ \bar{p^c}(x ) \gamma^\mu n(x ) - \bar{n^c}(x ) \gamma^\mu p(x ) ] d^{\dagger}_\mu ( x ) + \\ & & + \frac{g_{\rm t}}{m_{\rm d } } [ \bar{p}(x ) \sigma^{\mu\nu } n^c(x ) - \bar{n}(x ) \sigma^{\mu\nu } p^c(x ) ] d_{\mu\nu}(x ) + \nonumber\\ & & + \frac{g_{\rm t}}{m_{\rm d } } [ \bar{p^c}(x ) \sigma^{\mu\nu } n(x)- \bar{n^c}(x ) \sigma^{\mu\nu } p(x ) ] d^{\dagger}_{\mu\nu}(x ) + \nonumber\\ & & + ie\frac{2g_{\rm t}}{m_{\rm d } } [ \bar{p}(x ) \sigma^{\mu\nu } n^c(x ) - \bar{n}(x ) \sigma^{\mu\nu } p^c(x ) ] a_\mu(x ) d_{\nu}(x ) - \nonumber\\ & & - ie\frac{2g_{\rm t}}{m_{\rm d } } [ \bar{p^c}(x ) \sigma^{\mu\nu } n(x)- \bar{n^c}(x ) \sigma^{\mu\nu } p(x ) ] a_\mu(x ) d^{\dagger}_{\nu}(x ) + o(e^2 ) + ... .\nonumber\end{aligned}\ ] ] the ellipses stand for interactions of the proton and neutron with other fields such as photon , pions , etc . the effective lagrangian ( [ label20 ] ) differs from that obtained in ref . [ 1 ] by the coupling constants of the effective interactions of the physical deuteron with the electromagnetic field describing the anomalous magnetic dipole @xmath16 and electric quadrupole moment @xmath19 . we have to underline that the suggested field theory model of the deuteron is applicable only at the low energy limit . thereby all interactions of the deuteron with other hadrons should run via one nucleon loop exchange . this is due to the one nucleon loop origin of the deuteron in our model . this assertion is very similar to the approximation accepted within the njl model , where all interactions of hadrons run via one constituent quark loop exchange [ 69 ] . however , if the one constituent quark loop approximation in the extended njl quark model [ 9 ] can be justfied in large n ( number of quark colours ) expansion , where the perturbation theory is developed in powers of @xmath130 . in the case of the relativistic field theory model of the deuteron for the justification of the one nucleon loop approximation we do not have any perturbative parameter . therefore , our approximation can not be justified like that in the extended njl quark model and should be considered as an effective one . for the confirmation of the validity of this effective approximation we can only refer to ref.[1 ] , where the two nucleon loop contributions to the binding energy have been calculated and found much less than the one nucleon loop contributions . we deem that most likely the deuteron can not be inserted in an intermediate state of any process of low energy interactions . this is connected with a very sensitive structure of the deuteron as an `` extended '' bound state with a small binding energy . the representation of such a state in terms of any local quantum field is rather limited . the latter entails an undetermined character of the description of the deuteron in intermediate states in terms of green functions of these local fields . now we can proceed to the computation of the effective interactions describing the anomalous magnetic and electric quadrupole moments of the deuteron . the complete set of one nucleon loop diagrams describing effective interaction of the deuteron with the electromagnetic field is depicted in fig . 2 . in leading order in the long wavelength expansion the diagrams in figs . 2a , b do not contribute to the anomalous magnetic and electric quadrupole moments . the contributions of these diagrams are divergent and can be removed by the renormalization of the electric charge of the deuteron . the nontrivial contributions to the effective anomalous magnetic and electric quadrupole moments are defined only by the diagrams in figs . in this section we give the detailed derivation of the effective corben schwinger lagrangian defined by the one nucleon loop diagram in fig . 2c and describing the effective electromagnetic interactions of the deuteron . the effective lagrangian of the diagram in fig . 2c is defined @xmath131 where @xmath132 the 4vector @xmath133 , where @xmath134 and @xmath135 are arbitrary constants , displays the dependence of the @xmath136 integral in ( [ label21 ] ) on the shift of the virtual momentum . this ambiguity of the computation of the integral over @xmath136 , which has been found by gertsein and jackiw [ 15 ] , is used to remove undesirable contributions and make the effective lagrangian gauge invariant . we use the fields of the physical deuteron . this is because the renormalization ( [ label6 ] ) introduces divergent terms that are small compared with the convergent ones we are following for the computation of the corben schwinger lagrangian . in order to display the gertsein jackiw ambiguity we follow the gertsein jackiw method and compute the difference @xmath137 in accordance with the gertsein jackiw method the difference ( [ label23 ] ) can be represented by the integral @xmath138 this shows that the gertsein jackiw ambiguity is just the surface term . following gertsein and jackiw [ 15 ] and evaluating the integral over @xmath136 symmetrically , we obtain @xmath139 the brackets @xmath140 denote the averaging over directions of the 4-vector @xmath141 . due to the limit @xmath142 we can neglect all momenta with respect to @xmath141 . @xmath143 averaging over directions of the 4vector @xmath141 @xmath144 we obtain @xmath145 thus the surface ambiguity noticed by gertsein and jackiw contains only finite contributions . now we should proceed to the computation of @xmath146 . in order to pick up the ambiguity connected with @xmath147 one can not apply the feynman method of the computation of momentum integrals like ( [ label22 ] ) . this method involves the merger of the factors in the denominator with the subsequent shift of virtual momentum . on this way one can lose the @xmath147depenedence that is due to the shift at the intermediate stage . thereby we have to compute the integral over @xmath136 without any intermediate shift . one can carry out this by applying a long wavelength expansion and keeping to the leading terms . @xmath148\ , \gamma^{\alpha}\,\frac{m_{\rm n}+\hat{k}+\hat{q}+\hat{k}_1}{m^2_{\rm n}-k^{\,2}}\,\times \nonumber\\ & & \times\,\bigg[1\,+\,\frac{2\,k\cdot ( q+k_1)}{m^2_{\rm n}-k^{\,2}}\bigg]\,\gamma^{\mu}\,\frac{m_{\rm n}+\hat{k}+\hat{q}+\hat{k}_1+\hat{k}_2}{m^2_{\rm n}-k^{\,2}}\bigg[1\,+\,\frac{2\,k\cdot ( q+k_1+k_2)}{m^2_{\rm n}-k^{\,2}}\bigg]\bigg\}\,=\nonumber\\ & & = \,\int\frac{d^4k}{\pi^2i}\,\frac{1}{(m^2_{\rm n}-k^{\,2})^{\,3}}{\rm tr}\,\{m^2_{\rm n}\gamma^{\beta}(\hat{k}+\hat{q})\gamma^{\alpha}\gamma^{\mu}+m^2_{\rm n}\gamma^{\beta}\gamma^{\alpha}(\hat{k}+\hat{q}+\hat{k}_1)\gamma^{\mu}+\nonumber\\ & & + m^2_{\rm n}\gamma^{\beta}\gamma^{\alpha}\gamma^{\mu}(\hat{k}+\hat{q}+\hat{k}_1+\hat{k}_2)+\gamma^{\beta}(\hat{k}+\hat{q})\gamma^{\alpha}(\hat{k}+\hat{q}+\hat{k}_1)\gamma^{\mu}(\hat{k}+\hat{q}+\hat{k}_1+\hat{k}_2)\}\nonumber\\ & & \bigg[1\,+\,\frac{2\,k\cdot ( 3q+2k_1+k_2)}{m^2_{\rm n}-k^{\,2}}\bigg]\,=\\ & & = \,\frac{1}{2}\int\frac{d^4k}{\pi^2i}\bigg[\frac{1}{(m^2_{\rm n}-k^2)^2}+\frac{m^2_{\rm n}}{(m^2_{\rm n}-k^2)^3}\bigg]{\rm tr}\,\{\gamma^{\beta}\hat{q}\gamma^{\alpha}\gamma^{\mu}+\gamma^{\beta}\gamma^{\alpha}(\hat{q}+\hat{k}_1)\gamma^{\mu}+\nonumber\\ & & + \gamma^{\beta}\gamma^{\alpha}\gamma^{\mu}(\hat{q}+\hat{k}_1+\hat{k}_2)\}\,+\nonumber\\ & & + \,2\,\int\frac{d^4k}{\pi^2i}\frac{k\cdot ( 3q+2k_1+k_2)}{(m^2_{\rm n}-k^2)^4}{\rm tr}\,\{m^2_{\rm n}(\gamma^{\beta}\hat{k}\gamma^{\alpha}\gamma^{\mu}+\gamma^{\beta}\gamma^{\alpha}\hat{k}\gamma^{\mu}+\gamma^{\beta}\gamma^{\alpha}\gamma^{\mu}\hat{k})+\nonumber\\ & & + \gamma^{\beta}\hat{k}\gamma^{\alpha}\hat{k}\gamma^{\mu}\hat{k}\}\,=\,{\cal j}^{\beta\alpha\mu}_{(1)}(k_1 , k_2 ; q)\,+\,{\cal j}^{\beta\alpha\mu}_{(2)}(k_1 , k_2 ; q).\nonumber\end{aligned}\ ] ] for the computation of @xmath149 it is sufficient to calculate the trace of the dirac matrices and integrate over @xmath136 @xmath150\ , [ ( q\,+\,2\,k_1\,+\,k_2)^{\alpha}g^{\beta\mu}\,+\nonumber\\ & & + \,(q\,+\,2\,k_1\,+\,k_2)^{\beta}g^{\mu\alpha}\,+\,(q\,+\,2\,k_1\,+\,k_2)^{\mu}g^{\alpha\beta}\,-\\ & & -\,2\,(k_1\,+\,k_2)^{\alpha } g^{\beta\mu}\,-\,2\,k^{\beta}_1 g^{\mu\alpha}]\,,\nonumber\end{aligned}\ ] ] where @xmath69 describes a divergent contribution depending on the cut off @xmath5 . due to the inequality @xmath151 we can neglect @xmath69 with respect to the convergent contribution . in order to compute @xmath152 we have to integrate first over @xmath136 directions . this gives @xmath153\,(3\,q\,+\,k_1\,+\,k_2)_{\,\lambda}\times \nonumber\\ & & \quad\quad\times { \rm tr}\,(\gamma^{\beta}\gamma^{\lambda}\gamma^{\alpha}\gamma^{\mu}+\gamma^{\beta}\gamma^{\alpha}\gamma^{\lambda}\gamma^{\mu}+\gamma^{\beta}\gamma^{\alpha}\gamma^{\mu}\gamma^{\lambda})\,=\\ & & = \,-\,\frac{1}{9}[1\,+\,6\,j_2(m_{\rm n})]\,[(3\,q\,+\,2\,k_1\,+\,k_2)^{\alpha}g^{\beta\mu}\,+\,(3\,q\,+\,2\,k_1\,+\,k_2)^{\beta } g^{\mu\alpha}\,+\nonumber\\ & & \quad\,+\,(3\,q\,+\,2\,k_1\,+\,k_2)^{\mu } g^{\alpha\beta}].\nonumber\end{aligned}\ ] ] here we have used the integrals summarizing the contributions , we get @xmath154\times \nonumber\\ & & \times [ ( 2\,k_1\,+\,k_2)^{\alpha}g^{\beta\mu}\,+\,(2\,k_1\,+\,k_2)^{\beta}\,g^{\mu\alpha}\,+\,(2\,k_1\,+\,k_2)^{\mu } g^{\alpha\beta}]\,+\\ & & + \,[1\,+\,2\,j_2(m_{\rm n})][-\,2\,(k_1\,+\,k_2)^{\alpha}g^{\beta\mu}\,-\,2\,k_1^{\beta}g^{\mu\alpha}].\nonumber\end{aligned}\ ] ] it is seen that the @xmath147dependence coincides with that obtained by the gertsein jackiw method ( [ label28 ] ) . due to the arbitrariness of @xmath147 we can arbsorb by the @xmath147term the terms having the same lorentz structure . this brings up the r.h.s . of ( [ label33 ] ) to the form @xmath155\,.\end{aligned}\ ] ] also we have neglected the divergent contribution . this approximation is valid due to the inequality @xmath151 . the effective lagrangian @xmath156 defined by ( [ label34 ] ) reads @xmath157.\nonumber\end{aligned}\ ] ] now we can consider @xmath134 and @xmath135 as free parameters that can be fixed from the requirement of the gauge invariance of the effective lagrangian described by the one nucleon loop diagram in fig . 2c . due to the constraints @xmath158 the corresponding terms in the lagrangian ( [ label34 ] ) can be dropped @xmath159.\nonumber\end{aligned}\ ] ] the subsequent transformations are performed by applying the identity @xmath160 where we have dropped the total divergence and the terms proportional to @xmath161 and @xmath162 . then it is convenient to rewrite the lagrangian ( [ label36 ] ) as follows @xmath163.\nonumber\end{aligned}\ ] ] putting @xmath164 we bring up the effective lagrangian ( [ label38 ] ) to the irreducible form that contains two parts defining the renormalization of the electric charge of the deuteron and the gauge invariant interaction coinciding with that given by corben and schwinger [ 14 ] @xmath165.\end{aligned}\ ] ] now the effective lagrangian @xmath166 is represented in the irreducible form , and we can proceed to the analysis of gauge invariance . putting @xmath167 we remove the finite contributions to the renormalization constant of the electic charge of the deuteron coming from the one nucleon loop diagram in fig . 2c and get a gauge invariant interaction . as a result the effective gauge invariant interaction reads @xmath168 thus the one nucleon loop diagram in fig . 2c defines the effective corben schwinger lagrangian @xmath169 describing the interaction of the deuteron with the electromagnetic field . we should underscore that the ambiguities connected with the shift of the virtual momentum of the diagram in fig . 2c are the intrinsic peculiarity of fermion loop diagrams , caused by the computation of such diagrams within the cut off regularization so , one can not consider the parameters @xmath134 and @xmath135 connected to a shift of virtual momentum @xmath170 as parameters especially introduced in the model . the requirement of gauge invariance defines unambiguously the parameters @xmath134 and @xmath135 and , correspondingly , the effective lagrangian @xmath169 . the choice @xmath171 is unique , since it brings up the lagrangian ( [ label38 ] ) to the irreducible form . the next choice @xmath172 is also unique , for the effective lagrangian ( [ label39 ] ) is given by the irreducible form . one can show that the contributions of the anomalous magnetic dipole moments of the proton and neutron to the effective corben schwinger lagrangian are proportional to @xmath69 and due to inequality @xmath92 are small compared with that given by eq . ( [ label40 ] ) . in this section we give the detailed derivation of the effective aronson lagrangian defined by the one nucleon loop diagram in fig . 2d and describing the effective electromagnetic interactions of the deuteron . the effective lagrangian described by the diagram in fig . 2d is given by @xmath173 and @xmath174 where the 4vector @xmath175 is due to the shift of the virtual momentum of the nucleons in the diagram in fig . 2d . the parameters @xmath134 and @xmath135 are free . they are not connected with those having been intriduced for the computation of the diagram in fig . 2c . in the r.h.s . of eq.([label41 ] ) we have used the fields of the physical deuteron , for the renormalization ( [ label6 ] ) introduces divergent terms that due to inequality @xmath92 are small compared with the convergent ones we are keeping for the computation of the aronson lagrangian . the gertsein jackiw ambiguity is given by @xmath176 now we should proceed to the computation of @xmath177 . by analogy with ( [ label22 ] ) we get @xmath178\ , \sigma^{\mu\nu}\frac{m_{\rm n}+\hat{k}+\hat{q}+\hat{k}_1}{m^2_{\rm n}-k^2}\,\times \nonumber\\ & & \times\,\bigg[1\,+\,\frac{2\,k\cdot ( q+k_1)}{m^2_{\rm n}-k^2}\bigg]\,\gamma^{\lambda}\frac{m_{\rm n}+\hat{k}+\hat{q}+\hat{k}_1+\hat{k}_2}{m^2_{\rm n}-k^2}\bigg[1\,+\,\frac{2\,k\cdot ( q+k_1+k_2)}{m^2_{\rm n}-k^2}\bigg]\bigg\}\,=\nonumber\\ & & = \,\int\frac{d^4k}{\pi^2i}\,\frac{1}{(m^2_{\rm n}-k^2)^3}{\rm tr}\,\{m^2_{\rm n}[\sigma^{\alpha\,\beta}(\hat{k}+\hat{q})\sigma^{\mu\nu}\gamma^{\lambda}+\sigma^{\alpha\beta}\sigma^{\mu\nu}(\hat{k}+\hat{q}+\hat{k}_1)\gamma^{\lambda}+\nonumber\\ & & + \sigma^{\alpha\beta}\sigma^{\mu\nu}\gamma^{\lambda}(\hat{k}+\hat{q}+\hat{k}_1+\hat{k}_2)]+\sigma^{\alpha\beta}(\hat{k}+\hat{q})\sigma^{\mu\nu}(\hat{k}+\hat{q}+\hat{k}_1)\gamma^{\lambda}\times \nonumber\\ & & \times ( \hat{k}+\hat{q}+\hat{k}_1+\hat{k}_2)\}\ , \bigg[1\,+\,\frac{2\,k\cdot ( 3q+2k_1+k_2)}{m^2_{\rm n}-k^2}\bigg]\,=\\ & & = \,\int\frac{d^4k}{\pi^2\,i}\frac{1}{(m^2_{\rm n}-k^2)^3}{\rm tr}\,\{m^2_{\rm n}[\sigma^{\alpha\beta}\hat{q}\sigma^{\mu\nu}\gamma^{\lambda}+\sigma^{\alpha\beta}\sigma^{\mu\nu}(\hat{q}+\hat{k}_1)\gamma^{\lambda}+\nonumber\\ & & + \sigma^{\alpha\beta}\sigma^{\mu\nu}\gamma^{\lambda}(\hat{q}+\hat{k}_1+\hat{k}_2)]-\frac{1}{2}k^2 \sigma^{\alpha\beta}\hat{q}\sigma^{\mu\nu}\gamma^{\lambda}\}\,+\nonumber\\ & & + \,2\,\int\frac{d^4k}{\pi^2i}\frac{1}{(m^2_{\rm n}-k^2)^4}{\rm tr}\,\{\frac{1}{2}m^2_{\rm n}k^2 [ \sigma^{\alpha\beta}(3\,\hat{q}+2\,\hat{k}_1+\hat{k}_2)\sigma^{\mu\nu}\gamma^{\lambda}+\nonumber\\ & & + \sigma^{\alpha\beta}\sigma^{\mu\nu}(3\,\hat{q}+2\,\hat{k}_1+\hat{k}_2)\gamma^{\lambda}+\sigma^{\alpha\beta}\sigma^{\mu\nu}\gamma^{\lambda}(3\,\hat{q}+2\,\hat{k}_1+\hat{k}_2)]-\nonumber\\ & & -\frac{1}{6}\,k^4\sigma^{\alpha\beta}(3\,\hat{q}+2\,\hat{k}_1+\hat{k}_2)\sigma^{\mu\nu}\gamma^{\lambda}\}\,=\,{\cal j}^{\alpha\beta\mu\nu\lambda}_{(1)}(k_1 , k_2 ; q)\,+\,{\cal j}^{\alpha\beta\mu\nu\lambda}_{(2)}(k_1 , k_2 ; q)\,.\nonumber\end{aligned}\ ] ] integrating over @xmath136 we obtain @xmath179\,{\rm tr}(\sigma^{\alpha\beta}\hat{q}\sigma^{\mu\nu}\gamma^{\lambda})\,+\nonumber\\ & & + \frac{1}{2}\,{\rm tr}\,[\sigma^{\alpha\beta}\sigma^{\mu\nu}(\hat{q}+\hat{k}_1)\gamma^{\lambda}+\sigma^{\alpha\beta}\sigma^{\mu\nu}\gamma^{\lambda}(\hat{q}+\hat{k}_1+\hat{k}_2)]\,,\end{aligned}\ ] ] @xmath180\,{\rm tr}[\sigma^{\alpha\beta}(3\hat{q}+2\hat{k}_1+\hat{k}_2)\sigma^{\mu\nu}\gamma^{\lambda}]\,-\nonumber\\ & & -\frac{1}{6}\,{\rm tr}\,[\sigma^{\alpha\beta}(3\hat{q}+2\hat{k}_1+\hat{k}_2)\sigma^{\mu\nu}\gamma^{\lambda}+\\ & & + \sigma^{\alpha\beta}\sigma^{\mu\nu}(3\hat{q}+2\hat{k}_1+\hat{k}_2)\gamma^{\lambda}+\sigma^{\alpha\beta}\sigma^{\mu\nu}\gamma^{\lambda}(3\hat{q}+2\hat{k}_1+\hat{k}_2)].\nonumber\end{aligned}\ ] ] now we should summarize the contributions and collect like terms @xmath181\,+\\ & + & \frac{1}{6}\,{\rm tr}\ , [ \sigma^{\alpha\beta}\sigma^{\mu\nu}(\hat{k}_1-\hat{k}_2)\gamma^{\lambda}]\,+\,\frac{1}{6}\,{\rm tr}\,[\sigma^{\alpha\beta}\sigma^{\mu\nu}\gamma^{\lambda}(\hat{k}_1\,+\,2\,\hat{k}_2)]\,.\nonumber\end{aligned}\ ] ] it is seen that the @xmath147dependence coincides with that obtained by the gertsein jackiw method . due to the arbitrariness of @xmath147 the vector @xmath182 can removed by the redefinition of @xmath147 . thereby we get @xmath183\,+\,\frac{1}{6}\,{\rm tr}\ , [ \sigma^{\alpha\beta}\sigma^{\mu\nu}\gamma^{\lambda}(\hat{k}_1\,+\,2\,\hat{k}_2)].\nonumber\end{aligned}\ ] ] after the calculation of the traces of the dirac matrices we obtain @xmath177 leading to the following effective lagrangian @xmath184.\nonumber\end{aligned}\ ] ] for the derivation of the effective lagrangian ( [ label49 ] ) we have used the equation of motion @xmath185 the analogous equation of motion is valid for the conjugated field . by collecting like terms in ( [ label49 ] ) we get @xmath186.\nonumber\end{aligned}\ ] ] the third and the fourth terms can be reduced by applying the equation of motion @xmath187 and analoguous for the conjugated field . then putting @xmath188 , we obtain @xmath189 , that brings up the effective lagrangian ( [ label51 ] ) to the following irreducible form @xmath190\,+\\ & + & i\,e\,\frac{2\,g^2_{\rm t}}{3\,\pi^2}\,\frac{1}{m^2_{\rm d}}\,(1\,+\,a)\,d^{\dagger}_{\mu\nu}(x)\,d^{\nu\lambda}(x)\,{f_{\,\lambda}}^{\mu}(x)\,.\nonumber\end{aligned}\ ] ] putting @xmath191 we remove the finite contributions to the renormalization constant of the electric charge of the deuteron and gain the gauge invariant interaction @xmath192 due to the irreducibility of the lagrangian ( [ label53 ] ) the choice @xmath193 is unique . the effective lagrangian ( [ label54 ] ) coincides fully with that suggested by aronson [ 10 ] . this means that the one - nucleon loop diagram in fig . 2d defines unambiguously the effective lagrangian if one imposes the requirement of the gauge invariance . the same result can be gained if one requires the vanishing of finite contributions to the renormalization constant of the electric charge of the deuteron . by analogy with the computation of the lagrangian ( [ label54 ] ) one can obtain the contribution of the anomalous magnetic dipole moments of the proton and neutron to the effective aronson interaction @xmath194 as a result the complete expression of the aronson effective lagrangian is given by @xmath195\,\frac{1}{m^2_{\rm d}}d^{\dagger}_{\mu\nu}(x)d^{\nu\lambda}(x){f_{\lambda}}^{\mu}(x).\end{aligned}\ ] ] thus we have shown that in the relativistic field theory model of the deuteron and in one nucleon loop approximation one can unambiguously compute effective electromagnetic interactions of the deuteron in terms of anomalous magnetic and electric quadrupole moments . this can be obtained under requirement of gauge invariance to every one nucleon loop diagram separately . at low energies , the process of the radiative neutron proton capture n + p @xmath0 d + @xmath1 comes off via electric and magnetic dipole transitions . in the usual notations @xmath196 , the deuteron is a @xmath197 state , and the possible transitions are @xmath198 where @xmath199 magnetic dipole and @xmath200 electric dipole . if the energies are low enough , the nucleons are in the s wave state and thereby only magnetic dipole transitions are possible . in the s wave state the magnetic dipole moment operator acts only on spin variables . this implies that the transition @xmath201 is forbidden [ 16 ] . thereby the only allowed transition is @xmath202 . this means that the anomalous magnetic dipole moments of the proton and neutron should give the main contribution [ 16 ] . in the relativistic field theory model of the deuteron the interactions of the deuteron with other fields should be obtained through one nucleon loop diagrams . therefore , for the computation of the effective lagrangian of the radiative neutron proton capture we have first to compute the effective lagrangian , describing low energy neutron proton scattering . keeping to the one pion exchange we obtain [ 1 ] @xmath203[\bar{n^c}(x ) p(x)]+[\bar{p}(x)\gamma^5 n^c(x)][\bar{n^c}(x)\gamma^5 p(x)]+\nonumber\\ & & + [ \bar{p}(x)\gamma_{\mu}\gamma^5 n^c(x)][\bar{n^c}(x)\gamma^{\,\mu}\gamma^5 p(x)]+3[\bar{p}(x)\gamma_{\,\mu}n^c(x)][\bar{n^c}(x)\gamma^{,\mu}p(x)]+\\ & & + \frac{3}{2}[\bar{p}(x ) \sigma_{\mu\nu } n^c(x)][\bar{n^c}(x)\sigma^{\mu\nu}p(x)]\,\big\}.\nonumber\end{aligned}\ ] ] only terms @xmath204[\bar{n^c}(x)\gamma^5 p(x)]$ ] and @xmath205[\bar{n^c}(x)\gamma^{\mu}\gamma^5 p(x)]$ ] contribute to the s wave of the neutron proton scattering in the low energy limit . therefore , these terms should effect the @xmath206 transition in the radiative neutron proton capture at low energies . the corresponding one nucleon loop diagrams are depicted in figs . 3 and 4 . first let us consider the contribution of the diagram in fig . 3a . the corresponding lagrangian reads @xmath207\,d^{\dagger}_{\mu}(x_1)\,a_{\nu}(x_2)\,\times\\ & & \times \ , e^{-\,i\,k_1\cdot x_1}\,e^{-\,i\,k_2\cdot x_2}\,e^{\,i\,(k_1\,+\,k_2)\cdot x}\,i\,e\,\frac{g^2_{\pi\rm nn}}{m^2_{\pi}}\,\frac{g_{\rm v}}{32\pi^2}\,{\cal j}^{\mu\nu}_{\,5}(k_1 , k_2 ; q)\,,\nonumber\end{aligned}\ ] ] where @xmath208 and @xmath133 is an arbitrary shift of virtual momentum . fortunately , the integral @xmath209 does not depend on the shift of the virtual momentum and can be computed unambiguously @xmath210 where we have used the relation @xmath211 . the structure function ( [ label60 ] ) leads to the effective lagrangian @xmath212\,,\end{aligned}\ ] ] where @xmath213 . now let us consider the contribution of the diagram in fig . this contribution is caused by the anomalous mangnetic moment of the proton . the effective lagranigan is defined @xmath214\,d^{\dagger}_{\mu}(x_1)\,f_{\alpha\beta}(x_2)\,\times\\ & & \times \ , e^{-\,i\,k_1\cdot x_1}\,e^{-\,i\,k_2\cdot x_2}\,e^{\,i\,(k_1\,+\,k_2)\cdot x}\,e\,\frac{\kappa_{\,\rm p}}{4m_{\rm n}}\frac{g^2_{\pi\rm nn}}{m^2_{\pi}}\,\frac{g_{\rm v}}{32\pi^2}\,{\cal j}^{\mu\alpha\beta}_{\,5}(k_1 , k_2 ; q)\,,\nonumber\end{aligned}\ ] ] where @xmath215 unfortunately , the integral @xmath216 depends on the shift of the virtual momentum . therefore , the leading contribution in the momentum expansion is fully arbitrary and is given @xmath217 where @xmath134 and @xmath135 are arbitrary parameters . the contribution of the momentum @xmath218 produces in the lagrangian the operator @xmath219 that vanishes due to maxwell@xmath220s equation of motion , i.e. @xmath221 . thus only the contribution of the momentum @xmath222 matters . the effective lagrangian produced by @xmath223 given by ( [ label64 ] ) reads @xmath224\,.\end{aligned}\ ] ] the effective lagrangian corresponding to the diagrams in figs . 3a is given by @xmath225\,.\end{aligned}\ ] ] now let us discuss how we can fix the parameter @xmath134 . as has been mentioned above the radiative capture n + p @xmath0 d + @xmath1 at low energies is a magnetic dipole transition @xmath226 . this implies that the contribution of the proton and neutron to amplitude of the radiative capture n + p @xmath0 d + @xmath1 should be proportional to their magnetic dipole moments . in the case of the proton the total magnetic dipole moment equals @xmath227 , i.e. @xmath228 , whereas the total magnetic dipole moment of the neutron is just its anomalous magnetic dipole moment , i.e. @xmath229 . the effective lagrangian ( [ label66 ] ) describes the contribution of the proton , therefore the quantity @xmath230 should be nothing more than the total magnetic dipole moment of the proton @xmath231 . this fixes the arbitrariness of the contribution of the diagram in fig . thus , we have to put @xmath232 . as a result the complete contribution of the diagrams in figs . 3a , b reads @xmath233\,.\end{aligned}\ ] ] thus we have fixed the ambiguity introduced by the diagram in fig . 3b by applying the selection rule @xmath226 . by extending the procedure of the computation of the diagrams in figs . 3a , b on other diagrams in fig . 3 we get the following complete effective lagrangian @xmath234\,.\end{aligned}\ ] ] the contributions of the diagrams describing the interaction of the deuteron with the anomalous magnetic dipole moments of the proton and neutron via the tensor nucleon current , i.e. proportional to the constant @xmath15 , are divergent and due to the inequality @xmath92 are small compared with the convergent ones . thereby the contributions proportional to the coupling constant @xmath15 can be neglected and do not appear in the amplitude of the radiative neutron proton capture n + p @xmath0 d + @xmath1 . the contribution of the @xmath235[\bar{n^c}(x)\gamma^{\mu}\gamma^5 p(x)]$ ] interaction can be computed by analogy to that given above . as a result we obtain @xmath236.\end{aligned}\ ] ] the total effective low energy lagrangian , describing the radiative neutron proton capture , is defined by the sum of the effective lagrangians ( [ label68 ] ) and ( [ label69 ] ) and reads the amplitude of the radiative neutron proton capture is given by where @xmath237 and @xmath238 are the polarization 4vectors of the photon and the deuteron , respectively , then @xmath239 and @xmath240 are the bispinorial wave functions of the neutron and proton normalized by @xmath241 and @xmath242 . the cross section @xmath243 of the radiative neutron proton capture has been measured for thermal neutrons at laboratory velocities @xmath244 ( the absolute value is @xmath245 ) [ 17 ] @xmath246 the cross section @xmath243 is given by where @xmath247 and @xmath248 are the energies of the neutron , proton , photon and the deuteron , respectively , @xmath249 is the laboratory velocity of the neutron , and @xmath250 is the squared amplitude averaged over polarizations of the neutron and proton and summed over polarizations of the deuteron and photon @xmath251 ^ 2\,\frac{g^2_{\rm v}}{256\pi^4}\,\varepsilon^{\alpha\beta\mu\nu}\,\varepsilon_{\lambda\beta\rho\omega}\,k_{\alpha}k^{\lambda}\bigg({g_{\mu}}^{\rho } - \frac{q_{\mu}q^{\rho}}{m^2_{\rm d}}\bigg)\,\times \frac{1}{4 } \times\nonumber\\ & & \times{\rm tr}\big[(\hat{p}_1 - m_{\rm n})(2q_{\nu}-m_{\rm n}\gamma_{\nu})\gamma^5 ( \hat{p}_2 + m_{\rm n})(-2q^{\omega } - m_{\rm n } \gamma^{\omega})\gamma^5\big]=\\ & & = ( \mu_{\rm p}-\mu_{\rm n})^2\,\frac{\alpha}{m^2_{\rm n}}\bigg[\frac{g^2_{\pi\rm nn}}{m^2_{\pi}}\bigg]^2\,\frac{g^2_{\rm v}}{256\pi^3}\,\varepsilon^{\alpha\beta\mu\nu}\,\varepsilon_{\lambda\beta\rho\omega}\,k_{\alpha}k^{\lambda}\bigg({g_{\mu}}^{\rho } - \frac{q_{\mu}q^{\rho}}{m^2_{\rm d}}\bigg)\,\times\nonumber\\ & & \times[4(p_1\cdot p_2 + 2m^2_{\rm n})\,q_{\nu}q^{\omega } + m^2_{\rm n}(p_{1\,\nu}p^{\omega}_2 + p^{\omega}_1 p_{2\,\nu})-(p_1\cdot p_2 - m^2_{\rm n})\,g^{\omega}_{\nu}]\,,\nonumber\end{aligned}\ ] ] where @xmath252 is the fine structure constant . in the low energy limit when the 3momenta of the neutron and proton tend to zero we obtain @xmath253 ^ 2\,\varepsilon^{\alpha\beta\mu\nu}\,\varepsilon_{\lambda\beta\rho\omega}\,k_{\alpha}k^{\lambda}\times\nonumber\\ & & \times[4(p_1\cdot p_2 + 2m^2_{\rm n})\,q_{\nu}q^{\omega } + m^2_{\rm n}(p_{1\,\nu}p^{\omega}_2 + p^{\omega}_1 p_{2\,\nu})]\,=\\ & & = ( \mu_{\rm p}-\mu_{\rm n})^2\,\frac{25\alpha\,q_{\rm d}}{32\pi}\bigg[\frac{g^2_{\pi\rm nn}}{m^2_{\pi}}\bigg]^2\,m^4_{\rm n}\varepsilon^2_{\rm d}\,,\nonumber\end{aligned}\ ] ] where we have used the relation ( [ label16 ] ) . thus the cross section of the radiative neutron proton capture reads @xmath254 ^ 2\,m_{\rm n}\varepsilon^3_{\rm d}\,.\end{aligned}\ ] ] putting @xmath255 we get the following theoretical value of @xmath243 @xmath256 it is seen that this theoretical value of the cross section disagrees with the experimental data . however , it should be stressed that we do not have taken into account the resonance contibution to the @xmath257 low energy neutron proton scattering [ 1820 ] . the account of the resonance contribution gives the amplitude of the radiative neutron proton capture in the form [ 21 ] where @xmath258 [ 4 ] is the @xmath257 neutron proton scattering length that is fully due to the resonance contribution [ 1820 ] . as a result the cross section of the radiative neutron proton capture is given by the theoretical value agrees now with the experimental data within an accuracy better than 8@xmath128 . the cross section ( [ label79 ] ) predicted in the relativistic field theory model of the deuteron agrees also well with that given by the potential model [ 22 ] . the process of the two proton fusion p + p @xmath0 d + @xmath259 + @xmath3 plays an important role for the synthesis of deuterons in stars , where the deuterons are destroyed again by the reaction p + d @xmath0 @xmath260 + @xmath1 [ 23 ] . in nuclear physics the process p + p @xmath0 d + @xmath259 + @xmath3 is a gamow teller transition governed by the weak axial vector nucleon current [ 24 ] . in the relativistic field theory model of the deuteron the low energy process p + p @xmath0 d + @xmath259 + @xmath3 is closely connected with low energy proton proton scattering , i.e. p + p @xmath0 p + p. low energy proton proton scattering differs very much from the low energy neutron proton scattering . this is mostly due to the strong contribution of the coulomb repulsion [ 25 ] . the small relative velocities @xmath249 in the proton proton system are suppressed by the factor @xmath261 , where @xmath252 is the fine structure constant . the factor @xmath261 is the gamow penetration factor [ 25,26 ] . for the acquaintance with the earliest computation of the cross section of the p + p @xmath0 d + @xmath259 + @xmath3 scattering we refer readers to the paper by bethe and critchfield [ 27 ] and the book by rosenfeld [ 28 ] . for the computation of the effective lagrangian describing the low energy proton proton scattering we assume the one pion exchange approximation @xmath262\,[\bar{p}(x)\gamma^5 p(x)]\,.\end{aligned}\ ] ] here we have denoted @xmath263 and @xmath264 takes into account the coulomb repulsion between protons at low energies [ 27,29 ] . applying the fierz transformation we bring up the lagrangian ( [ label80 ] ) to the form @xmath265[\bar{p^c}(x)p(x)]+[\bar{p}(x)\gamma^5 p^c ( x)][\bar{p^c}(x)\gamma^5 p(x)]+\nonumber\\ & & + [ \bar{p}(x)\gamma_{\mu}\gamma^5 p^c(x)][\bar{p^c}(x)\gamma^{\mu}\gamma^5 p(x)]+\frac{1}{2}[\bar{p}(x)\sigma_{\mu\nu}p^c(x)][\bar{p^c}(x)\sigma^{\mu\nu } p(x)]\,\big\}.\end{aligned}\ ] ] the process p + p @xmath0 d + @xmath259 + @xmath3 should run via the intermediate w boson exchange , i.e. p + p @xmath0 d + w@xmath266 @xmath0 d + @xmath259 + @xmath3 . the lagrangian describing the electroweak interactions of the w boson with proton , neutron , positron and neutrino reads [ 2,19 ] @xmath267\,w_{\mu}(x ) + { \rm h.c.}\;.\end{aligned}\ ] ] here @xmath268 is the electroweak coupling constant connected with the fermi constant @xmath269 and the w boson mass @xmath270 by the relation [ 2,19 ] @xmath271 where @xmath272 is the axial vector coupling constant [ 4 ] describing the renormalization of the weak axial vector hadron current by strong interactions . only the interaction @xmath273[\bar{p^c}(x)\gamma^{\mu}\gamma^5 p(x)]$ ] gives in our approach the main contribution to the amplitude of the transition p + p @xmath0 d + w. the corresponding one nucleon loop diagrams are depicted in fig . 5 . the effective lagrangian defined by the diagrams in fig . 5 is given by @xmath274\,d^{\dagger}_{\mu}(x_1)\,w^{\dagger}_{\nu}(x_2)\,\times\\ & & \times \ , e^{-\,i\,k_1\cdot x_1}\,e^{-\,i\,k_2\cdot x_2}\,e^{\,i\,(k_1\,+\,k_2)\cdot x}\,i\,g_{\rm a}\,\frac{g_{\rm w}}{2\sqrt{2}}\,\frac{g^2_{\pi\rm nn}(v)}{m^{\,2}_{\pi}}\,\frac{g_{\rm v}}{64\pi^2}\,\bar{{\cal j}}^{\alpha\mu\nu}(k_1 , k_2 ; q),\nonumber\end{aligned}\ ] ] where @xmath275 the computation of this integral can be reduced to the computation of the integral ( [ label22 ] ) defining the corben schwinger lagrangian , i.e. @xmath276 where we have denoted @xmath277 the integral @xmath278 does not depend on the shift of virtual momentum , i.e. on @xmath147 , and in leading order in the long wavelength expansion is given by @xmath279\times\nonumber\\ & & { \rm tr}\{\gamma^{\alpha}(m_{\rm n } + \hat{k } + \hat{q})\gamma^{\mu}(m_{\rm n } + \hat{k } + \hat{q } + \hat{k}_1 ) \gamma^{\,\nu}\}=\nonumber\\ & & = -\int\,\frac{d^4k}{\pi^2 i}\frac{2 m^2_{\rm n}}{(m^2_{\rm n}-k^2)^3}{\rm tr}\{\gamma^{\alpha}\hat{q}\gamma^{\mu } \gamma^{\nu } + \gamma^{\alpha}\gamma^{\mu}(\hat{q } + \hat{k}_1 ) \gamma^{\nu}\}-\\ & & - \int\,\frac{d^4k}{\pi^2 i}\frac{m^2_{\rm n}k^2}{(m^2_{\rm n}-k^2)^4}{\rm tr}\{\gamma^{\alpha}(3\hat{q}+2 \hat{k}_1 + \hat{k}_2)\gamma^{\mu } \gamma^{\nu } + \gamma^{\alpha}\gamma^{\mu}(3 \hat{q } + 2 \hat{k}_1 + \hat{k}_2)\gamma^{\nu}\}=\nonumber\\ & & = - { \rm tr}\{\gamma^{\alpha}\hat{q}\gamma^{\mu } \gamma^{\nu } + \gamma^{\alpha}\gamma^{\mu}(\hat{q } + \hat{k}_1 ) \gamma^{\nu}\}+\nonumber\\ & & + { \rm tr}\{\gamma^{\alpha}(\hat{q}+ \frac{2}{3 } \hat{k}_1 + \frac{1}{3 } \hat{k}_2)\gamma^{\mu } \gamma^{\nu } + \gamma^{\,\alpha}\gamma^{\mu}(\hat{q } + \frac{2}{3 } \hat{k}_1 + \frac{1}{3 } \hat{k}_2)\gamma^{\nu}\}=\nonumber\\ & & = \frac{1}{3 } { \rm tr}\{\gamma^{\alpha}(2\hat{k}_1 + \hat{k}_2)\gamma^{\mu } \gamma^{\nu } + \gamma^{\alpha}\gamma^{\mu}(-\hat{k}_1 + \hat{k}_2)\gamma^{\nu}\}=\nonumber\\ & & = 4 ( k^{\alpha}_1 g^{\mu\nu } - k^{\nu}_1 g^{\alpha\mu } ) + \frac{4}{3 } ( k_1 + 2 k_2)^{\mu } g^{\alpha\nu}\,.\nonumber\end{aligned}\ ] ] summing up the contributions ( [ label88 ] ) and ( [ label34 ] ) we obtain in order to fix ambiguities produced by the shift of the virtual momentum let us note that @xmath218 , being the momentum of the w boson , coincides with the momentum of the leptonic pair and is small . therefore , we suggest to keep only the terms leading in @xmath222 expansion @xmath280 here we have dropped the term proportional to @xmath281 giving in the lagrangian the total divergence of the deuteron field that vanishes : @xmath282 . the structure function @xmath283,defined by the eq.([label90 ] ) , would be invariant under gauge transformations of the deuteron field if @xmath284 , i.e. @xmath285 the effective lagrangian defined by the structure function ( [ label91 ] ) has the following form @xmath286.\end{aligned}\ ] ] in turn the effective lagrangian describing the low energy process p + p @xmath0 d + @xmath259 + @xmath3 reads @xmath287,\end{aligned}\ ] ] where @xmath288 is the leptonic electroweak current . in the low energy limit when the 3momenta of protons tend to zero the effective lagrangian ( [ label93 ] ) can be reduced to the expression @xmath289\,,\end{aligned}\ ] ] where we have used the relations that is valid in the limit of low 3momenta of protons and the deuteron . the amplitude defined by the effective lagrangian ( [ label94 ] ) reads where @xmath290 and @xmath291 are the dirac bispinors of the neutrino and positron . we have separated the factor @xmath292 , describing the coulomb repulsion , that is convenient for the sequent calculations . the cross section of the low energy p + p @xmath0 d + @xmath259 + @xmath3 reaction is defined where @xmath249 is a relative velocity of protons , and @xmath293 is the squared amplitude averaged over polarizations of protons and summed over polarizations of final particles , i.e. @xmath294 ^ 2\bigg(g_{\alpha\beta}-\frac{q_{\alpha}q_{\beta}}{m^2_{\rm d}}\bigg)\times\nonumber\\ & & \times { \rm tr}\{(\hat{k}_{\rm e}-m_{\rm e})\gamma^{\alpha}(1-\gamma^5)\hat{k}_{\nu}\gamma^{\beta}(1-\gamma^5)\ } \times \frac{1}{4 } \times\\ & & \times \frac{1}{4}\times{\rm tr}\{(\hat{p}_1-m_{\rm n})\gamma^5 ( 1+\gamma^5\hat{s}_1 ) ( \hat{p}_2+m_{\rm n})\gamma^5(1+\gamma^5\hat{s}_2)\},\nonumber\end{aligned}\ ] ] where @xmath295 is the mass of positron , @xmath296 and @xmath297 are 4vectors of polarization of interacting protons such as @xmath298 and @xmath299 [ 2 ] . since protons interact at low energies , we calculate the traces over dirac matrices in the low energy limit @xmath300 substituting eq . ( [ label99 ] ) in eq . ( [ label98 ] ) we get now let us take into account the spinorial properties of the interacting protons . at low energies the two proton fusion proceeds through an intermediate s wave state with the total spin of the two protons equal to zero ( s=0 ) . the former means that the spins of the interating protons are antiparallel , that gives @xmath301 . however , @xmath301 can be also realized for the two proton state with a total spin s=1 and zero projection onto the z axis . therefore , the factor @xmath302 sould be expanded as @xmath303 . keeping only the s wave contribution we arrive at the expression now we should carry out the integration over the phase volume of the final d @xmath304state @xmath305 where @xmath306 and the function @xmath307 is defined by the integral @xmath308 and normalized to unity at @xmath309 , i.e. @xmath310 . the cross section of the s wave low energy two proton fusion p + p @xmath0 d + @xmath259 + @xmath3 is given by the cross section is calculated in units of @xmath311 . the appearance of the factor @xmath312 agrees well with the result obtained by bethe and critchfield [ 27 ] . now we should average the cross section over relative velocities of the proton proton system with a maxwell boltzmann distribution [ 27 ] where @xmath313 is the boltzmann constant , and @xmath314 is the temperature . following [ 27 ] and computing the integral by a saddle point approximation we get where @xmath315 is connected with the temperature [ 27 ] @xmath316 the temperature dependence of eq . ( [ label106 ] ) coincides with that obtained by bethe and critchfield [ 27 ] . setting @xmath317 , measured in unites of @xmath318 , where @xmath319 is the temperature of the solar core in the standard solar model [ 24 ] , that gives @xmath320 we get the following estimate @xmath321 this magnitude is larger by a factor of 2.9 compared to the potential approach [ 24 ] ( see also [ 30 ] ) . in order to reconcile our result with the solar luminisoty we should assume that the temperature in the solar core equals @xmath322 in @xmath318 [ 31 ] . as has been remarked in ref . [ 31 ] the enhancement of the magnitude of the cross section of the two proton fusion leads to a decrease of the temperature in the solar core . this gives a strong suppression of the solar neutrino fluxes . the increase of the amplitude of the p + p @xmath0 d + w transition found in our approach is connected with the computation of the amplitude in terms of one nucleon loop diagrams . indeed , the structure function eq . ( [ label91 ] ) defining the effective lagrangian of the p + p @xmath0 d + w transition is due to the contribution of the anomalous part of the one nucleon loop diagram of the aav kind , i.e. with two axial vector and one vector vertices [ 15,32 ] . one can not expect that such an anomalous contribution should coincide with the contribution described within potential approach in terms of the overlap integral of the wave functions of the deuteron and two protons . the ambiguity of the anomalous contribution has been fixed by the requirement of gauge invariance under the gauge transformations of the deuteron field . it is very similar to the fixing of the ambiguity of the well known axial adler jackiw anomaly [ 2,33 ] . in the present paper we have developed the relativistic field theory model of the deuteron that has been suggested in ref . we have given the elaborate computation of the corben schwinger and aronson effective lagrangians describing the interactions of the deuteron with electromagnetic fields and defining the anomalous magnetic and electric quadrupole moments of the deuteron . we have adjusted the model to the calculation of the low energy processes such as the radiative neutron proton capture n + p @xmath0 d + @xmath1 for thermal neutrons and the fusion of two protons p + p @xmath0 d + @xmath259 + @xmath3 . this reaction is very important for stellar nucleosynthesis , where the deuterons are destroyed again by the reaction p + d @xmath0 @xmath260 + @xmath1 [ 23 ] . unfortunately , our model is far from being induced by the dynamics of qcd and seems like an old fashion approach . it is based on the lagrangian eq . ( [ label1 ] ) written with the only guiding principle that the deuteron is a bound state of the proton and neutron with spin one [ 3,10 ] . we have shown that the model , supplemented by the dynamics of low energy neutron proton and proton proton scattering , is able to describe reasonably the cross sections of the radiative neutron proton capture in agreement with both experimental data and predictions of the potential model . for the cross section of p + p @xmath0 d + @xmath259 + @xmath3 we have obtained an enhancement by a factor 2.9 compared to the potential approach . this enhancement of the cross section could be of relevance for a suppression of the solar neutrino fluxes [ 31 ] . we have found that the coupling constant @xmath15 does not contribute to the amplitudes of the processes under consideration . thereby one can assume that the interaction of the deuteron with the antisymmetric tensor nucleon current is less important for the physics of low energy interactions of the deuteron than the interaction with a vector nucleon current . if it is true , this gives rise the question concerning the need of keeping a nonzero value for the coupling constant @xmath15 . what would happen if we would put @xmath323 ? putting @xmath323 we face the only problem of the fit of the binding energy of the deuteron by the cut off @xmath324 . using eq . ( [ label18 ] ) at @xmath323 we obtain the best fit of the binding energy at @xmath325 . recall that we have identified @xmath326 with the effective radius of the deuteron [ 3 ] . by using the new value of the cut off we get : @xmath327 . this value agrees well the average value of the deuteron radius , i.e. @xmath328 [ 34 ] . putting then @xmath323 we simplify our model reducing the number of free parameters and gain the cut off @xmath325 that should define the average value of the deuteron radius @xmath327 . thus the number of free parameters introduced for the definition of parameters of the physical deuteron are just three : @xmath14 , @xmath5 and @xmath66 and we could fix them unambiguously . in this connection we should underscore that additional free paprameters appearing due to the ambiguities connected with the shifts of virtual momenta in the diagrams in figs . 25 are intrinsic peculiarities of one fermion loop diagrams , caused by the computation of such diagrams within cut off regularization [ 15 ] . therefore , one can not consider these parameters as free parameters especially introduced in the model . of course , the appearance of these ambiguities can lower , from the first glance , the predictive abilities of the model . however , as we have shown , most of these parameters can be unambiguously fixed by the requirement of gauge invariance . finally , we want to estimate the theoretical uncertainty of the relativistic field theory model of the deuteron . unfortunately , in our model there is not a small parameter , like @xmath252 , the fine structure constant in qed . there is not a large parameter like the number of quark colours n in the multi colour extension of qcd with @xmath329 that allows to apply large @xmath330expansion in powers of @xmath130 to quark gluon diagrams describing the strong low energy interactions of hadrons . our approach is effective and based on the one nucleon loop diagram approximation for the description of both the self interactions of the deuteron and coupling of the deuteron to other particles . by describing the self interactions of the deuteron at the one nucleon loop approximation we have fitted all parameters characterising the physical deuteron . in this case the only way to estimate the theoretical uncertainty of the approach is to compare the one nucleon loop data with the two nucleon loop corrections . in ref . [ 1 ] we have calculated the two nucleon loop correction to the binding energy of the deuteron . setting @xmath95 , this correction reads [ 1 ] @xmath331 the numerical value of the two nucleon loop correction makes up about 16@xmath128 of the binding energy @xmath332 calculated in the one nucleon loop approximation ( see eq . ( [ label70 ] ) ) . thus the magnitude @xmath333 might be accepted as a theoretical uncertainty of the relativistic field theory model of the deuteron . this agrees well with our prediction for the cross section of the radiative neutron proton capture . indeed , the theoretical value @xmath334 reproduces the experimental data @xmath335 with an accuracy of about 8@xmath128 . by taking into account the theoretical uncertainty the cross section of the radiative neutron proton capture should read : @xmath336 . with regard to the two proton fusion cross section , we predict an enhancement of @xmath337 compared to the potential approach . we acknowledge fruitful discussions with prof . g. e. rutkovsky and dr . h. leeb , and dr . m. meinhart for discussions concerning effective radius of the deuteron . this work was partially supported by the fonds zur frderung der wissenschaftlichen forschung in sterreich ( project p10361phy ) [ 1 ] a. n. ivanov , n. i. troitskaya , m. faber and h. oberhummer , phys . lett . * b 361 * ( 1995 ) 74 . [ 2 ] c. itzykson and j. - b. zuber , in _ quantum field theory _ ( mcdraw hill ) , p.p . 614 , 1980 ; v. de alfaro , s. fubini , g. furlan and c. rossetti , in _ currents in hadron physics _ ( north holland ) , chapt . 5 , 1973 ; m. gell mann and m. levy , nuovo . * 16 * ( 1960 ) 705 ; t. hakioglu and m. d. scadron , phys . rev . * d 42 * ( 1991 ) 941 ; a. n. ivanov , m. nagy and m. d. scadron , phys . lett . * b 273 * ( 1991 ) 137 ; r. delbourgo and m. d. scadron , mod . . lett . * a 10 * ( 1995 ) 251 . [ 3 ] y. nambu and g. jona lasinio , phys . * 122 * ( 1961 ) 345 ; ibid . * 124 * ( 1961 ) 246 . [ 4 ] s. de benedetti , in _ nuclear reactions _ , john wiley @xmath338 sons , inc . : new york london sydney , 1967 , p. 46 . [ 5 ] m. m. nagels et al . , nucl . phys . * b 147 * ( 1979 ) 253 . [ 6 ] t. eguchi , phys . rev . * d 14 * ( 1976 ) 2755 ; k. kikkawa , prog . phys . * 56 * ( 1976 ) 947 ; h. kleinert , in zichichi a. ( ed . ) : proc . of int . school of subnuclear physics , ( 1976 ) p. 289 . [ 7 ] t. hatsuda and t. kumihiro , progr . * 74 * , ( 1985 ) 765 ; phys . * b 198 * , ( 1987 ) 126 ; t. kumihiro and t. hatsuda , phys . * b 206 * , ( 1988 ) 385 . [ 8 ] s. klint , m. lutz , v. vogl and w. weise , nucl . phys . * a 516(1990 ) 429 ; 469 and references therein . [ 9 ] a. n. ivanov , m. nagy and n. i. troitskaya , int . j. mod . phys . a 7 * ( 1992 ) 7305 ; a. n. ivanov , int . j. mod . * a 8 * ( 1993 ) 853 ; phys . lett . * b 275 * ( 1992 ) 450 ; a. n. ivanov , n. i. troitskaya and m. nagy , int . j. mod . phys . * a 8 * ( 1992 ) 2027 ; 3425 ; a. n. ivanov , n. i. troitskaya and m. nagy , phys . * b 308 * ( 1993 ) 111 ; a. n. ivanov and n. i. troitskaya , nuovo . cim . * a 108 * ( 1995 ) 555 ; a. n. ivanov , n. i. troitskaya , m. faber , m. schaler and m. nagy , nuovo cim . * a 107 * ( 1994 ) 1667 ; phys . b 336 * ( 1994 ) 555 . [ 10 ] b. sakita and c. j. goebel , phys . * 127 * ( 1962 ) 1787 ; b. sakita , phys . * 127 * ( 1962 ) 1800 . [ 11 ] h. aronson , phys . rev . * 186 * ( 1969 ) 1434 . [ 12 ] j. d. jackson , in _ klassische electrodynamik _ , walter de gruyter , berlin new york , 1983 . [ 13 ] particle data group , phys . * d 50 * ( 1994 ) 1177 , part 1 . [ 14 ] h. c. corben and j. schwinger , phys . * 58 * ( 1940 ) 953 . [ 15 ] i. s. gertsein and r. jackiw , phys . rev . * 181 * ( 1969 ) 1955 ; r. w. brown , c. c. shih and b. l. young , phys . * 186 * ( 1969 ) 1491 . [ 16 ] a. di giacomo , g. paffuti and p. rossi , in _ selected problems of theoretical physics ( with solutions ) _ , world scientific , singapore new jersey hong kong , problem 22 , p. 68 . [ 17 ] a. e. cox , a. r. wynchank and c. h. collie , nucl . * 74 * ( 1965 ) 497 and references therein . [ 18 ] l. r. b. elton , in _ intoductory nuclear physics _ , london : pitman , 1959 . [ 19 ] m. d. scadron , in _ advanced in quantum theory and its applications through feynman diagrams _ , verlag , new york heidelberg : 1979 , pp . 237239 . [ 20 ] g. breit and e. wigner , phys . * 49 * ( 1936 ) 519 . [ 21 ] see ref.[3 ] , p. 156 . [ 22 ] w. f. hornyak , in _ nuclear structure _ , academic press , new york san francisco london : 1975 , p. 495 . [ 23 ] g. brner , in _ the early universe ( facts and fiction ) _ , springer verlag , berlin heidelberg new york tokyo , 1988 , p.111 . [ 24 ] c. e. rolfs and w. s. rodney , in _ cauldrons in cosmos _ , the university of chicago press , chicago and london , 1988 , pp . 328338 . [ 25 ] g. gamow , phys * 53 * ( 1938 ) 595 . [ 26 ] j. n. bachall , in _ neutrino astrophysics _ , cambridge university press , cambridge new york new rochelle sydney , 1989 , p. 60 . [ 27 ] h. a. bethe and c. l. critchfield , phys * 54 * ( 1939 ) 248 . [ 28 ] l. rosenfeld , in _ nuclear forces _ , north holland publishing company , amsterdam , 1948 , pp . 155157 . [ 29 ] f. l. yost , j. a. wheeler and g. breit , phys * 49 * ( 1936 ) 174 . [ 30 ] m. kamionkowski and j. n. bahcall , ap . j. * 359 * ( 1991 ) 884 ; j. n. bahcall and m. h. pinsonneault , rev . * 64 * ( 1992 ) 885 . [ 31 ] v. castellani , s. degl@xmath339innocenti , g. fiorentini , m. lissia and b. ricci , phys * 281 * ( 1997 ) 309 . [ 32 ] a. n. ivanov , sov . * 33 * ( 1981 ) 904 . [ 33 ] r. jackiw , in _ lectures on current algebra and its applications _ , princton series in physics , princton university press , princeton new jersey , 1972 . [ 34 ] see ref.[22 ] , p. 147 . 1 one nucleon loop diagrams contributing to the binding energy of the physical deuteron , where @xmath340 is the field of anti neutron . * 2 one nucleon loop diagrams describing the effective corben schwinger and aronson interactions that are responsible for the anomalous magnetic and electric quadrupole moments of the physical deuteron , where @xmath340 is the field of anti neutron . * fig . 3 the contribution of the @xmath341[\bar{n^c}(x)\gamma^{\,5}p(x)]$ ] to the amplitude of the radiative neutron proton capture . * fig . 4 the contribution of the @xmath342[\bar{n^c}(x)\gamma^{\,\mu}\gamma^{\,5}p(x)]$ ] to the amplitude of the radiative neutron proton capture . * fig . 5 the contribution of the @xmath343[\bar{p^c}(x)\gamma^{\,\mu}\gamma^{\,5}p(x)]$ ] to the amplitude of the p + p @xmath0 d + @xmath2 + @xmath3 scattering . by ivanov et al . published in nucl . phys . a617 ( 1997 ) 414 . * 1 . * the enhancement factor 2.9 obtained in ref . [ 1 ] with respect to the potential model is mainly related to the neglect of the electromagnetic mass difference of the proton and the neutron by using a value of @xmath344mev instead of 0.932mev in the phase space factor @xmath345 . furthermore , we used a few erroneous combinatorial and dynamical factors connected with identical and dynamical properties of the interacting protons . all of these factors are included now correctly in the eqs . 1 and 2 given below . for the astrophysical factor we give the following new expression @xmath346 ^ 2\bigg(1 - \frac{8\sqrt{2}\pi\,m^2_{\pi}}{g^2_{\rm \pi nn}}\ , \frac{a_{\rm pp}}{m_{\rm n}}\bigg)^2\,(w + e)^5\,f\bigg(\frac{m_{\rm e}}{w + e}\bigg),\end{aligned}\ ] ] where @xmath347 is the kinetic energy of the relative movement of the protons , and the function @xmath348 is given by eq . ( 6.24 ) of ref . [ 1 ] . for @xmath349 we obtain @xmath350 ^ 2\bigg(1 - \frac{8\sqrt{2}\pi\,m^2_{\pi}}{g^2_{\rm \pi nn}}\,\frac{a_{\rm pp}}{m_{\rm n}}\bigg)^2\,w^5\,f\bigg(\frac{m_{\rm e}}{w}\bigg ) = \nonumber\\ & = & 2.01\times 10^{-25}\,(1 \pm 0.30){\rm mev}\,{\rm b},\end{aligned}\ ] ] where 30@xmath128 is the assumed theoretical uncertainty of our model . the real theoretical uncertainty of the approach can turn out to be much less . the value @xmath351 makes up ( 51.7 @xmath352 15.5)@xmath128 of the value @xmath353 obtained by kamionkowski and bahcall in the potential approach [ 2 ] . we stress that the astrophysical factor @xmath354 in eqs . ( [ label1e ] ) and ( [ label2e ] ) describes only the contribution of the weak and strong low energy interactions and neglects the coulomb repulsion between protons in the one nucleon loop . the coulomb repulsion in the initial state is included in terms of the gamow penetration factor @xmath355 [ 3 ] , where @xmath356 , multiplied by the amplitude defined by weak and strong low energy interactions of the protons and the deuteron . our result concerning the astrophysical factor @xmath357 is rather promising . indeed , as has been shown by kamionkowski and bahcall an important contribution to the astrophysical factor for pp fusion comes from the region where the coulomb repulsion between the protons is dominating [ 2 ] . since we have managed to describe about 50@xmath358 of the astrophysical factor obtained in the potential approach keeping only the contribution of weak and strong low energy interactions of the protons and the deuteron , we expect to get the contribution of the same order when including the coulomb repulsion in the one nucleon loop . these calculations are in progress now . we note that we have tested our approach by calculating the disintegration of the deuteron by reactor antineutrinos @xmath359 + d @xmath0 n + n + e@xmath266 . this process is governed by the same physics as the quantity @xmath360 . we find excellent agreement with the potential model results and the experimental data [ 4 ] . ( 5.22 ) of ref . [ 1 ] should read @xmath361 eq . ( 5.23 ) of ref . [ 1 ] should read @xmath362^{\,2}\,\times\\ & & \times\bigg(1\,-\,\frac{8\pi\ , m^2_{\pi}}{g^2_{\pi\rm nn}}\,\frac{a_{\rm np}}{m_{\rm n}}\bigg)^2\,m_{\rm n}\,\varepsilon^{\,3}_{\rm d}\,=\,(276\pm 83)\,{\rm mb}\,,\end{aligned}\ ] ] where @xmath363 in notations of ref . our result agrees reasonably well with the experimental data @xmath364 [ 1 ] . the change of the cross section for radiative capture is caused by the equal contributions of the pole on the unphysical sheet to the amplitudes defined by the @xmath365 and @xmath366 neutron proton interactions at low energies . in ref . [ 1 ] we have predicted an accuracy of the approach of about 16@xmath128 . since this accuracy has been obtained from the comparison of the one nucleon loop and two nucleon loop contributions to the binding energy of the deuteron , it should be understood as the accuracy of amplitudes . thus , the theoretical accuracy of cross sections should be about 30@xmath128 . however , in practice this accuracy can be much better . at the end of chapter 2 of ref . [ 1 ] the phrase `` the contributions of these diagrams are divergent and can be removed by the renormalization of the wave - function of the deuteron '' should read `` the contributions of these diagrams are divergent and can be removed by the renormalization of the electric charge of the deuteron '' . the phrase `` ... finite contributions to the renormalization constant of the wave function of the deuteron ... '' appearing above eqs.(3.20 ) , ( 4.14 ) and ( 4.15 ) should read `` ... finite contributions to the renormalization constant of the electric charge of the deuteron ... '' 9 a. n. ivanov , n. i. troitskaya , m. faber and h. oberhummer , nucl . a617 ( 1997 ) 414 . m. kamionkowski and j. n. bahcall , apj . 420 ( 1994 ) 884 . h. a. bethe and c. l. critchfield , phys . ( 1939 ) 248 . h. oberhummer , a. n. ivanov , n. i. troitskaya and m. faber , _ disintegration of the deuteron by photons and antineutrinos in the relativistic field theory model of the deuteron _ , submitted to phys .	the relativistic field theory model of the deuteron suggested in [ 1 ] is revised and applied to the calculation of the cross sections of the low energy radiative neutron proton capture n + p @xmath0 d + @xmath1 and the low energy two proton fusion p + p @xmath0 d + @xmath2 + @xmath3 . for the low energy radiative neutron proton capture n + p @xmath0 d + @xmath1 our result agrees well with both the experimental data and the potential model prediction . in the case of the two proton fusion the cross section obtained is 2.9 times as much as that given by the potential approach . the obtained result is discussed in connection with the solar neutrino problem . = 16truecm = 23.5truecm -1.5 cm -1.2 cm nuclear physics * a * ( in press )
based on penetrative analysis of metal - poor halo stars and globular clusters ( gcs ) in the galaxy , two canonical galaxy formation scenarios the monolithic collapse scenario @xcite and the accretion / merging one @xcite were proposed , that have long been influential for later observational and theoretical studies of disk and elliptical galaxies . although observational studies of stellar halos in galaxies beyond the local group of galaxies have just recently started revealing structural and chemical properties of the halos @xcite@xcite , physical properties of globular cluster systems ( gcss ) in these galaxies have long been investigated in much more details @xcite . wide - field imaging and spectroscopic studies with large ground - based telescopes ( e.g. , keck 10 m ) have recently revealed gcs structures and kinematics in galaxies with different hubble types @xcite . furthermore a growing number of theoretical / numerical studies have recently been accumulated which have investigated dynamical and chemical properties of gcs and gcss based on admittedly realistic and self - consistent models of gc formation during galaxy formation and evolution @xcite@xcite . in this review paper , we therefore try to derive physical meanings from the selected four observed properties of gcs and gcss by comparing our numerical simulations of gc and gcs formation with latest observations . ) which an elliptical experienced during its formation ( bf06 ) . for clarity , the density distributions are normalized to their central values . thin dotted lines represent power - law slopes ( @xmath0 ) of @xmath0 = @xmath1 , @xmath2 , @xmath3 , and @xmath4 . note that the density profiles of gcss become flatter for larger @xmath5 , i.e. more mergers . , height=377 ] it has long been known that the radial density profiles of gcss in elliptical galaxies ( es ) vary with the total luminosities of their host galaxies @xcite . if the projected gcs density profiles in es with @xmath6-band absolute magnitudes of @xmath7 are fitted to the power - law ones like @xmath8 , where @xmath9 is the distance from the center of the host galaxy of a gcs , the power - law index @xmath10 is smaller ( i.e. , the profiles are steeper ) for larger @xmath7 ( i.e. , fainter es ) . although two physical mechanisms gc destruction by galactic tidal fields @xcite@xcite and dynamics of galaxy merging ( bekki & forbes 2006 ; bf06 ; @xcite ) have been so far proposed for the origin of gcs density profiles , we focus on the latter case in this paper . bf06 numerically investigated the structural properties of gcss in es formed from a sequence of major dissipationless galaxy merging and thereby found that the radial density profiles of gcss in es become progressively flatter as the galaxies experience more major merger events ( see figure 1 ) . the simulated profiles of gcss are found to be well described as power - laws with @xmath10 ranging from @xmath2 to @xmath4 in es . they are flatter than , and linearly proportional to , the slopes ( @xmath11 ) of the stellar density profiles . by applying a reasonable scaling relation between luminosities and sizes of galaxies to the simulation results , bf06 showed that @xmath12 , @xmath13 , and @xmath14 . these correlations between gcs profiles and their host galaxy luminosities are consistent reasonably well observations @xcite@xcite which suggests that the origin of structural non - homology of gcss in es can be understood in terms of the growth of es via major dissipationless galaxy merging . estimated for @xmath9 @xmath15 @xmath16 and those for @xmath9 @xmath15 @xmath17 for metal - poor clusters these correlations are derived from 18 results of 6 major merger models with three different projections @xcite . , height=188 ] recent observations on gcs kinematics in e / s0s have revealed that gcs kinematics can be quite diverse : gcss in some galaxies like m87 @xcite ngc 4472 @xcite and ngc 5128 @xcite show rotation whereas those in some galaxies such as ngc 1399 @xcite do not . bekki et al . ( 2005 ; b05 ; @xcite ) first tried to understand the observed diversity in gcs kinematics by numerically investigating gcs kinematics of e / s0s formed from major / minor galaxy merging . b05 demonstrated that both metal - poor cluster ( mpcs ) and metal - rich ones ( mrcs ) in es formed from major mergers can exhibit significant rotation at large radii ( @xmath1820 kpc ) due to the conversion of initial orbital angular momentum into intrinsic angular momentum of the remnant . based on a wide parameter study of galaxy mergers , b05 found that mpcs show higher central velocity dispersions than mrcs for most major merger models . @xmath19 ( where @xmath20 and @xmath21 , are the gcs maximum rotational velocity and central velocity dispersion , respectively ) ranges from 0.2 1.0 and 0.10.9 for the mpcs and mrcs respectively , within @xmath17 for the remnant elliptical . figure 2 shows an interesting result that does not depend on merger parameters : @xmath19 of mpcs within @xmath17 are greater than those of mpcs within @xmath16 . b05 also revealed the alignment of the major axes in 2d distributions between stars , gcs , and dark matter halos in the simulated es . the aligned major axis between stars , gcs , and dark matter appears to be one of the principal characteristics of es formed by major merging , which implies that _ observational studies on 2d distributions of gcss in es can tell us about the shapes of their host dark matter halos_. we also showed in this meeting that the total masses of e / s0s estimated from the gcs kinematics can be much closer to the real masses than those from the pne systems owing to less anisotropic velocity dispersion in gcss , in particular , for face - on s0s . this suggests that ( 1 ) gcss are better mass - estimators in e / s0s and ( 2 ) kinematical data sets of pne systems in e / s0s @xcite should be more carefully interpreted for the total masses of e / s0s . although these simulations are not based on cosmological n - body ones , these results may well provide new insight on the origin of the observed diversity in gcs kinematics of e / s0s . possible candidates of young and metal - rich gcs were discovered in interacting and merging galaxies @xcite , and physical properties of these gcs have been discussed in different contexts of galaxy and gc formation , such as the observed color bimodality in gcss in es @xcite , the birth rate of young gcs as a function of time in m82 @xcite and the age - metallicity relation of gcs in nearby interacting galaxies like the lmc and the smc ( e.g. , bekki et al . 2004 ; b04 ; @xcite ) . in this paper , we focus on the lmc s gcs in which nearly all gcs are either very old ( @xmath22 gyr ) or younger than @xmath23 gyr the `` age - gap '' problem @xcite@xcite . b04 and bekki & chiba ( 2005 ; @xcite ) challenged this age gap problem by investigating chemodynamical evolution of the lmc interacting both with the galaxy and the smc for a long time scale ( @xmath24 gyr ) . they found that the first close encounter between the lmc and the smc about 4 gyr ago was the beginning of a period of strong tidal interaction witch likely induced dramatic gas cloud collisions , leading to an enhancement of the gc formation which has been sustained by strong tidal interactions to the present day . figure 3 , showing the simulated age distributions of field stars and gcs in a model , reveals that gc formation can be reactivated about @xmath23 gyr ago , when the lmc can start its strong tidal interaction with the smc . these results imply that the origin of the age gap can be closely associated with interaction histories of the lmc , the smc , and the galaxy . ( @xmath25band absolute magnitude ) for the vmscs in 40 models @xcite . projected central velocity dispersion ( @xmath26 ; _ upper left _ ) , half - light - averaged surface brightness ( @xmath27 ; _ upper right _ ) , effective radius ( @xmath28 ; _ lower left _ ) , and central surface brightness ( @xmath29 ; _ lower right _ ) are plotted against @xmath7 . here the central surface brightness @xmath29 is expressed as @xmath30 , where @xmath31 , @xmath32 are the total luminosity of a vmsc and the radius within which 10% of @xmath31 is included , respectively . the best fit scaling relation for the vmscs is derived for each panel using the least square fitting method and described as a _ dotted _ line with the derived relation ( e.g. , @xmath21 @xmath33 @xmath34 ) . , height=377 ] very massive star clusters ( vmscs ) such as @xmath35 centauri , ultra - compact dwarfs ( ucds ) , and massive nuclear star clusters have unique characteristics that are quite different from those of `` normal '' gcs . for example , ucds discovered in the fornax and the virgo clusters of galaxies @xcite@xcite@xcite have intrinsic sizes of less than 100pc , and have absolute @xmath36band magnitudes ranging from @xmath37 to @xmath38mag , which is more than 2 magnitudes brighter than the most massive gc in the galaxy ( i.e. , @xmath35 cen ) . two physical mechanisms for the vmsc formation have been so far proposed : the `` galaxy threshing '' scenario @xcite@xcite in which vmscs originate from nuclei of nucleated dwarf galaxies and the `` cluster merging '' one in which vmscs are formed from merging of smaller star clusters in tidal tails of merging galaxies @xcite . although more details on physical properties of vmscs have been recently revealed @xcite@xcite , it remains unclear which of the two scenarios is more convincing for the origin of vmscs . here we discuss what we can learn from vmsc properties _ if they were previously nuclei of nucleated galaxies . _ both dissipationless @xcite@xcite and dissipative @xcite formation scenario of stellar galactic nuclei have provided some interesting predictions on nuclear properties of galaxies @xcite . figure 4 shows the scaling relations between different physical parameters of merger remnants of star clusters initially with the observed gc scaling relations @xcite . the fact that the simulated relations deviate from the gc s ones implies that _ the scaling relations of vmscs can be used for understanding whether the stellar nuclei can be formed from merging of many smaller clusters in the central regions of galaxies _ @xcite . the dissipative nucleus formation model @xcite has predicted the spread of ages and metallicities in stellar populations of stellar galactic nuclei and accordingly can be discussed in the context of the observed spread of ages and metallicities in @xmath35 cen @xcite . thus dynamical and chemical properties of vmscs can tell us about nucleus formation histories in galaxies . thus , structural , kinematical , and chemical properties of gcss in galaxies have fossil information on dynamics of major / minor galaxy merging ( e.g , angular momentum redistribution processes in merging ) , interaction histories of galaxies , and the formation histories of stellar galactic nuclei in dwarfs . previous theoretical / numerical studies however did not discuss so extensively the observed correlations between gcs properties and their host ones @xcite in the context of a hierarchical clustering scenario of galaxy formation . several authors just recently have started their investigation on the gcs - host relations based on semi - analytic models @xcite and high - resolution numerical simulations in @xmath39cdm models @xcite@xcite . a number of observed gcs - host relations , such as the positive correlation between gcs metallicities and their host galaxy luminosities @xcite have not been clearly explained by any galaxy formation scenarios . since these gcs - host relations may have profound physical meanings on galaxy formation and evolution , it is doubtlessly worthwhile for future numerical simulations of gcs formation to explore the origin of these relations . eggen , d. lynden - bell , a.r . sandage : apj , * 136 * , 748 ( 1962 ) l. searle , r. zinn : apj , * 225 * , 357 ( 1978 ) j. dalcanton , r.a . bernstein : aj , * 123 * , 1328 ( 2002 ) s. zibetti , s.d.m . white , j. brinkmann : mnras , * 347 * , 556 ( 2004 ) w.e . harris : ara&a , * 29 * , 543 ( 1991 ) j.p . brodie , j. strader : ara&a in press ( astro - ph/0602601 ) ( 2006 ) k. bekki , d.a . forbes , m.a . beasley , w.j . couch : mnras , * 335 * , 1176 ( 2002 ) o.y . gnedin : in this volume ( 2006 ) k.m . ashman , s.e . zepf : in globular cluster systems , cambridge , u. k. ; new york : cambridge university press ( 1998 ) h. baumgardt : a&a , * 330 * , 480 ( 1998 ) e. vesperini , s. e. zepf , a. kundu , k.m . ashman : apj , * 593 * , 760 ( 2003 ) k. bekki , d.a . forbes : a&a , * 445 * , 485 ( 2006 ) w.e . harris : aj , * 91 * , 822 ( 1986 ) m. kissler - patig , k. gebhardt : aj , * 116 * , 2237 ( 1998 ) s.e . zepf , m.a . beasley et al : aj , * 120 * , 2928 ( 2000 ) e.w . peng , h.c . ford , k.c . freeman : apj , * 602 * , 705 ( 2004 ) t. richtler et al : aj , * 127 * , 2094 ( 2004 ) k. bekki , m.a . beasley , j.p . brodie , d.a . forbes , mnras , * 363 * , 1211 ( 2005 ) a.j . romanowsky , n.g . douglas et al : science , * 301 * , 1696 ( 2003 ) b.c . whitmore , f. schweizer : aj , * 109 * , 960 ( 1995 ) r. de grijs , n. bastian , h.j.g.l . lamers : mnras , * 340 * , 197 ( 2003 ) k. bekki , w.j . couch , w. j. et al : * 610 * , l93 ( 2004 ) g.s . da costa : in haynes r. , milne d. , eds , proc . 148 , the magellanic clouds , kluwer , dordrecht , p183 ( 1991 ) d. geisler : in this volume ( 2006 ) k. bekki , m. chiba : mnras , * 356 * , 680 ( 2005 ) m.j . drinkwater , m.d . gregg et al : nature , * 423 * , 519 ( 2003 ) j.b . jones , m.j . drinkwater et al : aj , * 131 * , 312 ( 2006 ) s. mieske , m. hilker , l. infante : , a&a , * 418 * , 445 ( 2004 ) k. bekki , w.j . couch , m.j . drinkwater : apj , * 552 * , l105 ( 2001 ) k. bekki , w.j . couch , m.j . drinkwater , y. shioya : mnras , * 344 * , 399 ( 2003 ) m. fellhauer , p. kroupa : mnras , * 330 * , 642 ( 2002 ) m. hilker , a. kayser , t. richtler , p. willemsen : a&a , * 422 * , l9 ( 2004 ) m. haegan et al : apj , * 627 * , 203 ( 2005 ) s.d . tremaine , j.p . ostriker , l. spitzer , jr : apj , * 196 * , 407 ( 1975 ) r. capuzzo - dolcetta , a. tesseri : mnras , * 308 * , 961 ( 1999 ) k. bekki , w.j . couch , y. shioya : apjl , in press ( astro - ph/064340 ) ( 2006 ) k. bekki , w.j . couch , m.j . drinkwater , y. shioya , apj , * 610 * , l13 ( 2004 ) s.g . djorgovski , r.r . gal et al : apj , * 474 * , l19 ( 1997 ) p. cte et al : 2006 , apjs in press ( astro - ph/0603252 ) ( 2006 ) m.a . beasley , c.m . baugh et al : mnras , * 333 * , 382 ( 2002 ) k.l . rhode , s.e . zepf , m.r . santos : apj , * 630 * , l21 ( 2005 ) k. bekki , h. yahagi , d.a . forbes : submitted to apjl ( 2006 )	we discuss the origin of physical properties of globular cluster systems ( gcss ) in galaxies in terms of galaxy formation and evolution processes . based on numerical simulations of dynamical evolution of gcss in galaxies , we particularly discuss ( 1 ) the origin of radial density profiles of gcss , ( 2 ) kinematics of gcss in elliptical galaxies , ( 3 ) transformation from nucleated dwarf galaxies into gcs ( e.g. , omega centauri ) , and ( 4 ) the origin of gcss in the large magellanic cloud ( lmc ) .
the microscopic study of atomic nuclei has gained in the recent years a noticeable level of reliability , following the improved knowledge of the interaction between the nucleons and the development of new and more sophisticated many body theories . it is now possible to exactly solve the schrdinger equation for nuclei containing up to @xmath2 nucleons@xcite , while the structure of medium and heavy nuclei can be to a large extent understood by means either of variational techniques@xcite or of highly refined perturbative expansions@xcite . these relatively recent developments have clearly stressed several inadequacies of the independent particle model ( ipm ) , on which the time honoured shell model is based . for instance , the ipm completely fails in describing the large energy behaviour of the one nucleon momentum distribution and the quenching of the spectroscopic factors@xcite . the scenario is not equally bright for hypernuclei , or bound systems of nucleons and one or more strange baryons , as the @xmath3 or @xmath4 hyperons . body techniques developed for the nuclei can be , in most of the cases , straightforwardly extended to the hypernuclei , at least from a formal point of view . however , the weak point of a consistent microscopic approach now lies in the poorer understanding of the hyperon nucleon ( yn ) and hyperon hyperon ( yy ) interactions with respect to the nuclear case . the limited amount of information on the yn scattering data is a consequence of the experimental difficulties due to the short lifetime of the hyperons and to the low intensity of the beam fluxes . for this reason phenomenological models and effective yn interactions have been often used in the study of hypernuclei . approaches based upon skyrme interactions @xcite and woods saxon @xcite potentials as well as relativistic mean field theories @xcite have been used . however , also microscopic yn potentials have been recently produced @xcite and adopted to study the structure of hypernuclei from @xmath5o to @xmath6pb @xcite . the @xmath0 single particle properties have been investigated in details in ref . @xcite , making use of the nuclear matter fermi hypernetted chain results in a local density approximation . in this paper we present a microscopic study of the structure of various hypernuclei containing a single @xmath3 hyperon . this study has been done by extending the correlated basis functions theory and the fermi hypernetted chain ( fhnc ) equations technique @xcite to describe nucleonic systems containing one hyperonic impurity . the nucleon nucleon ( nn ) interaction we have chosen is the semi - realistic two nucleon s3 potential of afnan and tang @xcite , as modified in ref.@xcite . in the present work we neglect the three nucleon potential . for the @xmath0nucleon ( @xmath0n ) potential we have chosen that proposed by bodmer and usmani @xcite . we have investigated the effects of many - body forces involving the @xmath0 by explicitly considering a @xmath7 interaction or , alternatively , by including a density dependence in the bare @xmath8 force . the interactions generate strong nn and @xmath0n correlations , beyond those described by a mean field approach . in our work the correlation effects are accounted for by a jastrow correlated wave function , where the two body correlations depend on the interparticle distances only . more realistic hamiltonians demand for a strong spin and isospin dependence in the correlation factor . however , we consider the simple jastrow ansatz as a first step towards a more complete description of the hypernuclei . the same strategy has been followed for doubly closed shell nuclei , where the accuracy of the variational method has now reached the same level as in nuclear matter@xcite . the fhnc equations for a @xmath0 embedded in infinite nuclear matter were developed in ref.@xcite . in ref.@xcite , doubly closed shell nuclei in @xmath9coupling were studied by fhnc and jastrow type correlations , while the @xmath1coupling was introduced in such a scheme in ref.@xcite . it was so possible to study nuclei ranging from @xmath10c to @xmath11pb within the correlated basis function theory , using cluster expansion at all orders and avoiding such shortcuts as low order cluster truncations and local density approximations . the results obtained in those papers are the basis of the present fhnc study of @xmath3hypernuclei . the plan of the paper is as follows : section [ hami ] is devoted to a brief description of the hamiltonian and of the correlated wave function ; section [ dens ] gives an outline of the fhnc theory for the hypernucleus ; in section [ vari ] the expressions for the variational energy are presented and discussed ; the results are given in section [ resu ] and the conclusions are drawn in section [ conc ] . the system under study is a hypernucleus composed by @xmath12 nucleons and one @xmath0 hyperon . the formalism we have developed is however general enough to describe any nuclear systems containing a non - nucleonic particle , treated as an impurity in the nucleonic fluid . we write the hamiltonian of the system as : @xmath13 where we have considered two and three body interactions between the hyperon and the nucleons and only two body ones between the nucleons . in the hamiltonian ( [ h ] ) the purely nucleonic part , @xmath14 , has been separated from that involving the @xmath0 , @xmath15 . the separation will be useful in the remainder of the paper . as already stated in the introduction , we use the modified nn s3 potential , which is fixed to reproduce the @xmath16 wave scattering nucleon - nucleon phase shifts at low energies , gives reasonable results both for light nuclei and nuclear matter and allows for utilizing only central two body correlations between the nucleons . the operatorial structure of s3 is : @xmath17 where @xmath18 , @xmath19 , @xmath20 , @xmath21 . modern nucleon - nucleon potentials are more complicated and contain the essential tensor terms , absent in s3 . on the other hand , we are mainly interested in presenting and testing the formalism . adopting the s3 potential is sufficient for this purpose and has the advantage of avoiding the technical complications deriving from the use of more realistic potentials . within the same philosophy , we have chosen the @xmath0nucleon potential proposed by bodmer and usmani @xcite . this potential has central , spin and space exchange components : @xmath22 with @xmath23 where @xmath24 is the spatial exchange operator and @xmath25 is defined as : @xmath26 the one - pion exchange term is @xmath27 with the pion mass @xmath28=0.7 @xmath29 and the cutoff parameter @xmath30=2.0 @xmath31 . the values of the various parameters @xmath32 are given in ref . @xcite . in this reference , a three - body @xmath0-nucleon - nucleon interaction was added to the just described @xmath0-nucleon one in order to fit the hypernucleus @xmath33 empirical @xmath34 by the variational monte carlo ( vmc ) method . the @xmath35 interaction has an important , attractive two - pion exchange part , which is the source of a tensor term . the tensor contribution , however , results to be zero if simple central correlations are adopted . as a consequence , such a semi - microscopic @xmath35 interaction is too repulsive in a jastrow correlated model , as we shall discuss later . for this reason , we simulate the effects of three and more body potentials by using an approach similar to that developed by friedman and pandharipande for pure nucleonic matter @xcite . we multiply @xmath36 by a density dependent function @xmath37/2\}~ , \label{frho}\ ] ] where @xmath38 is the nucleonic density . the value of the parameter @xmath39 is fixed to reproduce the empirical value of a single @xmath0 binding energy in nuclear matter , @xmath40 mev . the binding energy of the @xmath0 particle is defined as minus the difference between the energies of the nuclear systems with and without the @xmath0 : @xmath41 where we have used the notation : @xmath42 with @xmath43 or @xmath12 . in the previous equations we have averaged on the third component of the @xmath0 total angular momentum and we have explicitly singled out the difference between the center of mass energies of the two systems . our approach is based upon the variational principle , therefore the energy functionals given in eq.([blambda ] ) , @xmath44 and @xmath45 , should be independently minimized with respect to variations of the many - body wavefunctions , @xmath46 . the hypernucleus wavefunction is chosen as : @xmath47 where @xmath48 is the mean - field single particle wavefunction of the @xmath0 in @xmath1coupling , @xmath49 is a two - body scalar ( jastrow ) correlation between the hyperon and a single nucleon and @xmath50 is the correlated wave function describing the remaining @xmath12 nucleons . in our calculation this function is defined as : @xmath51 where @xmath52 is the slater determinant of a set of single particle wave functions of @xmath53 neutrons and @xmath54 protons . obviously , @xmath55 . the energy expectation values are evaluated by means of the fermi hypernetted chain@xcite cluster diagrams resummation technique . as already mentioned , our approach consists in considering the @xmath0 as an impurity in the nucleonic fluid . the fhnc equations for an impurity in homogeneous matter were derived in ref . @xcite for a @xmath3 hyperon in symmetric nuclear matter and in refs . @xcite for a @xmath56he atomic impurity in liquid @xmath57he . our aim is the evaluation the @xmath0 binding energy in finite nuclear systems . to this purpose , we extend the fhnc equations developed in ref.@xcite for doubly closed shell nuclei in the @xmath1coupling scheme with a jastrow correlated wave functions . the basic quantities to be analyzed are the one- and two - body densities ( obd and tbd ) . it is convenient to define densities for each type of particle , protons @xmath58 , neutrons @xmath59 and @xmath0 : @xmath60 where @xmath61 , @xmath62 is the projector operator over the particle of @xmath63 type . an averaged sum on the third components of the @xmath0 particle angular momentum is understood . as in refs.@xcite the densities are evaluated by using cluster expansion techniques . the nucleonic obd are divided in two parts : @xmath64 the first part is given by the sum of the cluster diagrams containing only nucleons , and represents the nuclear bulk contribution to the obd . one of the diagrams contributing to this part is the @xmath12 diagram of fig.[fig : diag ] . the second part of eq . [ ro1l ] , called rearrangement term , is obtained by summing all the diagrams containing the @xmath0 as an internal particle , like the diagram @xmath65 of fig.[fig : diag ] , and provides the modification of the nucleon obd due to the presence of the @xmath3 impurity . a mean - field description of the hypernucleus does not provide any rearrangement term . finally , the diagrams where the @xmath0 is an external particle , like the @xmath66 diagram of fig.[fig : diag ] , are summed in the -obd , @xmath67 . the separation of the one - body densities in bulk , rearrangement and @xmath0 terms will be very useful for the calculation of the @xmath0 binding energy in the nucleus and it will be exploited also for the calculation of other quantities occurring in the fhnc scheme . in analogy with the one body densities , we define the two body densities as : @xmath68 where @xmath69 and @xmath70 is one of the operators defined in eq.([vn ] ) . the @xmath0n two body densities are defined as : @xmath71 where @xmath72 are the operators of the @xmath0n interaction ( [ vl ] ) , and @xmath24 is the space exchange operator between the @xmath0 and the k - nucleon . in the fhnc approach the tbd are described in terms of two - body correlation functions @xmath73 and @xmath74 , and of uncorrelated densities . the latter are built from a single particle basis , defined in the present paper as a set of @xmath1 coupled wave functions of the type : @xmath75 where @xmath76 . @xmath77 are the spherical harmonics , @xmath78 are the clebsch gordan coefficients and @xmath79 the two - component pauli spinors . in the nucleonic case , the uncorrelated two body densities are : @xmath80 the explicit expressions of the parallel , @xmath81 , and antiparallel @xmath82 densities are given in ref . @xcite together with the other purely nucleonic fhnc quantities . we briefly discuss here the quantities involving the @xmath0 . the uncorrelated @xmath0 one body density is given by : @xmath83 the one - body @xmath0 uncorrelated density matrix is needed to calculate diagrams where the spatial coordinates between the @xmath3 and one nucleon are exchanged . its expression is : @xmath84 where we have indicated with @xmath85 the legendre polynomial of @xmath86th degree and @xmath87 the angle between the two vectors . a detailed description of the fhnc equations in the nucleonic matter , in a @xmath1 coupled single particle basis computational scheme , has been done in ref . we do not give here the equations for this case , but we rather present and discuss the changes related to the presence of the @xmath0 . the fhnc densities are : @xmath88~ , \\ c_{\lambda } & = & \int d { { \bf r}}~\rho_0^{\lambda } ( r)~\exp [ u_d^{\lambda } ( r)]~ , \\ \rho_{2,1}^{\lambda \beta } ( { { \bf r}}_1 , { { \bf r}}_2 ) & = & \xi_d^{\lambda } ( r_1 ) \left ( \xi_d^\beta ( r_2 ) g_{dd}^{\lambda \beta } ( { { \bf r}}_1 , { { \bf r}}_2 ) + \xi_e^\beta ( r_2 ) g_{de}^{\lambda \beta } ( { { \bf r}}_1 , { { \bf r}}_2 ) \right)~ , \\ \label{r=0 } \rho_{2,2}^{\lambda \beta } ( { { \bf r}}_1 , { { \bf r}}_2 ) & = & 0 ~. \end{aligned}\ ] ] the functions @xmath89 , @xmath90 and @xmath91 are analogous to their nucleonic partners whose expressions are given in the appendices a and b of ref . @xcite , and they are obtained with the substitution @xmath92 . in extending the expressions given in the reference , we should remember that there is no sum on @xmath0 , since it is just a single external impurity , and that we consider the same correlation function between @xmath0 and any type of nucleon : @xmath93 with @xmath94 . eq.([r=0 ] ) shows that , in our case , the spin component of the two - body density , @xmath95 , is zero . the reason lies in the fact that only diagrams corresponding to exchanges between identical particles contribute to the spin component of the two body density for jastrow correlated wave functions . as a consequence they do not affect the @xmath0n tbd . direct diagrams , where the particles are not exchanged , may contribute to @xmath95 only if the correlation contains spin dependent components ( which is not our case ) . moreover , this type of contribution has been found to be generally small @xcite . the evaluation of the diagrams containing the space exchange @xmath0n potential requires the introduction of a space exchange density , given by : @xmath96 where @xmath97 is defined as in appendix b of @xcite , with the substitutions @xmath98 in the construction of the @xmath99 functions . the expressions of the rearrangement terms are quite lengthy since the @xmath0 may be present in any of the elements involved in the densities . we give below the expression of the contributions to the one body density : @xmath100 + u_{e,\lambda}^\alpha ( r ) \right)~\exp [ u_d^\alpha ( r ) ] \\ & = & \xi_d^\alpha ( r ) u_{d,\lambda}^\alpha ( r ) + \xi_e^\alpha ( r ) u_{e,\lambda}^\alpha ( r ) = \xi_{d,\lambda}^\alpha ( r)~ , \\ \xi_{e,\lambda}^\alpha ( r ) & = & u_{d,\lambda}^\alpha ( r)~\exp [ u_d^\alpha ( r)]~. \end{aligned}\ ] ] the @xmath101 functions are given in ref.@xcite , while the @xmath102 and the rearrangement parts of the two body densities are given in the appendix . in this section we discuss the evaluation of the variational energy of a @xmath0 hypernucleus in the framework of the fhnc approach . the potential energy is separated into its nucleonic and @xmath0 pieces : @xmath103 the expectation values can be expressed in terms of the two body distribution functions as : @xmath104 where @xmath105 indicates the average on the third components of the @xmath0 angular momentum and @xmath106 is the bulk nucleonic potential energy . because the minimizations of @xmath44 and @xmath107 must be , in principle , carried on independently , the nucleon nucleon jastrow correlation , @xmath108 , might be different in the two cases , as well as the nucleon single particle wave functions , @xmath109 . however , we found that in our cases @xmath108 and @xmath109 do not practically change in going from the @xmath12 to the @xmath110 system . as a consequence , in our approach the nucleonic part of the hypernucleus wave function has been kept the same as in nucleus . this fact allows for obtaining the potential energy contribution to the binding energy of the @xmath3 by subtracting the pure nucleus @xmath111 from @xmath112 . the remaining part is separated in two contributions : the potential energy due to the interaction @xmath3nucleon , or interaction energy , and the modifications of the nucleon nucleon potential energy due to the presence of the @xmath3 , or rearrangement energy . we put in evidence these contributions by rewriting : @xmath113 with @xmath114 and @xmath115 the separation in interaction and rearrangement terms can also be done for the @xmath0 kinetic energy , evaluated here by means of the jackson feenberg expression @xcite . following an analysis similar to that done for the potential energy , we obtain : @xmath116 the interaction kinetic energy is given by : @xmath117 ~\rho_{2,1}^{\lambda \alpha}({{\bf r}}_1,{{\bf r}}_2 ) \,\,\ , , \end{aligned}\ ] ] with @xmath118 \,\,\ , .\ ] ] the expressions of @xmath119 $ ] and @xmath120 can be obtained from ref.@xcite ( eq.(16 ) and appendix c ) . the rearrangement part of the kinetic energy is : @xmath121 \rho_{2,1,\lambda}^{\alpha \beta}({{\bf r}}_1,{{\bf r}}_2 ) \nonumber \\ & & -\frac{\hbar^2}{4 m } \sum_{\alpha = p , n } \int d { { \bf r}}_1 \rho_{t1}^{\alpha}({{\bf r}}_1 ) \xi_{e,\lambda}^\alpha ( { { \bf r}}_1 ) \\ & & + \frac{\hbar^2}{4 m } \sum_{\alpha = p , n } \int d { { \bf r}}_1 d { { \bf r}}_2 \rho_{t2}^{\alpha}({{\bf r}}_1 , { { \bf r}}_2 ) \left [ \xi_e^\alpha ( { { \bf r}}_1 ) \xi_e^\alpha({{\bf r}}_2 ) g_{dd,\lambda}^{\alpha \alpha } ( { { \bf r}}_1,{{\bf r}}_2 ) + \right . \nonumber \\ & & \left . \left(\xi_{e,\lambda}^\alpha ( { { \bf r}}_1 ) \xi_e^\alpha({{\bf r}}_2)+ \xi_e^\alpha ( { { \bf r}}_1 ) \xi_{e,\lambda}^\alpha({{\bf r}}_2 ) \right ) g_{dd}^{\alpha \alpha } ( { { \bf r}}_1,{{\bf r}}_2 ) \right ] \nonumber \\ & & - \frac{\hbar^2}{2 m } \sum_{\alpha = p , n } \int d { { \bf r}}_1 d { { \bf r}}_2 \bigl [ \rho_{t3,p}^{\alpha}({{\bf r}}_1 , { { \bf r}}_2 ) h_{cc , p,\lambda}^{\alpha}({{\bf r}}_1 , { { \bf r}}_2 ) \nonumber \\ & & + \rho_{t3,a}^{\alpha}({{\bf r}}_1 , { { \bf r}}_2 ) h_{cc , a,\lambda}^{\alpha}({{\bf r}}_1 , { { \bf r}}_2 ) \bigr]~ , \nonumber \end{aligned}\ ] ] where @xmath122 \nonumber \\ & + & \xi_e^\alpha ( r_1)\xi_{e,\lambda}^\alpha ( r_2 ) g_{dd}^{\alpha \alpha } ( { { \bf r}}_1 , { { \bf r}}_2 ) \left ( n_{cc , d}^\alpha ( { { \bf r}}_1 , { { \bf r}}_2 ) + e_{cc , d}^\alpha ( { { \bf r}}_1 , { { \bf r}}_2 ) \right ) \nonumber \\ & + & \xi_{e}^\alpha ( r_1 ) \bigl [ \xi_e^\alpha ( r_2 ) \bigl ( \left ( g_{dd}^{\alpha \alpha } ( { { \bf r}}_1 , { { \bf r}}_2 ) -1 \right ) n_{cc , d,\lambda}^\alpha ( { { \bf r}}_1 , { { \bf r}}_2 ) \\ & + & g_{dd,\lambda}^{\alpha \alpha } ( { { \bf r}}_1 , { { \bf r}}_2 ) n_{cc , d}^\alpha ( { { \bf r}}_1 , { { \bf r}}_2 ) + n_{cc , d,\lambda}^{(x ) \alpha } ( { { \bf r}}_1 , { { \bf r}}_2 ) \bigr ) \nonumber \\ & + & \left(\xi_e^\alpha ( r_2 ) -1 \right ) n_{cc , d,\lambda}^{(\rho ) \alpha } ( { { \bf r}}_1 , { { \bf r}}_2 ) \nonumber \\ & + & \xi_e^\alpha ( r_2 ) \left ( g_{dd,\lambda}^{\alpha \alpha } ( { { \bf r}}_1 , { { \bf r}}_2 ) e_{cc , d}^\alpha ( { { \bf r}}_1 , { { \bf r}}_2 ) + g_{dd}^{\alpha \alpha } ( { { \bf r}}_1 , { { \bf r}}_2 ) e_{cc , d,\lambda}^\alpha ( { { \bf r}}_1 , { { \bf r}}_2 ) \right ) \bigr ] \nonumber .\end{aligned}\ ] ] again the expressions of @xmath123 for @xmath124 can be found in the appendix c of @xcite . the difference of the center of mass kinetic energies is given by : @xmath125 and the expression of @xmath126 is given in eq . ( 23 ) of ref.@xcite . we have studied the ground state structure of those hypernuclei having @xmath53 and @xmath54 values corresponding to doubly closed shell nuclei in @xmath1coupling , @xmath127c , @xmath5o , @xmath128ca , @xmath129ca , @xmath130zr and @xmath131pb . the same set of nuclei where the neutron in the highest neutron single particle level has been substituted by a @xmath0 hyperon ( @xmath132c , @xmath133o , @xmath134ca , @xmath135ca , @xmath136zr and @xmath6pb ) have been also considered . in order to deal in fhnc with the partially occupied neutronic level , we adopt the following procedure : the degeneracy factor @xmath137 , multiplying the partially occupied state contribution to the uncorrelated nucleonic obd , has been replaced by the factor @xmath138 . in this way , the obd results normalized to @xmath139 . moreover , in the spin parallel part of the uncorrelated nucleonic tbd , @xmath140 , the @xmath137 factor has been substituted by @xmath141 , while the ( much smaller ) antiparallel part , @xmath142 , remains unchanged . this choice ensures a correct normalization of the tbd . two different mean field potentials have been used to build the single particle nuclear wave functions . the first potential is a harmonic oscillator ( ho ) well with the same constant , @xmath143 , for protons and neutrons ; the second choice is the woods saxon ( ws ) potential used in the calculations of ref.@xcite . the parameters of the ws potential are different for protons and neutrons . we have always used a harmonic oscillator potential for the @xmath3 single particle potential . the correlation functions have a gaussian form : @xmath144 in order to calculate @xmath34 , we first minimize the bulk energy of the nucleus , @xmath145 . the minimization has been performed for each nucleus in two ways , depending on the nucleonic single particle used . in the case of the ho potential , the nuclear correlation , @xmath108 , has been taken from nuclear matter and the nucleus energy has been obtained by varying only the values of the oscillator constant of the ho mean field . for the woods saxon case , the mean field potential has been kept fixed as in ref . @xcite and the parameters of the correlation have been varied . we consider the ws model as the most realistic one , since its parameters have been determined to reproduce at best the nuclear densities . it must be noticed that , in the region of the variational space corresponding to the energy minimum , the energy itself is rather insensitive to small changes of the ws parameters . therefore it is possible to reproduce the densities and the radii without excessively spoiling the quality of the minimum . for the description of the nucleonic part of the hypernucleus we take the same @xmath108 and mean field potential as in the corresponding nucleus . then , @xmath146 is minimized by changing only the @xmath3ho constant and the parameters of the @xmath3nucleon correlation function , @xmath147 . the parameters of the nuclear correlation ( [ corr ] ) with the ho mean field are @xmath148 and @xmath149 @xmath31 . the nuclear matter binding energy per nucleon with this correlation and the s3 potential , at saturation density ( @xmath150 @xmath151 ) , is @xmath152=14.43 mev , close enough to the empirical value , @xmath153=16 mev , and comparable with the best , more sophisticated potentials on the market . for each nucleus the value of the oscillator constant , @xmath143 , has been fixed to get the energy minimum . in table [ tab : oscc ] we compare our ho and ws binding energies and rms charge radii with their experimental values . a general reasonable agreement is found , considering the relative simplicity of both wave functions and interactions . in particular , the ws radii are close the experimental ones . the @xmath10c nucleus represents somehow an anomaly , having a good estimate of the radius but a very small binding energy . the origin of this disagreement is still under investigation . it may be related to the inadequacy of a spherical model description of this nucleus . the parameter @xmath39 of the density dependent potential ( ddp ) , described in section ii , has been chosen to reproduce the @xmath3 binding energy in nuclear matter , @xmath154 . we find @xmath155 fm@xmath57 and @xmath157 mev and this value of the parameter @xmath39 has been used in the remaining finite systems calculations . the parameters of the @xmath3ho mean fields and of the @xmath158 correlations at the @xmath146 minimum are shown in table [ tab : osck ] . in table [ tab : detail ] we compare the contributions to the @xmath3 binding energy in the doubly closed shell hypernuclei calculated with the ho nucleonic mean field with those evaluated in nuclear matter we have also separated the interaction ( i ) and rearrangement ( r ) terms in the kinetic and potential energies . the upper part of the table gives the results with the two body @xmath3n interaction only and the related @xmath0 binding energy , @xmath159 . in the lower part of the table the contribution of the three - body @xmath0nn interaction of ref.@xcite and the corresponding binding energy @xmath160 are shown . finally we present the results obtained with the ddp potential . since the empirical value of @xmath34 in @xmath5o is estimated to be about 13 mev @xcite , the result obtained with the @xmath0n force only is too attractive . the vmc calculation of ref . @xcite gives @[email protected] mev , in qualitative agreement with our value . the empirical binding energy was reproduced in the vmc approach by the inclusion of an explicit @xmath0nn force . using the same three body potential , we obtain @xmath163=6.81 mev . we have already pointed out that central , jastrow , correlations underestimate the attractive contribution of the @xmath3nn interaction , as clearly appears from the @xmath164 value . this potential , when used in conjunction with jastrow correlated wave functions , does not even bind the @xmath3 in heavy nuclei . tensor like correlations are needed in order to effectively use potentials induced by one and more pion exchanges in spherically symmetric systems@xcite . the introduction of a ddp , fitted to the @xmath3 binding in nuclear matter , brings @xmath34 reasonably close to the empirical estimate , even with a relatively simple correlation . all the results presented hereafter have been obtained with the ddp . in table [ tab : sho ] we give the binding energies of a @xmath3 in its @xmath165 ground state , calculated with the ho single particle potential for several hypernuclei and nuclear matter . we explicitly show the overall interaction and rearrangement contributions . the difference of the center of mass kinetic energies has been included in the rearrangement part . the increase of the @xmath0 binding energy along a is mostly produced by the interaction energy and , specifically , by its potential energy part . in contrast , the dependence of the rearrangement energy on @xmath12 is much weaker . we found an analogous behavior the @xmath3 energies in the @xmath166 and @xmath167 states . to complete the information we give in table [ tab : pho ] , the results for the @xmath3 energies in the @xmath166 and @xmath167 states with the nucleonic ho mean field . in table [ tab : ws ] we compare the @xmath0 binding energies obtained with a woods saxon nucleonic mean field with the experimental energies . for the @xmath136zr the comparison is done with the energies measured in @xmath168y . the results for the @xmath0 binding energies are summarized in fig.[fig : ene ] , where they are presented as a function of @xmath169 and compared with the experimental energies of refs . @xcite ( dots ) , @xcite ( triangles ) and @xcite for the @xmath170c . the figure gives the @xmath0 energies in the @xmath165 , @xmath166 and @xmath167 states . the quality of the agreement with the experiment is rather good , in spite of the simplicity of the model . by inspecting the figure in more details , we find that all the calculations underbind the @xmath0 in carbon by @xmath171 , and provide a steeper variation of @xmath34 with respect to the experiments . this behavior resembles that already described in carbon nucleus . it is also worth noticing that , as expected , the heavier nuclei , zr and pb , are better described by the woods - saxon well than by the ho one . the effects of the @xmath0n correlations on the one body densities are shown in fig.[fig : dnuc ] , where the proton densities in four nuclei are compared . the dotted lines are the ipm densities , the dashed lines are the densities obtained in a purely nucleonic fhnc calculations , and , finally , the full curves represent the proton densities obtained when a @xmath0 hyperon is added in the @xmath16 wave to the doubly magic nucleonic core . in general , correlation effects are more important in oxygen and calcium than in the other two nuclei . moreover , the presence of the @xmath0 does not heavily modify the nucleonic densities . this is better shown in fig.[fig : core ] , where the differences between proton ( panel @xmath172 ) and neutron ( panel @xmath173 ) densities with and without the @xmath0 are given . the differences shown in the figure have been amplified by a factor 1000 . the heavier nuclei seem to be much more stable against deformations produced by the presence of the @xmath0 . this picture might change when the @xmath0 is inserted in the nucleonic core by substituting the last bound neutron . the differences , multiplied by 1000 , between the nucleonic densities of the isotopic hypernuclei with a+1 and a hadrons are given in fig.[fig : iso ] . the panels show the differences between the protons ( upper ) and neutron ( lower ) densities . the conservation of the charge and mass numbers for each hypernucleus implies that the curves in the upper panel are normalized to zero , while those in the lower panel are normalized to one . the analysis of this figure indicates that the nucleonic part is more perturbed if the @xmath0 substitutes a neutron having a low angular momentum , like in o , ca , and pb . in zr the neutron modified into a @xmath0 is lying on the @xmath174 level ( @xmath86=5 ) , and its wave function is peaked at the surface of the nucleus , as shown in panel @xmath175 of fig.[fig : iso ] . in any case , heavy nuclei have again densities more rigid against deformations induced by the hyperon . the @xmath0 densities in @xmath5o , @xmath128ca , @xmath130zr and @xmath131pb for @xmath16 and @xmath58 waves are shown in fig.[fig : dlam ] . the dashed lines are the ipm densities and the full ones those obtained in the full calculations . results very similar have been obtained for the @xmath133o , @xmath134ca , @xmath136zr and @xmath6pb hypernuclei . again , the differences between the ipm and fhnc approaches are smaller for the heavier systems . in this work we have studied some properties of hypernuclei containing one @xmath3 hyperon in the framework of the correlated basis functions theory . the @xmath3 binding energies in the @xmath16 , @xmath58 and @xmath176 states , its density and the rearrangement part of the one nucleon densities have been computed for hypernuclei whose nucleonic core has a doubly closed shell structure in @xmath1 coupling . the hyperon has been either added to the core ( @xmath127c , @xmath5o , @xmath128ca , @xmath129ca , @xmath130zr and @xmath131pb ) or substituted to a neutron in its highest shell ( @xmath132c , @xmath133o , @xmath134ca , @xmath135ca , @xmath136zr and @xmath6pb ) . the correlated wave function contains gaussian , jastrow correlations for the nn and @xmath0n pairs . the correlations act on a hypernuclear shell model wave function generated @xmath177 by a harmonic oscillator or woods saxon mean field for the nucleonic part and @xmath178 by a harmonic oscillator well for the @xmath3 single particle potential . cluster expansion and fermi hypernetted chain resummation technique have been used . the energy of the system has been minimized with respect to variations on the parameters of the wave function . we have employed semirealistic hamiltonians with the s3 two nucleon potential of afnan and tang , the bodmer and usmani @xmath3n potential and either a model of @xmath3nn three particle interaction , still proposed by bodmer and usmani , or a density dependent modification of the @xmath3n force , in order to take into account many hadron interactions . using the @xmath0nn potential with jastrow correlations severely underestimates the @xmath0 binding energy , since the important , attractive tensor component of the two pion exchange potential does not contribute in this model . in contrast , the density dependent modification of the @xmath0n potential , fitted to the @xmath0 binding in nuclear matter , provides results for the @xmath0 binding energy in nuclei in encouraging agreement with the experimental data . in fact , the disagreement is less than 10@xmath179 in all hypernuclei , for both @xmath16 and @xmath58 @xmath0 states . the only exception is provided by the carbon hypernuclei , showing a discrepancy with the experiments of @xmath18030@xmath179 that could be ascribed to the inefficiency of a spherically symmetric description of these systems . moreover , our analysis stresses the importance of adopting a good independent particle wave function as a starting point . in fact , the woods saxon nucleonic mean field , fitted to the experimental one nucleon densities , gives a better description of the @xmath0 binding than the harmonic oscillator model , especially in the heavy hypernuclei . the nucleonic core polarization effects due to the presence of the hyperon are more important in the lighter hypernuclei than in the heavier ones , both for the energy and for the one nucleon densities . the rearrangement contributions to the @xmath0 binding energy go from @xmath181 in c to @xmath182 in pb and nuclear matter . as far as the @xmath0density is concerned , the influence of the correlations is much more visible in o and ca than in the heavier zr and pb . the results shown in this paper have been obtained by means of relatively simple hamiltonians , which entitle to use equally simple wave functions . however , hyperon nucleon interactions may now be built on more microscopic grounds . it is to be expected that an effective use of these hamiltonians will ask for a more sophisticated correlation , with a strong dependence on the relative state of the correlated pair . the situation closely resembles that in doubly closed shell nuclei and in nuclear matter , where the introduction of modern potentials has prompted the extension of the correlated basis functions and fhnc theories in that direction . another interesting field of application of this technology is the study of hypernuclei with two or more hyperons , especially in view of a better determination of the hyperon hyperon interaction . this work has been partially supported by murst through the _ progetto di ricerca di interesse nazionale : fisica teorica del nucleo atomico e dei sistemi a molticorpi _ and by the spanish direccin general de ciencia y tecnologa under project pb98 - 1318 . in this appendix we collect the different contributions to the rearrangement part of the one and two body densities , where one of the internal particles must be a @xmath3 hyperon . following this rule , we obtain : the rearrangement parts of the two body distribution functions are : @xmath184 the equations for the nodal diagrams are : @xmath185~~ , \nonumber \end{aligned}\ ] ] with @xmath186 and @xmath187 finally , the equations for the @xmath188 nodal are : @xmath189 \nonumber \\ -n_{cc , d,\lambda}^{(\rho ) \alpha } ( { { \bf r}}_1,{{\bf r}}_2 ) & = & \bigl ( \rho_{d}^{\alpha } ( { { \bf r}}_1,{{\bf r}}_3 ) \xi_{e,\lambda}^{\alpha } ( r_3 ) \| g_{cc , p}^{\alpha } ( { { \bf r}}_3,{{\bf r}}_2 ) \bigr)+ \\ & & \bigl ( \rho_{d}^{\alpha } ( { { \bf r}}_1,{{\bf r}}_3 ) \xi_{e}^{\alpha } ( r_3 ) \| x_{cc , p,\lambda}^{\alpha } ( { { \bf r}}_3,{{\bf r}}_2)+n_{cc , p,\lambda}^{(x ) \alpha } ( { { \bf r}}_3,{{\bf r}}_2 ) \bigr)+ \nonumber \\ & & \bigl ( \rho_{d}^{\alpha } ( { { \bf r}}_1,{{\bf r}}_3 ) ( \xi_e^{\alpha } ( r_3)-1 ) \| n_{cc , p,\lambda}^{(\rho ) \alpha } ( { { \bf r}}_3,{{\bf r}}_2 ) \bigr)+ \nonumber \\ & s_d \bigl [ & \bigl ( \rho_{e}^{\alpha } ( { { \bf r}}_1,{{\bf r}}_3 ) \xi_{e,\lambda}^{\alpha } ( r_3 ) \| g_{cc , a}^{\alpha } ( { { \bf r}}_3,{{\bf r}}_2 ) \bigr ) + \nonumber \\ & & \bigl ( \rho_{e}^{\alpha } ( { { \bf r}}_1,{{\bf r}}_3 ) \xi_{e}^{\alpha } ( r_3 ) \| x_{cc , a,\lambda}^{\alpha } ( { { \bf r}}_3,{{\bf r}}_2)+n_{cc , a,\lambda}^{(x ) \alpha } ( { { \bf r}}_3,{{\bf r}}_2 ) \bigr ) + \nonumber \\ & & \bigl ( \rho_{e}^{\alpha } ( { { \bf r}}_1,{{\bf r}}_3 ) ( \xi_e^{\alpha } ( r_3)-1 ) \| n_{cc , a,\lambda}^{(\rho ) \alpha } ( { { \bf r}}_3,{{\bf r}}_2 ) \bigr ) \ \ \ \bigr]~~ , \nonumber\end{aligned}\ ] ] where @xmath190 . if @xmath191 , then @xmath192 and @xmath193 ; if @xmath194 , then @xmath195 and @xmath196 . 99 r. b. wiringa , s. c. pieper , j. carlson , and v. r. pandharipande , phys . c * 62 * , 014001 ( 2000 ) . a. fabrocini , f. arias de saavedra , and g. co , phys . c * 61 * , 044302 ( 2000 ) . j. h. heisenberg , and b. mihaila , phys . rev . c * 59 * , 1440 ( 1999 ) . a. fabrocini , and g. co , phys . c _ in press _ , arxiv : nucl - th/0012048 . m. rayet , ann . phys . ( n.y . ) * 102 * , 226 ( 1976 ) ; nucl . phys . * a367 * , 381 ( 1981 ) . d. j. millener , c. b. dover and a. gal , phys . rev . c * 38 * , 2700 ( 1988 ) . j. schaffner , c. greiner , and h. stcker , phys . c * 46 * , 332 ( 1992 ) ; j. schaffner , c. b. dover , a. gal , c. greiner , d. j. millener and h. stcker , ann . ( n.y . ) * 235 * , 35 ( 1994 ) . z. ma , j. speth , s. krewald , b. chen and a. reuber , nucl a608 * , 385 ( 1996 ) . k. tsushima , k. saito , and a. w. thomas , phys . lett . * b411 * , 9 ( 1997 ) ; d. vretenar , w. plsch , g. a. lalazissis , and p. ring , phys . c * 57 * , r1060 ( 1998 ) . p. m. m. maessen , t. a. rijken , and j. j. de swart , phys . c * 40 * , 2226 ( 1989 ) . b. holzenkamp , k. holinde , and j. speth , nucl . phys . * a500 * , 485 ( 1989 ) . t. a. rijken , v. g. j. stoks , and y. yamamoto , phys . rev . c * 59 * , 21 ( 1999 ) . i. vidaa , a. polls , a. ramos , and m. hjorth jensen , nucl . phys . * a644 * , 201 ( 1998 ) . j. cugnon , a. lejeune , and h.j . schulze , phys . c * 62 * , 064308 ( 2000 ) . q. n. usmani and a. r. bodmer , phys . c * 60 * , 055215 ( 1999 ) . s. rosati , in _ from nuclei to particles _ , proceedings of the international school of physics `` enrico fermi '' , course lxxix , edited by a. molinari ( north holland , amsterdam 1982 ) , p. 73 . i. r. afnan , and y. c. tang , phys . rev . * 175 * , 1337 ( 1968 ) . r. guardiola , a. faessler , h. mther , and a. polls , nucl . a371 * , 79 ( 1981 ) . bodmer , and q.n . usmani , nucl . a477 * , 621 ( 1988 ) . a. fabrocini , f. arias de saavedra , g. co , and p. folgarait , phys . c * 57 * , 1668 ( 1998 ) . q. n. usmani , nucl . * a340 * , 397 ( 1980 ) . g. co , a. fabrocini , s. fantoni , and i. e .. lagaris , nucl . phys . * a549 * , 439 ( 1992 ) . f. arias de saavedra , g. co , a. fabrocini , and s. fantoni , nucl . phys . * a605 * , 359 ( 1996 ) . usmani , s. pieper , and q.n . usmani , phys . c * 51 * , 2347 ( 1995 ) . b. friedman , and v.r . pandharipande , nucl . phys . * a361 * , 502 ( 1981 ) . a. fabrocini , and a. polls , phys . b25 , 4533 ( 1982 ) . j. boronat , f. arias de saavedra , e. buenda , and a. polls , jour . of low temp . phys . * 94 * , 325 ( 1994 ) . g. co , a. fabrocini , and s. fantoni , nucl . * a568 * , 73 ( 1994 ) . t. hasegawa _ et al . _ , c * 53 * , 293 ( 1996 ) . et al . _ , phys . rev . lett . * 66 * , 2585 ( 1991 ) . m. may _ et al . 78 * , 4343 ( 1997 ) . .optimum values of the nucleonic oscillator constant , @xmath143 ( in fm ) , of the rms radii @xmath197 ( in fm ) and of the binding energy per particle @xmath198 ( in mev ) for different nuclei . the woods - saxon potential parameters are those of ref . [ cols= " < , > , > , > , > , > , > , > " , ]	we study the properties of hypernuclei containing one @xmath0 hyperon in the framework of the correlated basis function theory with jastrow correlations . fermi hypernetted chain integral equations are derived and used to evaluate energies and one body densities of @xmath0hypernuclei having a doubly closed shell nucleonic core in the @xmath1 coupling scheme , from carbon to lead . we also study hypernuclei having the least bound neutron substituted by the @xmath0 particle . the semi - realistic afnan and tang nucleon - nucleon potential and bodmer and usmani @xmath0-nucleon potential are adopted . the effect of many body forces are considered by means either of a three body @xmath0-nucleon - nucleon potential of the argonne type or of a density dependent modification of the @xmath0-nucleon interaction , fitted to reproduce the @xmath0 binding energy in nuclear matter . while jastrow correlations underestimate the attractive contribution of the three body @xmath0 interaction , the density dependent potential provides a good description of the @xmath0 binding energies over all the nuclear masses range , in spite of the relative simplicity of the model . 16 cm 1.0 cm 1.cm
the highest redshift quasar known at the present time , at @xmath7 , is the first one to show a complete gunn - peterson trough ( becker et al . 2001 ; djorgovski et al . this suggests that the reionization of hydrogen was probably completed near the epoch @xmath2 . this conclusion is not yet completely certain because an ionized medium can produce a complete gunn - peterson trough if the neutral fraction is everywhere high enough ( owing to a low intensity of the ionizing background ) . however , the abrupt change with redshift near @xmath2 of the fraction of transmitted flux in the ly@xmath0 region of the spectrum ( becker et al . 2001 ) implies a rapid increase of the mean free path of ionizing photons with time ( fan et al . 2002 ) , at least in the region of the line of sight to the @xmath7 qso , just as expected when the last low - density regions of the intergalactic medium are ionized ( e.g. , gnedin 2000 ) . ending reionization by @xmath2 requires that at least one ionizing photon has been emitted to the intergalactic medium for each baryon in the universe , in addition to the amount needed to compensate for any recombinations that have taken place . this implies that the emissivity of ionizing photons ( defined as the mean number of photons emitted per unit time and volume ) should be greater than the mean number density of baryons divided by the age of the universe at @xmath8 , unless the emissivity had declined rapidly with time at @xmath9 . it is possible to infer the emissivity at lower redshift , by using the intensity of the ionizing background measured from the mean transmitted flux in the ly@xmath0 forest , and the mean free path of ionizing photons determined from the observed abundance of lyman limit systems . the first of these two quantities , the intensity of the ionizing background , can be measured in the context of the present theoretical understanding of the ly@xmath0 forest as arising from a photoionized intergalactic medium that follows the gravitational evolution of primordial fluctuations ( see the review by rauch 1998 and references therein ) . the observed mean transmitted flux can then be used to infer the ly@xmath0 scattering optical depth at the mean hydrogen density and in hubble expansion , @xmath10 : @xmath11 here , @xmath12 is the hydrogen neutral fraction at mean density , @xmath13 is the gas temperature in units of @xmath14 k , and @xmath15 is the photoionization rate due to the cosmic background in units of @xmath16 . when the value of @xmath17 was still highly uncertain , the value of @xmath10 inferred from observations of the mean transmitted flux , @xmath18 at @xmath19 ( see rauch et al . 1997 , and table 2 in mcdonald et al . 2000 ) implied a lower limit to the value of @xmath17 given by the lower limit on @xmath15 that was inferred from the observed qsos . now , @xmath17 has been more accurately measured from the primordial deuterium abundance ( omeara et al . 2001 ) and the cmb spectrum of fluctuations ( de bernardis et al . 2002 ; pryke et al . 2002 ; spergel et al . 2003 ) , and we can use the value of @xmath20 determined from the ly@xmath0 forest to constrain the intensity of the ionizing background and its evolution with redshift . the mean free path of ionizing photons is related to the abundance of lyman limit systems , which have column densities @xmath21 . the reason is that the column density distribution of absorption systems is approximately of the form @xmath22 ( e.g. , petitjean et al . 1993 ) , so the average rate of absorption is dominated by high column densities , up to the column density at which the absorption systems become optically thick . the abundance of lyman limit systems has been reasonably well measured up to redshift @xmath1 . in this paper , we combine the measurements of the intensity and the mean free path of ionizing photons to infer their rate of emission at @xmath1 ( 2 ) . we then compare this emissivity to the observations of known sources at this redshift ( 3 ) . then , in 4 , we find the consequences for the evolution of the emissivity at higher redshift to satisfy the requirement of reionizing the igm by @xmath2 . finally , in 5 the consequences for the thomson optical depth of the cosmic microwave background photons are discussed . we will use the flat cosmological model with @xmath23 , @xmath24 , and @xmath25 . in this section we evaluate the emissivity of ionizing radiation at redshift @xmath1 . in terms of the photoionization rate , @xmath26 , and the mean free path of the ionizing photons , @xmath27 , the photon emissivity is given by @xmath28 , where @xmath29 is a frequency - averaged cross section . the result we will find in this paper is that the emissivity at @xmath1 has a rather small value , implying that it must not have declined much with redshift , up to a substantially earlier epoch at @xmath30 , in order to satisfy the requirement that reionization was completed by @xmath31 . with that in mind , we will evaluate the maximum allowed value of @xmath26 and of @xmath32 at @xmath1 , which is the highest redshift at which the mean free path @xmath27 has been observationally determined . the most recent measurements of the mean transmitted flux through the ly@xmath0 forest , @xmath33 , were reported by bernardi et al . ( 2002 ) , from a large sample of qsos of the sloan digital sky survey . from their fig.4 , the effective optical depth [ defined as @xmath34 at @xmath1 is @xmath35 , where we have estimated the error from the errorbars and scatter of their measured points at various redshift bins . hence , @xmath36 at @xmath1 . note that previous values of @xmath33 from a small sample of high resolution spectra ( schaye et al . 2000 , mcdonald et al . 2000 ) were higher , but this was probably because their method underestimated the qso continuum flux , which becomes increasingly difficult to set by fitting regions between absorption lines as the redshift increases . we now use the results of mcdonald & miralda - escud ( 2001 , hereafter mm01 ) to infer the photoinization rate @xmath15 . according to their figure 2 , the value @xmath36 implies a photoionization rate of @xmath37 at @xmath38 . this value needs to be corrected from the model used in mm01 to the model we use here , and to @xmath1 , where we will choose all the parameters to maximize @xmath15 , and therefore infer an upper limit . first of all , the relationship between @xmath33 and @xmath10 in equation ( 1 ) depends on the amplitude of the mass power spectrum at the jeans scale of the ly@xmath0 forest : a lower amplitude of fluctuations implies that the gas is less concentrated in collapsed regions that correspond to saturated absorption lines , and a greater fraction of the gas is left in low - density regions , where it contributes more efficiently to decrease the mean transmitted flux . the measured amplitude of the power spectrum of the ly@xmath0 forest ( see croft et al . 2002 and references therein ) is @xmath39 times lower than the amplitude in the model that was used by mm01 ( as shown in mcdonald et al . 2000 ) , which had the parameters @xmath40 and @xmath41 for the power spectrum . in other words , the model used by mm01 would match the ly@xmath0 forest power spectrum amplitude found by croft et al . if its normalization had been changed to @xmath42 , leaving all other parameters unchanged . since the amplitude of mass fluctuations grows proportionally to the scale factor @xmath43 ( the effects of the vacuum energy and the radiation are negligible at @xmath1 ) , the model with normalization @xmath44 has the same relation of @xmath33 and @xmath10 at @xmath1 as the model with normalization @xmath45 at @xmath46 . it was also found in mm01 that the value of @xmath10 inferred from a fixed @xmath33 depends on the amplitude approximately as @xmath47 ( this can in fact be checked by comparing the values of @xmath15 at @xmath38 and @xmath48 for fixed @xmath33 in fig . 2 of mm01 ) . because the limit @xmath49 was derived at @xmath38 , and @xmath50 , the power spectrum amplitude correction is a factor @xmath51 . we note that here we have taken the central measured value of croft et al . of the power spectrum normalization , instead of a lower value within their errorbar ( which would further raise the upper limit to @xmath15 by about 20% ) , because other measurements of the power spectrum amplitude favor a higher value instead ( see seljak 2001 ; van waerbeke et al . 2001 ; croft et al . 2002 ; jarvis et al . 2002 ; spergel et al . 2003 ) . second , the inferred @xmath15 depends on the gas temperature . mm01 assumed @xmath52 , but measurements of the gas temperature from the ly@xmath0 forest spectra have given values in the range @xmath53 ( theuns et al . 2002 and references therein ) . although the optical depth of a uniform medium is proportional to @xmath54 ( eq . 1 ) because of the temperature dependence of the recombination coefficient , the relation between @xmath33 and @xmath10 is also affected by the temperature because thermal broadening spreads some of the absorption from saturated spectral regions to other regions with less absorption . we use the result found in mcdonald et al . ( 2001 , their 7.4 ) that thermal broadening causes a change in @xmath10 of 27% of that due to the recombination coefficient , in the opposite direction , therefore giving an effective dependence @xmath55 . if we assume the lowest value of the temperature among the reported measurements , @xmath56 , then the temperature correction for the upper limit of @xmath15 is @xmath57 . finally , we correct for all other parameters used in mm01 ( @xmath58 , @xmath24 , @xmath59 , @xmath38 ) to the ones used here ( @xmath25 , @xmath24 , @xmath60 , @xmath1 ) , which we choose within the current uncertainties to maximize the inferred @xmath15 . the result of all these corrections ( see eq . 1 ) for the upper limit to @xmath15 at @xmath1 is @xmath61 storrie - lombardi et al . ( 1994 ) found that the rate of lyman limit systems ( defined as absorption systems with column density @xmath62 ) at @xmath1 is @xmath63 . in other words , the mean spacing between lyman limit systems is @xmath64 at @xmath1 . to find the relationship of @xmath65 to the mean free path of ionizing photons , we assume that the column density distribution of the absorbers is @xmath66 ( petitjean et al . the absorption probability per unit length , @xmath67 , for a photon at the hydrogen ionization edge frequency @xmath68 , is @xmath69 where @xmath70 . at a higher frequency @xmath71 , and approximating the photoionization cross section as proportional to @xmath72 , the mean free path is increased to @xmath73 . we note here that our mean free path in ( [ mfp0 ] ) is 1.5 times longer than that adopted in madau , haardt , & rees ( 1999 ) , and in schirber & bullock ( 2002 ) . for some reason , their opacity corresponds to a lyman limit system abundance of 5 per unit redshift , larger than reported by storrie - lombardi et al . ( 1994 ) . the photoionization rate is related to the emissivity @xmath74 , the cross section @xmath75 , and the mean free path as a function of frequency by @xmath76 here , we use the approximation that the mean free path is much less than the horizon distance ( the exact relation between the emissivity and background intensity involves an integration over time with redshift effects , a full knowledge of the spectrum , and reemission ; see miralda - escud & ostriker 1990 ; haardt & madau 1996 ; fardal , giroux , & shull 1998 . we treat reemission approximately in 2.4 ) . assuming that @xmath77 , @xmath78 , and @xmath79 , we have @xmath80 , where @xmath81 , and @xmath82 is the emissivity per unit frequency at the ionization edge . the emissivity integrated over frequency is @xmath83 . the emissivity can be conveniently expressed as the number of ionizing photons emitted per hubble time and for each atom in the universe : @xmath84 where @xmath85 is the comoving number density of atoms ( adding hydrogen and helium ) , @xmath86 , and @xmath87 is the hubble constant at @xmath1 . using @xmath88 , taking the lower limit for the mean free path ( to obtain an upper limit for the emissivity ) , and assuming @xmath89 , the final result is @xmath90 in other words , there are fewer than eight ionizing photons being emitted per hubble time and for each atom at @xmath1 . some of the photons that contribute to the photoionization rate in the intergalactic medium arise from reemission by the ly@xmath0 absorption systems ( see haardt & madau 1996 ) . for every hydrogen atom that is photoionized by absorbing a photon , and assuming ionization equilibrium , a recombination will take place , and about 40% of the recombinations are direct to the ground state , producing another ionizing photon . however , the majority of the ionizing photons that are not very close to the hydrogen ionization edge are absorbed deep within lyman limit systems , and the photons produced by recombinations in these regions are close to the ionization edge [ with a typical frequency separation from the lyman limit of @xmath91 and not likely to escape from the lyman limit system . because of this , reemission by hydrogen recombination increases the background intensity by only @xmath92 ( as is found , for example , by using the approximation that the reemitted photons escape only when produced in an absorber of column density @xmath93 , which is the column density for which the typical optical depth for the photon to escape is unity ) . some photons are also emitted by helium recombinations . makes little difference to the total budget of ionizing photons , because for each absorption by , one photon that ionizes hydrogen is produced by the subsequent recombination . because we did not include absorption by in our previous estimate of the mean free path , we do not include the reemission either . finally , photons with energy above @xmath94 rydbergs are mostly used to photoionize , and each recombination can produce between one and three photons capable of ionizing hydrogen . about 25% of the recombinations that are not direct to the ground state go to @xmath95 , producing a balmer continuum photon , which can ionize hydrogen . in addition , all the recombinations produce either a lyman series photon , or two continuum photons when the last transition is from 2s to 1s ; in @xmath96 of the recombinations , two continuum photons that can both ionize hydrogen are produced ( see osterbrock 1989 , tables 2.8 , 4.11 , and 4.12 ) . hence , on average 1.33 ionizing photons are produced for each that is photoionized . for a source spectrum @xmath97 , with @xmath98 , the fraction of photons used to ionize is @xmath99 , so reemission increases the emissivity by a factor @xmath100 . this factor can be reduced during the reionization , when the rate of recombinations is less than the rate of photoionizations . based on this discussion of the effect of reemission from absorbers , we reduce our upper limit in ( [ emisi ] ) for the total emissivity @xmath32 by a factor @xmath101 , to obtain the upper limit for the direct emissivity from sources @xmath102 at @xmath1 ( rounded up to one digit given the various uncertainties discussed ) , @xmath103 fan et al . ( 2001 ) measured the following qso luminosity function at high redshift ( @xmath104 ) from sdss : @xmath105 where @xmath106 is the luminosity of a qso with absolute ab magnitude @xmath107 at @xmath108 , and @xmath109 . using a power - law spectral index @xmath110 for the qso spectrum at @xmath111 ( telfer et al . 2002 find @xmath112 at @xmath113 for radio - quiet qsos , and the spectral slope is shallower at longer wavelengths ; e.g. , schmidt , schneider , & gunn 1995 ) , the total number of ionizing photons ( with @xmath114 ) emitted is @xmath115 . the luminosity function in equation ( [ lfun ] ) was observed down to @xmath116 by fan et al . ( 2001 ) ; the total emissivity from qsos clearly depends on the luminosity at which the slope of @xmath117 becomes shallower than @xmath118 . if we assume the form @xmath119^{-1}$ ] , which produces a similar slope of @xmath120 at low luminosities ( @xmath121 ) as observed at lower redshift ( e.g. , boyle et al . 1988 , 2000 ) , then the comoving emissivity is @xmath122 expressing this as the number of photons emitted per hubble time and per atom , at @xmath1 , we find , @xmath123 a lower limit to the emissivity from qsos is obtained if @xmath124 , @xmath125 . in order not to exceed the upper limit obtained in the last section , @xmath126 < 7 $ ] , we must require @xmath127 . we note , however , that low - luminosity qsos may be intrinsically self - absorbed at the lyman limit , reducing their contribution to the ionizing emissivity ( see alam & miralda - escud 2002 ) . the number of faint agns has been constrained by steidel et al . ( 2002 ) , who find that their abundance at @xmath19 and at an apparent ab magnitude @xmath128 ( or luminosities @xmath129 ) is about 3% of the lyman break galaxies abundance at a magnitude @xmath130 at @xmath19 . using the lyman break galaxy luminosity function given in the next subsection , we find that this implies that the turnover of the qso luminosity function should occur at @xmath131 , assuming again the form @xmath132^{-1}$ ] . if the value of @xmath133 were the same at @xmath1 as at @xmath19 ( but with the qso abundance having declined as @xmath134 at all luminosities , as in eq . [ [ lfun ] ] ) , qsos would contribute about half of the total emissivity required at @xmath1 from equation ( [ emiss ] ) . there are also limits from the hubble deep field ( conti et al . 1999 ) , but these probe qsos at much lower luminosities and do not provide a stronger constraint on @xmath133 . in summary , the available observations are consistent with a contribution to the ionizing emissivity from qsos that does not exceed the upper limit in ( [ emiss ] ) . the galaxy ultraviolet luminosity function at high redshift has been measured by steidel et al . they fit the luminosity function to a schechter function , @xmath135 after converting to the flat cosmological model with @xmath60 , @xmath24 , the parameters they find are @xmath136 , @xmath137 , and @xmath138 at @xmath139 . to convert this to the ionizing luminosity , we use the models of bruzual & charlot ( 1993 ) , for which the luminosity per unit frequency decreases by a factor 10 from @xmath139 to @xmath140 , for an age of @xmath141 years ( these models assume a salpeter mass function up to a maximum mass of @xmath142 and solar metallicity ) . using also the approximation of @xmath143 at @xmath144 , we find the ionizing luminosity is @xmath145 . the form of the galaxy luminosity function ( [ gflum ] ) is highly uncertain at @xmath146 , because the observations of steidel et al . ( 1999 ) are not deep enough to reach lower luminosities , and galaxies down to @xmath147 are only detected in the hubble deep field ( e.g. , madau et al . the hubble deep field counts at @xmath1 are significantly below the prediction from the model in equation ( [ gflum ] ) , but this may be due to galaxy clustering ( steidel et al . 1999 ; see their fig . 8) . if we include only galaxies with luminosities larger than @xmath148 , the total emissivity inferred from ( [ gflum ] ) is @xmath149 ~.\ ] ] steidel , pettini , & adelberger ( 2001 ) found that lyman break galaxies may have escape fractions of ionizing photons as large as 50% , from a possible detection of lyman continuum flux after coadding spectra of several galaxies . if this result were true , the luminous galaxies with @xmath150 would already produce the maximum emissivity in equation ( [ emiss ] ) . alternatively , @xmath151 may be much lower ( see giallongo et al . 2002 ; note also that the value @xmath152 measured by steidel et al . 2001 was for the subset of bluest quartile of the lyman break galaxies , which may not be representative ) , and lower luminosity objects may be responsible for the bulk of the emissivity . more accurate determinations of the escape fraction , the abundance of lyman limit systems , and the inferred photoionization rate from the ly@xmath0 forest transmitted flux will be needed to resolve this question . the highest redshift qso known at the present time ( becker et al . 2001 ) shows a complete gunn - peterson trough at @xmath31 , but the spectra of all the sources at lower redshift have detectable transmitted flux blueward of ly@xmath0 , indicating that most of the volume in the intergalactic medium had to be reionized by @xmath2 . we now address the contraint implied by this fact on the evolution of the ionizing emissivity , in view of its value at @xmath1 ( eq . [ emiss ] ) . reionization proceeds by the creation of ionized regions around individual sources ( arons & wingert 1972 ) . the mean free path of ionizing photons through the neutral medium is typically very short compared to the distance between sources , so reionization proceeds by the growth of the ionized regions and can be modeled in terms of the fraction of the volume that is ionized at time @xmath153 , @xmath154 , assuming that the remaining fraction @xmath155 is completely neutral . moreover , because the emitted photons are used to ionize atoms almost immediately after their emission , it follows that the change in the fraction of atoms that are ionized in the universe is equal to the ionizing photon emission rate per atom minus the average recombination rate of atoms . if @xmath156 is the comoving ionizing photon emissivity per atomic nucleus , @xmath157 is the average recombination rate per atom in ionized regions ( which includes the effect of an effective clumping factor of the ionized gas ) , and @xmath158 is the fraction of baryons that actually need to be ionized in the ionized fraction @xmath159 of the universe ( with the remaining @xmath160 fraction being the baryons that are in dense systems , such as stars or dense gas in galaxies , that do not need to be ionized by external sources ) , the equation for the evolution of @xmath159 is ( madau , haardt , & rees 1999 ; miralda - escud , haehnelt , & rees 2000 ) @xmath161 this equation ( apart from the factor @xmath158 ) is analogous to the one obtained for the problem of an expanding cosmological region around an ionizing source , which was discussed by shapiro & giroux ( 1987 ) and donahue & shull ( 1987 ) , where @xmath159 behaves as the volume of the region produced by a source of luminosity proportional to @xmath162 . it is also identical to the equation of radiative transfer , where @xmath163 plays the role of the intensity , and @xmath164 and @xmath165 play the role of the absorption and emission coefficients , respectively . the solution is found through the substitutions @xmath166 , and @xmath167 : @xmath168 where @xmath169 , and @xmath170 is the initial time when the first sources turn on . assuming that any clumping factor is constant , and that @xmath171 [ e.g. , for the cosmological constant model with a present matter density @xmath172 , we have @xmath173 , the recombination rate varies as @xmath174 . choosing @xmath175 as a fiducial time , and @xmath176 , we have : @xmath177 ~. \label{reisol}\ ] ] before presenting solutions of the reionization equation , we discuss the value that we adopt for the recombination rate , @xmath178 . the average recombination rate is affected by the clumping factor of the ionized gas , which we define as : @xmath179 . the subscript @xmath180 is used here to denote that the average is to be taken only over those regions of space where matter is predominantly photoionized by photons that have escaped to the intergalactic medium . this restriction is necessary because if we did not impose it , the clumping factor would be dominated by regions of dense gas that are locally ionized by sources that were not included in the emissivity in equation ( [ emisi ] ) , precisely because the radiation from these sources is locally absorbed . in fact , in the absence of any restriction on the regions over which one must average to obtain the clumping factor of ionized matter , then the clumping factor would obviously be very large and dominated by the densest stars . once stars are eliminated from consideration , there are still regions that form in dense clouds in the interstellar medium of galaxies around massive young stars , in which all the emitted photons are immediately absorbed locally . this emission is irrelevant for the reionization of the intergalactic medium , and it is best not to include in equation ( [ emisi ] ) , and at the same time not to include the dense gas ionized locally in the clumping factor . this definition of the clumping factor inevitably involves an arbitrary choice in defining when a photon is considered to be escaped , and therefore when a region is predominantly ionized by escaped photons rather than local ones . in practice , a photon can be defined to have escaped once it reaches an intergalactic region with gas density below some critical value . the calculation of the clumping factor can only be obtained from full radiative transfer simulations of reionization , but its value can differ among studies by large factors depending on the definition that is adopted ( see gnedin 2000 and references therein ) . if the restriction that escaped photons dominate the radiation intensity is not imposed , then very large values of the clumping factor can be obtained , which could only be used consistently by including the locally absorbed photons in the emissivity . here , we use the simple model of miralda - escud et al . ( 2000 ) for the clumping factor , which uses the approximation that the gas is ionized only up to an overdensity @xmath181 , and all the gas at higher densities is neutral , or otherwise is ionized locally and therefore does not contribute to the clumping factor . the overdensity @xmath182 is related to the mean free path @xmath27 that a photon can traverse before being absorbed . during reionization , the mean free path @xmath27 should be of order the size of the regions , or of order the typical separation between neighboring sources . for @xmath183 , or @xmath184 at @xmath2 in comoving units , the critical overdensity is @xmath185 , and the clumping factor is very close to unity ( see figs . 2d and 6 in miralda - escud et al . 2000 ) . it has been pointed out by haiman , abel , & madau ( 2001 ) that before reionization is complete , a large number of low - mass halos should be present and should contain dense gas accreted before it was heated by reionization . the effect of these low - mass halos would be to decrease the mean free path at a fixed overdensity @xmath182 , and to widen the distribution of the overdensity . to estimate the plausible change in the clumping factor due to these low - mass halos , we can assume that the structure of the atomic intergalactic medium before reionization is self - similar to that of the ionized intergalactic medium , but with the scales reduced in proportion to the jeans scale @xmath186 , so that the clumping factor should depend only on the ratio @xmath187 , and the amplitude of the primordial density fluctuations at the jeans scale . if the temperature increases by a factor @xmath188 during reionization ( from @xmath189 k to @xmath190 k ) , the jeans length increases by a factor @xmath191 . the amplitude of primordial density fluctuations in cold dark matter with @xmath60 and @xmath24 increases by a factor @xmath192 between the scales of @xmath193 to @xmath194 mpc ( roughly corresponding to the jeans scales at these two temperatures ) . therefore , the clumping factor of a medium in which all the low - mass halos formed from scales intermediate between the jeans scales of the atomic and ionized medium have survived , at @xmath195 and @xmath196 , should be the same as the clumping factor of a medium with no such surviving halos , evaluated at @xmath19 and @xmath197 . from figures 2b and 6 of miralda - escud et al . ( 2000 ) , we find this clumping factor to be 7 . we note that this is an upper limit to the true clumping factor , because the gas in the low - mass halos that might increase the clumping factor above the value of @xmath198 found previously is likely to evaporate on a time short compared to the hubble time , once these halos have been reached by an ionization front ( shapiro & raga 2001 ) . the gas may also be expelled by a small amount of internal star formation even before the halo is reached by external ionizing radiation . finally , a discussion is needed to decide if the case a or case b recombination coefficient should be used . ionizing photons of frequency @xmath71 will typically be absorbed in regions of column density @xmath199 , where @xmath200 is the lyman limit frequency . the sources typically have power - law spectra and most of the emitted photons have frequencies substantially above @xmath200 , especially if the spectra are hardened by internal absorption . most of the photons emitted by direct recombinations to the ground state , which have frequencies very close to the lyman limit , will therefore be produced in high column density systems and will be reabsorbed within the same system ( in addition , among the small fraction of photons from direct recombinations to the ground state that escape the system in which they have been produced , some will be redshifted below the lyman limit and will not be absorbed again ) . hence , these photons do not contribute to increase the cosmic ionizing background intensity for the most part . an alternative way to understand this argument is that , even though in principle we are in a case b situation because the recombination photons are reabsorbed , in practice the inclusion of the photons from recombinations increases the clumping factor by the ratio of the case a to case b recombination coefficients because these photons increase the ionization in dense regions , therefore yielding a recombination rate equivalent to using case a. therefore , we use the case a recombination coefficient . we now compute @xmath154 assuming models where the emissivity varies as @xmath201 for @xmath202 , and has the value @xmath203 at @xmath1 , i.e. , the maximum allowed value found previously ( eq . [ emiss ] ) . we choose the recombination coefficient @xmath204 , valid for case a and @xmath205 k , and electron density @xmath206 at @xmath2 , valid for @xmath25 and if helium is only once ionized . this yields a recombination rate per hubble time @xmath207 . in some cases , we increase this recombination rate by a clumping factor @xmath208 . we fix the fraction of baryons in the ionized regions that actually need to be ionized to @xmath209 ( independent of redshift ) , in reasonable agreement with the fraction of baryons that seem to be present in damped ly@xmath0 systems at @xmath210 ( lanzetta et al . 1991 ; storrie - lombardi et al . 1996 ) , and the fraction of baryons above the critical overdensity @xmath185 at @xmath2 in the model discussed above for the clumping factor ( miralda - escud et al . 2000 , fig . we solve equation ( [ reisol ] ) , starting at @xmath211 at some initial redshift @xmath212 when the emissivity is assumed to turn on suddenly . the parameters of the model that we vary are @xmath212 , @xmath213 , and @xmath208 , and we require all models to reach @xmath214 ( i.e. , the end of reionization ) at @xmath2 . the results of five representative models are shown in figure 1 . when @xmath215 , assuming the emissivity to remain constant satisfies the requirement of completing reionization by @xmath2 when the emissivity is turned on at @xmath216 . with only a slight decline of the emissivity , the initial redshift needs to be pushed to much higher values . for @xmath217 , which corresponds to a decline of the comoving emissivity by a factor of only @xmath218 between @xmath1 and @xmath219 , the sources need to start at very high redshift to reach @xmath214 at @xmath2 . naturally , changes in the emissivity at redshifts higher than @xmath220 are irrelevant for the requirement that reionization ends at @xmath2 , because most of the atoms ionized at this high redshift will recombine again . we therefore conclude that , for the case of no clumping , _ the ionizing emissivity at @xmath221 can not be much lower than its maximum value allowed at @xmath1_. in other words , the sources responsible for the emission of ionizing radiation can not decrease their overall comoving emissivity by more than a factor @xmath218 between @xmath1 and @xmath219 . allowing now for a clumping factor , an increase to @xmath222 already requires the model with constant emissivity ( @xmath223 ) to maintain the emissivity up to a very high initial redshift . a model where the same amount of radiative energy is emitted at every hubble time [ i.e. , @xmath224 , since @xmath225 , starting at very high @xmath212 , results in complete reionization at @xmath2 when @xmath226 . therefore , for a clumping factor as large as @xmath227 , as suggested previously for a model where the ionized gas in all low - mass halos is retained , the emissivity would need to substantially increase with redshift . we have shown that by combining the measurement of the ionizing emissivity @xmath32 at @xmath1 with the requirement that the reionization must end by @xmath2 , we can infer that the emissivity does not decrease with redshift by more than a factor @xmath218 from @xmath1 to @xmath219 . we can now ask how this evolutionary constraint compares to the evolution at @xmath228 that is implied by observations of the ly@xmath0 forest decrement and the mean free path between lyman limit systems . from mcdonald & miralda - escud ( 2001 ) , the photoionization rate ( or proper intensity of the background ) declines roughly as @xmath229 from @xmath230 to @xmath1 . the abundance of lyman limit systems per unit redshift is proportional to @xmath231 ( storrie - lombardi et al . 1994 ) , implying that the proper mean free path varies as @xmath232 . hence , the comoving emissivity evolves as @xmath233 , essentially constant within the likely errors of measurement for both the proper intensity and mean free path evolution . therefore , the results of 4.3 show that if the clumping factor is low , this constant emissivity needs to roughly maintained at the same value up to @xmath234 , while if the clumping factor is much larger than one , the emissivity needs to modify its evolution and increase strongly with redshift at @xmath202 . the optical depth of cosmic microwave background photons to electron scattering , @xmath235 , is given by @xmath236 where @xmath237 is the thomson cross section for electron scattering . the values of @xmath235 for the five models in figure 1 are indicated in the figure . we have assumed @xmath209 at all redshifts ; the 10% of the matter not included is supposed to be either gas that remains atomic or molecular in self - shielded regions , or stars . these results show that the minimum allowed value of @xmath235 , which occurs for models where the emissivity turns on at a redshift not much larger than 6 , is @xmath238 . the maximum values of the optical depth depend on the emissivity at high redshift . if the emissivity does not increase with redshift , then the ionized fraction of the volume at high redshift needs to be small due to the fast rate of recombinations ( i.e. , for constant emissivity , @xmath239 , and the integral in eq . [ [ taues ] ] converges rapidly at high @xmath240 ) . for a model with a constant emissivity up to @xmath241 and clumping factor equal to one , the optical depth can be increased only up to @xmath242 ( in this model , reionization ends at @xmath243 , but the reionization end could always be delayed by increasing the clumping factor as the fraction @xmath159 reaches a value close to unity ) . an emissivity increasing with redshift can obviously produce larger optical depths , and as mentioned before the value of the emissivity at @xmath244 does not affect the epoch at which reionization ends , so in principle the optical depth could be arbitrarily high if enough ionizing sources were present at very high redshift . however , in the flat cold dark matter model with @xmath23 , which is strongly supported by all observations of large - scale structure , the first stars form only at @xmath245 from @xmath246 peaks collapsing into the first low - mass halos ( e.g. , couchman & rees 1986 ; bromm , coppi , & larson 1999 ; abel , bryan , & norman 2000 ; ricotti , gnedin , & shull 2002 ; venkatessan , tumlinson , & shull 2003 ) , and so the emissivity can not continue to increase up to a very high redshift . at the same time , observations indicate a cosmic star formation rate in galaxies that starts to decline with redshift at @xmath247 ( madau , pozzetti , & dickinson 1998 ) , although the ionizing emissivity might increase with redshift even when the star formation rate decreases because of an increasing escape fraction for ionizing photons . as this paper was being refereed , the results of the wmap mission were announced . kogut et al . ( 2003 ) give a result @xmath248 from a model - independent analysis , although the result and the error can change when fitted to different cdm models ( spergel et al . the value of @xmath249 implied by a constant emissivity with clumping factor of one up to very high redshift would represent a @xmath250 deviation from the value of kogut et al . a value as high as @xmath251 would clearly imply a substantial increase of the emissivity up to @xmath252 , when only a small fraction of the mass has collapsed into halos that can form stars in the cdm model . the observed mean transmitted flux of the ly@xmath0 forest , when combined with independently measured values of the mean baryon density , the amplitude of the power spectrum , and the gas temperature , yields a measurement of the intensity of the ionizing background . the abundance of lyman limit systems determines the mean free path of ionizing photons . these two quantities together determine the ionizing emissivity . we have argued in this paper that the emissivity measured in this way at @xmath1 is at most 7 ionizing photons per atom per hubble time . we have then shown that the emissivity over the redshift range @xmath253 can not be much lower than the value at @xmath1 in order that the universe can be reionized by @xmath2 , if the clumping factor of ionized gas during the reionization epoch is close to unity . if the clumping factor is larger than unity , then the ionizing emissivity must increase with redshift between @xmath1 and @xmath234 . the electron scattering optical depth must be at least @xmath254 , given the presence of ionized intergalactic gas at @xmath255 , and a reasonable minimum redshift range over which sources ionize the universe at @xmath30 . if the emissivity of ionizing photons does not increase with redshift at @xmath202 , then the optical depth must be less than @xmath256 . measurements of the optical depth from the cmb spectrum of temperature and polarization fluctuations therefore provide a powerful diagnostic to decide if the peak of the comoving ionizing emissivity occurred at @xmath257 ( with a broad plateau extending up to @xmath234 so that the end of reionization can be reached by @xmath2 ) , or at a much earlier epoch . the recent wmap result has an error that is still too large to reach a definite conclusion , but if a value @xmath258 is confirmed it implies that the comoving emissivity continues to rise up to @xmath245 . abel , t. , bryan , g. l. , & norman , m. l. 2000 , apj , 540 , 39 alam , s. m. k. , & miralda - escud , j. 2002 , apj , 568 , 576 arons , j. , & wingert , d. w. 1972 , apj , 177 , 1 becker , r. h. , et al . 2001 , aj , 122 , 2850 bernardi , m. , et al . 2002 , aj , in press ( astro - ph/0206293 ) boyle , b. j. , shanks , t. , croom , s. m. , smith , r. j. , miller , l. , loaring , n. , & heymans , c. 2000 , mnras , 317 , 1014 boyle , b. j. , shanks , t. , & peterson , b. a. 1988 , mnras , 235 , 935 bromm , v. , coppi , p. s. , & larson , r. b. 1999 , apj , 527 , l5 bruzual a. , g. , & charlot , s. 1993 , apj , 405 , 538 bryan , g. l. , & machacek , m. e. 2000 , apj , 534 , 57 conti , a. , kennefick , j. d. , martini , p. , & osmer , p. s. 1999 , aj , 117 , 645 couchman , h. m. p. , & rees , m. j. 1986 , mnras , 221 , 53 croft , r. a. c. , weinberg , d. h. , bolter , m. , burles , s. , hernquist , l. , katz , n. , kirkman , d. , & tytler , d. 2002 , apj , 581 , 20 de bernadis , p. , et al . 2002 , apj , 564 , 559 djorgovski , s. g. , castro , s. , stern , d. , & mahabal , a. a. 2001 , apj , 560 , l5 donahue , m. , & shull , j. m. 1987 , apj , 323 , l13 fan , x. , et al . 2001 , aj , 121 , 54 fan , x. , narayanan , v. k. , strauss , m. a. , white , r. l. , becker , r. h. , pentericci , l. , & rix , h .- w . 2002 , aj , 123 , 1247 giallongo , e. , cristiani , s. , dodorico , s. , & fontana , a. 2002 , apj , 568 , l9 gnedin , n. 2000 , apj , 535 , 530 haardt , f. , & madau , p. 1996 , apj , 461 , 20 haiman , z. , abel , t. , & madau , p. 2001 , apj , 551 , 599 hu , w. 2000 , apj , 529 , 12 jarvis , m. , bernstein , g. , jain , b. , fischer , p. , smith , d. , tyson , j. a. , & wittman , d. 2002 , submitted to aj ( astro - ph/0210604 ) kogut , a. , et al . 2003 , submitted to apj ( astro - ph/0302213 ) madau , p. , ferguson , h. c. , dickinson , m. e. , giavalisco , m. , steidel , c. c. , & fruchter , a. 1996 , mnras , 283 , 1388 madau , p. , haardt , f. , & rees , m. j. 1999 , apj , 514 , 648 madau , p. , pozzetti , l. , & dickinson , m. 1998 , apj , 498 , 106 mcdonald , p. , miralda - escud , j. , rauch , m. , sargent , w. l. w. , barlow , t. a. , cen , r. , & ostriker , j. p. 2000 , apj , 543 , 1 mcdonald , p. , & miralda - escud , j. 2001 , apj , 549 , l11 mcdonald , p. , miralda - escud , j. , rauch , m. , sargent , w. l. w. , barlow , t. a. , & cen , r. 2001 , apj , 562 , 52 miralda - escud , j. , & ostriker , j. p. 1990 , apj , 350 , 1 omeara , j. m. , et al . 2001 , apj , 552 , 718 osterbrock , d. 1989 , _ astrophysics of gaseous nebulae and active galactic nuclei _ ( university science books ) petitjean , p. , webb , j. k. , rauch , m. , carsell , r. f. , & lanzetta , k. m. 1993 , mnras , 262 , 499 pryke , c. , et al . 2002 , apj , 568 , 46 rauch , m. 1998 , ara & a , 36 . 267 rauch , m. , miralda - escud , j. , sargent , w. l. w. , barlow , t. a. , weinberg , d. h. , hernquist , l. , katz , n. , cen , r. , & ostriker , j. p. 1997 , apj , 481 , 601 ricotti , m. , gnedin , n. y. , & shull , j. m. 2002 , apj , 575 , 49 schaye , j. , theuns , t. , rauch , m. , efstathiou , g. , & sargent , w. l. w. 2000 , mnras , 318 , 817 schirber , m. , & bullock , j. s. 2002 , apj , submitted ( astro - ph/0207200 ) schmidt , m. , schneider , d. p. , & gunn , j. e. 1995 , aj , 110 , 68 seljak , u. 2002 , submitted to mnras ( astro - ph/0111362 ) shapiro , p. , & giroux , m. 1987 , apj , 321 . l107 shapiro , p. r. , & raga , a. c. 2001 , rev . astrof . , 10 , 109 spergel , d. n. , et al . 2003 , apj , in press ( astro - ph/0302209 ) steidel , c. c. , adelberger , k. , giavalisco , m. , dickinson , m. , & pettini , m. 1999 , apj , 519 , 1 steidel , c. c. , hunt , m. , shapley , a. , adelberger , k. , pettini , m. , dickinson , m. , & giavalisco , m. 2002 , apj , 576 , 653 steidel , c. c. , pettini , m. , & adelberger , k. 2001 , apj , 546 , 665 storrie - lombardi , l. , mcmahon , r. g. , irwin , m. j. , & hazard , c. 1994 , apj , 427 , l13 storrie - lombardi , l. , mcmahon , r. g. , & irwin , m. j. 1996 , mnras , 283 , l79 telfer , r. c. , zheng , w. , kriss , g. a. , & davidsen , a. f. 2002 , apj , 565 , 773 theuns , t. , zaroubi , s. , kim , t .- s . , tzanavaris , p. , & carswell , r. f. 2002 , mnras , 332 , 367 van waerbeke , l. , et al . 2001 , a & a , 374 , 757 venkatesan , a. , tumlinson , j. , & shull , j. m. 2003 , apj , 584 , 621	the average rate of emission of ionizing radiation per unit volume ( or emissivity ) in the universe can be calculated as the ratio of the intensity of the ionizing background to the mean free path of ionizing photons . the intensity of the background is measured from the mean transmitted flux of the ly@xmath0 forest , and the mean free path is measured from the abundance of lyman limit systems , which has been observed so far up to @xmath1 . this yields an emissivity that is not larger than 7 ionizing photons per hubble time for each atom in the universe at @xmath1 , which may reasonably arise from qsos and star - forming galaxies . in order for the reionization to end by @xmath2 , and assuming that the clumping factor of ionized gas during the reionization epoch is close to unity , this ionizing comoving emissivity can not decline from @xmath1 up to @xmath3 by more than a factor @xmath4 . if the clumping factor were much larger than unity , then the emissivity would need to rapidly increase with redshift . unless the ionizing emissivity increases substantially from @xmath1 to @xmath5 , the thomson optical depth to the cmb must be in the range @xmath6 .
mapping the location of stars and galaxies as a function of distance or redshift helps us understand not only the contents of the universe but also its structure and evolution and the physical principles shaping what we observe . quasars , or active galactic nuclei ( agn ; used synonymously in this work ) , are arguably the most useful extra - galactic source for mapping the universe at high redshift . these accreting super - massive black holes ( bhs ) can outshine their host galaxies by several orders of magnitude and are thus observable at much greater distances than their quiescent counterparts . moreover , quasar spectra are characterized by the presence of high equivalent width ( ew ) emission lines distributed across uv to nir wavelengths . these emission lines can be identified and redshifts determined even with relatively low s / n , resource - economic , survey - quality " spectra . several large surveys have been obtaining large numbers of quasar spectra for cosmological studies , such as measurements of baryon acoustic oscillations ( e.g. , * ? ? ? * ; * ? ? ? the sloan digital sky survey ( sdss ) , alone , with programs such as the baryon oscillation spectroscopic survey ( boss ) , has spectroscopically confirmed @xmath3370,000 quasars @xcite . these quasar redshifts are invaluable for studies on both large and small scales , both for studies directly related to the quasars and for those reliant upon the intervening absorption . however , making accurate redshift measurements of quasars , particularly at high redshifts , is surprisingly difficult an issue that is , perhaps , not broadly known or appreciated outside the direct quasar physics community . quasar spectra are a blended superposition of many emission and absorption components that arise from physically distinct sources at different distances from the bh . components include the thermal continuum from the accretion disk , narrow and broad emission lines from the narrow - line region ( nlr ) and broad - line region ( blr ) , respectively , intrinsic and intervening absorption lines , and host - galaxy starlight . of these emission- and absorption - line components , some are better - suited for redshift determinations than others . associated quasar absorption lines are usually attributed to outflows from the nucleus and are not expected to lie at the systemic redshift . host - galaxy stellar absorption features , on the other hand , are the most robust measure of the galaxy s redshift . unfortunately , these absorption lines are generally masked by the luminous quasar contributions . even if observable , rest - frame optical lines , such as caii h & k lines at @xmath5 3969 , 3934 , are inaccessible in optical spectra of high - redshift quasars . in the absence of reliable absorption lines , quasar redshifts are best determined from narrow emission lines that arise from the nlr or the host galaxy . of particular interest are lines due to forbidden transitions that can not arise in the high - density and high - temperature blr environment , and so are not blended with a broad - line component or severely susceptible to dynamics dominated by the nuclear activity . the narrow lines are still sometimes observed to have small blue - shifts ( @xmath310@xmath6100 ^ -1 km s@xmath4 ) compared to host - galaxy absorption lines , and this effect may have a luminosity dependence ( see , e.g. , * ? ? ? * ; * ? ? ? unfortunately , using high - ew , isolated , narrow forbidden lines also becomes difficult using optical spectra of high - redshift quasars because rest - frame uv forbidden emission lines have much smaller ews , and , at the highest redshifts ( @xmath73 ) , the only nlr emission lines visible in optical spectra are due to permitted transitions that are blended with blr emission , have relatively low ew , and are often resonance transitions susceptible to self - absorption , e.g. , ly@xmath8 , iv1549 civ@xmath9 , or heii @xmath10 . if absorption and narrow emission lines can not be used for redshifts , the broad emission lines are used . this is most often required at high redshifts , where the other methods are no longer suitable . difficulties in measuring redshifts based on these lines arise first from fundamental difficulties with either the specific transition or data quality ( or both ) . first , the strong uv lines ( ly@xmath8 , iv civ , iii ciii ] , and mgii ) are all either resonance lines susceptible to self - absorption , and/or are heavily blended with other species or within a multiplet . the line ratios within these blends and multiplets depend on physical properties of the nuclear environment , such as optical depth , density , and incident ionizing radiation ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? sdss - i / ii - pipeline redshifts ( hereafter sdss - pipeline ) are based on either emission - line matching or cross correlation with the @xcite composite quasar spectrum . while using a composite spectrum formed from many thousands of survey - quality quasar spectra will provide a very high s / n template , it does not account for the intrinsic diversity in the physical structure , environment , and spectral energy distributions ( sed ) among quasars . it will thus create biases in the redshifts of objects with properties different than the average properties of the quasars used in its creation . of particular importance is understanding ( i ) the physical properties that modify the structure of the emission lines , and ( ii ) on what observable variables those physical properties depend . one challenge is that the sdss pipeline assumes only a single wavelength for many multiplets and/or highly blended transitions in the uv for its emission - line matching and line identification . this assumption combined with diversity in the spectral structure of individual quasars ( i.e. , differences in multiplet ratios for those with relatively wide velocity separation ) could easily contribute systematic uncertainties in the determined redshifts beyond those typically quoted based on [ oiii ] or mgii ( @xmath350@xmath6300 ^-1 km s@xmath4 ; * ? ? ? known velocity shifts between different quasar emission lines are another challenge that contributes to biases in redshifts ( e.g. , see * ? ? ? * and discussion and references therein ) . several studies have already explored biases in the sdss - pipeline redshifts ( e.g. , @xcite , hereafter hw10 ) . while they still utilize cross - correlation with a master template , hw10 improve the sdss redshift estimates of quasars by building a redshift ` ladder ' as a function of increasing redshift , since the redshifts of more nearby quasars can be more accurately determined from host - galaxy stellar features and strong forbidden , narrow emission lines . hw10 and @xcite find trends in line - to - line velocity shifts with quasar luminosity . while luminosity is an easily measured observable , more in - depth analyses of the spectral diversity of quasars from eigenvector analysis ( cf . * ) suggest that the largest source of emission - line diversity in quasars , dubbed eigenvector 1 " ( ev1 ) , is more likely related to accretion rate and the quasar spectral energy distribution ( sed ) . ev1 analysis shows that the relative strengths of different emission lines and the velocity shifts between lines seem to be well - correlated ( e.g. , * ? ? ? * ) . in flux - limited surveys , luminosity can be a reasonable proxy for accretion rate , potentially leading to the luminosity correlation found by hw10 and others . this connection is a concern for redshifts based on composite quasar spectra formed from flux - limited samples , as the composite will be weighted toward the spectral properties associated with the relatively higher luminosity quasars . redshift estimates from the boss pipeline @xcite , which we utilize in this work , use an eigenvector principle component analysis ( pca ) method that is more sophisticated than the previous sdss - pipeline cross - correlation redshifts . cross - correlation and pca analysis should both be more robust against biases due to small , intrinsic velocity shifts between lines @xcite than using individual emission lines to determine the redshift because these methods average over all shifts . additionally , the pca - based redshifts should be even more robust than cross correlation with a single composite spectrum because a template built for each quasar is not as susceptible to the averaging " biases from using a single quasar template . however , the boss pca templates are built using training - set spectra with redshifts determined from the original sdss - method . as such , while the boss pca method is more sophisticated overall ( see also @xcite for continued improvements ) , redshift biases ( see , e.g. , * ? ? ? * ; * ? ? ? * ) may have propagated into the templates due to redshift inaccuracies in the training set spectra . a particular problem associated with accurate quasar redshift estimates is whether bh masses can be well - estimated using the iv civ emission line . in particular , the apparent blueshifts in the iv civ line have been used as evidence that the iv civ velocity widths are indicative of non - virial motions and therefore ill - suited for virial bh mass calculations ( see also * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? . however , since radiation - driven winds generally have velocities comparable to the bh escape velocity @xcite , they might still be quite reasonable virial estimates in this scenario . in any case , the inference of significant and/or generally ubiquitous blueshifts in iv civ emission depends on the reliability of the redshifts . here , we investigate the possibility of systematic errors in the redshifts of quasars using boss spectra taken as part of the sdss - rm project . we use two narrow emission lines observable in optical spectra of intermediate- and high - redshift quasars . our investigation is laid out as follows : in section [ s_data ] we present the sdss reverberation mapping project data that we use for this investigation . section [ s_specanal ] describes our analysis of these data for measuring the redshifts from heii@xmath11640 and the [ oii]@xmath13727 doublet , and in section [ s_discussion ] we discuss our results . final remarks on our results are made in section [ s_conclusions ] . the sdss reverberation mapping project ( sdss - rm ) is spectroscopically monitoring broad - line quasars in a single 7 deg@xmath11 field ( the cfht - ls w3 field ) with the sdss telescope s @xcite boss spectrograph @xcite . here , we utilize the data obtained from the sdss - iii @xcite ancillary program between 2014 january and july . covering redshifts over the range @xmath12 , the sdss - rm sample consists of 849 quasars with a flux limit of @xmath13 mag . each of the 32 epochs of observations was a @xmath32 hr exposure taken during dark / grey time , with an average cadence of @xmath34 days over this @xmath36-month period . the technical overview of this program is provided by @xcite . here we use the subsample used by @xcite to study the properties of the iv1549 civ@xmath9 emission - line region . the sample consists only of quasars with @xmath0 , where objects were removed that obstructed the iv civ analysis , such as a few with very low iv civ equivalent width ( ew ) and/or several additional objects with broad absorption lines ( see * ? ? ? * for further sample selection details ) . the final sample consists of 482 sources , and for this analysis , we only use the high - s / n coadded " spectrum , made by combining all good epochs using the latest boss spectroscopic pipeline _ idlspec2d _ ( see * ? ? ? * and schlegel et al . , in prep . ) . most ( 405 ) quasars have all 32 epochs of spectra included in the coadd , and only 9 have more than 3 ( 10% ) epochs discarded . figure [ f_physprop ] , reproduced from @xcite , shows the redshift and @xmath14 magnitude distribution of our sample . the redshifts for our sample , as determined from the boss pipeline @xcite , are likely already a moderate improvement over typical boss - pipeline redshifts because we are able to use the coadded spectra ( see * ? ? ? * ) , as opposed to the individual , lower - s / n single - epoch spectra ( although @xcite and @xcite find that line centroid measurements , and thus emission - line cross correlations , are relatively more robust in the presence of spectral noise than other measured emission - line properties ) . however , the real advantage afforded by this relatively higher - s / n sample is to enable the analysis of low - ew narrow emission lines , not otherwise reliably detected in low - s / n spectra , to study the effects of intrinsic quasar diversity on the redshift determinations . we compare redshifts determined using the heii@xmath10 emission line , which _ is _ susceptible to spectroscopic diversity and shifts in its observed line properties ( likely related to ev1 effects ; see , e.g. , * ? ? ? * hereafter r11 ) , and [ oii]@xmath13727 , which , presumably , is _ not_. heii is not a resonance line and therefore not susceptible to self - absorption , and intervening absorption affects only a small percentage of objects . it is also a more isolated transition than many uv lines , so blending is less of a concern . this makes it a more favorable line for studying the effects of quasar diversity on redshift determinations . however , there are two potential problems : first , while we would ideally isolate and study the nlr component of heii , it is not always observed , presumably due to the same effects we are trying to study . furthermore , while the blr component of heii typically has a low ew , when the nlr component is absent , any remaining heii emission appears significantly blueshifted and blended with ( or part of ) the iv civ red shelf.1600 that also varies in strength among quasars and has yet to be identified as uniquely due to any specific ionic species or blend ( see , e.g * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * for discussions and further references ) . ] second , heii is a high - ionization line , so an argument can be made that even the narrow component of this line arises , at least partially , in an outflow , as is suggested by the blueshifts of other high - ionization broad lines ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? [ oii ] , on the other hand , is a forbidden narrow emission line , so it is emitted predominantly from the extended nlr that is not as susceptible to kinematic effects and diversity regulating the spectral properties of emission components arising deep within the potential of the bh . [ oii ] can also be emitted throughout the host galaxy , which will also not be affected by the agn environment and will lead to even more robust systemic redshift determinations . @xcite find a tight correspondence between the peak of the [ oii ] emission and stellar absorption lines for the lower - redshift subset of the sdss - rm sample , with a systematic shift of only 8 ^ -1 km s@xmath4 and an intrinsic scatter of 46 ^ -1 km s@xmath4 . consequently , the redshifts determined by [ oii ] for the present sample serve , in effect , as our control . while [ oii ] is the shortest - wavelength unblended forbidden emission line that is relatively - strong and still present in quasar spectra for intermediate - to - high redshifts , it still falls beyond the boss wavelength coverage for @xmath71.78 , so only 154 of our 482 quasars have [ oii ] present in their spectra . thus , while it is reliable , it is not very applicable to high-@xmath15 quasars . because of the diversity in the observed properties of the heii line across our sample , it is not as straight - forward to perform automated emission - line fitting ( e.g. , * ? ? ? * ) and still control for the optimal number and relative shift of fit components . we calculate the heii centroid directly from the coadded spectra ( i.e. , not from functional - form fits to the heii line ) based on interactively selected wavelength boundaries . we attempted to choose these boundaries at approximately the half maximum flux level on either side of the heii peak selected by eye , and we use this centroid to define our heii - based " redshifts . since the intrinsic strength of the heii line , especially the nlr component , varies between objects , the boundary selection is often quite subjective . in some objects , no heii line is visible at all , even for very high s / n spectra , and in others , noise or absorption obscures the peak . in these cases , we also use other nearby emission features , such as the 1400 feature ( a blend of siiv and oiv ] ) and the oiii ] 1663 lines , or even absorption lines , which have a high probability of being at or very near the systemic redshift ( @xcite ; allen & hewett , in preparation ) , to help inform the choice of the wavelength range that provides the most reasonable redshift through visual inspection under these circumstances . when we could discern no reasonable criteria for measuring the heii centroid , the boss redshift was kept . because of these difficulties , we assigned heii redshift quality ( q ) flags to each object , defined by the following criteria : 1 . very weak , or very broad , apparently blueshifted blr component ( compared to other features such as the 1400 feature and oiii]@xmath11663 doublet ) with no nlr component , or no obvious line ; 2 . clean , narrow peak ; 3 . noisy , but a relatively narrow peak is still clear ; 4 . broader line , but a reliably determined peak ; 5 . broader line , and the peak can not be reliably contributed to a narrow component ; 6 . noisy spectrum , where the peak / centroid is uncertain or possibly contaminated by noise or intervening absorption . we take the redshift determinations from q = 1 , 2 , and 3 objects to be relatively more robust measurements of the systemic redshift , as a strong narrow component should not be significantly blueshifted , with q=3 the least robust due to the broader heii line . q = 4 , 5 , and 0 are less robust due to data - related issues , intrinsic quasar properties , or both . due to the overall subjectivity of these measurements , they are not to be taken as high precision redshifts , but they nonetheless provide sufficient evidence to support our investigation . we further test the robustness of these measurements against an independent measure of the heii line center in the next section . coincidentally , the sample is roughly evenly divided , with 237 objects flagged with q = 1 , 2 , or 3 , and 245 objects flagged with q = 4 , 5 , or 0 . figure [ f_redshiftquality ] shows an example spectrum for each of the six quality categories . to form a control sample of redshifts based on a line not susceptible to ev1 effects , we use the `` prepspec '' analysis of this sample . prepspec is a reverberation mapping spectral preparation and analysis software written by one of us ( kh ) and applied to the sdss - rm project sample ( see * ? ? ? * for details ) . as part of the prepspec analysis , all blr , nlr , and host - galaxy emission lines visible in each spectrum are modeled . the prepspec output includes a measure of the modeled emission - line centers and velocity offsets with respect to the input redshift values . while still dependent on automated modeling of the spectrum , this method provides an independent comparison for the heii line center , uncertainties on the measurements , and a homogeneous methodology for comparing to [ oii ] our control , forbidden narrow line . we use the prepspec heii and [ oii ] emission - line velocity offsets determined with respect to our heii - based redshifts used as the input redshift . we only focus on the heii and [ oii ] @xmath13727 emission lines as a means to compare with our own heii - based redshifts and to address the effects of quasar diversity on redshift ( see * ? ? ? * for a comprehensive analysis of relative emission - line velocity shifts for this sample ) . uncertainties in the measured line centers are determined from monte carlo simulations . the median 1@xmath2 uncertainties in the prepspec heii and [ oii ] velocities are 481 ^-1 km s@xmath4 and 253 ^-1 km s@xmath4 , respectively . we use the analysis of the iv civ emission - line region by @xcite to measure the iv civ blueshift with respect to both the boss - pipeline redshifts and our heii - based redshifts . we utilize the iv civ emission - line peak wavelength measured from the gauss hermite ( gh ) profile fits . the uncertainties in the centroid measurements are estimated for each object from monte carlo simulations that create 500 flux - resampled spectra on which the measurements are repeated . the median 1@xmath2 iv civ centroid uncertainty for this full sample is 183 ^ -1 km s@xmath4 ( see * ? ? ? * ) . to be consistent with a similar study presented by r11 , we calculate our iv civ blueshifts with respect to the estimated systemic redshift such that increasing blueshifts are larger , positive velocities . figure [ f_otherzcompare ] and table 1 show the differences between our heii - based redshifts and those inferred from prepspec for heii ( left ) and [ oii ] for the 154 objects that have this measurement ( middle ) . the prepspec heii redshifts are consistent with our simple , direct approach , given the uncertainties . the median difference between our heii centroid measurements and the prepspec heii line centers is 54 ^ -1 km s@xmath4 which only @xmath310% of the median prepspec heii measurement uncertainty . bootstrap monte carlo simulations that draw randomly from the distribution shown in the left panel of figure [ f_otherzcompare ] 10,000 times , with no weighting for measurements drawn multiple times , estimate a 1@xmath2 ( 3@xmath2 ) uncertainty in this median systematic shift of @xmath1622 ^ -1 km s@xmath4 ( @xmath1662 ^ -1 km s@xmath4 ) , further supporting the consistency between our rough , visually - determined centroid - based heii line centers and those determined through automated fitting . the comparison with [ oii ] ( middle panel of figure [ f_otherzcompare ] ) shows a relatively large scatter of 780 ^ -1 km s@xmath4 , although a small systematic blueshift of heii relative to [ oii ] is evident . we find a distribution median of 348 ^ -1 km s@xmath4 and a bootstrap 1@xmath2 ( 3@xmath2 ) uncertainty on the median of 104 ^ -1 km s@xmath4 ( 242 ^ -1 km s@xmath4 ) . the scatter is defined by the half of the 16@xmath684% inter - percentile range ( hipr ) of the distribution that would correspond to 1@xmath2 if the distribution were gaussian , and the uncertainty in the median is again based on 10,000 trials . nonetheless , this systematic shift is still within the 1@xmath2 quadrature summed heii and [ oii ] prepspec measurement uncertainties . while it is difficult to assign uncertainties to our direct heii redshift estimates based on the interactive centroid calculation , the systematic shift in the heii[oii ] redshift distribution is also within the observed scatter of the comparison between the interactive and the prepspec heii line centers , which as show above , were statistically consistent at the 3@xmath17 level . this suggests that the typical heii[oii ] line center shift is also likely within or near the order of the uncertainties of our direct heii centroid measurement method . however , since quasar diversity effects are expected to be imprinted on heii but not [ oii ] , we also measured the systematic shift after splitting the sample by quality rating . we find relatively larger systematic shifts and more scatter for the more uncertain q=4 , 5 , 0 heii redshifts than for q=1 , 2 , 3 ( see table 1 ) . despite the contributions of noise to the q=4 , 5 , 0 distribution , this is also as expected if the primary source of the observed velocity shift is ev1 effects within the diversity of quasar spectra , from which we expect the strength of narrow heii to anti - correlate with its the broad+narrow emission - line width and the velocity shift away from systemic . this physical effect is likely additionally exacerbated within the q=4 , 5 , 0 subsample by the uncertainty in our inability to isolate the heii narrow component when it is weak or absent . @xcite find similar but smaller relative velocity shifts , with a median heii shift of @xmath18 ^ -1 km s@xmath4 , i.e. , blueward , with respect to [ oii ] using yet another independent method for fitting the emission lines in the coadded spectra of the sdss - rm sample . @xcite use a slightly different sample than ours , including only 134 objects with both heii and [ oii ] due to a combination of spectral - fitting redshift and/or wavelength limits and the reliability of the automated fitting process in identifying and fitting the lines in cases where they are weak . this may contribute to the relatively smaller systematic shifts and scatter they find compared to the present results . in addition , @xcite find luminosity trends in relative line shifts in the sdss - rm sample , or equivalently , redshift determinations based on cross - correlation with different emission lines , consistent with other work ( hw10 ; r11 ; * ? ? ? * ; * ? ? ? hw10 discuss the systematic effects in cross - correlation redshift determination between lines as due to sed effects for example , the differing line ratios within the iii ciii ] blend . these are most certainly related to the ev1 effects we are interested in here . hw10 do not directly take these effects into account , although they do apply a correction as a function of quasar luminosity that is arguably related , but there is much scatter . we see from the right panel of figure [ f_otherzcompare ] that there is a weak correlation as a function of luminosity for our heii[oii ] velocity shift measurements , consistent with the results of @xcite . however , there is significant scatter about this weak correlation , arguably driven by the lack of scatter toward positive shifts at high l , rather than a consistent trend across the full luminosity range . nonetheless , this trend goes in the direction expected from the sed and ev1 effects , where the highest luminosity , highest accretion - rate sources are more likely to have signatures of quasar outflows in their spectra ( see also discussions by , e.g. , r11 and references therein ) . we investigate the possibility for biases in the boss - pipeline redshifts due to quasar diversity using our heii redshifts described in section [ s_analredshifts ] . table [ t_shifts ] quantifies the redshift differences we find between boss - pipeline and heii - based redshifts , which are shown in figure [ f_zbiases ] . we look separately at the full sample as well as the `` more '' ( q = 1,2,3 ) and less " ( q = 4,5,0 ) reliable subsamples . the peak of the full sample distribution is broad , but generally , we find a median shift in the boss redshifts of 1080 ^-1 km s@xmath4 relative to those based on heii , with a 1@xmath2 ( 3@xmath2 ) uncertainty on the median from 10,000 bootstrap trials of 90 ^ -1 km s@xmath4 ( 225 ^ -1 km s@xmath4 ) . this systematic difference is significantly larger ( greater than a factor of three ) than the median shift between heii and [ oii ] that can be explained by quasar diversity effects in the heii redshifts . the dependence of this redshift difference on redshift shows a relatively larger bias for @xmath72.8 , compared with lower redshifts , probably because this corresponds to the redshift where mgii exits the boss spectral range . there is also a large bias for systems with low s / n . this is generally consistent with results of the effects of data quality on measurements of emission line properties @xcite . the bottom panel of figure [ f_zbiases ] suggests that quasar diversity systematics remain present in the boss - pipeline redshifts , despite the improvements we expected by using pca analysis over strict emission line cross correlation with a composite spectrum . we find the smallest boss heii redshift differences at the highest luminosities , log @xmath19 , but extrapolation in figure [ f_otherzcompare ] suggests that these luminosities have the largest deviations between heii and [ oii ] up to @xmath32000 ^ -1 km s@xmath4 presumably owing to ev1 quasar diversity effects . this seems to indicate that _ both _ the boss and the heii redshifts are biased ( i.e. , underestimating the redshift ) at these high luminosities , which is the regime where the ev1 effects make heii a poorer redshift indicator . on the other hand , at lower luminosities , we find differences between heii and boss - pipeline redshifts of @xmath31000 ^ -1 km s@xmath4 , or more . yet , these lower luminosity sources have heii redshifts consistent with those from [ oii ] and are thus likely to be probing the systemic redshift . these results suggest that both high - luminosity and low - luminosity quasars are biased in opposite senses : boss redshifts of high luminosity quasars are underestimated , while those of low - luminosity quasars are overestimated . this effectively compresses the redshift luminosity plane or the parameter space of any other measured or physical properties with similar dependencies . in other words , cross - correlation redshifts have an averaging " bias in not accounting for the dependence of quasar line shifts on luminosity ( cf . hw10 ) or some more physical underlying source . this was in some sense done by design in sdss - pipeline redshifts , where the @xcite template was created with iv civ at 1546 , so that contributing quasars that were noted to have a range in iv civ peak wavelengths spanned the template wavelength @xmath201000 ^ -1 km s@xmath4 , rather than a template with iv civ at the expected 1549 , with objects showing shifts @xmath302000 ^ -1 km s@xmath4 ( g. t. richards , private communication ) . the best explanation we can determine for this effect remaining so large in boss - pipeline redshifts is the implicit dependence of the boss - pipeline redshifts on cross - correlation redshifts : the pca training sample redshifts are still based on comparison with a composite spectrum , despite the improvements afforded by pca . as discussed previously , hw10 attempted to mitigate quasar - diversity biases in sdss - pipeline redshifts by applying a luminosity correction . in figure [ f_hwzcompare ] we compare the heii redshifts to hw10 redshifts for the 47 quasars ( 17 with q=1 , 2 , 3 ) we have in common with hw10 . the left panel of figure [ f_hwzcompare ] and statistics in table [ t_shifts ] show a larger shift between our heii redshifts and hw10 redshifts than the small differences we measured between heii and [ oii ] . this is , at first glance , surprising , since hw10 made corrections to the sdss redshifts for the luminosity - dependent biases . one interpretation of these differences is that there are still quasar emission - line property diversity systematics in the hw10 redshifts , despite the improvements and corrections they employ . they do still make use of cross - correlation with template spectra and so may still be susceptible to averaging across the diversity to form their composite spectra . alternately , we may simply be `` unlucky '' in our overlap sample . admittedly , the number statistics of our comparison are small and the scatter is large . from a quasar physics standpoint , however , the sdss spectroscopic flux limit is shallower than that of boss and sdss - rm . so the overlap with our sdss - rm sample is biased to bright agn that probes a different part of quasar diversity parameter space . this implies the distribution of overlap with our sdss - rm sample is not unbiased . indeed , the open circles in the right panel of figure [ f_hwzcompare ] show that the 18 objects in our overlap sample with hw10 that also have [ oii ] redshifts are not distributed uniformly and , instead , exhibit larger than average shifts between heii and the systemic redshift ( assumed from [ oii ] measurements ) . with this biased sample - overlap , we can not make an objective and comprehensive investigation into the degree to which hw10 redshifts effectively correct for ev1/sed effects in their redshifts , and/or whether they may still be susceptible to quasar diversity biases . however , we try to gain a small amount of additional insight by looking at the 36 of the 47 objects in the sdssrm@xmath6hw10 overlap sample that have mgii present in the spectrum ( shown in figure [ mgiithumbnails ] ) . of these , 30 of the hw10 redshifts are based on mgii cross - correlation , which is generally taken to be the most reliable broad line for determining redshifts for [email protected] . for example , within the sdss - rm sample , @xcite find only a small systematic shift , @xmath22 ^ -1 km s@xmath4 , on average , between the location of the mgii peak and the caii stellar absorption features , with no luminosity dependence . there are 15 objects of this subsample that have q = 1 , 2 , 3 . from inspection of the observed mgii doublet peaks in figure [ mgiithumbnails ] relative to the predictions from the heii - based redshift ( solid lines ) , boss redshift ( dotted lines ) , and hw10 redshift ( dashed lines ) , there is not an obvious , uniform explanation for the redshift differences . for the three objects ( rmid 321 , rmid408 , and rmid770 ) with q=1 ( for which heii redshifts are the most robust ) , we find that the heii redshift more accurately predicts the peaks of the mgii doublet than the boss - pipeline or hw10 methods . all three of these objects display a shift between the heii and hw10 redshifts of @xmath31000 ^ -1 km s@xmath4 , consistent with the average heii boss redshift difference . yet , rmid321 and rmid408 , the two of these three objects that also have [ oii ] measurements , show a relative shift between [ oii ] and heii of only 11 ^ -1 km s@xmath4 and @xmath652 ^ -1 km s@xmath4 , respectively . at least in these cases , this provides strong evidence that the heii redshift for these objects is trustworthy . consistent with the large scatter in the left panel of figure 5 , there is a lot of inconsistency in the mgii peak - prediction accuracy of the q=2 and q=3 heii redshifts : in some objects , the heii - based redshift is better ( e.g. , rmid155 ) , while for others ( e.g. , rmid774 ) , it makes a much worse prediction than boss or hw10 . sometimes , the hw10 redshift is clearly superior to the boss or heii estimate , such as for rmid201 and rmid676 , q=0 cases , where heii is frequently asymmetric and is often blueshifted with respect to [ oii ] , and/or no discernible narrow heii emission line is present . these cases highlight the clear improvement in redshifts offered by the luminosity ( i.e. , quasar diversity ) corrections of hw10 over the sdss- or boss- pipeline . there are two other considerations when comparing results with mgii , which is likely the best uv broad line to use for redshifts . first , even mgii is sometimes blueshifted with respect to h@xmath23 and [ oiii]@xmath24 ( e.g. , * ? ? ? * ; * ? ? ? such objects tend to be the same that are argued to have high accretion rates , high luminosity , blueshifts , and low - ew , high - ionization lines ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? these objects are possibly also related to so - called weak - line quasars that also may have characteristically weak or absorbed x - ray properties and are suggested to lie at the extreme end of this parameter space ( e.g. , * ? ? ? * ; * ? ? ? furthermore , mgii likely has little or no emission contribution from the nlr because of the photoionization physics regulating this transition . as such , diversity in blr kinematics between objects will affect redshifts based on the mgii peak compared to redshifts determined from narrow emission lines or host - galaxy stellar absorption lines . this is likely a cause of scatter in the peak shifts between mgii and [ oii ] found by @xcite and others . second , the mgii doublet _ ratio _ is dependent on the physical conditions of the blr . while the doublet is often not fully resolved due to moderate spectral resolution or because of the large blr velocities , the line peak can still shift among differing physical environments that allow the ratio to vary between 1:1 and 2:1 . interpreting cross - correlation results therefore becomes non - trivial , and such effects will also affect the characteristics of the pca training set or template spectrum . clearly , this is a difficult and multi - faceted problem . while mgii is certainly better than other uv quasar emission lines that exhibit more blending or stronger ev1/quasar diversity effects , mgii will still exhibit diversity due to the physical environment of the nucleus . results from our heii - based redshift analysis suggest that the reliability of current high-@xmath15 quasar redshifts can be connected to spectroscopic quasar diversity related to the intrinsic , physical differences in the nuclear structure and/or sed , and is likely related to accretion rate ( see , e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? consequently , many of the iv civ blueshifts previously inferred for sdss quasars by r11 and others may be similarly biased if the quasar redshift determinations are unreliable . we still find evidence that iv civ exhibits line peak blueshifts ( as do other high - ionization broad lines , including heii ) , _ but only in some _ quasars . this is demonstrated in figure [ f_civblue ] , where the top panels show the measured iv civ centroid blueshifts based on the boss - pipeline redshifts ( top left ) and our heii - based redshifts ( top right ) . we find different trends for q=1 , 2 , 3 sources as for q=4 , 5 , 0 sources . the distinction between these subsamples , while technically set by our ability to reliably determine a heii - based redshift , is , as argued above , related to the intrinsic differences in the physical properties of these sources . objects with strong ( or relatively stronger ) , narrow heii emission components ( typically q = 1 , 2 , and 3 ) do not have systematically blueshifted iv civ emission , while objects without this narrow component can show very large iv civ blueshifts that are likely caused by sed effects , such as radiation line - driven outflows ( see , e.g. , r11 ; * ? ? ? * and references therein for further discussion ) . however , figure [ f_civblue ] and the discussions above provide evidence that boss - pipeline - based redshifts are even more biased than redshifts based on heii , even considering the shortcomings of using this line . the most likely explanation for these biases is due to the measurement of redshifts from cross - correlation with template spectra that are formed from a sample that covers the large physical property parameter space of quasars . this complicates analyses that try to understand this parameter space using iv civ diagnostics and any studies that require reliable redshifts as a means to interpret emission- or absorption - line velocity shifts in quasar spectra ( e.g. , * ? ? ? evidence for the averaging effect discussed in section 1 likely due to the ultimate dependence of the boss - pipeline redshifts on an average composite spectrum is shown in the top left panel of figure [ f_civblue ] . the distribution of iv civ blueshifts for q = 1 , 2 , and 3 redshifts is co - spatial with the q = 4 , 5 , and 0 distribution , and all objects are driven to an average " iv civ blueshift , likely coincident with that imposed by a bias intrinsic in the redshifts of the pca training set of quasar spectra . on the other hand , using the heii - based redshifts , we still find iv civ blueshifts ( some very large ) , but the lower limit is consistent with no blueshift , within the precision of our measurements . importantly , the overall dynamic range of iv civ blueshifts is much larger now that the averaging bias has been corrected . similar evidence is also seen by investigating the iv civ blueshift ew parameter space used by r11 ( bottom panels of figure [ f_civblue ] ) . r11 argue that this observational parameter space traces physical quasar properties with respect to the prevalence of disk - winds ( due to differences in mass accretion rate ) . our results are consistent with this picture , but we argue that the effect of disk winds on iv civ is not as ubiquitous as that indicated by r11 at a significance level that would statistically bias black hole mass estimates , i.e. , with a systematic blueshift larger than the typical velocity width measurement uncertainties . despite the use of the improved hw10 redshifts by r11 , the most likely explanation for the difference between their iv civ blueshift results and what we find here is a remaining bias because of the shallower flux - limited sdss sample compared to what we probe with sdss - rm . our analysis here suggests that because the hw10 correction is suboptimal for the lowest accretion - rate region of quasar parameter space , the results of r11 do not accurately describe iv civ blueshift trends across the full range of quasar properties . the left panels of figure [ f_civblue ] show that both subsamples those with stronger , narrow heii lines ( red ) and those without ( blue ) occupy the same part of this parameter space when basing the iv civ blueshifts on boss - pipeline redshifts but become separated when the iv civ blueshift is measured with respect to the heii redshifts . this further supports the physical picture that is presented by r11 : quasars with significant iv civ blueshifts are likely those with strong disk winds . however , the iv civ blueshift suggestive of a dominant disk wind is not ubiquitous among the quasar population ; a number of quasars show , on average , _ no _ systematic blueshift within our measurement uncertainties . these are predominantly objects with strong low - velocity ( narrow ) heii and iv civ emission - line cores . one argument that the disappearance of a systematic iv civ blueshift for a portion of our targets when using heii - based redshifts could simply indicate that heii and iv civ share similar blueshifts and that the boss redshifts are not biased . however , its important to note that the minimal iv civ blueshift objects are also those with heii emission clearly attributable to a nlr emission that show little to no velocity shift with respect to [ oii ] or mgii ( e.g. , rmid321 and rmid408 ) , so this interpretation is unlikely for this population of quasars . the _ broad _ heii component ( the only visible component for many objects ) shows similar shifts as the other high - ionization lines , so broad heii likely does trace iv civ . as such , the results shown in figure [ f_civblue ] are still not fully unbiased . our combined results suggest that instead , given accurate , unbiased redshifts , the bottom right panel of figure [ f_civblue ] would likely show ( i ) a somewhat broader distribution of q = 1 , 2 , 3 sources with a lower limit still at zero but a possible mean blueshift of a couple of hundred ^-1 km s@xmath4 to account for the difference between the observed median iv civ blueshift of @xmath25 ^ -1 km s@xmath4 ( i.e. , a redshifted iv civ peak ; see table 1 ) and the observed velocity shifts between heii and [ oii ] of 339 ^ -1 km s@xmath4 , likely predominantly attributable to the q=2 and 3 objects with broader narrow " heii lines , and ( ii ) a distribution of q = 4 , 5 , 0 quasars with a higher mean blueshift because the heii redshift is also likely biased low for high accretion - rate sources with strong winds and no narrow heii . detailed studies of quasars have clearly shown that there is significant diversity in the details of quasar spectra . the most prominent differences , termed the ev1 parameter space @xcite , have been linked to physical differences in the quasar environment , most likely determined by the accretion rate . one known observational consequence of this diversity is the relative shifting of the peaks of quasar emission lines , argued to be related to the strength of disk winds that can be driven at high accretion rates ( see r11 and references therein ) . the relation between the observed spectroscopic properties of quasars and the accretion rate leads to correlations of quasar properties with luminosity in flux - limited samples . as a consequence , determining redshifts for quasars is not as straightforward as determining redshifts for galaxies if the desired precision is @xmath26 ^ -1 km s@xmath4 . this is because the mean quasar spectrum changes as a function of redshift due to the observed luminosity dependence of the emission - line properties . in this work , we investigated the degree to which this quasar diversity affects redshift determinations for the sample of @xmath0 sdss - rm quasars by using redshifts based on the heii@xmath11640 emission - line centroid . we are not advocating that heii ultimately be used as a general tool for determining high redshifts . its reliability depends on the identification of the narrow component that is only present in some quasars , since the peak wavelength of broad heii is also susceptible to shifts across ev1 parameter space in the absence of a narrow - line component . furthermore , heii is often low - ew , and therefore can not always be cleanly isolated in lower - s / n survey - quality " spectra . nonetheless , it has benefits for estimating high-@xmath15 quasar redshifts within the constraints on its observed properties that are set by physically - motivated quasar diversity characteristics . we found that heii - based redshifts in the sdss - rm sample are consistent , on average , with [ oii]-based redshifts to within 1@xmath2 of the prepspec line center measurement uncertainties . observed shifts larger than these uncertainties are predominately found in the brightest quasars . [ oii]-based redshifts should not be susceptible to line shifts due to the diversity in quasar physical properties deep within the nucleus , which causes the velocity shifts primarily in the broad and/or high - ionization lines , because it is emitted from gas from much larger radii and/or from elsewhere in the quasar host . objects with q=1 showed a median heii shift with respect to [ oii ] of only @xmath27 ^ -1 km s@xmath4 , which is well within the median statistical uncertainty of the prepspec [ oii ] measurements . this suggests that the redshifts of high - redshift quasars with strong narrow heii can be reliably determined to this level of precision from this line alone . the distributions of heii - to-[oii ] velocity shifts for the full sample suggest that heii exhibits , on average , a small blueshift with respect to the systemic quasar rest frame , on the order of a couple hundred ^-1 km s@xmath4 , assuming that [ oii ] is a better proxy for this . however , the heii[oii ] shifts can be large if the emission - line peak of heii can not be attributed to emission from the nlr . these shifts are not enough to explain the full bias in the boss redshifts indicated by this investigation . by comparing boss - pipeline redshifts to our heii - based redshifts , we found evidence that the boss - pipeline redshifts may be biased , possibly overestimating the redshifts of some high-@xmath15 quasars by @xmath31000 ^-1 km s@xmath4 , on average ( see fig . [ f_zbiases ] ) , while underestimating the redshifts of a smaller population because of the imprint of the physical quasar diversity on observed spectroscopic properties . the main goal of the boss survey to measure baryonic acoustic oscillations in the ly@xmath8 forest of high-@xmath15 quasars ( e.g. , * ? ? ? * ) did not depend on quasar redshifts having a precision greater than this limit , so the great success of the boss survey in that regard is unaffected by these results . however , boss spectra are now being used for a myriad of other purposes that this apparent bias may impact if studies require redshifts to higher precision that this measured systematic difference . the likely source of this bias in the sdss - pipeline redshifts is their basis on a comparison to ( or in the case of the boss pipeline , from a pca training set that ultimately has redshifts determined from ) composite quasar spectra . in the creation of such spectra , objects that exhibit systematic emission - line shifts are coadded with those that do not . this creates an `` averaging '' effect in terms of the location of the peaks of quasar emission lines . because the shifts between the lines are dependent on luminosity for flux - limited samples , the degree of bias introduced depends on how closely the observed spectroscopic properties of the sample used for making the quasar composite resembles the sample to which it is applied . we additionally investigated the impact of these redshift biases on physical diagnostics of quasars based on iv civ emission - line shifts . adopting the heii - based redshifts for our sample results in a distribution of iv civ blueshifts that has a larger range than that found for the hw10-corrected sdss - pipeline redshifts ( cf . we find a median for our sample that is statistically consistent with zero blueshift within our measurement uncertainties , for objects with a strong heii narrow component , and an extended tail to large blueshifts , dominated by objects with little or no narrow heii line emission . this general picture is consistent with other current studies using more sophisticated redshift determination methods for high-@xmath15 quasars ( p. hewett , private communication ) and with a physical picture where sed - dependent effects can cause a range in iv civ blueshifts , but that iv civ blueshifts are not ubiquitous across the quasar population within the precision probed by our study . this result has implications for iv civ - based bh mass estimates , the reliability of which has been called into question partially because of the presumed ubiquitous presence of these blueshifts and their being attributed to non - virial motions in the blr . if instead , iv civ blueshifts are not ubiquitous and redshifts are reliably determined , carefully chosen samples can largely avoid objects observed to exhibit presumed non - virial blr motions much larger than the uncertainties in the line width measurements used for estimating the black hole masses , thus leading to reliable black hole mass estimates for such samples based on current calibrations . future work ( bizogni et al . , in preparation ) is aimed at formulating new calibrations for single - epoch iv civ - based masses that will account for these large iv civ blueshifts and better account for other non - virial components , such as the non - variable , low - velocity core ( see * ? ? ? our results suggest that future improvements in quasar template - based redshift determinations could potentially be made by creating multiple templates along the eigenvector 1 ( ev1 ) sequence of quasars properties that spans the observed spectroscopic diversity of quasars . template selection for each quasar can then be made by comparing as many measurable spectroscopic properties as possible that have been found to relate to ev1 properties , e.g. , line ratios , emission - line shapes , line ew , luminosity , continuum slope , x - ray soft excess and power - law slope , etc . @xcite . alternately , for pca - based redshifts , efforts should be made to formulate a pca training set composed of quasars covering this same ev1 parameter space , but for which the spectra of each training - set quasar is of high s / n and covers both uv and optical rest wavelengths . both requirements are important so that the redshifts of the training set are as robust as possible ( e.g. , based on multiple , unblended narrow emission lines and/or host - galaxy absorption lines , which are often of low ew and require high - quality data to isolate ) . we are grateful to yue shen for producing the coadded sdss - rm spectra used in this work and to c. s. kochanek for discussions and editorial contributions . we thank p. hewett and g. t. richards for stimulating discussions that improved the content and clarity of this work . kdd is supported by an nsf aapf fellowship awarded under nsf grant ast-1302093 . wnb acknowledges support from nsf grant ast-1516784 . cjg acknowledges support from nsf grant ast- 1517113 . bmp is grateful for support from nsf grant ast-10008882 . kh acknowledges support from stfc grant st / m001296/1 . lch thanks carnegie observatories for providing telescope access and acknowledges financial support from peking university , the kavli foundation , the chinese academy of science through grant no . xdb09030102 ( emergence of cosmological structures ) from the strategic priority research program , and from the national natural science foundation of china through grant no . jrt acknowledges support from nasa through hubble fellowship grant # 51330 awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc . , for nasa under contract nas 5 - 26555 . funding for sdss - iii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , and the u.s . department of energy office of science . the sdss - iii web site is http://www.sdss3.org/. sdss - iii is managed by the astrophysical research consortium for the participating institutions of the sdss - iii collaboration including the university of arizona , the brazilian participation group , brookhaven national laboratory , university of cambridge , carnegie mellon university , university of florida , the french participation group , the german participation group , harvard university , the instituto de astrofisica de canarias , the michigan state / notre dame / jina participation group , johns hopkins university , lawrence berkeley national laboratory , max planck institute for astrophysics , max planck institute for extraterrestrial physics , new mexico state university , new york university , ohio state university , pennsylvania state university , university of portsmouth , princeton university , the spanish participation group , university of tokyo , university of utah , vanderbilt university , university of virginia , university of washington , and yale university . & & & & & & & + boss@xmath6heii & all & 482 & 25.7 & 1080 & 1605 & 1057 & 2459 + boss@xmath6heii & q = 1,2,3 & 237 & 23 & 1320 & 1230 & 1366 & 1928 + boss@xmath6heii & q = 4,5,0 & 245 & 29.7 & 690 & 2085 & 758 & 2854 + boss@xmath6hw10 & hw10+sdssrm & 47 & 61.5 & -386 & 1781 & -618 & 2112 + boss@xmath6heii & hw10+sdssrm & 47 & 61.5 & 990 & 1755 & 948 & 2782 + hw10@xmath6heii & hw10+sdssrm & 47 & 61.5 & 1500 & 1898 & 1566 & 2644 + hw10@xmath6heii ; q=1,2,3 & hw10+sdssrm & 17 & 81.1 & 1500 & 1641 & 1310 & 1687 + hw10@xmath6heii ; q=4,5,0 & hw10+sdssrm & 30 & 59.7 & 1319 & 2231 & 1711 & 3076 + heii@xmath6psheii & all & 482 & 25.7 & 54 & 475 & 66 & 617 + heii@xmath6psheii & q = 1,2,3 & 237 & 23 & 98 & 402 & 130 & 472 + heii@xmath6psheii & q = 4,5,0 & 245 & 29.7 & 5 & 620 & 3 & 726 + heii@xmath6ps[oii ] & sdssrm w/ [ oii ] & 154 & 24.7 & -348 & 780 & -354 & 775 + heii@xmath6ps[oii ] ; q=1,2,3 & sdssrm w/ [ oii ] & 87 & 23.2 & -339 & 643 & -377 & 667 + heii@xmath6ps[oii ] ; q=4,5,0 & sdssrm w/ [ oii ] & 67 & 25.1 & -455 & 963 & -325 & 901 + + [ -1.5ex ] * iv civ blueshifts : * & & & & & & & + boss - based @xmath15 & all & 482 & 25.7 & 1258 & 664 & 1526 & 1874 + boss - based @xmath15 & q = 1,2,3 & 237 & 23 & 1217 & 615 & 1394 & 1740 + boss - based @xmath15 & q = 4,5,0 & 245 & 29.7 & 1301 & 729 & 1652 & 1989 + heii - based @xmath15 & all & 482 & 25.7 & 26 & 706 & 469 & 1899 + heii - based @xmath15 & q = 1,2,3 & 237 & 23 & -118 & 362 & 29 & 1323 + heii - based @xmath15 & q = 4,5,0 & 245 & 29.7 & 459 & 1111 & 895 & 2247	we use the coadded spectra of 32 epochs of sloan digital sky survey ( sdss ) reverberation mapping project observations of 482 quasars with @xmath0 to highlight systematic biases in the sdss- and boss - pipeline redshifts due to the natural diversity of quasar properties . we investigate the characteristics of this bias by comparing the boss - pipeline redshifts to an estimate from the centroid of heii@xmath11640 . heii has a low equivalent width but is often well - defined in high - s / n spectra , does not suffer from self - absorption , and has a narrow component that , when present ( the case for about half of our sources ) , produces a redshift estimate that , on average , is consistent with that determined from [ oii ] to within 1@xmath2 of the quadrature sum of the heii and [ oii ] centroid measurement uncertainties . the large redshift differences of @xmath31000 ^-1 km s@xmath4 , on average , between the boss - pipeline and heii - centroid redshifts suggest there are significant biases in a portion of boss quasar redshift measurements . adopting the heii - based redshifts shows that iv civ does not exhibit a ubiquitous blueshift for all quasars , given the precision probed by our measurements . instead , we find a distribution of iv civ centroid blueshifts across our sample , with a dynamic range that ( i ) is wider than that previously reported for this line , and ( ii ) spans iv civ centroids from those consistent with the systemic redshift to those with significant blueshifts of thousands of kilometers per second . these results have significant implications for measurement and use of high - redshift quasar properties and redshifts and studies based thereon .
the transition metal perovskite oxides exhibit many interesting physical phenomena . in some of these compounds , the orbital degrees of freedom play an important role in their magnetic properties due to the strong spin - orbital coupling @xcite . examples include the mott - hubbard type insulators @xmath5 and @xmath6 , which show very unusual magnetic properties . although the early experiments on @xmath5 and @xmath6 were reported back in the mid 1970 s @xcite , there has been renewed interest in the past decade on these materials @xcite . there are two magnetic phases in @xmath5 : c - type antiferromagnetic order ( ferromagnetic chains along the z - axis which stagger within the x - y plane ) at temperature @xmath7k @xmath8k , and g - type antiferromagnetic order ( staggered in all three directions ) at temperature @xmath9k @xcite . the magnetic order in @xmath6 is always c - type . the microscopic mechanism leading to the difference between these two compounds is still under investigation , and it might be related to the fact that at room temperature the cubic crystal structure is significantly distorted in @xmath5 but almost undistorted in @xmath6 . it is generally believed that the relevant orbital degrees of freedom , the degenerate or almost degenerate @xmath10 states are crucial to the observed magnetic properties . there have also been interesting theoretical studies related to these magnetic behaviors @xcite . in particular , khaliullin @xmath11 @xcite considered a spin - orbital hamiltonian starting with 3-fold degenerate @xmath12 orbitals , and compared the free energies between the c - type and g - type spin states in @xmath13 by including an explicit jahn - teller energy in the model . very recently , blake @xmath11 @xcite reported neutron diffraction experiment in @xmath0 , which shows clear evidence that the orbital ordering has a sudden change from high temperature g - type to low temperature c - type at the @xmath14k magnetic phase transition , manifested by a change in the jahn - teller type of distortion . the data also show clear evidence for the orbital transition from high temperature disordered phase to the g - type ordered phase at @xmath3k . this has motivated us to study the spin - orbital ordering in @xmath0 . in this paper , we consider a microscopic model for insulating @xmath5 , where each v - ion has two electrons with parallel spins favored by the large coulomb repulsion and the hund s coupling . the distortion in the cubic crystal structure already present at room temperature splits the degeneracy of the three @xmath12 orbitals so that the @xmath15 orbital is favored by the crystal field . in our model , we take it as always singly occupied , while the other electron occupies either the @xmath16 or the @xmath17 state . this description is consistent with the neutron diffraction experiment @xcite . we consider the superexchange interaction of the model and derive an effective hamiltonian for @xmath5 . we then study the mean field classical solutions of the model , and examine the spin and orbital orderings . we find a g - type orbital ordering at a higher temperature followed by an additional c - type spin ordering at a lower temperature . our result is consistent with the observed orbital phase transition at @xmath3 k , and spin phase transition at @xmath7k in @xmath5 . in this scenario , the orbital ordering at @xmath3k is of the electronic origin , and the lattice distortion at @xmath3k observed in the experiment is a consequence of the orbital ordering and the electron - lattice coupling . the superexchange interaction alone considered in our model does not explain the phase transition at @xmath14k , which may require other interactions such as the jahn - teller effect as proposed in previous articles @xcite . this paper is organized as follows . in section ii , we examine a multi - band hubbard model at electron density 2 electron per site , and consider the limit of large coulomb repulsion and the large hund s coupling . we then derive an effective hamiltonian based on the superexchange mechanism . in section iii , we discuss the mean field classical solutions of the model , and examine the phase diagram for the orbital and spin orderings . a brief summary is given in section iv . in @xmath5 , the vanadium electron configuration is @xmath18 . the compound has a cubic crystal structure , and each @xmath19 ion is surrounded by six oxygen ions . due to the cubic crystal field , the five - fold degenerate @xmath20 orbitals are split into a higher energy doublet of @xmath21 orbitals and a lower energy triplet @xmath12 orbitals . at low temperatures and for low energy physics , the relevant orbitals are the three - fold @xmath12 orbitals : @xmath15 , @xmath22,@xmath23 . in the strong coupling limit , the on - site coulomb repulsion between the two electrons in the @xmath20 states and the hund s coupling are much larger than the intersite electron hopping amplitudes , the system is a mott insulator with each v - ion having two localized electrons of parallel spins in two out of three degenerate @xmath12 orbitals . this scenario appears to be consistent with experiments . as indicated in the recent diffraction experiment @xcite , the cubic crystal is distorted at room temperature . as a result , the v - o bond distances are anisotropic . here we consider the structure at room temperature , where the v - o bond distance along @xmath24-axis ( perpendicular axis ) is the smallest(see figure 1 ) . this crystal structure further splits the @xmath25 states . the @xmath15 orbital has a lower energy , and becomes always singly occupied . the other d - electron is either in @xmath22 or @xmath23 orbital . in the diffraction experiment @xcite , the data also indicate a smaller difference in v - o bond lengths in the @xmath26 plane , which we shall neglect here for simplicity . = 8.5 cm the atomic hamiltonian @xcite is then given by @xmath27 , where the sum over @xmath28 runs all the v - sites , and @xmath29 in the above hamiltonian , @xmath30 creates ( annihilates ) an electron of orbital @xmath31 and spin @xmath32 at site @xmath28 , @xmath33 . @xmath34 , and @xmath35 , with @xmath36 representing orbitals @xmath37 , @xmath23 , @xmath15 respectively . @xmath38 is the on - site direct interaction , and j is the exchange interaction , or the hund s coupling . for the @xmath12 orbitals , @xmath39 for @xmath40 . in the case @xmath41 , this hamiltonian leads to an atomic ground state with each @xmath42 ion having a total spin @xmath43 with two - fold degenerate orbital configurations ( @xmath44 ) and ( @xmath45 ) . this last restriction in orbital configurations is valid for @xmath0 with strong lattice distortion but not for @xmath46 where the cubic structure is almost undistorted at room temperature . we next introduce the intersite hopping hamiltonian @xmath47 , given by @xmath48 where the sum runs over all the nearest neighbor v - v pairs , and @xmath49 is the electron hopping integral between two sites @xmath50 and @xmath51 from orbital @xmath31 to orbital @xmath52 . since the most important contribution to the hopping integrals is from the path via the @xmath53 state of the o - ion between the two neighboring v - ions , the hopping integrals are diagonal in the present problem due to the cubic symmetry . namely , we have @xmath54 . therefore , there are only two independent hopping parameters , @xmath55 , and @xmath56 , with the super - index indicating the direction of the two sites . in the limit @xmath57 , the system is an insulator with spin 1 on each v - ion . however , the virtual hopping introduces an effective intersite coupling of spins and the occupied orbitals . the effective hamiltonian for @xmath58 can be derived by applying perturbation theory to second order in @xmath59 or @xmath60 . let @xmath61 be a ground state of @xmath62 for two v - ions @xmath63 , where @xmath64 is the spin z - component , and @xmath65 is a pseudospin-1/2 operator for the orbitals : @xmath66 if @xmath22 is occupied , and @xmath67 if @xmath16 is occupied . the matrix elements between the unperturbed ground states of the two v - ions can be calculated within the second order perturbation theory , and it is given by , @xmath68 where the sum is over all the intermediate eigenstates @xmath69 of @xmath70 corresponding to the eigen energy @xmath71 , and @xmath72 is the ground state energy of @xmath70 . two - electron states with total spin @xmath43 are given in ref . the electronic configuration of the intermediate state @xmath73 is @xmath74 on one v - ion and @xmath75 on the other . in the appendix , we list all the states for @xmath76 , and the corresponding energy difference @xmath77 . the effective hamiltonian can be derived from these matrix elements . defining for each site a spin-1 operator @xmath78 and a pseudospin-1/2 operator @xmath65 that act on the @xmath79 and @xmath80 degrees of freedom , it can then be expressed as below , @xmath81\end{aligned}\ ] ] where @xmath82 gives the direction of the bond @xmath83 . the first term corresponds to spin - dependent orbital couplings while the second corresponds to orbital couplings which are spin independent . the first term also shows that the effective spin - spin couplings depend on orbital configuration . equivalently , by defining @xmath84 and @xmath85 , @xmath86 can be written as @xmath87\end{aligned}\ ] ] so that @xmath88 and @xmath89 are interpreted as the intersite orbital couplings for parallel spins ( @xmath90 , @xmath91 ) and antiparallel spins ( @xmath90 , @xmath92 ) respectively . we choose the energy unit to be @xmath93 , and denote @xmath94 , @xmath95 , @xmath96 , and @xmath97 . @xmath98 can be expressed in terms of parameters @xmath99 and @xmath100 , and are given below . @xmath101 in the above equations , @xmath102 , and @xmath103 and @xmath104 respectively . for a bond in the @xmath105 direction , where the @xmath15 orbital is inert due to zero hopping amplitude and the @xmath16 and @xmath22 hopping is isotropic , our model is similar to the original kugel - khomskii model@xcite with two differences . the first is the replacement of spin @xmath106 by spin @xmath107 the second is the effect of the @xmath15 occupation which changes the hund s coupling contribution to the on - site energies . we first discuss the intersite pseudospin couplings between two parallel spins . in this case , the psuedospin has a @xmath108 symmetry along @xmath105 -direction . along @xmath109- or @xmath110-direction , however , the virtual hopping integral for orbital @xmath111 or orbital @xmath112 vanishes , so there is no exchange term in the pseudospin , and @xmath113 is of the ising form . the pseudospin coupling between the two v - ions of antiparallel spins is quite different . there is a linear term @xmath114 along @xmath109- or @xmath110-direction , which either favors @xmath23 or @xmath22 orbital occupation to gain energy via the virtual hopping process . the pseudospin coupling along @xmath115-direction includes both the exchange term @xmath116 and the pair flip term @xmath117 . in spite of an isotropic matrix in the @xmath105-direction , the orbital hamiltonian is not @xmath108 symmetric because of the presence of hund s coupling . in particular , the pair flip term is related to superexchange processes involving those intermediate states listed in table ii that are split in energy due to hund s coupling . to illustrate this point further , we consider a pair of v - ions along @xmath105-direction with antiparallel spins and pseudospins are @xmath118 . the relevant intermediate states in the superexchange are the states listed in the second and the fifth rows in table ii . because these states have different energies , there is non - zero amplitude for the pseudospins to flip to @xmath119 . the pseudospin pair flip process is actually quite common in orbital physics . for example , there are pair flip terms in the effective hamiltonian for spin-1/2 systems with orbital degeneracy derived by castellani et al . @xcite . in the limit @xmath120 , @xmath121 , and @xmath122 , and so we have @xmath123 , and @xmath124 . the orbital coupling between the two v - ions of antiparallel spins vanishes , and the orbital coupling between the two v - ions of parallel spins remains to be pseudospin @xmath108 symmetric along @xmath105-direction and pseudospin ising symmetric along @xmath109- or @xmath110-directions . for @xmath120 , the lack of the global pseudospin @xmath108 symmetry is due to the anisotropic hopping integrals in the system . our effective hamiltonian here is similar to the hamiltonian proposed previously by khaliullin et al . @xcite with the following differences . these authors considered a model with 3-fold @xmath12 orbital degeneracy , while we consider a 2-fold orbital degeneracy with @xmath15 orbital being always singly occupied . below we shall compare the two hamiltonians by considering the the spin- orbital coupling between two v - ions along @xmath105-direction , namely @xmath125 , with @xmath126 . this may be carried out by imposing the orbital @xmath15 being always singly occupied in the hamiltonian of khaliullin et al . we find that the @xmath127 terms are the same in the two theories . however , the hamiltonian of khaliullin et al . does not include the pseudospin flip term @xmath128 . as we illustrated above , the pseudospin flip term is non - zero . as we shall see in the next section , this pair flip term does not affect the mean field results , which depend only on the z - components of pseudopsin in the present case . it will be an interesting question to examine if the pseudospin flip term is important to the orbital fluctuations . we start with the classical solutions of @xmath86 . the hamiltonian has a global su(2)symmetry in spin space , so that we can assume the spin ordering along the z - direction . the hamiltonian is invariant under the simultaneous transformation of global @xmath129 ( reversing orbitals at all sites ) and a @xmath130 rotation of the lattice about the @xmath105-axis . in general , we should consider orbital ordering along an arbitrary orientation . however , for the present problem , the orbital z - component terms are always larger or equal to the x- or y - component terms in @xmath98 of eq . therefore , we can discuss the classical solutions by considering the z - component of the orbital ordering only @xcite . in other words , the classical solutions are the same as the ising solutions in the present case . in table i , we show the energies per site in various classical states . note that @xmath131 , and @xmath132 , where the equality holds if and only if @xmath133 . we consider below the case with non - zero hund s coupling @xmath134 . in this case , the two states listed in table i with c - type antiferro - orbital ( co ) configuration have higher energies . also , we can see that the g - type antiferromagnetic spin ( gs ) and g - type antiferro - orbital ( go ) phase has a higher energy than the cs - go phase . therefore , the ground state is either ferromagnetic spin ( fs ) and go , or cs - go . in both fs - go and cs - go phases , the orbital is antiparallel in all three directions , favored by the combination of the symmetries in hopping integrals ( due to the cubic crystal symmetry ) and the hund s coupling . the asymmetry between the spin configuration along the z - axis and in the x - y plane is a result of the splitting of the @xmath15 orbital level from the other two @xmath12 orbitals . as expected , the fs - go phase is energetically more favored at a larger j where the hund s coupling dominates , and the cs - go phase is more favored at a smaller j. it may be helpful to understand these two possible ground states by examining the following limiting cases more explicitly . in the limit of large hund s coupling , @xmath135 , the terms in the energy expression in table i proportional to @xmath136 dominate . hence the fs - go phase has the lowest energy . in the limit @xmath137 , @xmath138 , and @xmath139 , so that the cs - go phase has the lowest energy . it should be noted that although the @xmath15 orbital is always singly occupied , the virtual hopping of the @xmath15 electron means that our model is not identical to a model with one electron occupying degenerate @xmath16 and @xmath22 orbitals . in that model , for g - type orbital ordering , the ground state with non - zero hund s coupling would always be ferromagnetic . this is also true if we compare our model to a model with one electron occupying triply degenerate @xmath12 orbitals @xcite . we have examined the linear orbital operator term in eq . ( 6 ) . for the spin configurations considered in this paper , namely the c- or g - type antiferromagnets or ferromagnet , the contribution of this linear term to the energy vanishes	motivated by the recent neutron diffraction experiment on @xmath0 , we consider a microscopic model where each @xmath1 ion is occupied by two @xmath2 electrons of parallel spins with two fold degenerate orbital configurations . the mean field classical solutions of the spin - orbital superexchange model predicts an antiferro - orbital ordering at a higher temperature followed by a c - type antiferromagnetic spin ordering at a lower temperature . our results are qualitatively consistent with the observed orbital phase transition at @xmath3k and the spin phase transition at @xmath4k in @xmath0 .
one of the objectives of systems control is _ performance regulation _ , namely the output tracking of a given setpoint reference despite modeling uncertainties , time - varying system s characteristics , noise , and other unpredictable factors having the effects of system - disturbances . a commonly - practiced way to achieve tracking is by a feedback control law that includes an integrator . an integral control alone may have destabilizing effects on the closed - loop system , and hence the controller often includes proportional and derivative elements as well thereby comprising the well - known pid control @xcite . recently there has been a growing interest in performance regulation of event - driven systems , including discrete event dynamic systems ( deds ) and hybrid systems ( hs ) , and a control technique has been proposed which leverages on the special structure of discrete - event dynamics @xcite . the controller consists of a standalone integrator with an adaptive gain , adjusted in real time as part of the control law . the rule for changing the gain is designed for stabilizing the closed - loop system as well as for simplicity of implementation and robustness to computational and measurement errors . therefore it obviates the need for proportional and derivative elements , and can be implemented in real - time environments by approximating complicated computations by simpler ones . in other words , the balance between precision and required computing efforts can be tilted in favor of simple , possibly imprecise computations . a key feature of the control law is that it is based on the derivative of the plant function , namely the relation between the system s control parameter and its output , which is computed or estimated by infinitesimal perturbation analysis ( ipa ) . this will be explained in detail in the following paragraphs . ipa is a well - known and well - tested technique for computing sample - performance derivatives ( gradients ) in deds , hs , and other event - driven systems with respect to controlled variables ; see @xcite for extensive presentations and surveys . its salient feature is in simple rules for tracking the propagations associated with a gradient along the sample path of a system , by low - cost algorithms . however , this simplicity may come at the expense of statistical unbiasedness of the ipa derivatives . in situations where ipa is biased , alternative perturbation - analysis techniques have been proposed , but they may require far - larger computing efforts than the basic ipa ( see @xcite ) . for the performance regulation technique described in this paper , it has been shown that ipa need not be unbiased and , as mentioned earlier , its most important requirement is low computational complexity @xcite . the control system we consider is depicted in figure 1 . assuming discrete time and one - dimensional variables , @xmath0 is the setpoint reference , @xmath1 denotes the time counter , the control variable @xmath2 is the input to the plant at time @xmath3 , @xmath4 is the corresponding output , and @xmath5 is the error signal at time @xmath3 . the control law is defined by the following equation , @xmath6 and we recognize this as an adder , the discrete - time analogue of an integrator , if the gain @xmath7 is a constant that does not depend on @xmath3 . the plant is an event - driven dynamical system whose output @xmath4 is related to its input @xmath2 in a manner defined in the next paragraph , and denoted by the functional term @xmath8 where @xmath9 is a random function . its ipa derivative @xmath10 is used to define the controller s gain @xmath11 via the equation @xmath12 and the error signal is defined as @xmath13 a recursive application of eqs . ( 1)-(4 ) defines the closed - loop system . as for the plant , it can have the following form . consider a continuous - time or discrete - time dynamical system whose input is @xmath14 , and its state is @xmath15 for some @xmath16 ; the notation @xmath17 designates continuous time or discrete time . let @xmath18 be a function that is absolutely integrable over finite - length intervals . partition the time - axis @xmath19 into contiguous left - closed , right - open intervals , @xmath20 , called _ control cycles_. suppose that the input to the latter dynamical system has constant values during each interval @xmath21 , and it can be changed only at the boundary of these intervals . in the setting of the system of figure 1 , @xmath22 is the value of the input @xmath23 during @xmath21 , and @xmath4 can be either @xmath24 or @xmath25 where @xmath26 is the duration of @xmath21 . in the case of discrete time , a sum - term of the form @xmath27 replaces the integral . we do not specify how to determine the control cycles @xmath21 , and they can have an a - priori constant length , or their termination can be the result of certain events . for example , let the plant consist of an m / d/1 queue , @xmath2 is the value of the service time during @xmath28 , and @xmath4 is the time - average of the sojourn times of all jobs arriving during @xmath28 . ipa can be used to compute the derivative @xmath29 via a well - known formula @xcite . generalizing this example , suppose that the plant - system is a stochastic event - driven system ( deds or hs ) , @xmath30 is a control variable , assumed to have a constant value ( @xmath22 ) during @xmath21 , @xmath31 , is a random function of @xmath22 as indicated by eq . ( 2 ) , and its derivative @xmath32 is computed by ipa . later we will be concerned with measurement and computational errors and hence modify eqs . ( 2 ) and ( 3 ) accordingly . the development of the proposed regulation technique has been motivated primarily by applications to computer cores , especially regarding regulation of power and instruction - throughput by adjusting the core s clock rate ( frequency ) @xcite . concerning throughput , there are no prescriptive , let alone analytic models for the frequency - to - throughput relationship , and a complicated , intractable queueing model has had to be used for simulation . nonetheless a simple ipa algorithm has been developed and used to approximate the sample derivative for determining the integrator s gain via eq . the regulation technique was extensively tested on programs from an industry - based suite of benchmarks , splash-2 @xcite , using a detailed simulation platform for performance assessment of computer architectures , manifold @xcite . we reported the results in @xcite , and deemed them encouraging and meriting a further exploration of the regulation technique . in the context of ipa research , this regulation technique represents two new perspectives . first , the traditional application of ipa throughout its development has been to optimization , whereas here it is applied in a new way , namely to performance regulation . second , much of the research of ipa has focused on its unbiasedness , whereas here , in contrast , the concern is with fast computation which may come at the expense of accuracy and unbiasedness . the main novelty of the paper as compared to references @xcite is in the fact that it concerns not simulation but an actual implementation . in this we were facing new challenges associated with real - time measurements , computations , and control . consequently we were unable to control each core separately as in @xcite , and hence applied the regulation method to a processor containing multiple cores . furthermore , due to issues with real - time computation , we were forced to take drastically cruder approximations to the ipa derivatives than in @xcite , and in fact it seems that we drove the degree of imprecision to the limit . how this worked on application programs will be seen in the sequel . in any event , the work described here is , to our knowledge , the first implementation ( beyond simulation ) of ipa in a real - time control environment . the rest of the paper is structured as follows . section 2 summarizes relevant convergence results of the regulation technique in an abstract setting . section 3 describes the system under study and its model , presents simulation results on manifold followed by implementation on a state - of - the - art computer processor , and compares the two . section 4 derives some lessons from these results and proposes directions for future research . this section recounts established results concerning convergence of the regulation technique defined by recursive applications of eqs . ( 1 ) to ( 4 ) , as summarized in ref . ideally convergence means that @xmath33 hence @xmath34 ( see figure 1 ) . this can be achieved under suitable assumptions ( mentioned below ) if the plant system is time invariant , and hence the function @xmath35 does not depend on @xmath36 . in that case the control loop comprised of equations ( 1)-(4 ) essentially implements newton s method for solving the equation @xmath37 , for which there are well - known sufficient conditions for convergence . these include situations where the derivative term @xmath38 in eq . ( 3 ) is computed in error , for which convergence in the sense of ( 5 ) is ascertained under upper bounds on the magnitude of the error @xcite . if the system is time varying , eq . ( 5 ) may not hold true , and in this case convergence can be characterized by the equation @xmath39 where @xmath40 depends on a measure of the system s variability . to make matters concrete let @xmath41 be a differentiable function , and suppose that the term @xmath42 in eq . ( 2 ) is a functional approximation of @xmath43 . assuming that @xmath44 is differentiable as well , we can view the term @xmath32 as an approximation to @xmath45 in ( 3 ) . however , for reasons that will be seen in the sequel , we add another layer of approximation to @xmath32 , denoted by @xmath46 , so that eq . ( 3 ) computes the term @xmath47 . defining @xmath48 and @xmath49 , eqs . ( 2 ) and ( 3 ) become @xmath50 and @xmath51 respectively . the regulation technique now is defined by recursive applications of eqs . ( 1),(7),(8),(4 ) . to analyze its convergence , suppose first , to simplify the discussion , that there exists a closed , finite - length interval , @xmath52 , such that every point @xmath22 computed by the regulation algorithm is contained in @xmath52 ; this assumption will be removed later . moreover , @xmath52 satisfies the following assumption . \(i ) the function @xmath53 and the functions @xmath54 are differentiable throughout @xmath52 . ( ii ) the function @xmath53 is either convex or concave , and monotone increasing or decreasing throughout @xmath52 . ( iii ) there exists @xmath55 such that @xmath56 . various ways to relax this assumption will be discussed shortly . the following result was proved in @xcite . _ proposition 2.3 and lemma 2.2 in @xcite _ : for every @xmath57 , @xmath58 , and @xmath59 , there exist @xmath60 , @xmath61 , and @xmath62 such that , for every interval @xmath52 satisfying assumption 1 and the following two additional conditions : ( i ) @xmath63 , and ( ii ) @xmath64 : 1 ) . if for some @xmath65 , and for @xmath66 , @xmath67 , @xmath68 , and @xmath69 , then @xmath70 2 ) . if for all @xmath71 , @xmath67 , @xmath68 , and @xmath69 , then @xmath72 @xmath73 in the context of the system considered in this paper , what we have in mind is a situation where the plant is an event - driven system controlled by a real - valued parameter @xmath74 , @xmath75 is an expected - value function defined over a finite horizon ( hence not in steady state and possibly dependent on an initial condition ) , @xmath76 is a sample - based approximation ( possibly biased ! ) of @xmath43 over the control cycle @xmath21 , and @xmath77 is a sample approximation of @xmath78 . a few remarks concerning assumption 1 are due . 1 ) . the differentiability assumption is unnecessary , convexity of @xmath75 and almost - sure differentiability of @xmath79 at a given point @xmath74 suffice . these conditions often arise in the context of ipa . under these weaker assumptions the proofs in @xcite can be carried out in the context of convex analysis rather than differentiable calculus . 2 ) . the condition that @xmath67 , @xmath36 , can be enforced in the case where @xmath52 is a constraint set for the sequence @xmath80 . in that case ( 1 ) would be replaced by @xmath81 where @xmath82 is the projection of @xmath83 onto i , i.e. , the point in @xmath52 closest to @xmath84 . the proof of convergence is unchanged . often the magnitude of the error terms @xmath85 and @xmath86 can be controlled by taking longer control cycles , but there is no way to ensure the inequalities @xmath68 and @xmath69 _ for every _ @xmath71 , which is stipulated as a condition for eq . practically , however , with long - enough control cycles we can expect those inequalities to hold for finite strings of @xmath87 , thus guaranteeing the validity of eq . . if these strings are long enough , equation ( 9 ) would result in @xmath88 approaching 0 at a geometric rate , then periodically jumping away due to the sporadic occurrence of larger errors , but again returning towards 0 rapidly , etc . this behavior has been observed in all of the examples where we tested the regulation technique for a variety of event - driven systems @xcite . another source of the jitters described in the previous paragraph is the time - varying nature of the system . this is particularly pronounced in the system tested in this paper , since the workload of programs processed by a core can vary widely in unpredictable ways . nonetheless we shall see that the regulation algorithm gives good results . it is possible to ascertain the assumptions underscoring the analysis in @xcite for simple systems ( e.g. , tandem queues and some marked graphs ) , but may be impossible to ascertain them for more complicated systems . for instance , it can be impossible to prove differentiability or even convexity of an expected - value function from characterizations of its sample realizations , or bounds on the errors associated with ipa . moreover , some of these assumptions , including convexity or concavity , were made in @xcite in order to enable an analysis but may not be necessary . the aforementioned convergence results serve to explain the behavior of the regulation method that was observed in all of our former experiments @xcite as well as in those described later in this paper . the system - architecture considered in this paper is based on an out - of - order ( ooo ) core technology whereby instructions may complete execution in an order different from the program order , hence the `` out of order '' designation . this enables instructions execution to be limited primarily by data dependency and not by the order in which they appear in the program . data dependency arises when an instruction requires variables that first must be computed by other instructions . a detailed description of ooo architectures can be found in @xcite , while a high - level description is contained in @xcite . here we provide an abstract functional and logical description , and refer the reader to @xcite for a more - detailed exposition . the functionality of an ooo core is depicted in figure 2 . instructions are fetched sequentially from memory and placed in the instruction queue , where they are processed by functional units , or servers in the parlance of queueing theory . the queue is assumed to have unlimited storage and there is a server associated with each buffer . the processing of an instruction starts as soon as it arrives _ and _ all of its required variables become available . the instruction departs from the queue as soon as its execution is complete _ and _ the previous instruction ( according to the program order ) departs . in the parlance of computer architectures , an instruction is said to be _ committed _ when it departs from the queue . the instruction - throughput of the core is defined and measured by the average number of instructions committed per second . [ ! t ] instructions generally are classified as _ computational instructions _ or _ memory instructions_. access times of external , off - chip , memory instructions are one - to - two orders of magnitude longer than those of computational instructions . therefore most architectures make use of a hierarchical memory arrangement where on - chip cache access takes less time than external memory such as dram . first the cache is searched , and if the variable is found there then it is fetched and the instruction is completed . if the variable is not stored in cache ( a situation known as _ cache miss _ ) then it is fetched from external memory ( typically dram ) and placed in the cache , whence it is accessed and the instruction is completed . external memory instructions can be thought of as being placed in a finite - buffer , first - in - first - out queue , designated as the _ memory queue _ in figure 2 . when this queue becomes full , the entire memory access , including cache , is stalled . thus , there are three causes for an instruction to be stalled : a computational or memory instruction waiting for variables computed by other instructions , a memory instruction waiting for the memory queue to become non - full , and any instruction waiting ( after processing ) for the previous instruction to depart from the queue . we point out that instructions involving computation and @xmath89 cache - access are subjected to the core s clock rate , while memory instructions involving external memory , such as dram , are not subjected to the same clock . this complicates the application of ipa and may cause it to be biased . a quantified discrete - event model of this process is presented in the appendix , and a more general description can be found in @xcite , which also contains a detailed algorithm for the ipa derivative of the throughput as a function of frequency . we use a cycle - level , full system discrete event simulation platform for multi - core architectures , manifold . the simulated model consists of a 16-core x86 processor die , where each core is in a separate clock domain and can control its own clock rate independently of other cores . for a detailed description of the manifold simulation environment and its capabilities , please see ref . @xcite . we simulated two application programs from the benchmark - suite splash-2 , barnes and water - ns @xcite . barnes is a computation - intensive , memory - light application while water - ns is memory intensive . for each execution , all of the 16 cores of the processor run threads of the same benchmark concurrently while each one of them is controlled separately . the control cycle is set to @xmath90 ms for both barnes and water - ns . the frequency range of the cores is set to @xmath91 $ ] . we assume a continuous frequency range for the simulations , but later will consider a realistic , discrete range for the implementation described in the sequel . the target instruction throughput is set to the same value for each core for both barnes and water - ns , and we experiment with the target throughput values of 1,200 mips ( million instructions per second ) , 1,000 mips , and 800 mips . in terms of instructions per control cycle , these target values correspond to @xmath92 , @xmath93 , and @xmath94 , respectively . the relationship between clock frequency and instruction throughput is determined by the manifold processor model , but its ipa derivative was computed according to the high - level instruction flow described above and in the appendix . for each application run we present , in the following paragraphs , the results for one of the 16 cores chosen at random . consider first the target throughput of 1,200 mips . the throughput simulation results for the barnes benchmark are shown in figure [ fig : mipsmanifoldbarnesco10t1200 ] , where the horizontal axis indicates time in ms and the vertical axis indicates instruction throughput . we discern a rise of the throughput from its initial value ( measured at 643.1 mips ) towards the target level of 1,200 mips , which it reaches for the first time in about 1.5 ms , or 15 control cycles . thereafter it oscillates about the target value , which is not surprising due to the unpredictable , rapidly - changing program workload . the average throughput computed over the time interval [ @xmath95,@xmath96 ( soon after the throughput has reached the target value ) is @xmath97 mips , which is 42.6 mips off the target level of 1,200 mips . figure [ fig : freqmanifoldbarnesco10target1200 ] depicts the graph of frequency vs. time ( in ms ) , and it shows some saturation at its highest level of 5.0 ghz , in the time - interval @xmath98 $ ] . saturation at the highest level can correspond to a negative offset of the average throughput from its target level , since it indicates that the system may be unable to raise the throughput to a desired level . during the period of frequency saturation indicated in figure 4 , the throughput shown in figure 3 it more jittery and sporadically attains slightly - lower values than after time 25ms . it also shows these characteristics between the end of the saturation period and time @xmath99 . therefore , the extent of the effects of the saturation on the aforementioned offset of 46.4 mips is not clear . nonetheless we mention this point since it will be more pronounced in some of the results on which we report later . also , we computed the average throughput in the intervals @xmath100 $ ] and @xmath101 $ ] , after the jittery behavior of the throughput has somewhat subsided . the results are 1,192.6 mips and 1,192.9 mips , respectively , corresponding to offsets of 7.4 mips and 7.1 mips from the target throughput of 1,200 mips . these results suggest that the frequency saturation plays some role in the larger , 46.4-mips offset that was computed over the interval @xmath102 $ ] . for the target throughput of 1,000 mips , the results ( not shown due to space limitations ) showed a rise in throughput from its initial value of 420.5 mips to 1,000 mips in 2.1 ms , or 21 control cycles . the average throughput in the @xmath103 $ ] interval is 990.2 mips , corresponding to an offset of 9.8 mips of the throughput from its target value of 1,000 mips . the frequency saturated at its upper limit only at 5 isolated control cycles with minimal effects on the throughput . for the throughput target of 800 mips , the results show a rise in the throughput from its initial value of 679.3 mips to 800 mips in 1.9 ms , or 19 iterations . the average throughput in the interval [ 1.9ms,100ms ] was 839.6 mips , which is 39.6 mips off the target value of 800 mips . there was a considerable frequency saturation at the lowest level of 0.5 ghz , which explains the positive offset . returning to the results for the target level of 1,200 mips , we considered a way to reduce the throughput oscillations and frequency saturation by scaling down the gain in eq . we did this by replacing eq . ( 1 ) by the following equation , @xmath104 for a suitably - chosen constant @xmath105 . after some experimentation on various benchmarks ( excluding those tested here ) we chose @xmath106 . the resulting frequencies did not saturate throughout the program s run , and yielded an average throughput of 1,198,5 mips , which is 1.5 mips off the target level of 1,200 mips . though working well for this example , this technique may be problematic when used with an implementation rather than simulation , as will be discussed in the next subsection . for water - ns , consider first the throughput target of 1,200 mips . simulation results of throughput and frequency are shown in figure [ fig : mipsmanifoldwaternsco10target1200 ] and figure [ fig : freqmanifoldwaternsco10target1200 ] , respectively . we notice greater fluctuations and more saturation than for barnes . in particular , figure [ fig : freqmanifoldwaternsco10target1200 ] shows three distinct periods of frequency saturations at its upper limit of 5.0 ghz , and figure [ fig : mipsmanifoldwaternsco10target1200 ] shows very low throughput during these periods . to explain this , recall that water - ns is a memory - heavy program , and execution times of memory instructions are longer ( typically by one or two orders of magnitude ) than computational instructions . during those periods the instructions of water - ns mainly concern memory access , which are low - throughput instructions . the controller is applying its highest frequency in order to push the throughput to its target value , but that frequency is not high enough to have much effect . this is why the periods of high - limit frequency saturation are characterized by very low throughput . this has a pronounced affect of lowering the average frequency measured during the program s execution . in fact , the throughput obtained from the simulation rises from its initial value of 429 to its target level of 1,200 mips in about 1.8 ms , or 18 control cycles ( this rise is not evident from figure [ fig : mipsmanifoldwaternsco10target1200 ] due to its insufficient granularity ) , and the average throughout from the time the target level is reached ( @xmath107 ) to the end of the program - run ( @xmath108 ) is @xmath109 mips , which is 73.2 mips off the target level of 1,200 mips . despite this offset , we observe that as soon as the program transitions from memory mode to computational mode , as indicated by the end of the saturation periods in figure [ fig : freqmanifoldwaternsco10target1200 ] , the throughput returns quickly to about its target level , as can be seen in figure [ fig : mipsmanifoldwaternsco10target1200 ] . for the target throughput of 1,000 mips , simulation results show the throughput increasing from its initial value of 472.1 mips to the target level on 1,000 mips in 2.3 ms , or 23 control cycles . there is considerable frequency saturation at the high limit of 5.0 ghz . the average throughput in the interval @xmath110 $ ] is 947.8 mips , which means an offset of 52.2 mips off the target throughput . for the target throughput of 800 mips , simulation results indicated a rise of the throughput from its initial value of 443.3 to its target level in about 2.3 ms , or 23 control cycles . there is considerable saturation of the frequency at its lower level of 0.5 ghz , and hence a positive offset between the computed average throughput and its target level . indeed , the average throughput in the interval @xmath110 $ ] is 862.6 , mips , hence meaning an offset of 62.6 mips of the throughput from its target value of 800 mips . all of these results are summarized in table i , showing the offset ( in mips ) of average throughput from target throughput , obtained from manifold simulations of barnes and water - ns with throughput targets of 1,200 , 1,00 , and 800 mips . returning to the throughput target of 1,200 mips , an application of the modified algorithm with @xmath106 in eq . ( 12 ) yielded the average throughput of 1,143.6 mips which is 56.4 mips off the target level of 1,200 mips . this is a smaller offset than the 73.2 mips obtained from the unmodified algorithm , and it is explained by the fact that there is still considerable , though less frequency saturation than with the unmodified algorithm . [ ! t ] [ ! t ] [ ! t ] [ ! t ] .manifold simulations : offset of average throughput from target levels [ cols="<,<,<,<",options="header " , ] comparing the data summarized in table i and table ii , we that the regulation technique performs slightly better on the haswell implementation platform than on the manifold simulation environment . the reason for this may be due to the fact that in the simulation experiment we regulate the throughput of each core separately , while in the implementation we control the average throughput of all the cores in the processor . this paper describes the testing of an ipa - based throughput regulation technique in multicore processors . the testing was performed on both a simulation environment and an implementation platform . despite crude approximations that have had to be made in the implementation setting , the proposed technique performed slightly better than in the simulation setting . future research will extend the regulation method from a centralized control of a single processor to a distributed control of networked systems . this section provides a quantitative description of the instruction - flow in the ooo - cache high - level model described at the beginning of section iii . denote by @xmath111 , @xmath112 , the instructions arriving at the instruction queue in increasing order . let @xmath74 denote the clock rate , or frequency , and let @xmath113 be the clock cycle . denote by @xmath114 the arrival time of @xmath111 relative to the arrival time of @xmath115 , namely @xmath116 , and let @xmath117 be the clock counter at which @xmath111 arrives . then , @xmath118 denote by @xmath119 the time at which execution of @xmath111 starts , and let @xmath120 denote the time at which execution of @xmath111 ends . we next describe a way to compute @xmath119 . consider first the case were @xmath111 is a computational instruction . if all of its required variables are available at its arrival time then @xmath121 . on the other hand , if @xmath111 has to wait for such variables , let @xmath122 denote the index ( counter ) of the instruction last to provide such a variable , then @xmath123 . next , if @xmath111 is a memory instruction , then @xmath119 is the time it starts a cache access . if the memory queue is not full at time @xmath114 , then @xmath121 . on the other hand , if the memory queue is full at time @xmath114 , let @xmath124 denote the index of the instruction at the head of the queue , then , @xmath125 . to compute @xmath120 , consider first the case where @xmath111 is a computational instruction . let @xmath126 denote the number of clock cycles it takes to execute @xmath111 . then , @xmath127 . on the other hand , if @xmath111 is a memory instruction , let @xmath128 denote the number of clock cycles it takes to perform a cache attempt . if the cache attempt is successful and the variable is found in cache , then @xmath129 . if the variable is not in cache , the instruction is directed to the memory queue . its transfer there involves a small number of clock cycles , @xmath130 , hence its arrives at the queue at time @xmath131 . the memory queue is a fifo queue whose service time represents an external - memory access , which is independent of the core s clock . denote by @xmath132 the sojourn time of @xmath111 at the memory queue . then @xmath133 . finally , the departure time of @xmath111 from the instruction queue , denoted by @xmath134 , is @xmath135 . given a control cycle consisting of @xmath136 instructions , the throughput is defined as @xmath137 . since @xmath138 , we can view the throughput as a function of @xmath74 and denote it by @xmath139 . its ipa derivative , @xmath140 , can be computed by following the above dynamics of the instructions flow . this , and a more detailed discussion of the model , can be found in @xcite . y. wardi , x. c. seatzu , chen , and s. yalamanchili , performance regulation of event - driven dynamical systems using infinitesimal perturbation analysis , under review in _ nonlinear analysis : hybrid systems _ , also in arxiv , ref . arxiv:1601.03799v1 [ math.oc ] , 2016 . woo , m. oharat , e. torriet , j. singhi and a. guptat , the splash-2 programs : characterization and methodological considerations , _ proceedings of the isca 22nd annual international symposium on computer architectures , ( isca95 ) _ , santa margherita ligure , italy , 1995 . j. wang , j. beu , r. behda , t. conte , z. dong , c. kersey , m. rasquinha , g. riley , w. song , h. xiao , p. xu , and s. yalamanchili , manifold : a parallel simulation framework for multicore systems , _ proc . ieee international symposium on performance evaluation of systems and software ( ispass ) _ , 2014 . x. chen , h. xiao , y. wardi , and s. yalamanchili , throughput regulation in shared memory multicore prtocessors , _ proc . 22nd ieee intl . conference on high performance computing ( hipc ) _ , bengaluru , india , december 16 - 19 , 2015 . c. seatzu , and y. wardi , performance regulation via integral control in a class of stochastic discrete event dynamic systems , _ proc . 12th ifac - ieee international workshop on discrete event systems ( wodes14 ) _ , paris , france , may 14 - 16 , 2014 . s. browne , j. dongarra , n. garner , g. ho , and p. mucci , a portable programming interface for performance evaluation on modern processors , _ the international journal of high performance computing applications _ , volume 14 , number 3 , pp . 189 - 204 , fall 2000 .	a new technique for performance regulation in event - driven systems , recently proposed by the authors , consists of an adaptive - gain integral control . the gain is adjusted in the control loop by a real - time estimation of the derivative of the plant - function with respect to the control input . this estimation is carried out by infinitesimal perturbation analysis ( ipa ) . the main motivation comes from applications to throughput regulation in computer processors , where to - date , testing and assessment of the proposed control technique has been assessed by simulation . the purpose of this paper is to report on its implementation on a machine , namely an intel haswell microprocessor , and compare its performance to that obtained from cycle - level , full system simulation environment . the intrinsic contribution of the paper to the workshop on discrete event system is in describing the process of taking an ipa - based design and simulation to a concrete implementation , thereby providing a bridge between theory and applications .
metal oxides demonstrate a variety of physical and chemical properties , sometimes in intriguing combinations . apart from being ubiquitous in nature , metal - oxide surfaces and interfaces find diverse technological applications and are being explored for potential use in future electronic devices . in particular , surfaces of magnetoelectric antiferromagnets such as possess an equilibrium surface magnetization , @xcite making them suitable for use as active layers in electrically - switchable magnetic nanostructures . @xcite the ( 0001 ) surface has been a subject of many experimental @xcite and theoretical @xcite studies , but its surface remains poorly understood . low - energy electron diffraction ( leed ) experiments for a thin ( 0001 ) film grown on a cr ( 110 ) single crystal revealed an unusual reentrant structural phase transition,@xcite in which the surface structure changes from @xmath5 to @xmath6 and back to @xmath5 under cooling from room temperature to 150 k and then further down to 100 k. the origin of these phase transitions is not understood . while the high - temperature transition may , as suggested by the leed data , @xcite be a conventional order - disorder transition , the second one is unusual in that a more symmetric phase appears at lower temperatures . the situation is further complicated by the fact that , as shown by takano _ @xcite , both phase transitions disappear for thicker films grown in a similar way . this suggests that the epitaxial strain has an important effect on the surface energetics . oxidation of the cr @xmath7 surface was investigated by leed and auger spectroscopy , @xcite and it was found that under growth conditions similar to those of ref . the thin ( 0001 ) film is subject to a tensile epitaxial strain of about 1.5% . in this paper we study the structure of the ( 0001 ) surface using first - principles electronic structure calculations and monte carlo simulations . our results suggest that the dynamics of the system is driven by the occupation of two competing surface cr sites . the system can be mapped to an ising model on a two - dimensional hexagonal lattice in external field . the first phase transition is clearly identified as a conventional ordering transition , and the theoretical transition temperature is found to be in good agreement with experiment . our calculations further reveal a strong effect of tensile epitaxial strain , which parametrically drives the system towards configurational frustration , particularly in combination with antiferromagnetic ordering . an explanation of the second phase transition is offered based on these results . the paper is organized as follows . in section [ method ] we describe the computational methods . section [ surf - struc ] presents the results on the configurational and magnetic energetics of the ( 0001 ) surface , including the identification of the competing surface cr sites , the construction of the configurational hamiltonian , the analysis of magnetic interactions , and the evaluation of the ground - state phase diagram . section [ thermo ] deals with configurational thermodynamics of the surface , and section [ sec - exper ] discusses the relation of the results to experiments . the electronic structure of the ( 0001 ) surface is presented in section [ el - str ] , and its magnetic properties in section [ surf - mag ] . section [ summary ] draws the conclusions . electronic structure calculations were performed using the projector - augmented wave method @xcite implemented in the vasp code . @xcite for the cr @xmath8 shell we employed the rotationally - invariant lsda@xmath9 method @xcite with @xmath10 ev and @xmath11 ev . this method was preferred over gga@xmath9 adopted in ref . due to its better description of the structural , electronic , and magnetic properties of bulk .@xcite different surface superstructures were modeled using supercells representing symmetric slabs with eight atomic layers of o and 16 atomic layers of cr stacked along the ( 0001 ) direction . the periodically repeating slab is separated from its image by 1.5 nm of vacuum . we considered @xmath5 , @xmath12 , @xmath13 , and @xmath6 surface supercells ( where @xmath5 corresponds to the hexagonal unit cell of bulk ) . the lateral dimensions of the unstrained supercell were fixed to the calculated equilibrium bulk values;@xcite for the strained case these values were used as a reference . apart from these constraints , the ionic positions were relaxed until the hellmann - feynman forces were converged to less than @xmath14 ev@xmath15 . the plane - wave energy cutoff was fixed to @xmath16 ev and the brillouin zone integration was performed using @xmath17-centered monkhorst - pack grids . @xcite for relaxation we used gaussian smearing of @xmath18 ev and a @xmath19-point mesh equivalent to or denser than @xmath20 for the @xmath5 surface supercell . we checked the convergence with respect to the number of @xmath19-points , the energy cutoff for the plane wave expansion , the size of the vacuum region , and the thickness of the slab . these tests indicate that the total energies are generally converged to within 1 mev . density of states ( dos ) calculations were performed using gaussian smearing of @xmath21 ev and a @xmath19-point mesh equivalent to or denser than @xmath22 for the @xmath5 supercell . the energy barriers for the thermally - activated jumping of cr ions between the two competing surface sites were calculated using the nudged elastic band method . @xcite seven images were inserted between the two energy minima , and in each image the ions were relaxed so that forces perpendicular to the reaction path were smaller than 0.05 ev / . crystalizes in the corundum structure with the @xmath23 space group . it can be viewed as a stacking of buckled honeycomb cr double layers along the ( 0001 ) direction with quasi - hexagonal closed - packed o layers in between , see fig.[slab ] . the ( 0001 ) surface is polar , and simple electrostatic arguments suggest that non - stoichiometric terminations by an o layer or by a cr double layer should lead to divergent electrostatic potential in the bulk . on the other hand , the surface can terminate in the middle of the buckled cr layer so that only half of the cr ions from this layer remain on the surface . although still polar , this termination is stoichiometric , and the electrostatic potential in the bulk is not divergent . it can therefore be expected that this termination is energetically favorable . indeed , surface termination by a single cr layer was consistent with leed @xcite and scanning tunneling microscope @xcite measurements of the ( 0001 ) surface in ultrahigh vacuum . further , first principles calculations by rohrbach _ @xcite based on the gga@xmath9 method have shown that this termination has the lowest surface energy ( compared to all others considered ) over the entire range of oxygen chemical potential where is stable . note that earlier results based on the gga method , which leads to grossly incorrect electronic and magnetic properties,@xcite were quite different.@xcite in this work we only consider the ( 0001 ) surface terminated by a single layer of cr . the location of the cr ions within the single cr terminating layer has been debated . within the double cr layer there are three possible octahedral sites , two of them being occupied in the bulk . they give rise to three nonequivalent surface sites ( a , c and d , see fig . [ slab ] ) that surface cr ions can occupy . occupation of site a corresponds to the continuation of the bulk structure . further , as pointed out by gloege _ et . al . _ , @xcite the surface cr ion can jump below the oxygen subsurface layer and occupy the empty octahedral site within the underlying cr double layer . this interstitial site is directly underneath the surface site a , and we denote it by b ( see fig . [ slab ] ) . surface terminations . gray and red spheres represent cr and o atoms , respectively.,scaledwidth=45.0% ] in order to identify the energetically preferable sites , we therefore considered four @xmath5 surface models corresponding to the exclusive occupation of sites a , b , c , or d , respectively . in all cases a significant inward relaxation was observed , as expected for a nominally polar surface . the relaxation data for models a and b are included in appendix [ app - a ] . we define the surface energy as @xmath24 here @xmath25 is the ground state energy of the slab for the given surface model with magnetic structure corresponding to bulk , @xmath26 is the ground state energy per unit cell of bulk , @xmath27 and @xmath28 are the numbers of atoms in the slab and in the bulk unit cell , and @xmath29 is the number of surface cr atoms on one side of the slab . the surface energies for the four @xmath5 surface terminations are given in table [ tab1 ] . the surface energy is the lowest when site a is occupied . occupation of sites c and d leads to much higher surface energies , and we therefore do not consider their occupation in the subsequent analysis . on the other hand , the surface energy of model b is only slightly higher than that of model a. thus , sites a and b can both be partially occupied , which can lead to non - trivial ordered terminations and phase transitions ; these issues are addressed in the following subsections . .surface energies of @xmath5 surface models with different surface sites occupied . [ cols="^,^,^,^,^",options="header " , ] [ tab4 ] the upper panel of fig . [ relax ] shows the vertical coordinate of surface cr ions occupying site a for different surface models as a function of @xmath30 defined after eq . ( [ model2 ] ) . this coordinate is referenced with respect to that of the cr ions occupying sites b averaged over different surface models . ( the subsurface o layer was used as an anchor for measuring the @xmath31 coordinate . ) for surface supercells with two inequivalent a - site cr ions their vertical coordinates were similar , and we used their average . one can see that the calculated data agree very well with eq . ( [ model2 ] ) for both unstrained and strained surfaces . of a cr ion at site a for different surface models as a function of @xmath30 defined after eq . ( [ model2 ] ) . circles ( squares ) correspond to the unstrained ( strained ) surface . solid lines are linear fits to the data . lower panel : @xmath32 as a function of @xmath33 ( see eq . ( [ ehat ] ) ) for different surface models . red ( light ) symbols correspond to the ground state , and blue ( dark ) symbols to the paramagnetic state . circles ( squares ) : data for unstrained ( strained ) surface . solid lines are fits to a quadratic function with zero constant term , see eq . ( [ ehat ] ) . from bottom to top , the curves are shifted upward by 0 , 0.01 , 0.01 and 0.03 ev , respectively.,scaledwidth=45.0% ] here we demonstrate the quality of the fit of _ ab initio _ energies to the configurational hamiltonian and explain the importance of the three - body term in eq . ( [ model3 ] ) . note that for all surface models for which _ ab initio _ energies were calculated , the a sites are equivalent ( ignoring the directionality of the bonds on the actual surface , see section [ config ] ) . in this case , the surface energy from eq . ( [ model3 ] ) can be rewritten as @xmath34 where @xmath33 is the value of @xmath30 for the a sites . setting @xmath35 , where @xmath36 and @xmath37 are the surface energies for models a ( @xmath5 ) and b ( @xmath5 ) , we can define @xmath38 in the lower panel of fig . [ relax ] we plotted @xmath32 as a function of r using the _ ab initio _ energies for all considered surface models . we included the data for both strained and unstrained surfaces using both ground state and paramagnetic surface energies . the resulting plots are very well fitted by the quadratic function with a zero constant term , demonstrating the high fidelity of the fit . the value of the parameter @xmath39 extracted from the fit ranges from @xmath40 to @xmath41 , which , as expected , is a small number . however , the relative importance of the three - body term compared to the two - body term is @xmath42 . since for the considered surface models @xmath33 varies between @xmath43 and @xmath44 , the relative importance of the three - body term is substantial and reaches 50% . in the diagrammatic cluster - expansion language one can say that although @xmath39 is small , the number of corresponding diagrams is large . h. jnsson , g. mills , and k. m. jacobsen , in nudged elastic band method for finding minimum energy paths of transitions in classical and quantum dynamics in condensed phase simulations , edited by b. j. berne , g. ciccotti , and d. f. coker ( world scientific , singapore , 1998 ) . the shaded areas in fig . [ temppd ] denote regions where the ground - state structure was not included in mfa . typically these structures have complicated orderings and should undergo first - order transitions to simpler orderings like a@xmath1b@xmath0 or ab at fairly low temperatures . therefore , with exception of the small rectangle on the left for @xmath45 , the high - temperature behavior in the shaded areas is correct within mfa .	the surface of a cr@xmath0o@xmath1 ( 0001 ) film epitaxially grown on cr undergoes an unusual reentrant sequence of structural phase transitions ( @xmath2 ) . in order to understand the underlying microscopic mechanisms , the structural and magnetic properties of the cr@xmath0o@xmath1 ( 0001 ) surface are here studied using first - principles electronic structure calculations . two competing surface cr sites are identified . the energetics of the surface is described by a configurational hamiltonian with parameters determined using total energy calculations for several surface supercells . effects of epitaxial strain and magnetic ordering on configurational interaction are also included . the thermodynamics of the system is studied using monte carlo simulations . at zero strain the surface undergoes a @xmath3 ordering phase transition at @xmath4 . tensile epitaxial strain together with antiferromagnetic ordering drive the system toward strong configurational frustration , suggesting the mechanism for the disordering phase transition at lower temperatures .
m dwarfs have received considerable attention regarding their internal stellar structure , their atmospheres , and their distribution in the galaxy ( see reviews by @xcite , @xcite and @xcite and references therein ) . however , their atmospheric structure , dynamics , and magnetic processes remain poorly understood . m dwarfs show clear evidence of chromospheric heating , as revealed by emission lines in optical and uv wavelengths @xcite that constrain models of chromospheric structure @xcite . they also contain hot coronae , as detected by their x - ray emission , with @xmath4 up to @xmath5 @xcite . @xcite and others have analyzed x - ray flaring in m dwarfs , and many observations of flaring activity have been made at other wavelengths @xcite . although hot outer atmospheres and flaring activity have been studied in stars of f , g , and k spectral types , m dwarfs deviate from chromospheric / coronal relations found in those stars , indicating a qualitative difference in atmospheric structure , velocity fields , and magnetic morphology @xcite . observations of chromospheres and coronae in m dwarfs have led to models that include significant amounts of nonradiative , magnetic heating in their outer atmospheres @xcite . the observed scatter in x - ray luminosities and chromospheric emission line fluxes can be attributed to temporally changing surface inhomogeneities @xcite or differences in filling factor between stars @xcite . @xcite showed that acoustic dissipation may provide enough coronal heating for stars dominated by quiet regions , but magnetic heating is required to account for the coronal heating in the more active m dwarfs that likely have sizable covering factors of kilogauss magnetic fields @xcite . here , we probe the atmospheres of m dwarfs with near - uv spectra at the resonant `` h and k '' transitions of at 3968 and 3933 respectively . we use sufficiently high resolution to resolve structure in the emission line profiles and to extract line widths , asymmetries , and central absorptions . we examine a sample of 147 late k and m dwarfs from spectra previously obtained by the california and carnegie planet search . multiple spectra for each star yield mean values over time and any temporal changes . we compare the h and k lines to h@xmath3 equivalent widths , providing a tie to coronal gas . in section [ sec : observations ] we describe the stellar sample and the keck spectra . in section [ sec : analysis ] we describe the double gaussian model of the lines . in section [ sec : results ] we present the resulting observed line parameters , and in section [ sec : discussion ] we offer discussion and summary . all spectra used in this study were taken from the california and carnegie planet search program , with a set - up described by @xcite and briefly summarized here . from 1997 - 2004 , the planet search program has been monitoring the velocities of 700 fgkm stars with the goal of detecting extrasolar planets , as described by @xcite . the spectra were all obtained with the hires spectrometer at keck observatory @xcite . they span the wavelength region , 3850 to 6200 with a resolution of @xmath6=55,000 corresponding to a spectrometer psf with fwhm = 5.5 or 2.8 pixels ( 24 @xmath7 each ) on the ccd . exposure times of typically 8 minutes yield a signal - to - noise ratio of 60 per pixel in the continuum around the ca ii h and k lines for the majority of the m dwarfs , most having v=9 - 11 studied here . but the quality can be higher or lower depending on seeing and b - v of the star . the spectra are not flux - calibrated . the stars included here are those from the planet search program having b - v @xmath8 1.3 corresponding to a spectral type of about k7 . all colors were taken from the hipparcos catalogue @xcite . they were originally selected for that search by sieving the gliese catalog and hipparcos catalog for late k - type and m dwarfs that were both brighter than v = 11.5 and devoid of known stellar companions within 2 arcsec , to avoid contamination at the spectrometer slit . the stars were chosen without regard to previously known emission lines in their spectra . indeed , most such stars exhibit emission at h and k ( but not necessarily at h@xmath3 ) . thus , this present sample is representative of the single , k7-m5 dwarfs within @xmath115 pc , though some are slightly farther away . the limit at spectral type m5 stems largely from their faintness for the purposes of high resolution spectroscopy , most such stars being fainter than v = 11.5 . the resulting sample of 147 stars is shown plotted in a color - magnitude diagram in figure [ fig : hr_diagram ] . the stars departing farthest from the main sequence are labeled , with the departures due most likely to stellar companions , extraordinary metallicity , or simply to poor photometry . the width of the main sequence in figure [ fig : hr_diagram ] indicates the limited value of b - v as a diagnostic for @xmath9 for m dwarfs , due to its sensitivity to metallicity caused by line blanketing @xcite . table 1 lists for each star its common names , spectral type , @xmath2 , b - v , the number of observations , and the time range over which it was observed . we obtained a total of 2408 spectra for the 147 stars . each spectrum was reduced as part of the planet - search automated pipeline . the raw ccd images were biased - subtracted , the flat - fielding was done with a quartz lamp through the same spectrometer setup , the spectral order locations were mapped and fitted with a third order polynomial , and the final one - dimensional spectra were extracted by summing along the column within each order . these spectra give the number of photons in each pixel of width 2 . inspection by eye of the h and k emission line profiles from the 147 late - k and m dwarfs reveals considerable structure . the widths of the h and k emission lines in the k7-m0 dwarfs are clearly wider than those of the m1-m5 dwarfs . many emission lines are double - peaked , reminiscent of the central `` absorption '' reversal well known to exist in h and k profiles of g and k dwarfs . many m dwarfs that do not exhibit double peaks appear to have flat - topped emission lines , as if weak central absorption were suppressing the peak of the emission . also visible to the eye are asymmetries in the double - peaked emission profiles , with the central absorption being shifted redward relative to the emission . the blueward peak is almost always taller than the redward one , providing a sensitive measure of relative doppler shift of the emission and central absorption . the h and k profiles stem from the non - lte line transfer and velocity fields in the low density chromospheres of m dwarfs . the central absorption stems from the dependence of the source function on height in the optically thick chromosphere rather than from actual absorption . the variety of visible structure in the line profiles calls for quantitative measurements , in advance of future modeling of the line transfer . our analysis of the h and k lines was performed by fitting them with the sum of two gaussian curves . we tried other functional forms , such as flat topped functions , but the double - gaussian yielded the smallest value of @xmath10 and ( remarkably ) reproduced all off the observed structure in the line profiles . the model consisted of a positive gaussian representing the chromospheric emission that emerges from the broad bottoms of the photospheric ca ii absorption features . added to that is a second gaussian of negative amplitude representing non - lte `` absorption '' at line center . a constant background level was left as a final parameter , representing the base of the photospheric absorption . the height , width , and central position of both gaussians were free to vary as well as the background level , for a total of 7 parameters used to minimize @xmath10 . this model provides excellent fits to those emission lines that exhibit either obvious central absorption or flat tops . for those emission lines without an obvious central absorption , we found that the double gaussian commonly fit the line profiles well , especially at the top , better than did a single gaussian . figure [ fig : example_fit ] shows our model fit to the h and k lines from spectra of three stars , illustrating different line shapes : obviously double - peaked lines , flat - topped lines , and thin lines . in each case the emission gaussian alone fits the edges of the line but overshoots the top , which can then be accommodated by the second gaussian of negative amplitude . from these fits we derived a set of parameters for each line including the equivalent width ( ew ) , the full width at half maximum ( fwhm ) , and the wavelength difference between the centers of the two gaussians ( @xmath11 ) , indicating a relative doppler shift between the emission and absorption . for those profiles that were clearly double - peaked , we determined the ratio of the maximum counts in the blue peak to the maximum counts in the red peak ( v / r ) . the equivalent width was determined by dividing the area in the line by the continuum around the line . the area under the line was easy to measure . we subtracted the constant background found by our fit from the line and integrated the counts within the line . the continuum determination presented challenges . the spectra of m dwarfs in this region contain many absorption lines , leaving no obvious choice for a constant value to assign to the continuum . we determined a continuum value by averaging the number of counts within two wavelength bins located on either side of the broad photospheric h line , as shown in figure [ fig : example_cont ] . these continuum regions are analogous to those used in the determination of the well - known mt . wilson @xmath12 values . ( the locations of these ranges were set with reference to the center of the emission line so as to be constant within the rest frame of each star . ) we performed a spline on the double - gaussian fitted curve to find the width at half of the background - subtracted maximum . emission at 3970 , just redward of the h line.,scaledwidth=75.0% ] we used @xmath11 and v / r as measures of the asymmetry of the lines . we defined @xmath11 as the center of the absorption gaussian in our model minus the center of the emission gaussian , so that a value greater than zero corresponds to a redshifted absorption relative to the emission . as previously used by @xcite , v / r is only applicable for double - peaked lines , i.e. those with a visible absorption at line center . for these lines , we found the ratio of the maximum number of counts in the blue side of the emission line to that on the red side . lines having a blue peak higher than the red peak ( as typically seen here ) have v / r @xmath8 1 . segregation of red and blue peaks was formally made ( without operator intervention ) using the location of the center of the emission gaussian . the parameter @xmath11 is applicable to more stars than was v / r since it can be calculated for any line , whereas v / r requires a double - peaked line . only 40% of the stars here were double - peaked . our automated fitting procedure occasionally failed to find good fits to the h and k lines in our 2408 spectra . the process obtained good fits for both of the lines for 63% of the spectra and was not able to find a good fit for either line in 15% of the cases . in many cases the poor fits were simply due to spectra of such low signal - to - noise ratio that the @xmath10 minimization algorithm failed . since we had multiple observations of each star , we ignored the poorest fits from our subsequent analysis . four stars had no adequate fit of either line , with @xmath10 well above unity . two of them , l 758 - 107 and gl 745b , have emission lines that were too weak to fit with any integrity . the other two , gl 388 and hip 112460 , have very strong h and k lines , but the reduced @xmath10 values were above 10 due to their distinctly non - gaussian profiles , and therefore we consider these fits dubious . of the 147 original stars , 136 have at least one good measurement of both the h and the k lines , and another 7 stars have a good quality measurement of at least one line . from each spectrum of the h and k lines for each star , we compiled the mean values of the best - fit parameters and their standard deviations . the latter is a measure of internal precision and temporal variability , to be discussed in each subsequent section . tables 2 and 3 list the final measurements for the h and k lines , respectively , for each star . columns 1 - 3 list the common names of each star , and column 4 gives the number of spectra analyzed with good fits . columns 5 and 6 give the equivalent width and its standard deviation . columns 7 and 8 give the fwhm and its standard deviation . columns 9 and 10 give @xmath13 and its standard deviation and columns 11 and 12 give the mean and standard deviations for @xmath14 . ideally , one would like to measure chromospheric parameters as a function of stellar mass . unfortunately , we do not have secure photometric calibration of stellar mass for m dwarfs that would be immune to variations in chromospheric activity . nor do we have a mass calibration immune to the unknown metallicities of the stars . thus b - v and @xmath2 remain as modest proxies of stellar mass here , with due concern about their integrity as such . we proceed to examine how the measured chromospheric parameters depend on @xmath2 which is likely a better proxy for mass than b - v . clearly there is a great need for accurate masses for m dwarfs . previous studies of m dwarfs have reported measurements of either line fluxes ( @xcite , @xcite ) or equivalent widths as the diagnostics stellar activity , most often using unresolved lines . as our spectra are not flux calibrated , we report the equivalent width of the h and k lines , obtained from the double - gaussian fit . such measurements are explicitly dependent on the color , and hence @xmath2 and mass , of the star . the nearby uv continuum decreases with increasing color and lower mass . thus for a fixed luminosity in the lines , the h and k equivalent widths would increase with increasing stellar color , simply due to the decreasing continuum . figure [ fig : ew_sptype ] shows the measured values of ew for the h and k lines ( diamonds and crosses , respectively ) versus @xmath2 . the figure shows that the ews of h and k for late k and early m dwarfs ( @xmath2= 7 - 9 ) are typically @xmath11 . however , the later type m dwarfs ( @xmath2= 9 - 13 ) have smaller ews of h and k , typically @xmath10.3 . thus , the ew of h and k increases by a factor of 3 toward decreasing @xmath2 , and thus toward higher stellar mass . the increase in ew with stellar mass is in the opposite direction from that which would be caused simply by the increasing uv continuum flux with stellar mass . _ thus , the luminosity in the h and k lines increases , by at least a factor of 3 , with increasing mass from m5 to k7 . _ . top : all stars . bottom : a close - up of the majority of stars that have ew @xmath15 3 . the errors of @xmath10.025 are less than the size of the symbols.,title="fig:",scaledwidth=65.0% ] . top : all stars . bottom : a close - up of the majority of stars that have ew @xmath15 3 . the errors of @xmath10.025 are less than the size of the symbols.,title="fig:",scaledwidth=65.0% ] of some interest is the transition from partially to fully convective stars that is expected to occur near 0.3 @xmath16 @xcite . indeed , @xcite found an increased level of activity ( as measured by h@xmath3 ews ) at m2.5/3 , which corresponds roughly to this mass . @xcite and @xcite note that 0.3 m@xmath0(where full convection sets in ) may correspond to spectral types m4-m5 . it is tempting to associate the remarkable factor of 3 rise in the chromospheric h and k lines seen here with a dynamo explanation . such an explanation may be appropriate . note , however , that @xcite found an increase in h@xmath3 ew for later type stars while here we find a decrease in the h and k emission for those stars . nonetheless , figure [ fig : ew_sptype ] does not reveal a sharp discontinuity in ew at any @xmath2 . thus we find no convincing evidence of a discontinuous change in the dynamo in the domain k7-m5 . since the plotted values of ew are the means of multiple measurements for each star , we can compare the vertical spread of the points to the standard deviations seen for each star individually . the standard deviations of ew ( listed in tables 2 and 3 ) are typically @xmath10.1 while the spread in ews at any given @xmath17 is much larger . thus it appears that the spread in ew at a given @xmath2 is greater than can be accounted for by temporal variability . the observed spread may be caused by the degeneracy in mass at a given @xmath17 ; however , we suspect that this degeneracy alone is not adequate to account for all of the spread . rather , the chromospheric activity of a star is likely influenced by other stellar parameters , notably rotation , that vary from star to star at a given mass . indeed similar analysis in h@xmath3 emission has shown that activity is related to both age and rotation @xcite , which is likely the cause of variance of emission as well . the fwhm of the h and k emission lines is plotted in figure [ fig : width_sptype ] versus @xmath2 . this plot shows a tight relation between h and k line width and @xmath2 , such that fwhm increases with brighter @xmath2 and hence with stellar mass . this is reminiscent of the wilson - bappu effect in fgk stars as discussed in @xcite and modeled by @xcite , where the logarithm of the width of chromospheric emission lines is observed to increase linearly with decreasing absolute magnitude . this remarkably tight relationship between line width and @xmath2 for m dwarfs is not understood , to our knowledge . the conventional wilson bappu effect in fgk stars represents a correlation between emission line width with luminosity which in turn is directly related to evolutionary state ( luminosity class ) and hence surface gravity . in contrast , the m dwarfs all reside nearly on the main sequence , none having evolved at all , albeit having different masses , radii , and gravities . nonetheless , it is apparent that the h and k line width in m dwarfs could similarly serve to determine distances . ultimately , a line - transfer explanation of the wilson - bappu effect in m dwarfs must include the variation of chromospheric pressure and mass column as a function of stellar mass . . the fwhm increases with brighter @xmath2 , in a remarkably tight correlation . the fwhm increases by a factor of @xmath13 from over this narrow range in spectral types , m5-k7 . this correlation is reminiscent of the wilson bappu effect known for fgk stars . the errors in fwhm are @xmath10.3 , smaller than the size of the symbols.,scaledwidth=75.0% ] slight inflections are visible in figure [ fig : width_sptype ] , but the limitations discussed earlier associated with the use of @xmath17 as a proxy for stellar mass may be the cause . figure [ fig : width_sptype ] exhibits a lower limit of 9 for the observed fwhm of the h and k lines . this limiting width represents the convolution of the lowest intrinsic line widths and the width of the psf of the spectrometer , 5.5 . thus , considering a quadrature sum of widths , the intrinsic line widths must have a lower limit of 7.1 for k7-m5 dwarfs . there appear to be few , if any , stars from k7-m5 having lines narrower than 7.1 , even among those fainter than @xmath2= 14 . such a lower limit to the h and k line widths would indicate a departure from the monotonic wilson - bappu relation set by the brighter m dwarfs . perhaps some other broadening mechanism besides line transfer effects sets a floor on the fwhm of these lines . spectra with very high resolution of a sample of m3-m8 dwarfs at h and k could verify this possible discontinuity or breakdown of the wilson bappu effect in m dwarfs . the values plotted in figure [ fig : width_sptype ] are the means of multiple measurements from each star , and reveal considerable spread in fwhm at a given @xmath2 . we examined the standard deviations of the multiple measurements for individual stars to determine if vertical spread in figure [ fig : width_sptype ] could be due to temporal changes in the stars . the maximum standard deviation for individual stars is @xmath10.76 , the same variance as the observed vertical spread among the different stars . but the typical standard deviation of 0.35 is considerably less than the vertical spread in figure [ fig : width_sptype ] . thus the distribution in fwhm among stars at a given mass is due in part to temporal variability , but the spread is dominated by the distribution of rotation and perhaps by the degeneracy of mass for a given @xmath17 . close inspection of figure [ fig : width_sptype ] shows that the k lines are generally wider than the h lines . figure [ fig : width_correlation ] shows a histogram of the difference between the mean fwhm of the k line and that of the h line for each star . the distribution is displaced toward positive values by 1.5 . this greater width of the k line undoubtedly results from the slight difference in the radiative transfer of the h and k emission lines . the h and k emission lines arise from two upper states , @xmath18 and @xmath19 , respectively , and a common lower state , @xmath20 . the different 2j+1 degeneracies of the two upper levels cause a slight enhancement of the source function of the k line over the h line at each height in the low density , upper chromosphere . in particular , radiative de - excitation of the k line occurs more commonly than for the h line . this difference causes the k line to approach the lte emissivity limit ( the planck function ) more closely than does the h line allowing the wings of the k line to grow faster , broadening its final profile . however , a full description of the differing widths of the h and k lines requires a detailed radiative transfer model , including the pressure , temperature , mass columns and velocity fields as a function of height in the chromosphere , along with the partial redistribution of photon frequencies during transitions . due to the high resolution of these spectra , many of the h and k line profiles are visibly asymmetric . figure [ fig : see_redshift ] shows representative spectra from three stars with asymmetry in the h line that is visible to the eye as a redshift of the central absorption ( relative to the main emission ) . our v / r and @xmath11 measurements permit quantitative analysis of this effect for lines with obvious asymmetry and more subtle cases . = 0.4 , 0.2 , and 0.1 ( left , middle , right ) is clearly observable in the line profile data and model fits . the left spectrum is of gj 413 , middle spectrum of gj 424 , and right spectrum of gj 338a.,scaledwidth=75.0% ] in figures [ fig : vr_sptype ] and [ fig : shift_sptype ] we plot v / r and @xmath11 versus @xmath2 . the two figures show that there is little correlation between asymmetry and absolute magnitude . figure [ fig : shift_sptype ] shows an increased spread in @xmath13 with higher @xmath2 , i.e. for lower mass stars . this scatter is most likely due to the fact that these less luminous stars have noisier lines and therefore often have lower quality fits . visual inspection of the spectra with @xmath21 confirms that these spectra tend to have a lower signal - to - noise ratio , and that the noise may be detrimentally affecting the fits . however , these plots clearly show that most stars have v / r greater than 1 and @xmath11 greater than 0 , both indicating the presence of redshifted central absorption . . the solid line represents no asymmetry , above which the domain represents lines having redshifted central absorption . most k7-m5 dwarfs apparently have redshifted absorption reversals . representative errors are shown for v / r.,scaledwidth=75.0% ] , vs. @xmath2 . top : the full range of values . bottom : the domain within @xmath221.5 of symmetry . the predominance of points _ above _ the horizontal line shows that most k7-m5 dwarfs have redshifted absorption reversals . representative errors in the top plot are about the size of the plot symbols , and are shown in the bottom plot.,title="fig:",scaledwidth=68.0% ] , vs. @xmath2 . top : the full range of values . bottom : the domain within @xmath221.5 of symmetry . the predominance of points _ above _ the horizontal line shows that most k7-m5 dwarfs have redshifted absorption reversals . representative errors in the top plot are about the size of the plot symbols , and are shown in the bottom plot.,title="fig:",scaledwidth=68.0% ] as illustrated in the bottom plot of figure [ fig : shift_sptype ] , the central absorption redshifts for these stars tend to be @xmath10.1 . such small redshifts may seem difficult to detect , but they correspond to 0.05 pixels , easily detectable with signal - to - noise ratios above 20 , characteristic of these spectra . the measured redshifts for the representative h lines in figure [ fig : see_redshift ] are 0.1 - 0.4 and clearly visible to the eye . the atmospheres of these stars must have gas with low source function that either has lower outward velocity or higher inward velocity than the higher source function gas below it . the relative speed of 0.1 is remarkable both in its consistency among the m dwarfs and in it small magnitude . this speed is much slower than the escape velocity , indicating that the gas is not falling ballistically . instead , it implies gas slowly decelerating upward or sinking downward slowly , with velocities controlled by gas pressure differences or magnetic forces . just as figure [ fig : width_sptype ] showed possible evidence of wider k lines than h lines , figure [ fig : shift_sptype ] shows that the k line central absorption is typically more redshifted than the h line counterpart . to quantify this , we plot h line asymmetry measurements versus k line asymmetry measurements for each star in figure [ fig : asym_corr_hk ] . ( bottom ) . the solid lines separate measured redshifted absorption from blueshifted absorption . the dashed lines show an identity relation between the two variables . the asymmetries in the h and k lines are clearly correlated , albeit with scatter due to the measurement errors . the standard deviation of the residuals to a fit of the v / r data to the identity relation is 0.024 , which is roughly equal to the v / r error . for the @xmath11 data , the standard deviation is 0.7 , while the measurement error is @xmath10.23 .,title="fig:",scaledwidth=63.0% ] ( bottom ) . the solid lines separate measured redshifted absorption from blueshifted absorption . the dashed lines show an identity relation between the two variables . the asymmetries in the h and k lines are clearly correlated , albeit with scatter due to the measurement errors . the standard deviation of the residuals to a fit of the v / r data to the identity relation is 0.024 , which is roughly equal to the v / r error . for the @xmath11 data , the standard deviation is 0.7 , while the measurement error is @xmath10.23 .,title="fig:",scaledwidth=63.0% ] if every star had an h line central absorption redshifted by the same amount as its k line central absorption , we would expect to see an identity relation in a plot of these parameters . although our errors may be overestimated , visual inspection of figure [ fig : asym_corr_hk ] shows an excess of points above the identity line , indicating a higher redshift for the absorption reversal of the k line . this implies that the k line central absorption forms in a region with slightly slower outward speeds ( or greater inward speeds ) than both the gas where the h abosorption reversal forms and the gas where the h emission forms . this result is consistent with a progression of decreasing outward velocities , from lower to upper chromosphere . lastly , we find an interesting possible trend between the average asymmetry and ews of the h and k lines . figure [ fig : asym_ew ] shows more scatter in the asymmetry of a line at ews less than about 1 , while at ews up to 2 the asymmetry tends to a smaller range of slightly redshifted values . lines with low ews may be weak and therefore more likely to have noise disrupt their fits , but visual inspection of these lines shows that this can not account for all of the spread seen in figure [ fig : asym_ew ] . this indicates a possible connection between the physics governing the chromospheric activity of a star and the velocity gradient in the chromosphere , such that as a star becomes more active its central absorption becomes redshifted . here we compare the h and k equivalent widths to previously published h@xmath3 equivalent widths taken from @xcite . h@xmath3 equivalent widths have long been a standard measure of chromospheric heating in m dwarfs . by comparing the ca ii resonance lines to h@xmath3 emission , the non - radiative heating and gas conditions in the lower and upper chromosphere may be compared . h@xmath3 has been extensively studied with regard to chromospheric and coronal heating in m dwarfs , indicating a peculiar behavior @xcite . with increasing magnetic fields , the balmer lines first become stronger _ in absorption _ before they fill in and eventually go into emission . in contrast , the ca ii lines increase monotonically with increased magnetic heating @xcite . due to this monotonic behavior , the ca ii emission strength provides a less ambiguous diagnostic of the magnetic state of the stellar surface than does h@xmath3 @xcite . nonetheless , the most magnetically active m dwarfs , notably those with strong coronal emission , always show h@xmath3 in emission . in figure [ fig : caew_haew ] we plot the h@xmath3 equivalent width against the h and k equivalent width for our sample of m dwarfs . the h@xmath3 equivalent width remains negative ( absorption ) near a value of about -0.2 for a wide range range of ca ii ews less than 2 . the h@xmath3 absorption varies little with increasing h and k equivalent width up to 2 . ews for the same stars . the two emission lines are correlated only for large h and k emission above ew = 3 . this indicates that a magnetic threshold exists for h@xmath3 while h and k remains a useful chromospheric diagnostic at all activity levels . ew values were taken from @xcite . ca ii ew errors are 0.025 .,scaledwidth=70.0% ] however , for h and k equivalent widths above @xmath12 , h@xmath3 abruptly goes into emission and increases monotonically with increasing h and k . the uncorrelated behavior between ca ii and h@xmath3 for stars with h@xmath3 in absorption has been seen before and attributed to highly segregated regions of line formation @xcite . @xcite found that the ca ii k line is formed in the lower chromosphere , while the h@xmath3 line is formed in the upper chromosphere with a different cooling rate . the data presented here agree with such previously reported behavior and reflect this segregation . the suggestion is that the effect of magnetic fields on the two emitting regions remains separate until a certain level of magnetic heating is reached , at which point the lower and upper portions of the chromosphere behave similarly . the h and k line profiles of the 143 k7-m5 dwarfs examined here revealed various properties of their chromospheres and variations with stellar mass . the h and k line widths and fluxes increase with increasing stellar mass and luminosity , reminiscent of the wilson - bappu effect . the latest type stars here ( @xmath1m4-m5 ) have h and k emission line widths no more than 7.1 ( fwhm ) , and possibly smaller as limited by our spectral resolution of 5.5 . in contrast , the higher mass m dwarfs have h and k emission line widths of 30 . thus , the h and k line widths vary with stellar mass by a factor of @xmath14 ( from 7.1 to @xmath130 ) over the spectral type range from m5-k7 , corresponding to a range of stellar mass from 0.3 - 0.55 m@xmath0 . it is remarkable that the line widths vary roughly as the square of stellar mass . the broadening seen in the h and k lines is caused either by curve of growth effects in the line wings as the core approaches optically depth unity , or by broadening due to velocity fields @xcite . apparently the structure of chromospheres changes significantly over this relatively small range in stellar mass . the k line is consistently wider than the h line in all k7-m5 dwarfs , and its central absorption is consistently more redshifted ( relative to the chromospheric emission ) . while the unequal degeneracies in the upper levels producing these lines cause somewhat differing optical depths , these data emphasize that the h and k lines are produced in slightly different regions of the chromosphere and are therefore governed by slightly different line transfer and kinematics . most of the k7-m5 stars exhibited redshifted central `` absorption '' reversals superimposed on the emission core , with an typical shift of 0.1 . as the absorption occurs at line center , that part of the line profile must form higher in the atmosphere near optical depth unity at line center . the redshifted `` absorption '' indicates that this higher material has a lower source function and that it is moving outward more slowly ( or inward more quickly ) than the lower material that has a higher source function . the redshifted reversal can be explained with a simple heuristic model . prominence - like or micro - flare activity results in chromospheric gas being magnetically heated and thrust outward . the outward moving , dense gas produces the emission while the material above it has expanded and cooled , and is moving more slowly . that upper material has lower density ( and perhaps slightly lower temperature ) , yielding a lower source function . in such a picture , the gas decelerates after the initial outward kick . however , the observed 0.1 is much slower than free - fall , ballistic speeds . thus , the modest velocity difference of 0.1 implies a fluid velocity gradient that is governed by magnetic fields or small gas pressure differences . these small and steady velocity gradients may be compared to the more violent kinematics that occur in major flares in m dwarfs @xcite . full ntle models with kinematics will be constructed by lucianne walkowicz , in progress . stars with lower h and k equivalent widths exhibited a larger spread in the redshift of the absorption component . between ews of @xmath112 the redshift was confined to a small range near 0.1 ; stars with weaker emission exhibited a wider range of redshifts . thus there appears to be some relationship between the physical effects that govern the ew and the velocity fields . the standard deviations from the mean values measured for both ews and fwhms were greater than our measurement errors , confirming that temporal changes do occur for these stars . unfortunately , the degeneracy between @xmath2 and stellar mass produces spread in our plots that can not be distinguished from spread due to temporal variability or the influence of individual stellar parameters ( e.g. age , rotation rate , metallicity ) on chromospheric heating . as previously found , there is clear evidence here of segregation between the region of the chromosphere that produces h@xmath3 and that which produces h and k for stars with h@xmath3 in absorption . however , at ca ii ews greater than about 2 we find a strong correlation between h and k and h@xmath3 . apparently the upper and lower regions of chromospheres in m dwarfs behave independently until some magnetic activity threshold is reached , at which point they become governed by the same physical mechanism . the results in this paper offer empirical suggestions of structural and kinematic attributes of m dwarf chromospheres that depend on stellar mass and vary with height . models will be needed that include the non - lte , partial redistribution line transfer of the h and k lines in atmospheres that contain velocity fields , to thoroughly explain the observed properties of chromospheric line profiles . in addition , stellar mass estimates are required to study the relation between chromospheric activity and mass . finally , contemporaneous high resolution optical and uv spectra along with x - ray measurements would help constrain the relationship between chromospheres and coronae in m dwarfs . we thank r. paul butler , steve vogt , and jason wright for obtaining most of the spectra used here . we thank gibor basri and lucianne walkowicz for valuable conversations . we appreciate many valuable comments provided by the anonymous referee , making the paper much richer . we acknowledge support from nasa grants ncc5 - 96 and ncc5 - 511 as well as nsf grant hrd-9706268 . we wish to especially thank those of hawaiian ancestry on whose sacred mountain of mauna kea we are privileged to be guests . without their generous hospitality , the keck observations presented herein would not have been possible . lllcrrccc & 428 & & m1 & 9.67 & 1.48 & 21 & 11757.10 & 12989.71 + 1 & 439 & hd 225213 & m1.5 & 10.37 & 1.46 & 3 & 12600.79 & 12899.00 + 14 & 1368 & & k7 & 8.06 & 1.42 & 10 & 12097.03 & 12924.89 + 15 & 1475 & hd 1326 & m1.5 & 10.31 & 1.56 & 25 & 10461.77 & 12924.89 + 15b & & hd 1326b & m3.5 & 13.28 & 1.80 & 19 & 10667.00 & 13197.08 + 1009 & 1734 & & m1.5 & 9.86 & 1.48 & 4 & 12600.80 & 13196.06 + 26 & & & m2.5 & 10.55 & 1.54 & 26 & 11581.71 & 12988.81 + 27 & 3143 & & m0.5 & 9.62 & 1.46 & 4 & 12600.80 & 13196.08 + 48 & 4856 & & m3 & 10.40 & 1.46 & 16 & 10667.04 & 13197.11 + 49 & 4872 & & m1.5 & 9.55 & 1.50 & 18 & 10667.05 & 13197.11 + 54.1 & 5643 & & m4.5 & 14.25 & 1.85 & 15 & 11012.05 & 13196.09 + & 8051 & & m2 & 10.70 & 1.52 & 10 & 11581.76 & 12926.06 + 83.1 & & & m4.5 & 13.99 & 1.82 & 18 & 11411.04 & 12989.75 + & & g 244 - 047 & m3 & 11.23 & 1.54 & 9 & 11581.79 & 12850.07 + 87 & 10279 & & m1.5 & 9.95 & 1.43 & 18 & 11410.13 & 13045.71 + & 11048 & & m0.5 & 8.96 & 1.51 & 18 & 11581.80 & 12897.96 + 105b & & & m4 & 12.37 & 1.62 & 2 & 12830.12 & 12988.77 + 109 & 12781 & & m3 & 11.19 & 1.55 & 13 & 11410.10 & 12854.11 + 156 & & sao 130811 & k7 & 8.10 & 1.37 & 19 & 10461.85 & 12897.99 + & 20917 & hd 28343 & k7 & 8.05 & 1.38 & 13 & 11226.80 & 12898.05 + & 21553 & hd 232979 & k7 & 8.59 & 1.43 & 10 & 11543.96 & 12898.04 + & 21556 & & m1.5 & 10.13 & 1.52 & 10 & 11580.78 & 12899.04 + 176 & 21932 & hd 285968 & k5 & 10.10 & 1.51 & 19 & 10839.76 & 12898.12 + & 22627 & & m3.5 & 11.54 & 1.60 & 14 & 11580.83 & 13016.83 + 180 & 22762 & & m2 & 10.44 & 1.52 & 7 & 11580.79 & 12898.10 + & 23512 & & m3 & 11.87 & 1.66 & 10 & 11543.90 & 12899.10 + 191 & 24186 & hd 33793 & m1 & 10.92 & 1.54 & 11 & 11170.92 & 12988.87 + & 24284 & & m2 & 10.29 & 1.55 & 6 & 12899.08 & 13045.74 + 205 & 25878 & hd 36395 & m1.5 & 9.14 & 1.52 & 19 & 10420.00 & 12924.99 + & & g 097 - 054 & m3.5 & 11.34 & 1.59 & 6 & 11580.84 & 12899.11 + 208 & 26335 & hd 245409 & k7 & 8.61 & 1.44 & 21 & 10420.01 & 12899.12 + & 26801 & hd 233153 & m0.5 & 9.39 & 1.44 & 9 & 11580.77 & 12925.00 + 213 & 26857 & & m4 & 12.67 & 1.67 & 8 & 11550.89 & 12924.07 + & & g 192 - 013 & m3.5 & 12.32 & 1.60 & 10 & 11543.99 & 12925.03 + & 29052 & & m4 & 11.58 & 1.60 & 8 & 11580.85 & 12988.93 + 226 & 29277 & & m2 & 10.62 & 1.51 & 13 & 10839.95 & 12925.05 + 229 & 29295 & hd 42581 & m1 & 9.33 & 1.51 & 21 & 10420.06 & 12924.15 + 239 & 31635 & hd 260655 & k7 & 9.67 & 1.46 & 13 & 10839.94 & 12924.11 + 250b & & & m2 & 10.35 & 1.42 & 1 & 12924.09 & 12924.09 + 251 & 33226 & hd 265866 & m3 & 11.18 & 1.58 & 16 & 10784.07 & 12925.08 + 273 & 36208 & & m3.5 & 11.99 & 1.53 & 14 & 10805.89 & 12988.95 + & 36338 & & m3 & 10.99 & 1.42 & 6 & 11580.87 & 12988.96 + & 36834 & & m0.5 & 10.50 & 1.70 & 10 & 11580.89 & 13045.80 + & 37217 & & m3 & 11.52 & 2.00 & 6 & 11580.86 & 12925.13 + 285 & 37766 & & m4.5 & 12.25 & 1.61 & 9 & 11544.07 & 13076.97 + 2066 & 40501 & & m2 & 10.24 & 1.50 & 13 & 10805.88 & 13017.87 + & 41689 & & k7 & 8.88 & 1.35 & 5 & 12989.02 & 13077.03 + & 42220 & & k7 & 8.57 & 1.42 & 5 & 12989.03 & 13077.02 + 324 & & hd 75732b & m4 & 12.58 & 1.65 & 11 & 11544.12 & 12988.99 + 338a & 45343 & hd 79210 & k7 & 8.68 & 1.43 & 18 & 10607.75 & 12989.04 + 338b & 120005 & hd 79211 & k7 & 8.75 & 1.42 & 19 & 10607.76 & 12989.04 + & 46655 & & m3.5 & 11.80 & 1.60 & 14 & 11544.11 & 13017.88 + & 46769 & & m0 & 9.53 & 1.48 & 4 & 12989.06 & 13077.04 + 357 & 47103 & & m2.5 & 11.14 & 1.57 & 13 & 10839.97 & 12989.11 + & 47513 & & m1.5 & 10.12 & 1.50 & 10 & 11580.94 & 13017.97 + & 47650 & & m3 & 10.89 & 1.80 & 8 & 11580.91 & 12989.08 + & 48714 & & m0 & 8.90 & 1.47 & 13 & 11552.03 & 13017.94 + 382 & 49986 & sao 137411 & m1.5 & 9.80 & 1.51 & 20 & 10607.76 & 13017.96 + 388 & & sao 81292 & m3.5 & 11.07 & 1.54 & 11 & 12064.85 & 13044.88 + 390 & 51007 & & m1 & 9.69 & 1.45 & 12 & 11580.98 & 13045.88 + 393 & 51317 & & m2 & 10.33 & 1.52 & 19 & 10607.77 & 13045.86 + 394 & & hd 237903 & k7 & 8.56 & 1.34 & 2 & 12804.82 & 13044.89 + & 51468 & hd 90875 & k5 & 6.86 & 1.18 & 13 & 11200.05 & 13044.90 + 397 & 51525 & & k5 & 7.88 & 1.33 & 18 & 11200.06 & 13044.90 + 402 & 53020 & & m4 & 12.91 & 1.68 & 10 & 11552.11 & 13017.99 + 406 & & & m5.5 & 16.68 & 2.00 & 16 & 11172.07 & 13018.00 + 408 & 53767 & & m2.5 & 10.92 & 1.55 & 18 & 10607.79 & 13077.05 + 410 & 53985 & hd 95650 & m0 & 9.24 & 1.44 & 18 & 11581.03 & 13077.05 + 411 & 54035 & hd 95735 & m2 & 10.46 & 1.51 & 36 & 10602.91 & 13018.16 + 412a & 54211 & sao 43609 & m0.5 & 10.26 & 1.54 & 19 & 10605.79 & 13077.08 + 413 & 54532 & & m2 & 10.31 & 1.55 & 11 & 11552.09 & 13017.17 + 414b & & hd 97101b & m1.5 & 9.57 & 1.57 & 6 & 12064.77 & 13044.98 + 414a & 54646 & hd 97101 & k8 & 7.93 & 1.36 & 21 & 10463.01 & 13044.98 + 424 & 55360 & sao 15475 & m0 & 9.53 & 1.42 & 10 & 11552.06 & 13018.15 + 433 & 56528 & sao 202602 & m1.5 & 10.01 & 1.49 & 11 & 11200.04 & 13018.04 + & 57050 & & m4 & 11.79 & 1.60 & 10 & 11581.04 & 13077.10 + 436 & 57087 & & m2.5 & 10.63 & 1.52 & 78 & 11552.07 & 13196.77 + 445 & 57544 & & m3.5 & 12.12 & 1.59 & 15 & 10840.15 & 13018.11 + 447 & 57548 & & m4 & 13.46 & 1.77 & 14 & 10983.75 & 13018.03 + 450 & 57802 & sao 43609 & m1 & 10.12 & 1.52 & 17 & 10840.13 & 13018.11 + & 59406 & l 758 - 107 & m4 & 12.08 & 1.40 & 11 & 11581.11 & 13018.06 + 465 & 60559 & & m4 & 11.57 & 1.62 & 13 & 11552.13 & 13018.08 + 486 & 62452 & & m3.5 & 11.83 & 1.56 & 19 & 10839.14 & 13046.07 + & 62687 & hd 111631 & k7 & 8.32 & 1.42 & 13 & 11582.07 & 13046.04 + 494a & 63510 & & m0.5 & 9.43 & 1.45 & 22 & 11582.07 & 13046.05 + 514 & & & m0.5 & 9.63 & 1.50 & 26 & 10546.01 & 13196.82 + & 66459 & sao 63647 & k5 & 8.88 & 1.40 & 11 & 11581.07 & 13071.99 + 526 & 67155 & hd 119850 & m1.5 & 9.79 & 1.45 & 19 & 10546.02 & 13071.96 + & 67164 & & m3.5 & 11.53 & 1.54 & 13 & 11582.05 & 13179.80 + & 68469 & hd 122303 & m1 & 9.68 & 1.49 & 12 & 11410.73 & 13071.95 + & 70865 & & m2 & 9.90 & 1.47 & 5 & 12897.72 & 13179.79 + & 70975 & & m3.5 & 11.71 & 1.61 & 13 & 11582.09 & 13153.87 + 555 & 71253 & & m4 & 12.42 & 1.62 & 16 & 11410.73 & 13153.87 + 9492 & 71898 & sao 17568 & m3 & 10.91 & 1.62 & 17 & 10840.16 & 13077.11 + 569a & 72944 & & m2.5 & 10.24 & 1.48 & 12 & 12062.87 & 13153.93 + 581 & 74995 & & m3 & 11.57 & 1.61 & 15 & 11409.76 & 13077.15 + & 79755 & hd 147379a & k7 & 8.46 & 1.41 & 16 & 11679.96 & 13179.94 + & & hd 147379b & m3 & 10.54 & 1.51 & 15 & 11679.97 & 13179.94 + 623 & 80346 & & m3 & 10.74 & 1.48 & 2 & 12062.92 & 12099.83 + 625 & 80459 & & m1.5 & 11.04 & 1.59 & 17 & 10862.14 & 13179.92 + 628 & 80824 & & m3.5 & 11.97 & 1.60 & 21 & 11409.75 & 13153.92 + 638 & 82003 & hd 151288 & k5 & 8.16 & 1.37 & 24 & 10601.99 & 13179.97 + 649 & 83043 & & m1 & 9.55 & 1.59 & 27 & 11409.82 & 13195.81 + & 83762 & & m3 & 10.95 & 1.56 & 3 & 12897.79 & 13181.93 + & 84099 & & m3.5 & 11.14 & 1.50 & 16 & 11678.94 & 13074.09 + 667c & & & m1.5 & 11.00 & 1.57 & 14 & 11704.98 & 13153.97 + & 84790 & & m2.5 & 10.91 & 1.56 & 18 & 11704.94 & 13181.92 + 673 & 85295 & hd 157881 & k5 & 8.10 & 1.36 & 21 & 10275.88 & 13153.98 + 678 & 85665 & sao 122446 & m0 & 9.32 & 1.45 & 18 & 10955.06 & 13153.98 + 687 & 86162 & sao 17568 & m3 & 10.87 & 1.50 & 27 & 10603.96 & 13179.98 + 686 & 86287 & & m1 & 10.07 & 1.53 & 20 & 10604.94 & 13181.85 + 694 & 86776 & & m2.5 & 10.61 & 1.55 & 20 & 10956.10 & 13179.99 + 2130a & 86961 & & m2 & 11.54 & 1.46 & 13 & 11702.95 & 13153.99 + 699 & 87937 & & m4 & 13.24 & 1.74 & 74 & 10602.00 & 13195.82 + 701 & 88574 & hd 165222 & m1 & 9.91 & 1.52 & 29 & 10602.02 & 13072.12 + & & lhs 462 & m3 & 11.58 & 1.58 & 14 & 11704.96 & 13180.01 + & & g 205 - 028 & m3.5 & 11.61 & 2.01 & 12 & 11705.98 & 13180.02 + 4063 & & & m3.5 & 12.12 & 1.42 & 16 & 11409.86 & 13180.03 + & 91699 & & m3 & 11.00 & 1.63 & 15 & 11706.00 & 13181.91 + 725a & 91768 & hd 173739 & m3 & 11.15 & 1.54 & 23 & 10602.05 & 13179.88 + 725b & 91772 & hd 173740 & m3.5 & 11.96 & 1.59 & 21 & 10602.06 & 13179.97 + 729 & 92403 & & m3.5 & 13.58 & 1.60 & 25 & 10984.03 & 13154.00 + 745a & 93873 & & m1.5 & 11.07 & 1.55 & 13 & 10982.94 & 13180.07 + 745b & 93899 & hd 349726 & m1.5 & 11.01 & 1.55 & 14 & 10983.00 & 13180.06 + & & g 207 - 019 & m3 & 11.33 & 1.80 & 12 & 11706.04 & 13180.05 + 752a & 94761 & hd 180617 & m2.5 & 10.29 & 1.50 & 35 & 12061.95 & 13196.85 + 1245b & & & m5.5 & 15.56 & 1.98 & 3 & 12804.00 & 13154.10 + & 99427 & hd 193202 & k5 & 7.75 & 1.32 & 14 & 11342.00 & 13181.97 + 793 & 101180 & & m2.5 & 10.94 & 1.56 & 15 & 10983.98 & 13181.98 + 806 & 102401 & & m1.5 & 10.31 & 1.49 & 18 & 10602.10 & 13182.00 + 809 & 103096 & & m0.5 & 9.30 & 1.49 & 20 & 10602.11 & 13182.02 + & 104432 & & m1 & 10.45 & 1.49 & 13 & 11704.05 & 13182.05 + 846 & 108782 & hd 209290 & m0 & 9.09 & 1.46 & 18 & 11410.00 & 13196.92 + 849 & 109388 & & m3.5 & 10.70 & 1.52 & 22 & 11410.01 & 13196.93 + & 109555 & & m2 & 10.11 & 1.52 & 10 & 12095.09 & 13195 . 90 + 860a & 110893 & hd 239960 & m3 & 11.58 & 1.65 & 20 & 10606.07 & 13180.98 + 873 & 112460 & & m3.5 & 11.77 & 1.54 & 19 & 12094.98 & 13196.95 + 876 & 113020 & & m4 & 11.81 & 1.60 & 144 & 10602.09 & 13196.99 + 880 & 113296 & hd 216899 & m1.5 & 9.47 & 1.51 & 22 & 10666.04 & 13195.96 + 884 & 113576 & hd 217357 & k5 & 8.34 & 1.39 & 41 & 10365.88 & 13197.01 + 887 & 114046 & hd 217987 & m0.5 & 9.75 & 1.48 & 34 & 10984.12 & 13197.01 + 891 & 114411 & & m2 & 10.29 & 1.57 & 4 & 12600.73 & 13197.00 + & 115332 & & m4 & 11.58 & 1.50 & 12 & 11899.78 & 13197.02 + & 115562 & & m1 & 9.44 & 1.47 & 3 & 12897.87 & 12988.80 + 905 & & & m5 & 14.77 & 1.91 & 16 & 11010.01 & 13197.06 + 908 & 117473 & sao 128397 & m1 & 10.10 & 1.48 & 28 & 10366.88 & 13196.01 + 911 & 117886 & & k7 & 8.91 & 1.42 & 4 & 12600.78 & 13197.04 + lllcrrrrrrrr & 428 & & 20 & 1.25 & 0.13 & 19.49 & 0.31 & 0.12 & 0.08 & 1.020 & 0.019 + 1 & 439 & hd 225213 & 3 & 0.27 & 0.06 & 12.66 & 0.30 & 0.26 & 0.14 & & + 14 & 1368 & & 10 & 1.48 & 0.10 & 23.82 & 0.25 & 0.09 & 0.07 & 1.017 & 0.018 + 15 & 1475 & hd 1326 & 22 & 0.33 & 0.07 & 14.42 & 0.36 & 0.07 & 0.07 & 1.006 & 0.021 + 15b & & hd 1326b & 10 & 1.42 & 1.04 & 10.17 & 0.98 & 0.05 & 0.25 & & + 1009 & 1734 & & 3 & 1.35 & 0.09 & 19.34 & 0.28 & 0.13 & 0.07 & 1.017 & 0.029 + 26 & & & 25 & 0.42 & 0.09 & 14.20 & 0.51 & -0.02 & 0.24 & 1.032 & 0.034 + 27 & 3143 & & 3 & 0.95 & 0.15 & 20.01 & 0.53 & -0.17 & 0.23 & 0.952 & 0.096 + 48 & 4856 & & 14 & 0.51 & 0.05 & 15.12 & 0.44 & 0.16 & 0.17 & 1.017 & 0.029 + 49 & 4872 & & 18 & 2.00 & 0.24 & 19.92 & 0.28 & 0.00 & 0.05 & 1.001 & 0.011 + 54.1 & 5643 & & 11 & 6.10 & 3.05 & 10.49 & 0.42 & 0.12 & 0.48 & & + & 8051 & & 10 & 0.87 & 0.22 & 12.95 & 0.41 & 0.08 & 0.39 & 1.054 & 0.074 + 83.1 & & & 9 & 6.33 & 1.45 & 10.85 & 0.32 & 0.14 & 0.59 & & + & & g 244 - 047 & 7 & 0.58 & 0.08 & 15.32 & 0.25 & 0.18 & 0.15 & 1.022 & 0.132 + 87 & 10279 & & 17 & 0.39 & 0.05 & 16.80 & 0.44 & 0.12 & 0.14 & 1.022 & 0.041 + & 11048 & & 18 & 1.29 & 0.15 & 22.06 & 0.50 & 0.11 & 0.07 & 1.021 & 0.019 + 105b & & & 1 & 0.33 & & 10.26 & & 0.36 & & & + 109 & 12781 & & 10 & 0.54 & 0.11 & 11.75 & 0.38 & 0.21 & 0.35 & 1.013 & 0.005 + 156 & & sao 130811 & 19 & 0.79 & 0.08 & 25.67 & 0.48 & 0.12 & 0.07 & 1.042 & 0.028 + & 20917 & hd 28343 & 13 & 1.11 & 0.12 & 24.94 & 0.70 & 0.07 & 0.08 & 1.021 & 0.020 + & 21553 & hd 232979 & 10 & 1.21 & 0.05 & 22.88 & 0.36 & 0.11 & 0.05 & 1.032 & 0.018 + & 21556 & & 10 & 0.38 & 0.04 & 17.78 & 0.50 & 0.29 & 0.23 & 1.065 & 0.072 + 176 & 21932 & hd 285968 & 19 & 0.97 & 0.10 & 17.23 & 0.40 & 0.07 & 0.08 & 1.013 & 0.026 + & 22627 & & 12 & 0.75 & 0.08 & 11.51 & 0.40 & -0.82 & 2.78 & 0.957 & 0.063 + 180 & 22762 & & 7 & 0.51 & 0.13 & 12.70 & 0.49 & 0.41 & 0.54 & & + & 23512 & & 3 & 0.46 & 0.07 & 9.92 & 0.30 & 0.56 & 0.23 & & + 191 & 24186 & hd 33793 & 3 & & & 10.86 & 0.80 & -0.07 & 2.79 & & + & 24284 & & 6 & 0.38 & 0.03 & 15.76 & 0.42 & 0.12 & 0.06 & 1.021 & 0.022 + 205 & 25878 & hd 36395 & 18 & 1.70 & 0.12 & 21.70 & 0.33 & 0.02 & 0.05 & 1.007 & 0.012 + & & g 097 - 054 & 4 & 0.38 & 0.03 & 11.15 & 0.63 & 0.00 & 0.22 & & + 208 & 26335 & hd 245409 & 15 & 2.46 & 0.13 & 21.40 & 0.31 & -0.03 & 0.13 & 0.995 & 0.014 + & 26801 & hd 233153 & 9 & 1.81 & 0.19 & 21.06 & 0.36 & 0.06 & 0.08 & 1.014 & 0.019 + 213 & 26857 & & 5 & 0.46 & 0.21 & 9.82 & 0.44 & -0.55 & 0.88 & & + & & g 192 - 013 & 8 & 0.67 & 0.15 & 10.21 & 0.51 & -0.13 & 0.72 & & + & 29052 & & 3 & 0.40 & 0.16 & 10.61 & 0.78 & 3.82 & 4.02 & 0.993 & + 226 & 29277 & & 5 & 0.61 & 0.10 & 13.65 & 0.46 & 0.02 & 0.48 & 0.960 & + 229 & 29295 & hd 42581 & 21 & 1.24 & 0.14 & 21.10 & 0.43 & 0.09 & 0.05 & 1.023 & 0.015 + 239 & 31635 & hd 260655 & 13 & 0.59 & 0.09 & 17.65 & 0.45 & 0.12 & 0.09 & 1.014 & 0.019 + 250b & & & 1 & 0.88 & & 15.43 & & -0.05 & & 0.977 & + 251 & 33226 & hd 265866 & 14 & 0.34 & 0.06 & 11.66 & 0.42 & 0.01 & 0.19 & 0.980 & 0.034 + 273 & 36208 & & 12 & 0.42 & 0.11 & 10.69 & 0.31 & 0.03 & 0.08 & 0.999 & 0.052 + & 36338 & & 5 & 0.52 & 0.17 & 12.02 & 0.60 & 0.10 & 0.56 & 1.019 & + & 36834 & & 10 & 0.55 & 0.06 & 14.20 & 0.43 & 0.11 & 0.15 & 1.009 & + & 37217 & & 1 & 0.40 & & 11.31 & & 0.05 & & & + 285 & 37766 & & 1 & 20.16 & & 15.89 & & 2.94 & & & + 2066 & 40501 & & 12 & 0.63 & 0.09 & 14.90 & 0.34 & 0.09 & 0.10 & 1.014 & 0.030 + & 41689 & & 3 & 0.54 & 0.04 & 22.97 & 0.33 & 0.22 & 0.12 & 1.073 & 0.057 + & 42220 & & 5 & 0.81 & 0.05 & 23.83 & 0.30 & 0.04 & 0.14 & 1.013 & 0.039 + 324b & & hd 75732b & 1 & 0.73 & & 9.23 & & -2.78 & & & + 338a & 45343 & hd 79210 & 16 & 1.28 & 0.11 & 23.04 & 0.41 & 0.11 & 0.05 & 1.032 & 0.014 + 338b & 120005 & hd 79211 & 19 & 1.36 & 0.09 & 22.72 & 0.33 & 0.08 & 0.05 & 1.016 & 0.009 + & 46655 & & 8 & 0.34 & 0.05 & 10.39 & 0.82 & 0.53 & 0.81 & 0.989 & + & 46769 & & 4 & 0.92 & 0.07 & 20.02 & 0.41 & 0.04 & 0.11 & 1.021 & 0.013 + 357 & 47103 & & 6 & 0.20 & 0.04 & 10.71 & 0.51 & 0.51 & 2.97 & 0.988 & 0.103 + & 47513 & & 10 & 1.03 & 0.13 & 17.30 & 0.35 & 0.10 & 0.16 & 1.022 & 0.030 + & 47650 & & 7 & 2.87 & 0.43 & 13.79 & 0.20 & 0.06 & 0.17 & & + & 48714 & & 13 & 1.58 & 0.09 & 22.20 & 0.31 & -0.02 & 0.08 & 1.001 & 0.015 + 382 & 49986 & sao 137411 & 20 & 1.57 & 0.13 & 18.69 & 0.31 & 0.04 & 0.05 & 1.008 & 0.010 + 388 & & sao 81292 & 0 & & & & & & & & + 390 & 51007 & & 11 & 1.27 & 0.12 & 19.88 & 0.40 & 0.15 & 0.11 & 1.029 & 0.027 + 393 & 51317 & & 19 & 0.68 & 0.08 & 14.51 & 0.34 & 0.04 & 0.08 & 1.004 & 0.014 + 394 & & hd 237903 & 2 & 0.93 & 0.01 & 24.63 & 0.01 & 0.14 & 0.02 & 1.029 & 0.017 + & 51468 & hd 90875 & 11 & 0.36 & 0.04 & 30.00 & 0.48 & 0.11 & 0.07 & 1.029 & 0.018 + 397 & 51525 & & 18 & 1.14 & 0.09 & 24.60 & 0.62 & 0.08 & 0.16 & 1.023 & 0.027 + 402 & 53020 & & 6 & 0.90 & 0.15 & 10.18 & 0.50 & -0.34 & 0.71 & & + 406 & & & 4 & 33.39 & 14.65 & 11.56 & 0.29 & 0.09 & 0.20 & & + 408 & 53767 & & 17 & 0.54 & 0.08 & 12.24 & 0.27 & 0.03 & 0.12 & 0.990 & 0.014 + 410 & 53985 & hd 95650 & 11 & 3.13 & 0.33 & 20.31 & 0.35 & 0.00 & 0.08 & 0.997 & 0.011 + 411 & 54035 & hd 95735 & 32 & 0.18 & 0.03 & 13.32 & 0.28 & 0.09 & 0.13 & 1.018 & 0.022 + 412a & 54211 & sao 43609 & 10 & 0.17 & 0.03 & 14.40 & 0.45 & 0.12 & 0.07 & 1.057 & 0.037 + 413 & 54532 & & 10 & 0.61 & 0.06 & 15.93 & 0.55 & 0.22 & 0.25 & 1.036 & 0.036 + 414b & & hd 97101b & 6 & 1.10 & 0.07 & 21.06 & 0.46 & 0.14 & 0.22 & 1.022 & 0.020 + 414a & 54646 & hd 97101 & 20 & 0.87 & 0.09 & 25.17 & 0.40 & 0.08 & 0.07 & 1.024 & 0.020 + 424 & 55360 & sao 15475 & 10 & 0.61 & 0.08 & 19.52 & 0.50 & 0.20 & 0.10 & 1.036 & 0.021 + 433 & 56528 & sao 202602 & 10 & 0.47 & 0.05 & 18.01 & 0.50 & 0.16 & 0.16 & 1.038 & 0.057 + & 57050 & & 4 & 0.33 & 0.05 & 10.54 & 0.67 & -0.10 & 1.00 & 1.062 & 0.021 + 436 & 57087 & & 67 & 0.38 & 0.03 & 14.64 & 0.46 & 0.00 & 0.67 & 1.003 & 0.034 + 445 & 57544 & & 8 & 0.24 & 0.07 & 9.98 & 0.73 & 0.12 & 1.05 & 0.960 & 0.065 + 447 & 57548 & & 10 & 0.53 & 0.19 & 9.76 & 0.42 & 0.06 & 0.18 & 0.956 & + 450 & 57802 & sao 43609 & 17 & 1.20 & 0.12 & 16.82 & 0.34 & 0.09 & 0.13 & 1.021 & 0.015 + & 59406b & l 758 - 107 & 0 & & & & & & & & + 465 & 60559 & & 1 & 0.16 & & 10.26 & & -0.11 & & & + 486 & 62452 & & 12 & 0.28 & 0.04 & 10.65 & 0.55 & 0.07 & 0.55 & 0.999 & 0.082 + & 62687 & hd 111631 & 13 & 1.41 & 0.12 & 23.31 & 0.34 & 0.05 & 0.07 & 1.009 & 0.020 + 494a & 63510 & & 2 & 7.48 & 0.17 & 22.08 & 0.00 & 5.89 & 0.82 & & + 514 & & & 26 & 0.90 & 0.08 & 19.34 & 0.28 & 0.11 & 0.06 & 1.029 & 0.016 + & 66459 & sao 63647 & 11 & 1.10 & 0.07 & 22.19 & 0.39 & 0.09 & 0.06 & 1.019 & 0.013 + 526 & 67155 & hd 119850 & 18 & 0.61 & 0.09 & 18.04 & 0.44 & 0.12 & 0.04 & 1.027 & 0.019 + & 67164 & & 3 & 0.37 & 0.07 & 10.94 & 0.31 & 2.94 & 5.14 & 1.066 & + & 68469 & hd 122303 & 11 & 0.78 & 0.09 & 19.58 & 0.36 & 0.15 & 0.06 & 1.036 & 0.019 + & 70865 & & 5 & 0.55 & 0.07 & 19.31 & 0.45 & 0.08 & 0.02 & 1.018 & 0.019 + & 70975 & & 2 & 0.19 & 0.04 & 11.03 & 0.38 & -1.24 & 1.97 & & + 555 & 71253 & & 10 & 0.37 & 0.05 & 10.83 & 0.29 & -0.08 & 0.27 & 1.068 & 0.057 + 9492 & 71898 & sao 17568 & 13 & 0.44 & 0.11 & 12.39 & 0.51 & 0.06 & 0.50 & 1.015 & 0.052 + 569a & 72944 & & 1 & 4.75 & & 16.43 & & -0.15 & & 0.980 & + 581 & 74995 & & 12 & 0.18 & 0.02 & 11.13 & 0.71 & 0.02 & 0.28 & 1.041 & 0.032 + & 79755 & hd 147379a & 16 & 1.19 & 0.07 & 24.38 & 0.26 & 0.09 & 0.06 & 1.021 & 0.015 + & & hd 147379b & 14 & 0.83 & 0.06 & 15.48 & 0.40 & 0.11 & 0.13 & 1.023 & 0.023 + 623 & 80346 & & 2 & 0.31 & 0.07 & 11.06 & 0.37 & 0.21 & 0.47 & & + 625 & 80459 & & 16 & 0.34 & 0.09 & 10.69 & 0.47 & 0.14 & 0.41 & 0.953 & + 628 & 80824 & & 16 & 0.45 & 0.10 & 10.59 & 0.41 & -0.33 & 1.55 & 1.055 & + 638 & 82003 & hd 151288 & 24 & 0.87 & 0.07 & 24.62 & 0.26 & 0.15 & 0.04 & 1.040 & 0.016 + 649 & 83043 & & 27 & 1.14 & 0.08 & 19.66 & 0.37 & 0.10 & 0.07 & 1.017 & 0.017 + & 83762 & & 2 & 0.40 & 0.03 & 12.35 & 0.03 & 0.65 & 0.40 & & + & 84099 & & 12 & 0.48 & 0.07 & 12.30 & 0.41 & 0.85 & 2.66 & 0.989 & + 667c & & & 9 & 0.36 & 0.05 & 11.30 & 0.36 & -0.03 & 0.65 & 1.020 & + & 84790 & & 9 & 0.34 & 0.04 & 11.42 & 0.50 & 0.68 & 2.92 & 0.972 & 0.021 + 673 & 85295 & hd 157881 & 21 & 1.27 & 0.10 & 23.67 & 0.53 & -0.15 & 0.08 & 0.976 & 0.014 + 678 & 85665 & sao 122446 & 17 & 0.99 & 0.13 & 20.22 & 0.40 & 0.09 & 0.07 & 1.021 & 0.026 + 687 & 86162 & sao 17568 & 26 & 0.44 & 0.08 & 12.36 & 0.57 & 0.00 & 0.13 & 1.001 & 0.030 + 686 & 86287 & & 20 & 0.42 & 0.06 & 16.64 & 0.34 & 0.16 & 0.08 & 1.031 & 0.024 + 694 & 86776 & & 18 & 0.72 & 0.13 & 14.25 & 0.39 & 0.14 & 0.17 & 0.994 & 0.017 + 2130a & 86961 & & 10 & 1.53 & 0.16 & 19.32 & 0.26 & 0.08 & 0.16 & 1.025 & 0.028 + 699 & 87937 & & 51 & 0.33 & 0.11 & 9.11 & 0.35 & 0.02 & 0.20 & & + 701 & 88574 & hd 165222 & 29 & 0.65 & 0.04 & 17.57 & 0.33 & 0.13 & 0.08 & 1.023 & 0.025 + & & lhs 462 & 3 & 0.24 & 0.04 & 10.27 & 0.52 & 0.07 & 0.15 & 1.003 & + & & g 205 - 028 & 6 & 0.74 & 0.21 & 11.74 & 0.39 & 0.22 & 0.45 & 1.025 & + 4063 & & & 14 & 0.69 & 0.10 & 14.23 & 0.59 & 0.16 & 0.31 & 1.026 & 0.064 + & 91699 & & 12 & 0.26 & 0.09 & 11.99 & 0.71 & 0.56 & 2.97 & 1.073 & 0.050 + 725a & 91768 & hd 173739 & 19 & 0.20 & 0.05 & 10.81 & 0.36 & 0.02 & 0.16 & & + 725b & 91772 & hd 173740 & 13 & 0.38 & 0.14 & 9.88 & 0.38 & -0.03 & 0.19 & 0.992 & + 729 & 92403 & & 10 & 5.82 & 1.24 & 11.35 & 0.44 & -0.06 & 2.37 & & + 745a & 93873 & & 0 & & & & & & & & + 745b & 93899 & hd 349726 & 0 & & & & & & & & + & & g 207 - 019 & 4 & 0.31 & 0.10 & 10.15 & 0.51 & 1.94 & 1.42 & 1.078 & + 752a & 94761 & hd 180617 & 35 & 0.66 & 0.06 & 17.19 & 0.29 & 0.13 & 0.07 & 1.014 & 0.015 + 1245b & & & 1 & 5.52 & & 14.38 & & 0.28 & & & + & 99427 & hd 193202 & 14 & 0.59 & 0.14 & 25.58 & 5.75 & -0.10 & 0.94 & 1.054 & 0.022 + 793 & 101180 & & 12 & 1.19 & 0.12 & 11.85 & 0.46 & 0.11 & 0.24 & & + 806 & 102401 & & 17 & 0.63 & 0.08 & 14.66 & 0.45 & 0.11 & 0.21 & 1.026 & 0.049 + 809 & 103096 & & 20 & 1.10 & 0.09 & 21.66 & 0.30 & 0.15 & 0.04 & 1.039 & 0.014 + & 104432 & & 10 & 0.18 & 0.02 & 11.48 & 0.59 & -0.05 & 0.59 & 1.012 & 0.096 + 846 & 108782 & hd 209290 & 18 & 1.31 & 0.12 & 21.75 & 0.47 & 0.07 & 0.05 & 1.017 & 0.014 + 849 & 109388 & & 21 & 0.62 & 0.04 & 16.35 & 0.42 & 0.10 & 0.13 & 1.021 & 0.026 + & 109555 & & 9 & 1.38 & 0.17 & 18.98 & 0.41 & 0.06 & 0.08 & 0.998 & 0.016 + 860a & 110893 & hd 239960 & 15 & 0.35 & 0.07 & 10.63 & 0.42 & 0.07 & 0.12 & 1.026 & 0.032 + 873 & 112460 & & 0 & & & & & & & & + 876 & 113020 & & 132 & 0.51 & 0.11 & 11.44 & 0.39 & 0.03 & 0.24 & 0.996 & 0.027 + 880 & 113296 & hd 216899 & 22 & 1.39 & 0.14 & 20.33 & 0.35 & 0.06 & 0.05 & 1.012 & 0.011 + 884 & 113576 & hd 217357 & 40 & 1.14 & 0.08 & 23.29 & 0.33 & 0.07 & 0.06 & 1.013 & 0.015 + 887 & 114046 & hd 217987 & 34 & 0.77 & 0.13 & 18.00 & 0.44 & 0.10 & 0.06 & 1.017 & 0.017 + 891 & 114411 & & 4 & 0.76 & 0.08 & 14.65 & 0.51 & 0.04 & 0.28 & 1.004 & + & 115332 & & 11 & 0.67 & 0.13 & 11.40 & 0.39 & 0.02 & 0.41 & 1.016 & + & 115562 & & 3 & 1.52 & 0.19 & 20.57 & 0.51 & 0.01 & 0.03 & 1.006 & 0.005 + 905 & & & 9 & 2.37 & 1.07 & 9.44 & 0.37 & 1.63 & 3.31 & & + 908 & 117473 & sao 128397 & 28 & 0.27 & 0.04 & 14.12 & 0.43 & 0.08 & 0.10 & 1.020 & 0.018 + 911 & 117886 & & 4 & 1.27 & 0.08 & 22.08 & 0.42 & 0.07 & 0.21 & 1.016 & 0.050 + lllcrrrrrrrr & 428 & & 20 & 0.91 & 0.10 & 20.83 & 0.58 & 0.21 & 0.21 & 1.022 & 0.031 + 1 & 439 & hd 225213 & 3 & 0.20 & 0.04 & 13.84 & 0.01 & -0.17 & 0.25 & & + 14 & 1368 & & 10 & 1.07 & 0.07 & 25.71 & 0.47 & 0.16 & 0.16 & 1.011 & 0.019 + 15 & 1475 & hd 1326 & 24 & 0.24 & 0.04 & 15.42 & 0.45 & 0.39 & 0.35 & 1.032 & + 15b & & hd 1326b & 8 & 0.81 & 0.54 & 10.04 & 0.86 & -1.07 & 1.45 & & + 1009 & 1734 & & 4 & 0.97 & 0.07 & 21.04 & 0.44 & 0.27 & 0.49 & 0.977 & 0.005 + 26 & & & 21 & 0.33 & 0.06 & 15.33 & 0.77 & 0.46 & 1.17 & 1.016 & 0.016 + 27 & 3143 & & 0 & & & & & & & & + 48 & 4856 & & 12 & 0.38 & 0.04 & 16.44 & 0.36 & 0.38 & 0.53 & 1.037 & 0.053 + 49 & 4872 & & 18 & 1.36 & 0.15 & 21.24 & 0.27 & 0.12 & 0.15 & 1.018 & 0.017 + 54.1 & 5643 & & 6 & 5.94 & 3.92 & 11.07 & 0.35 & -1.90 & 0.93 & & + & 8051 & & 7 & 0.69 & 0.17 & 14.01 & 0.53 & -0.60 & 0.76 & & + 83.1 & & & 5 & 4.54 & 0.74 & 11.76 & 0.24 & -1.17 & 2.79 & & + & & g 244 - 047 & 2 & 0.51 & 0.03 & 16.94 & 0.35 & 1.66 & 2.07 & 0.965 & + 87 & 10279 & & 16 & 0.29 & 0.04 & 17.97 & 0.64 & 0.26 & 0.53 & 1.028 & 0.039 + & 11048 & & 18 & 0.90 & 0.10 & 23.63 & 0.58 & 0.17 & 0.09 & 1.029 & 0.017 + 105b & & & 0 & & & & & & & & + 109 & 12781 & & 8 & 0.40 & 0.08 & 12.87 & 0.58 & 1.04 & 2.43 & & + 156 & & sao 130811 & 18 & 0.58 & 0.04 & 27.75 & 0.63 & 0.22 & 0.16 & 1.045 & 0.038 + & 20917 & hd 28343 & 13 & 0.78 & 0.09 & 26.83 & 0.82 & 0.20 & 0.09 & 1.031 & 0.015 + & 21553 & hd 232979 & 10 & 0.87 & 0.05 & 24.52 & 0.26 & 0.16 & 0.09 & 1.034 & 0.013 + & 21556 & & 8 & 0.28 & 0.04 & 19.26 & 0.65 & 0.36 & 0.17 & 1.112 & 0.043 + 176 & 21932 & hd 285968 & 19 & 0.69 & 0.08 & 18.44 & 0.41 & 0.23 & 0.32 & 1.028 & 0.021 + & 22627 & & 6 & 0.59 & 0.06 & 12.75 & 0.69 & 0.53 & 2.70 & 1.015 & + 180 & 22762 & & 6 & 0.41 & 0.08 & 13.85 & 0.50 & 0.30 & 1.54 & & + & 23512 & & 1 & 0.26 & & 10.52 & & -1.49 & & & + 191 & 24186 & hd 33793 & 3 & & & 12.28 & 0.84 & 4.25 & 6.25 & 0.946 & + & 24284 & & 6 & 0.30 & 0.04 & 17.37 & 0.45 & 0.29 & 0.17 & 1.061 & 0.078 + 205 & 25878 & hd 36395 & 18 & 1.19 & 0.09 & 23.45 & 0.46 & 0.14 & 0.10 & 1.026 & 0.012 + & & g 097 - 054 & 2 & 0.33 & 0.02 & 12.46 & 0.39 & 0.03 & 0.48 & & + 208 & 26335 & hd 245 - 409 & 15 & 1.70 & 0.10 & 22.99 & 0.29 & 0.12 & 0.13 & 1.006 & 0.015 + & 26801 & hd 233153 & 8 & 1.24 & 0.17 & 22.44 & 0.41 & 0.13 & 0.10 & 1.025 & 0.020 + 213 & 26857 & & 4 & 0.26 & 0.27 & 10.23 & 0.55 & 2.11 & 3.88 & & + & & g 192 - 013 & 3 & 0.52 & 0.03 & 11.26 & 0.30 & -2.10 & 4.24 & & + & 29052 & & 1 & 0.29 & & 10.52 & & 0.08 & & 1.082 & + 226 & 29277 & & 5 & 0.41 & 0.08 & 13.74 & 0.60 & 0.04 & 1.45 & 0.977 & 0.026 + 229 & 29295 & hd 42581 & 19 & 0.84 & 0.10 & 22.65 & 0.48 & 0.22 & 0.15 & 1.038 & 0.022 + 239 & 31635 & hd 260655 & 13 & 0.40 & 0.05 & 19.11 & 0.81 & 0.39 & 0.24 & 1.030 & 0.030 + 250b & & & 1 & 0.59 & & 16.63 & & 0.19 & & & + 251 & 33226 & hd 265866 & 12 & 0.27 & 0.04 & 13.06 & 0.68 & 0.24 & 0.86 & 1.085 & 0.091 + 273 & 36208 & & 12 & 0.29 & 0.07 & 11.82 & 0.43 & 0.73 & 2.48 & 1.027 & + & 36338 & & 3 & 0.47 & 0.10 & 12.93 & 0.33 & -0.88 & 2.62 & & + & 36834 & & 8 & 0.42 & 0.04 & 15.26 & 0.73 & 0.03 & 0.50 & 0.977 & + & 37217 & & 1 & 0.23 & & 11.65 & & -0.17 & & 0.961 & + 285 & 37766 & & 3 & 11.94 & 1.88 & 18.27 & 0.55 & 1.23 & 0.81 & & + 2066 & 40501 & & 7 & 0.49 & 0.07 & 16.28 & 0.90 & -0.21 & 1.07 & 1.008 & 0.031 + & 41689 & & 4 & 0.40 & 0.02 & 24.84 & 0.30 & 0.42 & 0.14 & 1.081 & 0.032 + & 42220 & & 4 & 0.59 & 0.03 & 25.90 & 0.27 & 0.11 & 0.11 & 1.019 & 0.025 + 324b & & hd 75732b & 0 & & & & & & & & + 338a & 45343 & hd 79210 & 16 & 0.87 & 0.09 & 24.93 & 0.46 & 0.23 & 0.09 & 1.038 & 0.017 + 338b & 120005 & hd 79211 & 19 & 0.95 & 0.09 & 24.46 & 0.38 & 0.24 & 0.14 & 1.043 & 0.017 + & 46655 & & 2 & 0.25 & 0.04 & 11.90 & 1.97 & -0.47 & 1.99 & & + & 46769 & & 4 & 0.69 & 0.03 & 20.91 & 0.28 & 0.07 & 0.32 & 1.033 & 0.042 + 357 & 47103 & & 5 & 0.18 & 0.02 & 12.42 & 1.41 & -0.52 & 0.26 & 0.950 & + & 47513 & & 7 & 0.75 & 0.04 & 18.59 & 0.44 & 0.11 & 0.27 & 1.023 & + & 47650 & & 3 & 2.07 & 0.37 & 14.96 & 0.01 & 0.40 & 0.41 & 1.015 & + & 48714 & & 13 & 1.10 & 0.09 & 23.86 & 0.35 & 0.06 & 0.09 & 1.02 & 0.016 + 382 & 49986 & sao 137411 & 20 & 1.10 & 0.10 & 20.05 & 0.46 & 0.19 & 0.16 & 1.014 & 0.013 + 388 & & sao 81292 & 0 & & & & & & & & + 390 & 51007 & & 9 & 0.92 & 0.07 & 21.53 & 0.33 & 0.15 & 0.12 & 1.033 & 0.010 + 393 & 51317 & & 18 & 0.47 & 0.06 & 15.51 & 0.50 & 0.15 & 0.29 & 1.021 & 0.021 + 394 & & hd 237903 & 2 & 0.73 & 0.03 & 26.58 & 0.00 & 0.22 & 0.13 & 1.055 & 0.030 + & 51468 & hd 90875 & 13 & 0.26 & 0.03 & 32.59 & 0.44 & 0.16 & 0.09 & 1.048 & 0.029 + 397 & 51525 & & 15 & 0.84 & 0.08 & 26.43 & 0.65 & 0.16 & 0.20 & 1.017 & 0.022 + 402 & 53020 & & 0 & & & & & & & & + 406 & & & 3 & 16.22 & 5.29 & 12.56 & 0.33 & 0.21 & 1.03 & 0.991 & + 408 & 53767 & & 13 & 0.40 & 0.06 & 13.38 & 0.63 & -0.19 & 0.94 & 1.014 & 0.008 + 410 & 53985 & hd 95650 & 16 & 2.22 & 0.22 & 21.54 & 0.27 & 0.06 & 0.16 & 0.992 & 0.014 + 411 & 54035 & hd 95735 & 30 & 0.14 & 0.03 & 14.60 & 0.44 & 0.54 & 0.47 & & + 412a & 54211 & sao 43609 & 16 & 0.15 & 0.03 & 15.51 & 0.93 & 0.48 & 0.49 & 1.071 & 0.040 + 413 & 54532 & & 9 & 0.44 & 0.05 & 17.49 & 0.74 & 0.38 & 0.39 & 0.999 & 0.012 + 414b & & hd 97101b & 5 & 0.84 & 0.04 & 22.38 & 0.62 & 0.38 & 0.22 & 1.071 & 0.015 + 414a & 54646 & hd 97101 & 20 & 0.62 & 0.08 & 27.18 & 0.66 & 0.16 & 0.15 & 1.036 & 0.022 + 424 & 55360 & sao 15475 & 9 & 0.44 & 0.04 & 21.14 & 1.02 & 0.49 & 0.62 & 1.037 & 0.025 + 433 & 56528 & sao 202602 & 4 & 0.33 & 0.02 & 19.67 & 0.96 & 0.43 & 0.38 & 1.088 & 0.054 + & 57050 & & 1 & 0.28 & & 11.09 & & -0.54 & & & + 436 & 57087 & & 51 & 0.31 & 0.03 & 15.98 & 0.57 & 0.18 & 0.53 & 1.008 & 0.064 + 445 & 57544 & & 0 & & & & & & & & + 447 & 57548 & & 6 & 0.38 & 0.09 & 11.00 & 0.64 & 1.13 & 2.88 & & + 450 & 57802 & sao 43609 & 15 & 0.82 & 0.09 & 17.88 & 0.33 & 0.40 & 0.27 & 1.023 & + & 59406b & l 758 - 107 & 0 & & & & & & & & + 465 & 60559 & & 2 & 0.10 & 0.00 & 12.17 & 1.55 & -4.35 & 6.95 & & + 486 & 62452 & & 5 & 0.17 & 0.10 & 13.29 & 2.50 & -1.11 & 2.88 & 1.059 & 0.025 + & 62687 & hd 111631 & 11 & 1.02 & 0.06 & 25.18 & 0.38 & 0.13 & 0.11 & 1.023 & 0.012 + 494a & 63510 & & 2 & 5.21 & 0.16 & 22.98 & 0.39 & 0.90 & 12.48 & & + 514 & & & 26 & 0.65 & 0.08 & 20.79 & 0.32 & 0.30 & 0.14 & 1.039 & 0.020 + & 66459 & sao 63647 & 11 & 0.77 & 0.06 & 23.82 & 0.35 & 0.25 & 0.09 & 1.041 & 0.014 + 526 & 67155 & hd 119850 & 17 & 0.44 & 0.06 & 19.45 & 0.58 & 0.35 & 0.19 & 1.034 & 0.023 + & 67164 & & 3 & 0.23 & 0.03 & 11.45 & 0.63 & 1.84 & 1.91 & & + & 68469 & hd 122303 & 12 & 0.58 & 0.09 & 20.88 & 0.42 & 0.38 & 0.28 & 1.040 & 0.034 + & 70865 & & 5 & 0.45 & 0.06 & 21.27 & 0.83 & 0.38 & 0.35 & 1.060 & 0.043 + & 70975 & & 1 & 0.16 & & 13.29 & & -0.12 & & 1.001 & + 555 & 71253 & & 7 & 0.33 & 0.04 & 12.11 & 0.81 & 0.05 & 0.70 & 1.016 & + 9492 & 71898 & sao 17568 & 6 & 0.30 & 0.05 & 13.66 & 1.09 & -0.66 & 1.06 & 1.046 & + 569a & 72944 & & 1 & 3.74 & & 17.72 & & 0.19 & & & + 581 & 74995 & & 12 & 0.13 & 0.05 & 12.42 & 1.07 & 0.25 & 0.82 & 1.073 & 0.136 + & 79755 & hd 147379a & 16 & 0.86 & 0.08 & 26.40 & 0.28 & 0.21 & 0.14 & 1.028 & 0.023 + & & hd 147379b & 11 & 0.63 & 0.08 & 16.73 & 0.42 & -0.15 & 0.35 & 0.970 & 0.015 + 623 & 80346 & & 1 & 0.19 & & 11.67 & & 6.48 & & & + 625 & 80459 & & 11 & 0.25 & 0.07 & 11.79 & 0.67 & 0.52 & 1.88 & & + 628 & 80824 & & 19 & 0.32 & 0.06 & 11.84 & 0.59 & -0.39 & 1.23 & & + 638 & 82003 & hd 151288 & 22 & 0.62 & 0.06 & 26.48 & 0.42 & 0.28 & 0.05 & 1.053 & 0.019 + 649 & 83043 & & 26 & 0.83 & 0.07 & 21.09 & 0.35 & 0.22 & 0.14 & 1.026 & 0.016 + & 83762 & & 2 & 0.33 & 0.06 & 14.97 & 0.78 & 0.22 & 0.79 & 1.077 & + & 84099 & & 13 & 0.41 & 0.14 & 13.22 & 0.93 & 0.36 & 1.03 & 1.018 & 0.066 + 667c & & & 4 & 0.28 & 0.05 & 11.91 & 0.71 & -0.45 & 0.92 & & + & 84790 & & 8 & 0.26 & 0.02 & 12.81 & 0.70 & 0.00 & 1.01 & 1.063 & 0.078 + 673 & 85295 & hd 157881 & 21 & 0.88 & 0.07 & 25.60 & 0.52 & -0.33 & 0.06 & 0.963 & 0.016 + 678 & 85665 & sao 122446 & 17 & 0.65 & 0.09 & 21.41 & 0.51 & 0.25 & 0.14 & 1.032 & 0.025 + 687 & 86162 & sao 17568 & 26 & 0.34 & 0.05 & 13.50 & 0.56 & 0.20 & 0.67 & & + 686 & 86287 & & 19 & 0.30 & 0.05 & 17.68 & 0.48 & 0.34 & 0.27 & 1.032 & 0.031 + 694 & 86776 & & 19 & 0.51 & 0.11 & 15.34 & 0.61 & 0.51 & 0.59 & & + 2130a & 86961 & & 8 & 1.14 & 0.16 & 20.65 & 0.58 & 0.22 & 0.51 & 1.023 & 0.023 + 699 & 87937 & & 34 & 0.22 & 0.07 & 10.10 & 0.50 & 0.18 & 1.53 & 1.015 & + 701 & 88574 & hd 165222 & 28 & 0.48 & 0.04 & 18.82 & 0.47 & 0.32 & 0.23 & 1.026 & 0.022 + & & lhs 462 & 3 & 0.20 & 0.03 & 11.45 & 1.40 & 0.89 & 0.12 & 1.033 & + & & g 205 - 028 & 2 & 0.52 & 0.02 & 12.21 & 0.80 & -1.93 & 5.02 & & + 4063 & & & 10 & 0.54 & 0.09 & 15.78 & 0.79 & 1.36 & 3.80 & 0.974 & 0.048 + & 91699 & & 7 & 0.23 & 0.03 & 12.52 & 0.43 & 0.92 & 1.67 & 0.931 & 0.095 + 725a & 91768 & hd 173739 & 19 & 0.14 & 0.03 & 12.01 & 0.64 & 0.50 & 2.63 & & + 725b & 91772 & hd 173740 & 14 & 0.26 & 0.08 & 10.80 & 0.47 & -0.35 & 0.49 & & + 729 & 92403 & & 5 & 3.93 & 0.31 & 13.07 & 0.64 & -2.22 & 1.84 & & + 745a & 93873 & & 2 & & & 11.63 & 1.57 & -0.26 & 0.22 & 1.022 & + 745b & 93899 & hd 349726 & 0 & & & & & & & & + & & g 207 - 019 & 2 & 0.33 & 0.12 & 11.08 & 1.58 & -0.31 & 0.98 & & + 752a & 94761 & hd 180617 & 35 & 0.52 & 0.05 & 18.54 & 0.39 & 0.33 & 0.17 & 1.032 & 0.018 + 1245b & & & 0 & & & & & & & & + & 99427 & hd 193202 & 11 & 0.46 & 0.05 & 29.16 & 0.57 & 0.34 & 0.09 & 1.045 & 0.027 + 793 & 101180 & & 9 & 0.84 & 0.09 & 12.92 & 0.28 & 0.23 & 0.71 & & + 806 & 102401 & & 12 & 0.46 & 0.07 & 15.52 & 0.62 & -0.09 & 0.64 & 1.036 & 0.034 + 809 & 103096 & & 20 & 0.76 & 0.06 & 23.23 & 0.35 & 0.29 & 0.08 & 1.052 & 0.016 + & 104432 & & 7 & 0.16 & 0.03 & 12.44 & 0.54 & -0.04 & 1.70 & & + 846 & 108782 & hd 209290 & 18 & 0.94 & 0.09 & 23.38 & 0.48 & 0.18 & 0.09 & 1.030 & 0.019 + 849 & 109388 & & 20 & 0.47 & 0.03 & 17.65 & 0.45 & 0.22 & 0.35 & 1.003 & 0.029 + & 109555 & & 9 & 1.02 & 0.11 & 20.40 & 0.45 & 0.21 & 0.12 & 1.011 & 0.009 + 860a & 110893 & hd 239960 & 15 & 0.24 & 0.06 & 11.57 & 0.41 & -0.55 & 2.22 & 1.005 & 0.003 + 873 & 112460 & & 0 & & & & & & & & + 876 & 113020 & & 108 & 0.38 & 0.08 & 12.68 & 0.51 & 0.13 & 1.76 & 1.027 & 0.033 + 880 & 113296 & hd 216899 & 22 & 0.96 & 0.09 & 21.72 & 0.36 & 0.16 & 0.13 & 1.029 & 0.014 + 884 & 113576 & hd 217357 & 40 & 0.79 & 0.06 & 25.06 & 0.46 & 0.16 & 0.09 & 1.029 & 0.013 + 887 & 114046 & hd 217987 & 33 & 0.54 & 0.06 & 19.37 & 0.53 & 0.28 & 0.17 & 1.018 & 0.010 + 891 & 114411 & & 3 & 0.55 & 0.08 & 15.51 & 0.56 & -0.26 & 0.66 & & + & 115332 & & 8 & 0.52 & 0.10 & 12.54 & 0.58 & -0.11 & 1.83 & & + & 115562 & & 3 & 1.08 & 0.12 & 22.17 & 0.54 & 0.21 & 0.10 & 1.044 & 0.054 + 905 & & & 2 & 1.59 & 0.28 & 11.10 & 0.01 & 0.08 & 3.62 & & + 908 & 117473 & sao 128397 & 27 & 0.20 & 0.03 & 15.15 & 0.53 & 0.29 & 0.62 & 1.049 & 0.012 + 911 & 117886 & & 4 & 0.89 & 0.05 & 23.69 & 0.69 & 0.27 & 0.34 & 1.061 & 0.054 + , s. s. , allen , s. l. , bigelow , b. c. , bresee , l. , brown , b. , cantrall , t. , conrad , a. , couture , m. , delaney , c. , epps , h. w. , hilyard , d. , hilyard , d. f. , horn , e. , jern , n. , kanto , d. , keane , m. j. , kibrick , r. i. , lewis , j. w. , osborne , j. , pardeilhan , g. h. , pfister , t. , ricketts , t. , robinson , l. b. , stover , r. j. , tucker , d. , ward , j. , & wei , m. z. 1994 , in proc . spie instrumentation in astronomy viii , david l. crawford ; eric r. craine ; eds . , volume 2198 , p. 362	we have measured the profiles of the h and k chromospheric emission lines in 147 main sequence stars of spectral type m5-k7 ( masses 0.30 - 0.55 m@xmath0 ) using multiple high resolution spectra obtained during six years with the hires spectrometer on the keck 1 telescope . remarkably , the average fwhm , equivalent widths , and line luminosities of h and k _ increase _ by a factor of 3 with increasing stellar mass over this small range of stellar masses . we fit the h and k lines with a double gaussian model to represent both the chromospheric emission and the non - lte central absorption . most of the sample stars display a central absorption that is typically redshifted by @xmath10.1 relative to the emission . this implies that the higher - level , lower density , chromospheric material has a smaller outward velocity ( or higher inward velocity ) by 0.1 than the lower - level material in the chromosphere , but the nature of this velocity gradient remains unknown . the fwhm of the h and k emission lines increase with stellar luminosity , reminiscent of the wilson - bappu effect in fgk - type stars . both the equivalent widths and fwhm exhibit modest temporal variability in individual stars . at a given value of @xmath2 , stars exhibit a spread in both the equivalent width and fwhm of h and k , due both to a spread in fundamental stellar parameters including rotation rate , age , and possibly metallicity , and to the spread in stellar mass at a given @xmath2 . the k line is consistently wider than the h line , as expected , and its central absorption is more redshifted , indicating that the h and k lines form at slightly different heights in the chromosphere where the velocities are slightly different . the equivalent width of h@xmath3 correlates with h and k only for stars having equivalent widths above @xmath12 , suggesting the existence of a magnetic threshold above which the lower and upper chromosphere become thermally coupled .
deep neural networks have recently been applied to various tasks @xcite , including image processing @xcite , speech recognition @xcite , natural language processing @xcite , and bioinformatics @xcite . although they have simple layered structures of units and connections , they outperform other conventional models by their ability to learn complex nonlinear relationships between input and output data . in each layer , inputs are transformed into more abstract representations under a given set of the model parameters . these parameters are automatically optimized through training so that they extract the important features of the input data . in other words , it does not require either careful feature engineering by hand , or expert knowledge of the data . this advantage has made deep neural networks successful in a wide range of tasks , as mentioned above . however , the inference provided by a deep neural network consists of a large number of nonlinear and complex parameters , which makes it difficult for human beings to understand it . more complex relationships between input and output can be represented as the network becomes deeper or the number of units in each hidden layer increases , however interpretation becomes more difficult . the large number of parameters also causes problems in terms of computational time , memory and over - fitting , so it is important to reduce the parameters appropriately . since it is difficult to read the underlying structure of a neural network and to identify the parameters that are important to keep , we must perform experimental trials to find the appropriate values of the hyperparameters and the random initial parameters that achieve the best trained result . in this paper , to overcome such difficulties , we propose a new method for extracting a global and simplified structure from a layered neural network ( for example , figure [ fig : exp1 m ] and [ fig : exp3 m ] ) . based on network analysis , the proposed method defines a modular representation of the original trained neural network by detecting communities or clusters of units with similar connection patterns . although the modular neural network proposed by @xcite has a similar name , it takes the opposite approach to ours . in fact , it constructs the model structure before training with multiple split neural networks inside it . then , each small neural network works as an expert of a subset task . our proposed method is based on the community detection algorithm . to date , various methods have been proposed to express the characteristics of diverse complex networks without layered structures @xcite , however , no method has been developed for detecting the community structures of trained layered neural networks . the difficulty of conventional community detection from a layered neural network arises from the fact that an assumption commonly used in almost all conventional methods does not hold for layered neural networks : to detect the community structure of network , previous approaches assume that there are more intra - community edges that connect vertices inside a community than inter - community edges that connect vertices in mutually different communities . a network with such a characteristic is called assortative . this seems to be a natural assumption , for instance , for a network of relationships between friends . in layered neural networks , however , units in the same layer do not connect to each other and they only connect via units in their parent or child layers . this characteristic is similar to that of a bipartite graph , and such networks are called disassortative . it is not appropriate to apply conventional methods based on the assumption of an assortative network to a layered neural network . a basic community detection method that can be applied to either assortative or disassortative networks has been proposed by newman et al @xcite . in this paper , we propose an extension of this method for extracting modular representations of layered neural networks . the proposed method can be employed for various purposes . in this paper , we show its effectiveness with the following three applications . 1 . * decomposition of layered neural network into independent networks * : the proposed method decomposes a trained neural network into multiple small independent neural networks . in such a case , the output estimation by the original neural network can be regarded as a set of independent estimations made by the decomposed neural networks . in other words , it divides the problem and reduces the overall computation time . in section [ sec : decomposition ] , we show that our method can properly decompose a neural network into multiple independent networks , where the data consist of multiple independent vectors . * generalization error estimation from community structure * : modularity @xcite is defined as a measure of the validity of a community detection result . section [ sec : gee ] reveals that there is a correlation between modularity and the generalization error of a layered neural network . it is shown that the appropriateness of the trained result can be estimated from the community structure of the network . * knowledge discovery from modular representation * : the modular representation extracted by the proposed method serves as a clue for understanding the trained result of a layered neural network . it extracts the community structure in the input , hidden , and output layer . in section [ sec : kd ] , we introduce the result of applying the proposed method to practical data . the remaining part of this paper is composed as follows : we first describe a layered neural network model in section [ sec : lnn ] . then , we explain our proposed method for extracting a modular representation of a neural network in section [ sec : communitydetect ] . the experimental results are reported in section [ sec : experiment ] , which show the effectiveness of the proposed method in the above three applications . in section [ sec : discussion ] , we discuss the experimental results . section [ sec : conclusion ] concludes this paper . we start by defining @xmath0 and a probability density function @xmath1 on @xmath2 . a training data set @xmath3 with a sample size @xmath4 is assumed to be generated independently from @xmath1 . let @xmath5 be a function from @xmath6 to @xmath7 of a layered neural network that estimates an output @xmath8 from an input @xmath9 and a parameter @xmath10 . for a layered neural network , @xmath11 , where @xmath12 is the weight of connection between the @xmath13-th unit in the depth @xmath14 layer and the @xmath15-th unit in the depth @xmath16 layer , and @xmath17 is the bias of the @xmath13-th unit in the depth @xmath14 layer . a layered neural network with @xmath18 layers is represented by the following function : @xmath19 where a sigmoid function is defined by @xmath20 the training error @xmath21 and the generalization error @xmath22 are respectively defined by @xmath23 where @xmath24 is the euclidean norm of @xmath7 . the generalization error is approximated by @xmath25 where @xmath26 is a test data set that is independent of the training data set . to construct a sparse neural network , we adopt the lasso method @xcite in which the minimized function is defined by @xmath27 where @xmath28 is a hyperparameter . the parameters are trained by the stochastic steepest descent method , @xmath29 where @xmath30 is the training error computed only from the @xmath13-th sample @xmath31 . here , @xmath32 is defined for training time @xmath33 such that @xmath34 which is sufficient for convergence of the stochastic steepest descent . ( [ eq : dw ] ) is numerically calculated by the following procedure , which is called error back propagation @xcite : for the @xmath18-th layer , @xmath35 for @xmath36 , @xmath37 algorithm [ alg_bp ] is used for training a layered neural network based on error back propagation . with this algorithm , we obtain a neural network whose redundant weight parameters are close to zero . randomly sample @xmath38 from uniform distribution on @xmath39 . @xmath40 , @xmath41 , where @xmath42 and @xmath43 is the @xmath15-th element of @xmath38-th sample . ( 1 ) output calculation of all layers : let @xmath45 be an output of the @xmath15-th unit in the depth @xmath14 layer . @xmath46 . ( 2 ) update weight @xmath12 and bias @xmath48 based on back propagation , where @xmath49 is a small constant . @xmath50 . here we propose a new community detection method , which is applied to any layered neural networks ( figure [ fig : com ] ( a ) ) . the proposed method is an extension of the basic approach proposed by newman et al @xcite . it detects communities of assortative or disassortative networks . the key idea behind our method is that the community assignment of the units in each layer is estimated by using connection with adjacent layers . as shown in figure [ fig : com ] ( b ) , a partial network consisting of the connections between every layer and its adjacent layers is represented in the form of two matrices : @xmath58 and @xmath59 . the matrix @xmath60 and @xmath61 represent the connections between two layers of depth @xmath62 and @xmath14 , and two layers of depth @xmath14 and @xmath16 , respectively . an element @xmath63 is defined as @xmath64 if the absolute value of the connection weight between the @xmath13-th unit in the depth @xmath62 layer and the @xmath15-th unit in the depth @xmath14 layer is larger than @xmath65 , otherwise @xmath66 , where @xmath65 is called a weight removing hyperparameter . in a similar way , @xmath67 is defined from the connection weight between the @xmath13-th unit in the depth @xmath14 layer and @xmath15-th unit in the depth @xmath16 layer . for simplicity , we denote @xmath60 and @xmath61 as @xmath68 and @xmath69 , respectively , in the following explanation . our method is based on the assumption that units in the same community have a similar probability of connection from / to other units . this assumption is almost the same as that in the previous method @xcite , except that our method utilizes both incoming and outgoing connections of each community , and it detects communities in individual layers . therefore , the community detection result is derived in a similar way to the previous method @xcite , as explained in the rest of this section . as shown on the right in figure [ fig : com ] ( b ) , the statistical model for community detection has three kinds of parameters . the first parameter @xmath70 represents the prior probability of a unit in the depth @xmath14 layer that belongs to the community @xmath71 . the conditional probability of connections for a given community @xmath71 is represented by the second and third parameters @xmath72 and @xmath73 , where @xmath74 represents the probability that a connection to a unit in the community @xmath71 is attached from the @xmath13-th unit in the depth @xmath62 layer . similarly , @xmath75 represents the probability that a connection from a unit in the community @xmath71 is attached to the @xmath15-th unit in the depth @xmath16 layer . these parameters are normalized so that they satisfy the following condition : @xmath76 our purpose is to find the parameters @xmath77 that maximize the likelihood of given matrices @xmath78 . to solve this problem , we introduce the community assignment @xmath79 , where @xmath80 is the community of the @xmath38-th unit in the depth @xmath14 layer . the parameters are optimized so that they maximize the likelihood of @xmath78 and @xmath81 : @xmath82 where @xmath83 then , the log likelihood of @xmath78 and @xmath81 is given by @xmath84 here , the community assignment @xmath81 is a latent variable and is unknown in advance , so we can not directly calculate the above @xmath85 . therefore , we calculate the expected log likelihood @xmath86 over @xmath81 instead . @xmath87 where @xmath88 is the number of units in the depth @xmath14 layer . by defining @xmath89 the above equation can be rewritten as follows : @xmath90 the parameter @xmath91 represents the probability that the @xmath38-th unit is assigned to the community @xmath71 . in other words , the community detection result is given by the estimated @xmath92 . the optimal parameters for maximizing @xmath86 of eq . ( [ eq : exploglh ] ) are found with the em algorithm . the parameters @xmath77 with given @xmath92 are recursively optimized . if @xmath93 maximizes @xmath86 , then they satisfy @xmath94 \left [ \prod_j { \tau'_{c , j}}^{b_{k , j } } \right]}{\sum_s \pi_s \left [ \prod_i { \tau_{s , i}}^{a_{i , k } } \right ] \left [ \prod_j { \tau'_{s , j}}^{b_{k , j } } \right ] } , \label{eq : q}\end{aligned}\ ] ] and @xmath95 the denominator and numerator in the last term of eq . ( [ eq : qdefine ] ) are given by @xmath96 \left [ \prod_j { \tau'_{g_l , j}}^{b_{l , j } } \right ] \right\}\\ & = & \left\ { \pi_c \left [ \prod_i { \tau_{c , i}}^{a_{i , k } } \right ] \left [ \prod_j { \tau'_{c , j}}^{b_{k , j } } \right ] \right\ } \left\ { \prod_{l\neq k } \sum_s \pi_s \left [ \prod_i { \tau_{s , i}}^{a_{i , l } } \right ] \left [ \prod_j { \tau'_{s , j}}^{b_{l , j } } \right ] \right\},\end{aligned}\ ] ] and @xmath97 \left [ \prod_j { \tau'_{s , j}}^{b_{k , j } } \right],\end{aligned}\ ] ] where @xmath98 is the kronecker delta . therefore , @xmath91 is given by eq . ( [ eq : q ] ) . the problem is to maximize @xmath86 of eq . ( [ eq : exploglh ] ) with a given @xmath92 under the condition of eq . ( [ eq : normalization ] ) . this is solved with the lagrangian undetermined multiplier method , which employs @xmath99 and @xmath100 from eq . ( [ eq : lag1 ] ) , the following equations are derived : @xmath101 using eq . ( [ eq : exploglh ] ) and eq . ( [ eq : lag2 ] ) , we obtain @xmath102 from eq . ( [ eq : lag3 ] ) and the condition of eq . ( [ eq : normalization ] ) , lagrange s undetermined multipliers @xmath103 are determined , and eq . ( [ eq : lag3 ] ) is rewritten as eq . ( [ eq : pitau ] ) . from the above theorem , the optimal parameters @xmath104 and the probability of community assignment @xmath105 for the optimized parameters are recursively estimated based on eq . ( [ eq : q ] ) and ( [ eq : pitau ] ) . in this paper , the community assigned to the @xmath38-th unit is determined by the @xmath71 that maximizes @xmath91 ( figure [ fig : com ] ( c ) ) . finally , we use the following methods to determine a modular representation of a layered neural network that summarizes multiple connections between the pairs of communities ( figure [ fig : com ] ( d ) ) . + + * four algorithms for determining bundled connections * : * method @xmath64 : community @xmath106 and @xmath107 have a bundled connection iff there exists at least one connection between the pairs of units @xmath108 . * method @xmath109 : let the number of units in communities @xmath106 and @xmath107 be @xmath110 and @xmath111 , respectively , and let the number of connections between the pairs of units @xmath108 be @xmath112 . communities @xmath106 and @xmath107 have a bundled connection iff @xmath113 holds , where @xmath114 is a threshold . * method @xmath115 : among the bundled connections defined by method @xmath109 , only those that satisfy the following ( 1 ) or ( 2 ) are kept and the others are removed . ( 1 ) for any community @xmath116 in the same layer as community @xmath106 , @xmath117 . ( 2 ) for any community @xmath118 in the same layer as community @xmath107 , @xmath119 . * method @xmath120 : among the bundled connections defined by method @xmath109 , only those that satisfy the above ( 1 ) and ( 2 ) are kept and the others are removed . by these procedures , we obtain the modular representation of a layered neural network . in this section , we show three applications of the proposed method : ( 1 ) the decomposition of a layered neural network into independent networks , ( 2 ) generalization error estimation from a community structure , and ( 3 ) knowledge discovery from a modular representation . here we verify the effectiveness of the proposed method in the above three applications . the following processing was performed in all the experiments : 1 . the input data were normalized so that the minimum and maximum values were @xmath121 and @xmath122 , respectively . the output data were normalized so that the minimum and maximum values were @xmath123 and @xmath124 , respectively . the initial parameters were independently generated as follows : @xmath125 . the connection matrix @xmath127 if the absolute value of the connection weight between the @xmath13-th unit in the depth @xmath62 layer and the @xmath15-th unit in the depth @xmath14 layer is larger than a threshold @xmath65 , otherwise @xmath128 . note that @xmath124 and @xmath123 are used instead of @xmath64 and @xmath129 for stable computation . similarly , @xmath67 is defined from the connection weight between the @xmath13-th unit in the depth @xmath14 layer and the @xmath15-th unit in the depth @xmath16 layer . all units were removed that had no connections to other units . 5 . for each layer in a trained neural network , @xmath130 community detection trials were performed . we defined the community detection result as one that achieved the largest expected log likelihood in the last of @xmath131 iterations of the em algorithm . 6 . in each community detection trial , the initial values of the parameters @xmath77 were independently generated from a uniform distribution on @xmath132 , and then normalized so that eq . [ eq : normalization ] held . 7 . in visualization of modular representation , all communities with no output bundled connections from them were regarded as unnecessary communities . in the output layer , the communities with no input bundled connections were regarded in the same way as above . the bundled connections with such unnecessary communities were also removed . these unnecessary communities and bundled connections were detected from depth @xmath18 to @xmath64 , since the unnecessary communities in the shallower layers depend on the removal of unnecessary bundled connections in the deeper layers . we show that the proposed method can properly decompose a neural network into a set of small independent neural networks , where the data set consists of multiple independent dimensions . for validation , we made synthetic data of three independent parts , merged them , and applied the proposed method to decompose them into the three independent parts . the method we used to generate the synthetic data is shown in figure [ fig : exp1 ] . in the following , we explain the experimental settings in detail . first , three sets of input data were independently generated . all the sets contained input data with five dimensions , and their values followed : @xmath133 . then , three neural networks were defined , each of which has independent weights and biases . in each neural network , all the layers consisted of five units , and the number of hidden layers was set at one . the sets of weights and biases for the first , second and third neural networks are denoted as @xmath134 , @xmath135 , and @xmath136 , respectively . these parameters were randomly generated as follows : @xmath137 for the weights @xmath138 and @xmath139 , the connections with smaller absolute values than one were replaced by @xmath129 . finally , three sets of output data were generated by using the above input data and neural networks by adding independent noise following @xmath140 . the three generated sets of input and output data were merged into one set of data , as shown in figure [ fig : exp1 ] . we trained another neural network with @xmath141 dimensions for input , hidden and output layer using the merged data . then , a modular representation of the trained neural network was made with the proposed method . the results of the trained neural network , its community structure , and its modular representation are shown in figures [ fig : exp1o ] , [ fig : exp1c ] , and [ fig : exp1 m ] , respectively . the numbers above the input layer and below the output layer are the indices of the three sets of data . these results showed that the proposed method could decompose the trained neural network into three independent networks . [ fig : exp1o ] [ fig : exp1c ] in general , a trained result of a layered neural network is affected by different hyperparameters and initial parameter values . here , we show that the appropriateness of a trained result can be estimated from the extracted community structure by checking the correlation between the generalization error and the modularity @xcite . the modularity is defined as a measure of the validity of the community detection result , and it becomes higher with more intra - community connections and fewer inter - community connections . let the number of communities in the network be @xmath142 , and @xmath143 be a @xmath144 matrix whose element @xmath145 is the number of connections between communities @xmath13 and @xmath15 , divided by the total number of connections in the network . the modularity @xmath146 of the network is defined by @xmath147 this is a measure for verifying the community structure of assortative networks , so it can not be applied directly to layered neural networks . in this paper , we define a modified adjacency matrix based on the original adjacency matrix of a layered neural network , and use it for measuring modularity . in the modified adjacency matrix , an element indexed by row @xmath13 and column @xmath15 represents the number of common units that connect with both the @xmath13-th and @xmath15-th units ( figure [ fig : modularitycalc ] ) . we set the diagonal elements of modified adjacency matrix at @xmath129 , resulting that there are no self - loops . if the data consist of multiple independent sets of dimensions like the synthetic data used in the experiment in section [ sec : decomposition ] , the generalization error is expected to be smaller when the weights of the connections between independent sets of input and output are trained to be smaller . therefore , a higher modularity indicates a smaller generalization error . + + + + + captypetable c\|c\|\|c\|c\|c & + & & & + & @xmath148 & @xmath149 & @xmath150 & @xmath151 + @xmath65 & @xmath152 & @xmath153 & @xmath154 & @xmath155 + & @xmath156 & @xmath157 & @xmath158 & @xmath159 + captypetable [ tab : cc ] in the experiment , we iterated the neural network training and community detection from the trained network @xmath160 times using the same data as in the experiment described in section [ sec : decomposition ] . the generalization error and modularity results for nine pairs of hyperparameters @xmath161 are shown in figure [ fig : exp2 ] , where @xmath28 is the lasso hyperparameter and @xmath65 is the weight removing hyperparameter . for the smaller @xmath28 and @xmath65 , the overall modularities were lower , which indicates that there were more connections between mutually independent neural networks . it was experimentally shown for some hyperparameters that better trained results were obtained ( with smaller generalization errors ) when the trained neural networks had clearer community divisions ( with higher modularity ) . table [ tab : cc ] shows the correlations for given @xmath161 . in order to show that the modular representation extracts the global structure of a trained neural network , we applied the proposed method to a neural network trained with practical data . we used data that represent the characteristics of each municipality in japan @xcite . the characteristics shown in table [ tab : notationsdata ] were used as the input and output data , and the data of municipalities that had any missing value were removed . there were @xmath162 training data . before the neural network was trained , all the dimensions for all sets of data were converted through the function of @xmath163 , because the original data are highly biased . the results are shown in figure [ fig : exp3data ] . the trained neural network and the modular representation using the above data are shown in figures [ fig : exp3o ] and [ fig : exp3 m ] , respectively . figure [ fig : exp3 m ] shows , for example , that the number of births , marriages , divorces , and people who engage in tertiary industry work ( a1 ) were inferred from the population aged 65 and older , the number of out - migrants , establishments , employees with employment and so on ( a4 ) . [ fig : exp3o ] from the extracted modular representation , we found not only the grouping of the input and output units , but also the relational structure between the communities of the input , output and hidden layers . for instance , the third community from the right in the depth @xmath109 layer ( b4 ) and the second community from the right in the depth @xmath115 layer ( b3 ) only connected to partial input and output units : they were used only for inferring the number of primary industrial employees and employed workers in their municipalities ( b1 , b2 ) , from the total population , the population under 15 years of age , the population of transference , and the numbers of family workers and executives ( b5 , b6 ) . in this section , we discuss the proposed algorithm from three viewpoints , the community detection method , the validation of the extracted result , and the scalability of our method . firstly , to extract a modular representation from a trained neural network , we employed a basic iterative community detection method for each layer . it is possible to modify this method , for example , by using the weights of connections or the connections in further layers . utilizing the output of each unit might also improve preciseness of the community detection result . the optimization of community detection methods and hyperparameters according to the task is future work . secondly , knowledge discovered from a modular representation depends on both the data and the analyst who utilizes the proposed method . there are both sensitive and robust communities which do and do not depend on them . therefore , it becomes important to separate the essential results from fluctuations . we anticipate that our method will form the basic analytic procedure of such a study . and lastly , in this paper , we experimentally evaluated the relationship between modularity and generalization error . it is well known that a community detection technique can be employed for large size networks . the analysis of larger datasets with higher dimensions would provide further information on layered neural networks . for such large datasets , it would also be important to evaluate the effectiveness of parallel computation , using the independent neural networks extracted with our proposed method . deep neural networks have achieved a significant improvement in terms of classification or regression accuracy over a wide range of applications by their ability to capture the complex hidden structure between input and output data . however , the discovery or interpretation of knowledge using deep neural networks has been difficult , since its internal representation consists of many nonlinear and complex parameters . in this paper , we proposed a new method for extracting a modular representation of a trained layered neural network . the proposed method detects communities of units with similar connection patterns , and determines the relational structure between such communities . we demonstrated the effectiveness of the proposed method experimentally in three applications . ( 1 ) it can decompose a layered neural network into a set of small independent networks , which divides the problem and reduces the computation time . ( 2 ) the trained result can be estimated by using a modularity index , which measures the validity of a community detection result . and ( 3 ) providing the global relational structure of the network would be a clue to discover knowledge from a trained neural network . table [ tab : notationsdata ] shows the notations of the data @xcite used in the experiment described in section [ sec : kd ] . the experimental settings of the parameters are shown in table [ tab : notations ] . c\|p5cm\|\|c\|p5 cm name & & name & + i1 & & i22 & + i2 & & i23 & + i3 & & i24 & + i4 & & o1 & + i5 & & o2 & + i6 & & o3 & + i7 & & o4 & + i8 & & o5 & + i9 & & o6 & + i10 & & o7 & + i11 & & o8 & + i12 & & o9 & + i13 & & o10 & + i14 & & o11 & + i15 & & o12 & + i16 & & o13 & + i17 & & o14 & + i18 & & o15 & + i19 & & o16 & + i20 & & o17 & + i21 & & o18 & + [ tab : notationsdata ] c\|p6cm\|c\|c\|c name & meaning & exp.1 & exp.2 & exp.3 + @xmath164 & & + @xmath4 & number of training data sets & @xmath165 & @xmath166 & @xmath162 + @xmath167 & number of test data sets & @xmath129 & @xmath166 & @xmath129 + @xmath168 & & & @xmath169 + @xmath18 & & & @xmath120 + @xmath28 & hyperparameter of lasso & @xmath170 & * & @xmath171 + @xmath172 & & + @xmath65 & weight removing hyperparameter & @xmath152 & * & @xmath173 + @xmath142 & & & @xmath174 + method & & & @xmath115 + @xmath114 & & + @xmath121 & & & @xmath175 + @xmath122 & & & @xmath64 + * : the nine parameters shown in the caption of figure [ fig : exp2 ] are used . [ tab : notations ] we would like to thank akisato kimura for his helpful comments on this paper .	deep neural networks have greatly improved the performance of various applications including image processing , speech recognition , natural language processing , and bioinformatics . however , it is still difficult to discover or interpret knowledge from the inference provided by a deep neural network , since its internal representation has many nonlinear and complex parameters embedded in hierarchical layers . therefore , it becomes important to establish a new methodology by which deep neural networks can be understood . in this paper , we propose a new method for extracting a global and simplified structure from a layered neural network . based on network analysis , the proposed method detects communities or clusters of units with similar connection patterns . we show its effectiveness by applying it to three use cases . ( 1 ) network decomposition : it can decompose a trained neural network into multiple small independent networks thus dividing the problem and reducing the computation time . ( 2 ) training assessment : the appropriateness of a trained result with a given hyperparameter or randomly chosen initial parameters can be evaluated by using a modularity index . and ( 3 ) data analysis : in practical data it reveals the community structure in the input , hidden , and output layers , which serves as a clue for discovering knowledge from a trained neural network . + _ keywords _ : layered neural networks , network analysis , community detection
the vela pulsar is one of the isolated neutron stars with the widest observational database , spanning from radio waves to high - energy @xmath2-rays . nevertheless , the value of its distance is still in debate . the canonical value of 500 parsec , derived by milne ( 1968 ) from the radio signals dispersion measure , has been recently questioned by several independent investigations . studies of both the kinematics of the host supernova remnant ( cha et al . 1999 ; bocchino et al . 1999 ) and constraints on the neutron star radius imposed by the pulsar soft x - ray spectrum ( page et al . 1996 ) have suggested a significant downward revision of the distance . in both cases , a value of @xmath3 250 parsec appears more likely than the canonical one . a model - free evaluation of the distance which could settle the question can be obtained only by measuring the annual parallax of the pulsar . to this aim we have been granted a triplet of consecutive hst / wfpc2 observation to be performed six months apart at the epochs of the maximum parallactic displacement . .previous estimates of the vela pulsar proper motion in literature [ cols="^,^,^,^ " , ] the best way to gauge the angular displacement of an object between different epochs is to perform relative optical astrometry measurements . this needs : * a set of `` good '' reference stars , accurately positioned , to provide the relative reference frame for each image * a reliable procedure to align the reference frames in our relative astrometry analysis we have used all the images of the vela pulsar field collected so far by the hst ( see tab.2 ) . the observations , all taken through the f555w `` v filter '' , have been obtained either using the original wfpc and the wfpc2 , with the pulsar positioned in one of the wfc chips or in the pc one , respectively . observation # 1 has been retrieved from the st - ecf database and recalibrated on - the - fly by the archive pipeline using the most recent reference data and tables , while observations # 2 and # 3 were obtained as part of our original parallax program , included in cycle 6 but never completed . the program is now being repeated in cycle 8 and image # 4 is indeed the first of the new triplet of wfpc2 observations . for each epoch , cosmic - ray free images were obtained by combining coaligned exposures . the choice of reference stars was limited to the chip containing the pulsar optical counterpart ( see caraveo & mignani 1999 for a qualitative discussion of the inter - chip astrometry ) . 25 common stars were selected in the pc images of 1997 , 1998 and 1999 ; 19 of them were identified in the 1993 wfc image . this defines our set of reference stars . their coordinates were calculated by 2-d gaussian fitting of the intensity profile . the evaluation of the positioning errors was conservative ; for each reference star the fit was repeated several times on centering regions of growing areas , till errors showed no more dependence on the background conditions ( see caraveo et al . uncertainties on the centroid positions were typically of order [email protected] pixel ( i.e. 1@xmath43 mas ) per coordinate in the wfpc2 images and [email protected] pixel ( i.e. 3@xmath46 mas ) per coordinate in the wfc one . + the values of the coordinates of reference stars have been then corrected for the effects of the significant geometrical distortion of the wfc and the wfpc2 ccds . this correction has been applied following two different mappings of the instrument field of view : ( i ) the solution determined by gilmozzi et al . ( 1995 ) , implemented in the stsdas task _ metric _ and ( ii ) the solution determined by holtzman et al . the procedures turned out to be equivalent , with the rms residuals on the reference stars coordinates after image superposition ( see below ) consistent within few tenths of mas . + as a reference for the image superposition , we have choosen the 1997 one . given the abundance of reference stars , the alignement has been performed following the traditional astrometric approach , consisting of a linear transformation with 5 free parameters i.e. 2 independent translation factors , 2 scale factors for @xmath5 and @xmath6 and a rotation angle . the residuals of reference stars positions clustered around 0 and appeared randomly distributed on the field of the image , showing no systematic effect . thus , we are confident that our analysis is bias free and reliable . as expected , the average rms residual on the reference stars positions is higher in the wfc - to - pc superposition ( of order 4 mas ) than in the pc - to - pc case ( of order 2 mas ) . we note that in all cases the relative orientation of the images coincide , within few hundredths of degree , with the difference between the corresponding telescope roll angles . the alignement procedure has been repeated after applying different image sharpening algorithms ( e.g. gaussian filtering ) , yielding very similar results . the displacement of the vela pulsar position can be immediately appreciated in the isophotal plot of fig.2 , which shows the zoomed superposition of all the frames taken at the different epochs . to measure the pulsar proper motion , we have performed linear fits to its ra / dec displacements vs time . as a first step , we used all the available points to have a longer time span . however , owing to the coarser angular resolution of the 1993 image , taken with the wfc , almost indistinguishable results are obtained using only the 1997 - 1999 points , all obtained with the pc . we also tried a direct comparison of the 1997 and 1999 points , taken exactly at the same day of the year i.e. with identical parallax factors , which should lead to the clearest measurement of the proper motion . all of the steps described above yielded results largely consistent within the errors . since all the proper motion values obtained through different image processing / frame superposition / displacement fit are consistent within @xmath7 mas yr@xmath8 from each other , we have conservatively assumed the value of 2 mas yr@xmath8 ( per coordinate ) as our overall error estimate . thus , our final estimate of the vela pulsar proper motion is : @xmath9 @xmath10 for a total proper motion of : @xmath11 with a position angle of @xmath12 . analysis of hst data yielded the most accurate measure of the vela pulsar proper motion . our value is in excellent agreement with the one of nasuti et al . ( 1997),but the errors are now reduced . what is worth mentioning here is that our data refer to a time span much shorter than the 20 years interval of the previous work . indeed , two years of hst observations are vastly enough to improve over 20 years observations from the ground . once more , this is a clear demonstration of the excellent potentialities of hst astrometry . + transforming proper motion into a transverse velocity requires the knowledge of the object s distance . at the canonical , although uncertain , 500 pc distance the implied velocity would be @xmath13 km s@xmath8 , somewhat on the low side for the fast moving pulsar family ( lorimer , bailes & harrison 1997 ) . a reduction of the pulsar distance will similarly reduce the transverse speed to an embarassingly low value . nailing down the pulsar proper motion is the first step to assess the annual parallactic displacements of the source . our next observations will , hopefully , yield a direct measurement of the vela pulsar distance . part of this work was done at the st - ecf of garching . adl wishes to thank ecf for the hospitality and the support during that period . adl is sincerely grateful to the collegio ghislieri of pavia ( italy ) , to the pii quinti sodales association of pavia and to the maximilianeum stiftung of munich , which offered a truly essential support and a warm hospitality for his stay in germany . references bailes , m. et al . , 1989 , apj 343 , l53 bignami , g.f . and caraveo , p.a . , 1988 , apj 325 , l5 bocchino , f. , maggio , a. and sciortino , s. , 1998 , a&a 342 , 839 caraveo , p.a . and mignani , r.p . , 1999 , a&a 344 , 367 caraveo , p.a . , bignami , g.f . , mignani , r.p . e taff , l.g . , 1996 , apj 461 , l91 cha , a.n . , sembach , k.r . e danks , a.c . , 1999 , apj 515 , l25 fomalont , e.b . , goss , w.m . , manchester , r.n . and lyne , a.g . , 1997 , mnras 286 , 81 fomalont , e.b . et al . , 1992 , mnras 258 , 497 gilmozzi , r. , ewald , s. and kinney , e. , wfpc2 instrument science report 95 - 02 holtzman , j. et al . , 1995 , p.a.s.p . 107 , 156 legge , d. , 1999 , poster presented at the iau colloqium 177 pulsar astronomy -2000 and beyond lorimer , d.r . , bailes , m. and harrison , p.a . , 1997 , mnras , 289 , 592 markwardt , c.b . and gelman , h. , 1994 , baas 26 , 871 milne , d.k . , 1968 , aust.j.phys 21 , 201 nasuti f.p . , mignani , r.p . , caraveo , p.a . and bignami , g.f . , 1997 , a&a 323 , 839 gelman , h. , koch - miramond , l. and aurire , m. , 1989 , apj 342 , l83 page , d. , shibanov , y.a . , and zavlin , v.e . , mpe report 263 , 173	using all the hst wfc / wfpc2 images of the field collected so far , we have performed an accurate relative astrometry analysis to re - assess the value of the vela pulsar proper motion . although covering a much shorter time span , our measurement clearly confirms the previous result obtained by nasuti et al . ( 1997 ) using ground - based optical data and nails the proper motion value to @xmath0 mas @xmath1 .
in this work we calculate very high lying eigenvalues of a billiard system using the so - called scaling method @xcite . this method has two main advantages , it is formulated on the boundary of the billiard ( allowing matrices of order @xmath0 , the wavenumber ) and it avoids zeroes searching algorithms . we have solved the helmholtz equation with dirichlet boundary conditions on the billiard boundary . a point worth to mention is that any eigenfunction is @xmath1 at domain points . at straight segments of the boundary , eigenfunctions are reflected as odd functions , so that the result satisfies the dirichlet condition . though being also @xmath1 at the boundary by straight pieces , they are not analytical at a vertex when two segments of the boundary join ( with inner angle @xmath2 , where @xmath3 is a noninteger number ) . this feature leads to what we address as diffractive phenomena . the main idea behind the scaling method is that the trial functions of the variational problem can be parametrised by energy ( alternatively , by the wavenumber @xmath0 ) . with this idea in mind these functions are expanded in a suitable scaling basis ( plane waves with different propagation directions are one example and these are the elements we use here ) . asking the function to be zero at the boundary of the billiard is the same as asking its norm there to be zero as well . then , eigenvalues ( and eigenfunctions ) can be obtained by solving a generalised eigenvalue problem that involves the quadratic form associated with the norm of the function on the boundary . in applying the method to this system , evanescent waves were needed . this is because real plane wave solutions to the helmholtz equation can not represent all the features of diffraction . it has been shown @xcite that an evanescent plane wave , which oscillates along propagation direction faster than the wavenumber @xmath0 , can be constructed by means of real plane waves . nevertheless , the corresponding superposition is a singular one , suggesting the direct use of the evanescent functions in the basis . this is the way we have dealt with diffraction . for simplicity we restrict our considerations to an angle of @xmath4 . a particularly suitable example to study the effect of this kind of corner on the quantum properties is the polygonal billiard shown in fig . [ fig : billiard ] . this billiard is called the truncated triangle . it has been studied in a variety of contexts @xcite . the truncated triangle belongs to the class of _ rational polygons_. that are polygons where all angles @xmath5 between sides are rationally related to @xmath6 , where @xmath7 are relatively prime integers . the free motion inside a rational polygon is integrable if @xmath8 for all @xmath9 , which is the case for rectangles , the equilateral triangles , the @xmath10-triangles and the @xmath11-triangles . all other rational polygons are _ pseudointegrable _ @xcite . like in integrable systems , the phase space is foliated by two - dimensional invariant surfaces @xcite . however , the genus of the surfaces is greater than one due to _ critical corners _ with @xmath12 @xcite . in the case of the truncated triangle , the genus is @xmath13 . roughly speaking , the invariant surface is a torus with an additional handle . quantum signatures of pseudointegrability can be found in the energy eigenfunctions @xcite and in the statistical properties of energy levels @xcite . the energy levels of pseudointegrable systems are correlated in contrast to those of integrable systems which are generically well described by the poissonian random processes @xcite . for example , the nearest - neighbor spacing distribution of pseudointegrable systems generically displays level repulsion @xcite , resembling the gaussian orthogonal ensemble ( goe ) of random - matrix theory @xcite which describes fully chaotic systems with time - reversal symmetry @xcite . significant deviations from goe have been observed first in refs . it has been suggested that the spectral statistics of pseudointegrable systems is another example of critical or intermediate statistics @xcite . critical statistics appear in many condensed matter problems such as in mesoscopic disordered systems at the critical point of the metal - insulator transition @xcite , in systems with a few interacting electrons @xcite , and in incommensurate multiwalled carbon nanotubes @xcite . in refs . @xcite it has been proposed to use the semi - poisson ( sp ) statistics as a reference point for critical statistics . the sp statistics is defined by a simple construction : remove every other level from an ordered poisson sequence @xcite . the sp statistics is useful because it provides explicit formulas for a number of statistical quantities which can be compared to the statistical properties of a given system . for several pseudointegrable systems @xcite it has been confirmed that the sp statistics indeed describes the short - range level correlations rather well . comparing the long - range level correlations is numerically a difficult task because the statistical properties of rational polygons converge extremely slowly as energy is increased @xcite . fortunately , semiclassical periodic - orbit theory allows to compute analytically the long - range level correlations , in terms of the level compressibility @xmath14 , of a few special systems , like certain right triangles @xcite and the barrier billiard @xcite . in the former case @xmath14 differs in general from the sp result , whereas in the latter case @xmath14 is in agreement with the sp statistics . nothing is known analytically about the generic case . our numerical analysis will show that the truncated triangle is well described by the sp statistics but deviations are not negligible . our analysis extends that reported in ref . @xcite in many respects : ( i ) the scaling method allows us to compute more levels , giving a better statistics . ( ii ) moreover , high - lying energy levels can be computed . this puts us in a position to study the relevant asymptotic regime . ( iii ) more statistical quantities are computed . ( iv ) the numerical results are compared to the sp statistics . the paper is organized as follows : in section [ sec : scaling ] , we describe the features that are considered when applying the scaling method to the system under investigation . in section [ sec : numerics ] , we present the statistical studies carried out with the data that has been obtained . finally , section [ sec : conclusion ] is devoted to conclusions . we are not going to explain the scaling method here , and we address the reader to the appendix and the given references for details . an eigenfunction of a billiard can be constructed as a plane wave superposition . this can include evanescent waves , i.e. , plane waves with complex wave vectors . these types of waves should be present in quantum billiards and can be associated with diffractive phenomena . discontinuities at the boundaries seem to be strongly related to the way evanescent functions must be considered to solve the problem . several authors have focused their attention on this issue @xcite . for polygonal billiards , in the generic case , theory suggests that there is no real plane wave superposition that can be an eigenfunction @xcite . on the other hand , good numerical results using only plane waves could be found @xcite , but working within the region of low energies . evanescent waves were studied in detail by berry @xcite in the context of quantum billiards . in this approach , the main reason to do it comes from the idea of constructing them by continuation of an external scattering superposition , containing only real plane waves . in principle , this does not seem to be possible . anyway , he showed how evanescent waves can be expressed as the singular limit of an angular superposition of real plane waves . in our system , and working at high energies , we need to consider evanescent waves explicitely , because a real plane wave representation is singular . in the semiclassical limit , the only way to obtain eigenfunctions that include evanescent waves is by considering them in the variational problem @xcite . in the remaining part of this section we are going to show some examples of the waves considered and the idea behind their selection . plane wave solutions to the helmholtz equation can be written as @xmath15}\ ] ] with @xmath16 and @xmath17 real , using coordinates @xmath18 in the plane . we can express this wave in a slightly different fashion by distinguishing the real and imaginary parts in the exponent , @xmath19 } \ ; \exp{[-k \sinh{(\alpha})\tilde y]}.\ ] ] here , the propagation direction @xmath20 implies an angle @xmath16 over the @xmath21 axis . in this direction the function has a wavelength that is given by @xmath22 . however , in the orthogonal direction @xmath23 the function is an exponential with coefficient @xmath24 . an example of these functions applied to our system can be found in fig . [ evadom ] , where we show an evanescent wave corresponding to the truncated triangle of parameter value @xmath25 ( note that since the absolute size of the system is irrelevant , we measure length scales as dimensionless quantities ) . in this case we chose @xmath26 in order to have a visible wave . for the higher - lying wave numbers ( @xmath27 or greater ) these functions are difficult to see in the domain since they decay very fast . nevertheless , the slower oscillating functions we have used in the latter energy region can be seen over the boundary . on the truncated triangle domain with parameter value @xmath25 . the propagation direction @xmath28 ( given by @xmath29 in this case ) and the corresponding decaying direction @xmath23 are both shown by the arrows . also , we point out values @xmath30 and @xmath31 of the arclength coordinate @xmath32.,width=226 ] now that the general ideas related to evanescent waves have been exposed , we are going to explain the way we selected them , and we are also going to show some examples for @xmath33 . the first thing to point out is that diffraction is generated at the critical corner with angle @xmath34 that is shown in fig . [ fig : billiard ] . so , our sets or families of evanescent waves are centered at this corner . we consider three different kinds of functions , all sharing the previous property , but decaying to one or the other side " of the corner ( in terms of the arclength boundary coordinate @xmath32 ) in the first place and also an additional one that decays to both sides . we use a family of 12 waves whose propagation direction @xmath16 is slightly smaller than @xmath35 , i.e. , around the right hand side of the boundary as seen from fig . [ fig : billiard ] , another set of 8 waves with almost horizontal propagation and finally one that goes along an intermediate direction ( for this one we took a value of @xmath36 , as can be seen in fig . [ evadom ] ) . in fact , in order to appreciate their contribution , it is convenient to look at them on the boundary of the billiard . for this reason we show in fig . [ evbound ] three examples of evanescent waves that are good representatives of these three families . as already mentioned , the exponential decay is from the critical corner , which corresponds to @xmath37 in terms of the boundary coordinate . this is the effect of considering an angle slightly different from @xmath38 angle for the propagation direction . the same happens with the second family whose example is directed almost along the horizontal ( left ) segment of the boundary . finally , we show the evanescent wave that decays to both sides of the corner . . vertical axis corresponds to the real part of the evanescent wave in arbitrary units . horizontal axis corresponds to the boundary arclength coordinate @xmath32 ( dimensionless ) . in ( a ) we show a wave decaying to the right ( propagation direction given by @xmath39 ) , in ( b ) one to the left ( @xmath40 ) , and in ( c ) to both sides ( @xmath29 ) from @xmath37 , the position of the critical corner.,width=264 ] this approach to the problem has proven to be very efficient . as a matter of fact , studying carefully several billiard eigenfunctions that show the greatest norm ( error ) over the boundary ( without considering evanescent waves in their calculation ) , we could check that these are the main components of the diffracted field . then we apply the symmetries of our system in order to get the right contributions . this is easy to implement by using the symmetry operations of the @xmath41 group . we have taken only up to 21 evanescent waves in order to get our results . this is a small number compared to the roughly 1000 real plane waves that are used in the highest - lying energy window we have obtained @xcite . but they are key to lower the error of the eigenvalues . this shows that , even though the relevance of this contribution goes to zero in the semiclassical limit we need to consider it in order to resolve individual states . we here examine the spectral statistics of the quantized truncated triangle . we consider the generic case where @xmath31 is an irrational number . in the nongeneric case of @xmath31 being rational , the energy spectrum contains a subset of poisson distributed levels @xcite . we focus on the parameter value @xmath25 , the reciprocal of the golden mean . for other parameter values of @xmath31 [ @xmath42 , @xmath43 , and @xmath44 similar results have been obtained . we use energy windows of length @xmath45 in five different regimes starting with level number @xmath46 , @xmath47 , @xmath48 , @xmath49 , and @xmath50 corresponding to the wave number @xmath51 , @xmath52 , @xmath53 , @xmath54 , and @xmath55 . to study the local fluctuations in the level sequence @xmath56 it is necessary to remove the systematic global energy dependence of the average density . to do so , we `` unfold '' the spectra in the usual way by setting @xmath57 ; see , e.g. , ref . @xmath58 is the smooth part of the integrated density of states , i.e. , the number of levels up to energy @xmath59 . we approximate @xmath58 by the generalized weyl s law including perimeter and corner corrections @xcite @xmath60 where @xmath61 is the area of the billiard , @xmath62 is the perimeter , and @xmath63 is the corner correction . the unfolded spectra @xmath64 have unit mean level spacing and are dimensionless . henceforth , the tilde will be suppressed . the most popular statistical quantity in the field of quantum chaos is the nearest - neighbor spacing distribution . it is defined as the probability density of the spacing @xmath65 ( in units of the mean level spacing ) between adjacent levels , @xmath66 clearly , the nearest - neighbor spacing distribution is a measure of short - range level correlations . we will compute its integral , the cumulative spacing distribution @xmath67 for the poisson statistics @xmath68 and @xmath69 , the goe is well described by the wigner surmise @xmath70 and @xmath71 , and for the sp statistics @xcite @xmath72 at small @xmath65 , @xmath73 exhibits a linear increase from zero ( level repulsion ) similar to the wigner surmise . at large @xmath65 , @xmath73 has an exponential fall - off as the poisson statistics . in fig . [ fig : cnnzero ] we plot the difference between the cumulative spacing distribution to the sp result for the lowest - energy window . good agreement with the sp statistics can be observed . figure [ fig : cnn ] shows a magnification , containing also the other energy windows . we see increasing deviations from sp for @xmath74 and @xmath75 , and then decreasing deviations for @xmath76 and @xmath77 . no clear trend to sp is visible as one goes to higher energies . similar small deviations from the sp statistics have been observed at the metal - insulator transition in the three - dimensional anderson model . in that case , an ensemble average over certain boundary conditions removes the deviations considerably @xcite . to see whether this is possible also in our case , we consider neumann boundary conditions on the two boundary segments which do not touch the critical corner ; see fig . [ fig : billiard ] . in this way we obtain four energy spectra corresponding to dirichlet / dirichlet , dirichlet / neumann , neumann / dirichlet , and neumann / neumann - boundary conditions . in contrast to the case of the anderson model we find that averaging over these boundary conditions does not reduce the deviation from the sp statistics . we now consider the next - to - nearest spacing distribution ( second - neighbor - spacing distribution ) and its integral . the sp statistics gives @xcite @xmath78 figure [ fig : nextcnn ] shows that the cumulative next - to - nearest spacing distribution is well described by the sp statistics . the maximal deviations from sp are smaller than in the case of the nearest - neighbor spacing distribution in fig . [ fig : cnn ] . but , again , there is no clear convergence to the sp statistics as one goes to higher energies . so far we have studied short - range level correlations ( nearest and next - to - nearest spacing distributions ) . long - range level correlations are conveniently studied with the help of the number variance @xmath79 @xmath80 is the local variance of the number @xmath81 of energy levels in the interval @xmath82 $ ] . for the sp statistics we have @xcite @xmath83 figure [ fig : sigma ] shows the number variance computed for the five energy windows . note that the regime is well below the crossover region where @xmath80 begins to saturate at a value determined by the shortest periodic orbit @xcite . with increasing energy the number variance comes closer to the sp result , without showing a clear stabilization . to estimate the limit curve as @xmath84 we use the extrapolation procedure described in ref . @xcite : extrapolate point by point ( with @xmath85 fixed ) the four highest curves with a fit @xmath86 . the limit curve , @xmath87 , is shown as dashed curve in fig . [ fig : sigma ] . it is closer to the sp result . to measure the difference we concentrate on the slope in the region of large @xmath85 , the so - called level compressibility @xmath88 we get @xmath89 from the extrapolated curve this is close to the sp result @xcite of @xmath90 ( @xmath91 for poisson @xcite and @xmath30 for goe @xcite ) . for @xmath51 , @xmath52 , @xmath53 , @xmath54 , and @xmath55 ( from below).,width=283 ] another measure of long - range level correlations is the form factor @xmath92 , the fourier transform of the two - point correlation function . the limit @xmath93 is related to the number variance by means of @xmath94 ( see , e.g. , ref . @xcite ) with @xmath14 from eq . ( [ eq : chi ] ) . the form factor can be approximated numerically by ( see , e.g. , ref . @xcite ) @xmath95 in our case @xmath96 . note that @xmath97 is dimensionless . figure [ fig : kav ] shows @xmath98 averaged over small intervals of size @xmath99 in the high - energy regime , i.e. , @xmath100 . it is difficult to estimate @xmath101 from such kinds of noisy data , but it is justified to say that @xmath101 is below the sp prediction @xmath90 , in agreement with our former numerical results on the number variance . a better way to compare the form factor to the sp statistics is introduced in ref . fit @xmath98 to the function @xmath102 if @xmath103 then function ( [ eq : kfit ] ) is the sp form factor . we use the quantity @xmath104 to measure the difference to the sp statistics . keep in mind that @xmath105 , in general , differs from @xmath106 since it depends also on @xmath98 with @xmath107 . figure [ fig : kav ] shows the result obtained by fitting eq . ( [ eq : kfit ] ) to the smoothed data over the range @xmath108 . we get @xmath109 . for the lower - energy windows with @xmath74 , @xmath75 , and @xmath110 we find @xmath111 , @xmath112 , and @xmath113 , respectively . again , as for the other spectral quantities described in the previous subsections , we find that the spectral statistics is roughly described by the sp statistics . however , there are small but significant deviations which show no clear trend to zero as the energies are increased . ) for the highest - energy window averaged over small intervals of size @xmath99 . the smooth curve is the fit ( [ eq : kfit ] ) with @xmath114.,width=283 ] we have shown how the scaling method can be applied to convex billiards with corners . to gain insight into the strong diffractive phenomena that appear in these systems we have studied the truncated triangle . as a result of our investigations of the diffracted field we could identify its main components . these were directly introduced in the function basis in order to obtain the spectral data of this system . evanescent waves conveniently selected and associated with the centers of diffraction ( in our case the two straight segments junction at @xmath4 ) have been successfully identified as a very efficient way to deal with these kinds of phenomena . this allowed us to obtain a great number of highly excited eigenvalues . we have studied the statistical properties of high - lying energy levels in this pseudointegrable billiard . we have found that the nearest - neighbor spacing distributions , next - to - nearest spacing distributions , number variance , and the spectral form factor are roughly described by the semi - poisson ( sp ) statistics . whether the sp statistics is asymptotically the exact statistics can not be decided . we would like to thank eduardo vergini for his helpful suggestions during the writing of the manuscript . also for the ideas he gave us regarding the evanescent wave approach to diffraction at corners . moreover , we would like to thank h. schomerus and t. gorin for discussions . in this appendix we are going to briefly explain the scaling method @xcite . the main point that one should have in mind is that the boundary norm can be written as a function of energy because scaling is used . if @xmath115 satisfies the helmholtz equation with eigenvalue @xmath116 and we associate with it the scaling function @xmath117 then , these functions satisfy the same equation with eigenvalue @xmath118 . if we have a billiard defined by a star - shaped domain and a @xmath119 exists such that @xmath120 at the boundary @xmath121 , then this is a scaling eigenfunction or , equivalently , an eigenfunction of the billiard with dirichlet boundary conditions . the boundary norm defined by the expression @xmath122 can be expanded up to third order around @xmath123 , independently on the exact shape of @xmath124 . we remind that @xmath125 ( where @xmath126 is the unit outgoing normal vector to @xmath121 ) is always positive for star - shaped domains . taking into account this result we can evaluate the norm and its derivative in @xmath0 at value @xmath127 , obtaining @xmath128 this useful expansion , dropping terms of order @xmath129 , turns out to be our quantization condition . then , all the scaling eigenfunctions with eigenvalues close to @xmath130 can be found by solving a generalized eigenvalue problem : @xmath131 \xi^{\mu } = 0 \ .\ ] ] in this equation and for numerical calculations , the quadratic form @xmath132 associated with @xmath133 can be evaluated in a basis of scaling functions @xmath134 ( like plane waves for instance ) by means of @xmath135 the eigenfunctions are @xmath136 , and the eigenvalues can be found as @xmath137 .	what we are going to call in this paper diffractive phenomena " in billiards is far from being deeply understood . these are sorts of singularities that , for example , some kind of corners introduce in the energy eigenfunctions . in this paper we use the well - known scaling quantization procedure to study them . we show how the scaling method can be applied to convex billiards with corners , taking into account the strong diffraction at them and the techniques needed to solve their helmholtz equation . as an example we study a classically pseudointegrable billiard , the truncated triangle . then we focus our attention on the spectral behavior . a numerical study of the statistical properties of high - lying energy levels is carried out . it is found that all computed statistical quantities are roughly described by the so - called semi - poisson statistics , but it is not clear whether the semi - poisson statistics is the correct one in the semiclassical limit .
ultra - luminous x - ray sources ( ulxs ) , also named super - eddington sources or super - luminous sources , were discovered by einstein observatory in 1980 s ( fabbiano 1989 ) . they are usually thought as the point sources with luminosity about @xmath5 or above in the accreting systems . assuming that the isotropic emission is with eddington limit , ulxs may be black holes whose mass go beyond those of the stellar black holes ( makeshima et al . 2000 ; feng and soria 2011 ; pasham et al . 2014 ; weng et al . recently , an ulx nustar j095551 + 6940.8 in the nearby galaxy m82 was discovered by bachetti et al . ( 2014 ) , which may be the counterpart of nustar j095551 + 6940.8 ( feng & kaaret 2007 ; bachetti et al . the spin period ( @xmath1 ) and period derivative ( @xmath2 ) prove that it is an accreting neutron star ( ns ) in high mass binary system ( hmxb ) ( bachetti et al . 2014 ) , which sheds a doubt on that all ulxs are accreting black holes . studying on nustar j095551 + 6940.8 would help us understand various distinctive characteristics of ulxs , and how an accreting neutron star can be explained as an ulx . to this point , many researchers have paid much attentions . shao and li ( 2015 ) modeled the formation history of nustar j095551 + 6940.8 . they suggested that ns in x - ray binary system might significantly contribute to ulx population . high - mass and intermediate - mass x - ray binary pulsars dominated the ulx population in m82 and milky way like galaxies , respectively . klu@xmath6niak and lasota ( 2015 ) pointed out that nustar j095551 + 6940.8 would become a millisecond pulsar within @xmath7 , which provided a new way of millisecond pulsar formation in hmxbs . dallosso et al . ( 2015 ) proposed that the favorite magnetic field of nustar j095551 + 6940.8 was @xmath8 based on the physical considerations and the observed properties of the source . mushtukov et al . ( 2015 ) pointed out that the surface magnetic field strength was @xmath9 for nustar j095551 + 6940.8 with the observed luminosity @xmath10 , but could not be too large due to the propeller effect . eksi et al . ( 2015 ) showed that its dipole magnetic field was as strong as @xmath11 according to the torque equilibrium condition . as expected , it could hold even stronger multi - pole field that suppresses the scattering cross section and increases the eddington luminosity ( eksi & alpar 2003 , canuto et al . moreover , tong ( 2015a , 2015b ) proposed that nustar j095551 + 6940.8 might be a magnetar in an accreting system , which favored the idea that magnetars were descendants of high mass x - ray binaries ( bisnovatyi - kogan & ikhsanov 2015 ) . magnetar is always an isolated ns with the super - strong magnetic field about @xmath12 , which is always thought to be the candidate of soft gamma repeater or anomalous x - ray pulsar ( duncan & thompson 1992 ; kouveliotou et al . 1998 ; mereghetti 2008 ; bisnovatyi - kogan & ikhsanov 2014 ; olausen & kaspi 2014 ) . until now , there is no evidence that indicates magnetar exists in binary system , though it could not be excluded from the theoretical ground . it has been proposed that the accreting magnetar in the close binary system could be used for interpreting the slow rotating ns in 4u 2206 + 54 ( ikhsanov & beskrovnaya 2010 ) . however , magnetars in binary systems should be the rare case , since the birth rate of magnetar is only ten percent of the normal ns ( lorimer 2008 ) . therefore , it can be imagined that a successful explanation of the status of nustar j095551 + 6940.8 can help us understand ns as an ulx that exists in binary system . by means of the accretion induced ns magnetic field evolution model ( zhang & kojima 2006 ) and ns spin - up formula by ghosh and lamb ( 1979 ) , we calculated the evolution of the magnetic field and spin period of a magnetar with the appropriate initial condition to make the evolved spin period fit the status of nustar j095551 + 6940.8 . with the result , we compare its evolutionary track of the magnetic field and spin period with those of the observed pulsar in binary systems . the paper is organized as below : the model is described in section 2 ; calculations and b - p diagram are presented in section 3 ; and the discussion and summary are included in section 4 . the model of accretion induced magnetic field decay of a ns was proposed by zhang and kojima ( 2006 ) . it is assumed that the accretion material is channeled by the strong magnetic field to the polar cap , where the field lines are expelled towards the out - polar regime and enter into the magnetic equator region . as a result , the strength of the polar field decays and the piled - up stronger out - polar field lines at equator region are mostly squeezed into the inner crust of the ns . the analytical form of the polar field decay of a ns is obtained , b = , [ zk ] where @xmath13 is a ratio parameter of the accretion mass @xmath14 and the crust mass @xmath15 of a ns . @xmath16 is with @xmath17 , where @xmath18 is the initial field at the beginning of the accretion , and @xmath19 is the bottom magnetic field while the magnetosphere is compressed onto the surface of ns , b_f=4.510 ^ 8(g)m_18^m^r_6 ^ -(/0.5 ) ^- , [ bf ] where @xmath20 , m and @xmath21 are the ns accretion rate in units of @xmath22 , the mass in solar mass and radius in units of @xmath23 , respectively . in detail , with the condition of the magnetosphere radius @xmath24 matching the ns radius r , i.e. , @xmath25 , the bottom field @xmath26 is achieved . generally , @xmath24 is related to the alfv@xmath27n radius @xmath28 r_m = r_a , r_a=3.210 ^ 8(cm)m_17 ^ -_30^m^- , [ ra ] [ rm ] where @xmath29 is the magnetic moment in units of @xmath30 and the preferred parameter @xmath31 is usually considered ( ghosh & lamb 1979 ; shapiro & teukolsky 1983 ; frank et al . 2002 ) . the evolution of the spin period of a ns in the accretion phase described by ghosh & lamb ( 1979 ) is -=5.810 ^ -5(syr^-1 ) [ m^-r_6^i_45 ^ -1]b_12^ ( pl_37^)^2n(_s ) , [ pdot1 ] where @xmath32 is the moment of inertia in units of @xmath33 and @xmath34 is luminosity of the accreting ns in units of @xmath35 ; @xmath36 is the polar magnetic field in units of @xmath37 ; the dimensionless torque @xmath38 is n(_s ) = 1.4 ( ) , [ ns ] which is associated with the fastness parameter @xmath39 , where @xmath40 is the spin frequency of the ns . @xmath41 is the keplerian frequency at the radius @xmath24 , @xmath42 is the critical value in the range of @xmath43 ( ghosh & lamb 1992 ) . during the ns spin - up process , the spin frequency will be constrained by the keplerian orbital frequency of the accreting material at the magnetosphere radius . therefore , the minimum spin period corresponding to the field strength can be described by the equation of the spin - up line ( or equilibrium period line ) with condition of the spin frequency being the same as the keplerian frequency at the magnetosphere radius ( bhattacharya and van den heuvel 1991 ) : p_eq=0.89(s)b^6/7_12m^-5/7_18 ^ -3/7r_6 ^ 16/7 . [ spin ] the corresponding line of eq.([spin ] ) has been plotted as the solid line in fig.[md1850 ] with the eddington rate and standard ns parameters , e.g. @xmath44 and @xmath45 . as an accreting ns , nustar j095551 + 6940.8 is a unique source since its luminosity and period variation ( @xmath46 and @xmath2 ) are higher than that of ns in hmxbs ( @xmath47 and @xmath48 ) ( ghosh & lamb 1979 ; shapiro & teukolsky 1983 ; bildsten et al . 1997 ; frank et al . 2002 ; liu et al . so its evolution history may be different from that of the usual x - ray pulsars in hmxbs . the low field magnetar sgr 0418 + 5729 had been discovered with the dipole magnetic field less than @xmath49 and the strong field of @xmath50-@xmath51 which might be internal ( rea et al . 2010 , 2012 ; turolla et al . 2011 ) . therefore , we simulate the magnetic field and spin period evolution of a magnetar based on the model of accretion induced polar magnetic field decay of a ns ( zhang & kojima 2006 ) . during the simulation , we set the initial condition ( accretion rate , initial magnetic field and spin period ) and try to find a magnetar with the evolved values of spin period and period derivative to be the same as those of nustar j095551 + 6940.8 that is spinning up now . the spin period of the observed magnetar is usually about 10 second with the estimated age @xmath52 ( mereghetti 2008 ; bisnovatyi - kogan & ikhsanov 2014 ; mereghetti et al . 2015 ) . since the companion mass of nustar j095551 + 6940.8 is more than @xmath53 , its main sequence stage before evolving to feed the accretion disk would last for about one million years , during which the magnetar might spin down to about 100 seconds due to the period and lifetime relation of the magnetic dipole emission as @xmath54 ( shapiro & teukolsky 1983 ) . so the initial spin period ( @xmath55 ) of the accreting magnetar is chosen to be @xmath56 . the initial magnetic field is set to be the usual values of magnetar as @xmath57 . employing the evolution formula of b and p given in section 2 , the evolved path of b and p for nustar j095551 + 6940.8 are calculated and plotted in fig.[md1850 ] . in order to calculate the present magnetic field of nustar j095551 + 6940.8 , the spin period @xmath1 and period derivative @xmath2 are used in eq.([pdot1 ] ) . we find that the simulated model curve in the b - p diagram ( fig.[md1850 ] ) requires the accretion rate to be @xmath58 , and the obtained magnetic field is @xmath59 . for such a result , the b - p position of nustar j095551 + 6940.8 is included in its evolutionary path . from the b - p evolutionary history , the source might accrete @xmath60 within about @xmath61 , which feeds the field reduction and spins up the magnetar to the present state . the fastness parameter is @xmath62 , by which we get the parameter @xmath63 under the condition of the critical value @xmath42 being @xmath64 ( ghosh & lamb 1992 ) . the value of the fastness parameter is less than @xmath42 which means nustar j095551 + 6940.8 is now in the spin - up process . as the continuing decreasing of the spin period , @xmath65 increases to the critical spin period @xmath42 and the spin - up torque vanishes . , @xmath58 and the observed spin period @xmath1 . the dash line represents the simulated evolutionary path of nustar j095551 + 6940.8 with the initial field of @xmath57 and spin period of 100s , as depicted in the figure . the dash - dotted line and solid line are the spin up line with the accretion rate @xmath58 and the eddington accretion rate , respectively . with the data from atnf pulsar catalogue ( manchester et al . 2005 ) , we plot the b - p position of observed binary pulsars with dots and two isolated millisecond pulsars above the eddington spin up line with triangles . , we deduce that nustar j095551 + 6940.8 could be an evolving magnetar which is now spinning up . with the initial magnetic field and spin period ( @xmath66 and @xmath67 ) , the simulation shows that the current magnetic field of nustar j095551 + 6940.8 has been deducted to @xmath3 with @xmath68 in about @xmath7 , while the accretion material is about @xmath69 . the subsequent evolution of nustar j095551 + 6940.8 depends on its companion mass . in the binary system , the b - p evolution of a ns is mainly affected by the companion mass , which is associated with the accretion lifetime ( pan et al . the companion mass of nustar j095551 + 6940.8 is more than @xmath70 ( bachetti et al . 2014 ) , which will lead to two kinds of possible evolutionary descendants : ( a ) if the companion is a @xmath71 star , the disk accretion can last @xmath72 . the accretion material will be about @xmath73 which can make nustar j095551 + 6940.8 to be a millisecond pulsar with the polar magnetic field of @xmath74 and the spin period about @xmath75 , while the super - strong out - polar field is squeezed into the ns core region . there will be a millisecond pulsar and a heavy white dwarf left in the binary system after the accretion . \(b ) if the companion is a @xmath76 star , the ns may accrete about @xmath77 within @xmath78 . a double ns system will * exist * at the end of the accretion , where the recycled pulsar is similar to the hulse - taylor pulsar psr 1913 + 16 with the dipole magnetic field of @xmath79 and spin period of @xmath80 . the evolutionary track of nustar j095551 + 6940.8 in the b - p diagram above the eddington spin - up line has been simulated ( the dash line in fig.[md1850 ] ) . from the theoretical consideration , the distribution of binary pulsars should be below the eddington spin up line ( pan et al . 2015 ) , since the minimum spin period of binary pulsar is arisen by the equilibrium period line with the eddington accretion rate . however , the radio observations show that 9 binary pulsars ( out of 250 ones ) lie above the eddington spin - up line ( see fig.[md1850 ] in this paper , also see fig.4 in pan et al . we proposed that some of the 9 binary pulsars have experienced the similar evolutions with the super - eddintion accretion to that of nustar j095551 + 6940.8 . that is to say , some of them may be the descendants of magnetars after accretion . moreover , the evolved track of nustar j095551 + 6940.8 can also be used to explain the formation of the millisecond pulsar with the magnetic field as high as @xmath81 , e.g. , two isolated millisecond pulsars b1821 - 24a and b1820 - 30a marked in fig.[md1850 ] , whose companions are believed to be evaporated after formations ( hessels 2008 ; kluzniak 1988 ) . now , we may have a whole picture of the evolutionary history of accreting ns in the b - p diagram : ns in high mass x - ray binary system and ns in low mass x - ray binary system evolve to the recycled pulsars below the eddington spin - up line , while the descendants of ultra - luminous accreting magnetars stop at the b - p positions above the eddington spin - up line , from where they begin their further spin - down evolutions by electromagnetic emission . the birth rate of magnetar is about @xmath82 compared with that of normal nss ( lorimer 2008 ) , so the number of accreting magnetar systems shown as the ultra - luminous neutron star in hmxbs should be also @xmath83 of the normal ns in hmxbs . moreover , from the radio observations , only 9 of 250 binary pulsars lie above the eddington spin - up line , thus we infer that the ratio of accreting magnetars to ns in x - ray binaries should be at the scale of 4% . until now , about a hundred of x - ray ns binary systems are observed ( liu et al . therefore , it can be inferred that there would be a couple of ulxs to be accreting magnetars that are similar to nustar j095551 + 6940.8 . based on the model of accretion induced polar magnetic field decay of the ns ( zhang & kojima 2006 ) , the polar magnetic field would decay one half order of magnitude after accreting @xmath84 , e.g. , from @xmath8 to @xmath85 . at the same time , the out - polar field is slightly increased , which keeps almost the same as the initial value of @xmath8 . for the model acquires the magnetic flux conserving globally , the dilution of the polar field lines will be compensated by the expanding out - polar field lines . the strong out - polar field lines will be totally squeezed into the ns core after accreting the mass about 0.1 . the result shows that the polar field of nustar j095551 + 6940.8 decays about two magnitude orders , from @xmath86 to @xmath87 after accreting 0.005 . the out - polar field is almost unchanged compared with the initial value of @xmath86 . the difference of the magnetic field strength between the polar region and equator region will intensely arise the asymmetry of the magnetic field structure of the ns , which makes the accretion structure of the evolved magnetar in hmxb far from the spherical geometry . the eddington luminosity , which requires a condition of the spherical accretion ( shapiro & teukolsky 1983 ; frank et al . 2002 ) , will be destroyed by the exotic magnetic field structure . in addition , the super - strong multipole field structure induces the reduction of the scattering cross section of photon - electron , and in turn increases the eddington critical luminosity ( eksi et al 2015 ; laycock et al . 2015 ) . in the simulation , the accretion rate @xmath58 is obtained by consisting the observed spin period and derivative of nustar j095551 + 6940.8 with the calculation values . to settle the ultra luminosity difficulty , we propose two types of accretions for magnetar in hmxb , i.e. , the radial random accretion @xmath88 without bringing orbital angular momentum and the disk accretion @xmath89 that contributes to the ns spin - up , thus the total accretion rate can be divided into two components , = + . the components of the two type accretions should be correlated to the ns magnetic structure , disk thickness and accretion environment . for the usual ns in hmxb , the two accretion components are comparable @xmath90 , so the source luminosity is comparable with the disk accretion luminosity @xmath91 , which is no bigger than the eddington limit . however , for the magnetar as an ulx in hxmb , the local super - strong field destroys the spherical accretion geometry and breaks the limit of the eddington luminosity . the structure of the super strong multipole field dominates the radiation mechanism and leads the emission to be beamed . so , the beaming factor @xmath92 can be introduced for the accretion luminosity @xmath93 that is associated with the x - ray luminosity @xmath94 : @xmath95 , where @xmath96 ( feng and soria , 2011 ) . for ulx nustar j095551 + 6940.8 , when assuming the pulsed fraction to be 0.05 ( king and lasota , 2016 ) , then the accretion rate @xmath97 of the simulation would be reasonable to infer the radial random accretion rate as @xmath98 , which can account for the observed ultra luminosity ( @xmath99 ) of the source . therefore , the source accretes about @xmath100 through the polar channel as calculated from the simulation , which implies @xmath101 to be accreted through the non - polar random accretion . this work is supported by the strategic priority research program on space science , the chinese academy of sciences , grant no.xda04010300 , the national natural science foundation of china ( nsfc 11173034 ) , and the national basic research program of china ( 2012cb821800 ) .	ultra - luminous x - ray sources are usually believed to be black holes with mass about @xmath0 . however , the recent discovery of nustar j095551 + 6940.8 in m82 by bachetti et al . shows that it holds the spin period @xmath1 and period derivative @xmath2 , which provides a strong evidence that some ultra - luminous x - ray sources could be neutron stars . we obtain that the source may be an evolved magnetar according to our simulation by employing the model of accretion induced the polar magnetic field decay and standard spin - up torque of an accreting neutron star . the results show that nustar j095551 + 6940.8 is still in the spin - up process , and the polar magnetic field decays to about @xmath3 after accreting @xmath4 , while the strong magnetic field exists in the out - polar region , which could be responsible for the observed low field magnetar . the ultra luminosity of the source can be explained by the beaming effect and two kinds of accretion radial random accretion and disk accretion . since the birth rate of magnetars is about ten percent of the normal neutron stars , we guess that several ultra - luminous x - ray sources should share the similar properties to that of nustar j095551 + 6940.8 . [ firstpage ] accretion : accretion disks binaries : close x - rays : stars stars : magnetars
in this section , we define the notion of a rudin - keisler mapping and associate to each mapping a dedekind cut . then we state some results of blass in @xcite characterizing , under ch , the types of cuts obtainable for rudin - keisler mappings from a p - point or a weakly - ramsey ultrafilter on @xmath1 . in last part of this section , we provide an outline of the rest of the article and highlight its main results . we remind the reader of the rudin - keisler reducibility relation . if @xmath0 is an ultrafilter on the base set @xmath7 and @xmath8 is an ultrafilter on the base set @xmath9 , then we say that @xmath8 is _ rudin - keisler reducible to _ @xmath0 and write @xmath10 if there there exists a function @xmath11 such that @xmath12 , where @xmath13 a _ rudin - keisler mapping from @xmath0 to @xmath8 _ is a function @xmath11 such that @xmath12 . associated to each ultrafilter @xmath0 on @xmath7 is an equivalence relation on @xmath14 . if @xmath15 and @xmath16 are two functions from @xmath7 to @xmath1 then we say that _ @xmath15 and @xmath16 are equivalent mod @xmath0 _ if there exists @xmath17 such that @xmath18 . the _ ultrapower _ @xmath19 is the collection of all equivalence classes with respect to this equivalence . all operations and relations defined on @xmath1 have natural extensions making the ultrapower an elementary extension of the standard model of @xmath1 . in particular , @xmath19 forms a linearly ordered set . ( in this case , @xmath20\le[g]$ ] if and only if @xmath21 . ) recall that , a _ dedekind cut of a linearly ordered set _ is a partition @xmath22 of the linear order such that no element of @xmath23 precedes any element of @xmath24 . we follow the work of blass in @xcite and associate to each rudin - keisler mapping from @xmath0 on @xmath7 to @xmath8 on @xmath9 a dedekind cut in the ultrapower @xmath25 . a cut @xmath22 in the ultrapower is said to be _ proper _ if @xmath23 is nonempty and @xmath24 contains the equivalence class of each constant map . the cut given by taking @xmath24 to be the set of equivalence classes of constant maps is called _ the standard cut_. let @xmath26 be an ultrafilter on the base set @xmath7 and @xmath27 . for any @xmath28 , we define _ the cardinality function of @xmath29 relative to @xmath5 _ by @xmath30 the set of all equivalence classes of cardinality functions of sets in @xmath26 , and all larger elements of @xmath31 , constitute the upper part @xmath23 of a cut @xmath22 of @xmath32 , which we call _ the cut associated to @xmath5 and @xmath26_. ( if @xmath33 is infinite for some @xmath34 then @xmath35 and has no equivalence class , so it makes no contribution to @xmath23 ; it is entirely possible for @xmath23 to be empty . ) the cut associated to @xmath5 and @xmath26 is proper if and only if @xmath5 is finite - to - one but not one - to - one on any set in @xmath26 . additionally , the existence of a proper cut in @xmath36 implies that @xmath26 is non - principal . the next three theorems are due to blass and appear as theorems 1 , 2 and 4 in @xcite . the first theorem shows that certain dedekind cuts are not obtainable from rudin - keisler mappings . in the remaining theorems of this section , @xmath22 is assumed to be a proper cut . [ cut - thm1 ] @xmath22 is the cut associated to some map of some ultrafilter to @xmath37 if and only if @xmath24 is closed under addition . before stating the next two theorems we remind the reader of the definitions of some special types of ultrafilters on @xmath1 . let @xmath0 be an ultrafilter on @xmath1 . 1 . @xmath0 is a _ p - point ultrafilter _ , if for each sequence @xmath38 of members of @xmath0 there exists @xmath39 such that for each @xmath40 , @xmath41 . ( here @xmath42 denotes the almost - inclusion relation . ) @xmath0 is a _ weakly - ramsey ultrafilter _ , if for each partition of the two - element subsets of @xmath1 into three parts there exists an element of @xmath0 all of whose two - element subsets lie in two parts of the partition . 3 . @xmath0 is a _ ramsey ultrafilter _ , if for each partition of the two - element subsets of @xmath1 into two parts there exists an element of @xmath0 all of whose two - element subsets lie in exactly one part of the partition . it clear that every ramsey ultrafilter is weakly - ramsey . in theorem 5 of @xcite , blass has shown that every weakly - ramsey ultrafilter is a p - point . [ cut - thm2 ] assume ch . @xmath22 is the cut associated to some map of some p - point ultrafilter to @xmath26 if and only if @xmath26 is a p - point , @xmath24 is closed under addition , and every countable subset of @xmath23 has a lower bound in @xmath23 . [ cut - thm4 ] assume ch . @xmath22 is the cut associated to some map of some weakly ramsey ultrafilter to @xmath26 if and only if @xmath26 is ramsey , @xmath24 is closed under exponentiation , and every countable subset of @xmath23 has a lower bound in @xmath23 . blass has remarked in @xcite that many of the ultrafilter - theoretic concepts involved in the previous theorems have natural model - theoretic interpretations in terms of ultrapowers . following blass , we consider models that are elementary extensions of the standard model whose universe is @xmath1 and whose relations and functions are all the relations and functions on @xmath1 . suppose that @xmath43 is such a model . an element @xmath44 is said to _ generate @xmath43 over the submodel @xmath45 _ if and only if no proper submodel of @xmath43 includes @xmath46 . in @xcite , blass notes that , if @xmath11 and @xmath0 is an ultrafilter on @xmath7 , then @xmath47 is an elementary embedding . furthermore , @xmath15 is an isomorphism of ultrafilters if and only if @xmath48 is an isomorphism of models . the image of @xmath48 is cofinal in @xmath49 if and only if @xmath15 is finite - to - one on some set of @xmath0 . hence , @xmath0 is a p - point ( ramsey ultrafilter ) if and only if every nonstandard submodel of @xmath50 is cofinal in ( equal to ) @xmath19 . an element @xmath51 is in the upper half of the cut associated to @xmath15 and @xmath0 if and only if @xmath52 is greater than some generator of @xmath19 over @xmath53 . [ isomorphismremark ] let @xmath26 be an ultrafilter on the base set @xmath7 and @xmath27 . suppose that @xmath54 is a bijection and @xmath55 is an ultrafilter on @xmath56 such that @xmath57 . then the cut associated to @xmath58 and @xmath55 is exactly the cut associated to @xmath5 and @xmath0 . additionally , suppose that @xmath59 is a bijection . since @xmath60 is an isomorphism of models , it follows that if @xmath22 is the cut associated to @xmath61 and @xmath0 then @xmath62 is the cut associated to @xmath5 and @xmath0 . in particular , if @xmath22 is the standard cut in @xmath63 then @xmath64 is the standard cut in @xmath65 . the purpose of this paper is to prove analogous results for ultrafilters satisfying similar properties . in section @xmath66 we introduce the setting for the main results of this article , namely the topological ramsey space @xmath4 . in section @xmath67 we introduce the generalization of the notion ramsey ultrafilter we study in later sections . in section @xmath68 , we characterize the cuts obtainable , under ch , from the generalization of ramsey we defined in section @xmath67 . the next theorem which we prove in section @xmath68 is one of the two main results of this manuscript . [ mainresult1 ] assume ch . @xmath22 is the cut associated to some map of some ramsey for @xmath4 ultrafilter on @xmath69 $ ] to @xmath70 if and only if @xmath70 is selective and @xmath22 is the standard cut in @xmath71 . in section @xmath72 we introduce the basic definitions associated with the tukey theory of ultrafilters . applying theorems of dobrinen and todorcevic in @xcite about ultrafilters generated from generic subsets of @xmath73 we prove the second main result of this paper . [ mainresult2 ] suppose @xmath6 is a ramsey for @xmath4 ultrafilter on @xmath74 generated by a generic subset of @xmath73 . if @xmath22 is the cut associated to some map from some p - point ultrafilter in the tukey type of @xmath6 to some ultrafilter @xmath8 then @xmath8 is a p - point ultrafilter and @xmath22 is the standard cut . in section @xmath75 we show that the main results imply the existence of special ultrafilters . we then conclude with some questions about the types of cuts obtainable from ultrafilters defined from other similar topological ramsey spaces . the author would like to express his deepest gratitude to natasha dobrinen for valuable comments and suggestions that helped make this article and its proofs more readable . we begin this section with the definition given by dobrinen and todorcevic in @xcite of the triple @xmath76 . the construction of @xmath4 was motivated by the work of laflamme in @xcite which uses forcing to adjoin a weakly - ramsey ultrafilter satisfying complete combinatorics over @xmath77 ( see @xcite for a definition of complete combinatorics . ) for each @xmath40 , let @xmath78 @xmath79 ] @xmath4 consists of all subtrees of @xmath74 which are isomorphic to @xmath74 . more precisely , @xmath80 is a member of @xmath4 if and only if there exists a strictly increasing sequence @xmath81 of natural numbers such that the following conditions hold : @xmath82 for each tree @xmath83 and @xmath40 , let @xmath84 and @xmath85 . let @xmath86 and define @xmath87 by letting @xmath88 . for each @xmath40 , let @xmath89 and @xmath90 . for @xmath91 , @xmath92 if and only if there exists a strictly increasing sequence @xmath81 of natural numbers such that for each @xmath40 , @xmath93 . for @xmath91 , @xmath94 if and only there exists @xmath40 such that @xmath95 . the relation @xmath96 is called the _ almost - reduction relation on @xmath4_. if @xmath97 then we let @xmath98 $ ] denote the maximal nodes of @xmath24 , that is the length two sequences in @xmath24 . similarly , for @xmath99 and @xmath100 we let @xmath101 $ ] and @xmath102 $ ] denote the collection of maximal nodes of @xmath103 and @xmath5 respectively . if @xmath97 and @xmath40 then we let , @xmath104 and @xmath105 for each @xmath106 , we let @xmath107 be the least natural number @xmath108 such that @xmath109 . topological ramsey theory was initiated by ellentuck in @xcite with his proof of an infinite - dimensional analogue of the ramsey theorem called the ellentuck theorem . after similar examples where discovered , the abstract notion of a topological ramsey space was introduced in by carlson and simpson in @xcite . carlson and simpson also introduced a finite set of axioms sufficient for proving an abstract version of the ellentuck theorem . the text @xcite by todorcevic is now the standard reference for topological ramsey spaces and proves the abstract ellentuck theorem using four axioms . for each @xmath99 and @xmath110 we let @xmath111=\ { t\in \mathcal{r}_{1 } : ( \exists n ) \ s = r_{n } ( t ) \ \ & \ t \le s\}$ ] . the _ ellentuck topology on @xmath4 _ is the topology generated by @xmath112 : s\in \mathcal{ar } \ \ & \ s \in \mathcal{r}_{1}\}$ ] . a subset @xmath113 of @xmath114 is _ ramsey _ if for every @xmath115,$ ] there is a @xmath116 $ ] such that @xmath117 \subseteq { \mathcal x}$ ] or @xmath117\cap { \mathcal x } = \emptyset.$ ] a subset @xmath113 of @xmath118 is _ ramsey null _ if for every @xmath115,$ ] there is a @xmath116 $ ] such that @xmath111\cap { \mathcal x } = \emptyset.$ ] the triple @xmath119 forms a _ topological ramsey space_. that is , every subset of @xmath120 with the baire property is ramsey and every meager subset of @xmath4 is ramsey null . the next theorem , which we use in section @xmath75 , can be thought of as a finite version of the ramsey theorem for the space @xmath4 . the result is theorem 3.5 of mijares in @xcite applied to the topological ramsey space @xmath4 . [ finiteramseytheorem ] let @xmath121 with @xmath122 be given . then , there exists @xmath123 such that for each @xmath124 and each partition of @xmath125 into two parts there exists @xmath126 such that @xmath127 lies in exactly one one part of the partition . next we define a collection of natural projection maps related to the space @xmath4 . dobrinen and todorcevic in @xcite have used these projections to completely characterize the tukey ordering below weakly - ramsey ultrafilters obtained from forcing with @xmath4 using almost - reduction . each @xmath128 $ ] is a sequence of natural numbers of length two , which we denote by @xmath129 . let @xmath130 \rightarrow \omega$ ] be the map which sends @xmath131 $ ] to @xmath132 . for each @xmath133 , define @xmath134 to be the map that removes the right - most @xmath135 branches of a given element of @xmath136 . in this section we introduce the generalizations of ramsey ultrafilter that we study in this manuscript and prove some theorems needed to for the main results in sections @xmath137 and @xmath75 . first we remind the reader of the definition of selective ultrafilter on @xmath1 . let @xmath0 be an ultrafilter on @xmath1 . @xmath0 is _ selective _ , if for each sequence @xmath38 of members of @xmath0 there exists @xmath138 enumerated in increasing order such that for each @xmath40 , @xmath139 . by a result of kunen in @xcite , an ultrafilter is ramsey if and only if it is selective . for each topological ramsey space @xmath140 , mijares in @xcite has introduced the notions of ramsey for @xmath140 and selective for @xmath140 ultrafilters . if @xmath140 is taken to be the ellentuck space the notions are equivalent and reduce to ramsey and selective given above . below , we follow the presentation of ramsey for @xmath4 and selective for @xmath4 given by dobrinen and todorcevic in @xcite . let @xmath0 be an ultrafilter on @xmath69 $ ] . @xmath141 is _ generated by @xmath142 _ if and only if @xmath143 : s\in{\mathcal c}\}$ ] is cofinal in @xmath144 . an ultrafilter @xmath0 generated by @xmath145 is _ selective for @xmath114 _ , if for each decreasing sequence @xmath146 of members of @xmath147 , there is another @xmath148 such that for each @xmath40 , @xmath149 [ selectthm ] suppose that @xmath0 is an ultrafilter on the base set @xmath69 $ ] and generated by a subset @xmath150 of @xmath151 . the following statements are equivalent : 1 . @xmath0 is selective for @xmath4 . 2 . @xmath0 is a p - point and @xmath152 is selective . 3 . for each decreasing sequence @xmath146 of members of @xmath147 , there is another @xmath148 such that for each @xmath153 , @xmath154 first we show that @xmath155 suppose that @xmath0 is selective for @xmath4 . to show that @xmath0 is a p - point consider a sequence @xmath156 of elements of @xmath0 . since @xmath0 is generated by @xmath157 there exists a sequence @xmath158 of elements of @xmath157 such that for all @xmath40 , @xmath159 \subseteq a_{i}$ ] . since @xmath0 is selective for @xmath4 there exists @xmath160 such that for all @xmath40 , @xmath161 . as each @xmath162 is finite for each @xmath40 , @xmath98 \subseteq^ { * } [ s_{i}]\subseteq a_{i}$ ] . hence @xmath0 is a p - point ultrafilter on @xmath69 $ ] . to show that @xmath152 is selective consider a sequence @xmath163 of elements of @xmath152 . since @xmath0 is generated by @xmath157 there exists a sequence @xmath158 of elements of @xmath157 such that for all @xmath40 , @xmath164 \subseteq x_{i}$ ] . since @xmath0 is selective for @xmath4 there exists @xmath160 such that for all @xmath40 , @xmath161 . let @xmath165 be the increasing enumeration of @xmath166 $ ] . then for each @xmath40 , @xmath166 \setminus\{x_{0 } , x_{1 } , \dots , x_{i-1}\ } = p''([s]\setminus [ r_{i}(s ) ] ) \subseteq \pi''[s_{i } ] \subseteq x_{i}$ ] . @xmath152 is selective as @xmath167 is in @xmath152 . next we show that @xmath168 suppose that @xmath26 is a p - point and @xmath169 is selective . to show that @xmath0 satisfies condition @xmath170 , consider an arbitrary decreasing sequence @xmath146 of members of @xmath150 . there is an @xmath171 such that for each @xmath153 , @xmath98 \subseteq^ { * } [ s_{n}]$ ] . let @xmath81 be the strictly increasing sequence such that for all @xmath40 , @xmath172 . define @xmath173 recursively by letting , @xmath174 : \pi(x ) > k_{j}\ } \subseteq x_{k'_{i}}$. } \end{cases}\ ] ] now define @xmath175 by letting @xmath176 note that @xmath16 can not be constant mod @xmath169 as @xmath152 is non - principal . since @xmath169 is selective , it must be the case that there is a @xmath177 $ ] such that @xmath16 is increasing on @xmath178 . enumerate @xmath9 in increasing order , as @xmath179 . then either @xmath180 or @xmath181 is a member of @xmath169 . let @xmath182 denote which ever is in @xmath169 . by construction , we find that for each pair @xmath183 of natural numbers there exists @xmath184 such that @xmath185 . since @xmath186 and @xmath98 \in { \cal u}$ ] , there is a @xmath187 such that @xmath188\subseteq \pi^{-1}(z ) \cap [ s]$ ] . let @xmath189 be the strictly increasing sequence such that for all @xmath40 , @xmath190 . since @xmath191 \subseteq z$ ] , we find that for each @xmath153 , each @xmath192 and each @xmath193 $ ] , there exists @xmath194 such that @xmath195 by definition of the sequence @xmath196 , it follows that @xmath197 $ ] . on the other hand , @xmath198 implies that @xmath199 . so @xmath200 $ ] and @xmath201 . hence @xmath202 holds . next note that for each @xmath97 , @xmath203 . hence @xmath204 holds trivially . let @xmath0 be an ultrafilter on @xmath69 $ ] generated by @xmath205 @xmath0 is _ ramsey for @xmath114 _ , if for every @xmath40 and every partition of @xmath206 into two parts there exists @xmath207 such that @xmath208 lies in one part of the partition . [ equivalentramsey ] suppose that @xmath0 is an ultrafilter on @xmath69 $ ] generated by @xmath209 . @xmath0 is ramsey for @xmath4 if and only if @xmath0 is selective for @xmath4 and for each @xmath153 , @xmath210 forms an ultrafilter on @xmath211 . ( @xmath212 ) by a lemma 3.8 of mijares in @xcite , every ramsey for @xmath4 ultrafilter is selective for @xmath4 . one the other hand , if @xmath0 is ramsey for @xmath4 then for each @xmath153 , @xmath210 forms an ultrafilter on @xmath211 . ( @xmath213 ) suppose @xmath0 is selective for @xmath4 and for each @xmath153 , @xmath214 forms an ultrafilter on @xmath211 . since @xmath215 and @xmath216 forms an ultrafilter on @xmath217 , it follows that if @xmath218 then every partition of @xmath219 into two parts there exists @xmath207 such that @xmath208 lies in one part of the partition . we proceed by induction on @xmath220 to show that @xmath0 is ramsey for @xmath4 . the previous remarks show that the base case of the induction holds . let @xmath220 be a natural number and suppose that every partition of @xmath206 into two parts there exists @xmath207 such that @xmath208 lies in one part of the partition . let @xmath221 be a partition of @xmath222 . we show that there exists @xmath207 such that @xmath223 lies in one part of the partition . for each @xmath224 , let @xmath225 . let @xmath226 since @xmath227 forms an ultrafilter on @xmath228 it follows that @xmath229 is a partition of @xmath206 . by the inductive hypothesis , there exists @xmath207 and @xmath230 such that @xmath231 . we first consider the case when @xmath232 . in particular , for each @xmath233 , @xmath234 . for each @xmath153 , let @xmath235 . hence there exists a sequence @xmath236 of elements of @xmath157 such that @xmath158 and for each @xmath153 , @xmath237 . by theorem @xmath238 there exist @xmath207 such that for each @xmath153 , @xmath239 . suppose that @xmath240 . then @xmath241 and @xmath242 . if @xmath243 then @xmath244 . hence , @xmath245 . so in the case when @xmath232 , @xmath246 . by an identical argument in the case when @xmath247 , there exists @xmath248 such that @xmath246 . by induction we find that for each @xmath40 and each partition of @xmath219 into two parts there exists @xmath207 such that @xmath249 lies in one part of the partition . in other words , @xmath0 is a ramsey for @xmath4 ultrafilter on @xmath69 $ ] . in this section , assuming ch , we characterize the types of proper dedekind cuts that can be obtained from a map @xmath250 \rightarrow \omega$ ] and a ramsey for @xmath4 ultrafilter . in the following theorems , all cuts are assumed to be proper . [ cut1 ] let @xmath0 be a ramsey for @xmath4 ultrafilter on @xmath251 $ ] generated by @xmath252 and @xmath5 be a map from @xmath69 $ ] to @xmath1 . the cut associated to @xmath5 and @xmath0 is the standard cut in @xmath253 . suppose that @xmath22 is the cut associated to @xmath5 and @xmath26 . let @xmath254 be given and suppose that @xmath15 is not constant mod @xmath255 . for each @xmath256 let @xmath257 be the lexicographically increasing enumeration of @xmath101 $ ] . let @xmath258 be the partition of @xmath259 given by letting @xmath260 since @xmath4 is ramsey for @xmath4 there exists @xmath261 and @xmath262 such that @xmath263 . if @xmath264 then @xmath5 is bounded by @xmath265 mod @xmath0 where @xmath266 is the lexicographically least element of @xmath188 $ ] . so if @xmath264 then @xmath5 is constant mod @xmath0 and @xmath22 is not a proper cut . if @xmath247 then @xmath5 is one - to - one mod @xmath0 , so @xmath22 is not a proper cut . hence , if @xmath22 is a proper cut then @xmath267 . in particular , if @xmath268 and @xmath81 is the increasing enumeration of @xmath269 $ ] , then for each @xmath40 , @xmath270}(k_{i } ) = \| [ t ] \cap p^{-1}\{k_{i}\}\| = i$ ] . for each @xmath153 , let @xmath271 : f ( p(x ) ) \ge n\}.\ ] ] since @xmath26 is an ultrafilter on @xmath69 $ ] we find that for each @xmath153 , either @xmath272 or @xmath69 \setminus x_{n } \in { \cal u}$ ] . if there exists @xmath273 such that @xmath69 \setminus x_{n } \in { \cal u}$ ] then @xmath15 would be bounded by @xmath108 mod @xmath255 . however , this can not happen since we assumed that @xmath15 is not constant mod @xmath255 . hence for each @xmath153 , @xmath272 . note that @xmath274 is a decreasing sequence of members of @xmath26 . since @xmath26 is selective for @xmath4 there exists @xmath275 such that for each @xmath153 , @xmath276 \subseteq x_{n}$ ] . let @xmath277 be the increasing enumeration of @xmath278 $ ] . since @xmath279 , we find that for each @xmath153 and each @xmath280\cap p^{-1}\{k_{n}\}=[s''(n)]$ ] . so for each @xmath153 , @xmath281\cap p^{-1}\{k_{n}\}\subseteq [ s''\setminus r_{n}(s '' ) ] \subseteq x_{n}$ ] . hence , for each @xmath153 and each @xmath282\cap p^{-1}\{k_{n}\}$ ] , @xmath283\cap p^{-1}(k_{n})\|= c_{[s'']}(k_{n}).\ ] ] since @xmath284 \in p({\cal u})$ ] we find that @xmath20 \ge [ c_{[s'']}]$ ] . so @xmath20 \in l$ ] as @xmath275 . additionally , note that the cardinality function of any member of @xmath150 is not constant mod @xmath255 . therefore the cut @xmath22 is the standard cut in @xmath285 . in the next lemma and theorem , @xmath70 is an ultrafilter on @xmath1 , and @xmath22 is a proper cut in @xmath71 . [ cut2 ] assume ch . if @xmath70 is selective and @xmath22 is the standard cut in @xmath286 then there exists a ramsey for @xmath4 ultrafilter @xmath26 such that @xmath22 is the cut associated to @xmath26 and @xmath287 . let @xmath70 be selective and @xmath22 be the standard cut in @xmath288 . we will construct a @xmath96-decreasing sequence @xmath289 that generates a ramsey for @xmath4 ultrafilter @xmath0 on @xmath69 $ ] such that @xmath290 . let @xmath291 be an enumeration of the elements of @xmath292 . we impose the following requirements on the sequence , for each @xmath293 : @xmath294 where @xmath295 is the natural number such that @xmath296 since @xmath217 is in bijective correspondence with @xmath69 $ ] and @xmath289 is an almost - decreasing sequence the sequence will generate a p - point ultrafilter on @xmath69 $ ] . each @xmath297 we be _ large _ in the sense that @xmath298\in \mathcal{v}$ ] ; this suffices to guarantee that @xmath299 . by theorem @xmath238 this is enough to show that @xmath0 is selective for @xmath4 . the conditions @xmath300 for @xmath293 , guarantee that for each @xmath301 , @xmath302 is an ultrafilter on @xmath303 . by theorem @xmath304 this is enough to show that @xmath0 is a ramsey for @xmath4 ultrafilter . by previous lemma this guarantees that the cut associated to @xmath0 and @xmath5 in @xmath288 is standard . first let @xmath305 . suppose that @xmath306 and @xmath307 have been defined so that @xmath308 satisfies the @xmath309 condition . assume that for each @xmath310 , @xmath297 is large and these @xmath297 s form a @xmath96-decreasing chain . c consider the case when @xmath311 is a successor ordinal . let @xmath312 be the ordinal such that @xmath313 . let @xmath295 be the natural number such that @xmath314 . let @xmath315 be the increasing enumeration of @xmath316)\in{\cal v}$ ] . by theorem @xmath317 there exists a subsequence @xmath318 such that for each @xmath123 there exists a set @xmath319 such that either @xmath320 @xmath321 or ( @xmath322 @xmath323 . since @xmath70 is an ultrafilter there exists a strictly increasing sequence @xmath324 such that @xmath325 and either for all @xmath40 , @xmath326 or for all @xmath40 , @xmath327 . let @xmath328 so for each @xmath40 , @xmath329}(k'_{m_{i } } ) & = \|[b ] \cap \pi^{-1}(k'_{m_{i}})\| , \\ & = \|[a'_{m_{i}}]\|,\\ & \ge i.\end{aligned}\ ] ] consequently , the following construction of @xmath330 is well - defined : @xmath331 by construction @xmath332 and either @xmath333 or @xmath334 note that @xmath335 is large since @xmath336=\{k'_{m_{i } } : i<\omega\}\in \mathcal{v}$ ] . let @xmath337 , then @xmath338 is large satisfies the condition ( @xmath312 ) . next consider the case when @xmath311 is a limit ordinal . since ch holds the cofinality of @xmath311 is @xmath1 . let @xmath339 be a @xmath96-cofinal sequence in @xmath340 . so for each @xmath40 , there exists @xmath341 such that for all @xmath342 , @xmath343 . without loss of generality , we may assume that @xmath344 since @xmath70 is selective there exist @xmath345 , @xmath346 and for all @xmath40 , @xmath347 . therefore for all @xmath40 , @xmath348 . hence for all @xmath40 , @xmath349 . for each @xmath153 , let @xmath350 . let @xmath351 , then @xmath352 is large because @xmath353 = \{h_{1},h_{2 } , h_{3 } , \dots\ } \in{\cal v}$ ] . moreover , for each @xmath153 , @xmath354 . so @xmath355 is a @xmath96-decreasing sequence of large sets . this completes the inductive construction of @xmath356 and thus the proof of the lemma . lemma @xmath357 and @xmath238 show that necessity holds and lemma @xmath358 shows that sufficiency holds . if @xmath360 is a ramsey for @xmath4 ultrafilter generated by a generic subset of @xmath361 and @xmath40 , then we let @xmath362 denote the ultrafilter @xmath363 . in this section , we show that for any rudin - keisler mapping between any two p - points within the tukey type of @xmath6 the only cut obtainable is the standard cut . notice that for each @xmath133 and each ramsey for @xmath4 ultrafilter @xmath6 , @xmath364 additionally , notice that if @xmath365 then the cardinality function of @xmath366 with respect to @xmath359 is defined as @xmath367 a simple counting argument shows that for each @xmath110 , each @xmath153 and each @xmath368 , @xmath369 next we outline the basic definitions of the tukey theory of ultrafilters . the tukey types of ultrafilters have been studied by many authors ( see @xcite , @xcite , @xcite , @xcite , @xcite and @xcite ) . for a survey of the area see @xcite by dobrinen . suppose that @xmath0 and @xmath8 are ultrafilters . a function @xmath15 from @xmath0 to @xmath8 is _ tukey _ if every cofinal subset of @xmath370 is mapped by @xmath15 to a cofinal subset of @xmath371 . we say that @xmath8 is _ tukey reducible to @xmath0 _ and write @xmath372 if there exists a tukey map @xmath373 . if @xmath374 and @xmath375 then we write @xmath376 and say that @xmath0 and @xmath8 are _ tukey equivalent . _ the relation @xmath377 is an equivalence relation and @xmath378 is a partial order on its equivalence classes . the equivalence class are also called _ tukey types _ or _ tukey degrees_. the tukey reducibility relation is a generalization of the rudin - keisler reducibility relation . if @xmath379 then the map sending @xmath380 to @xmath381 is tukey . so if @xmath382 , then @xmath383 . this leads to the following question : for a given ultrafilter @xmath0 , what is the structure of the rudin - keisler ordering within the tukey type of @xmath0 ? dobrinen and todorcevic in @xcite , have given an answer to this question if @xmath0 is a ramsey for @xmath4 ultrafilter on @xmath69 $ ] generated by a generic subset of @xmath73 . furthermore , dobrinen and todorcevic describe the rudin - keisler structure of the p - points within the tukey type of a ramsey for @xmath4 ultrafilter generated by a generic subset of @xmath73 . in particular , the tukey type of such an ultrafilter @xmath6 consists of a strictly increasing chain of rapid p - points of order type @xmath1 : @xmath384 where @xmath385 for @xmath40 . the next proof is the main result of this section . suppose that @xmath6 is a ramsey for @xmath4 ultrafilter on @xmath69 $ ] generated by a generic subset of @xmath73 . let @xmath386 be a p - point ultafilter in the tukey - type of @xmath6 and @xmath16 be a map from the base of @xmath386 to @xmath1 . let @xmath22 be the cut associated to @xmath386 and @xmath16 . by theorem 5.10 and example 5.17 from @xcite there exists @xmath387 such that @xmath388 , @xmath389 and @xmath390 . notice that since @xmath22 is not proper we have @xmath183 . by remark [ isomorphismremark ] , it is enough to prove the this theorem in the case when @xmath391 and @xmath392 . in particular , we may assume without loss of generality that @xmath393 . since @xmath6 is generated by a generic subset of @xmath361 , we find that theorem 4.25 and proposition 5.8 in @xcite imply that there exists @xmath394 and @xmath395 such that for all @xmath396 , @xmath397 if and only if @xmath398 . so if @xmath399 , then for each @xmath153 and each @xmath400 , @xmath401 let @xmath402 be given and suppose that @xmath15 is not constant mod @xmath403 . for each @xmath153 , let @xmath404 since @xmath363 is an ultrafilter on @xmath228 we find that for each @xmath153 , either @xmath405 or @xmath406 . if there exists @xmath273 such that @xmath407 then @xmath15 would be bounded by @xmath108 mod @xmath408 . however , this can not happen since we assumed that @xmath15 is not constant mod @xmath408 . hence for each @xmath153 , @xmath405 . note that @xmath274 is a decreasing sequence of members of @xmath363 . since @xmath6 is selective for @xmath4 there exists @xmath409 such that for each @xmath153 , @xmath410 . for each @xmath153 and each @xmath411 , @xmath412 . so ( @xmath413 ) implies that for each @xmath153 and each @xmath414 , @xmath415 so @xmath416 and @xmath20 \ge [ c_{\mathcal{r}_{1}(i)\upharpoonright(t\cap s')}]$ ] . so @xmath20 \in l$ ] as @xmath417 . additionally , note that the cardinality function of any member of @xmath418 is not constant mod @xmath419 . therefore the cut @xmath22 is the standard cut in @xmath420 . in this section we use the two main results to prove that , under ch , certain special ultrafilters exists . additionally , we ask some questions about the types of cuts that can be obtained from similarly defined topological ramsey spaces . the two main results of this paper only associate the standard cut to a given ultrafilter mapping . however the results of blass from @xcite only require the lower half of the cuts to be closed under certain operations . these differences allow us to show that , under ch , certain special ultrafilters exists which are not ramsey for @xmath4 . recall that under ch , p - point ultrafilters exist . let @xmath0 be a p - point ultrafilter on @xmath1 , @xmath20 $ ] be a nonstandard element of @xmath421 and @xmath422 \in \omega^{\omega}/\mathcal{u } : [ g]\le n\cdot[f]\}.\ ] ] notice that @xmath24 is closed under addition and contains the standard part of @xmath421 . let @xmath23 be the complement of @xmath24 . the condition that @xmath23 is nonempty and every countable subset of @xmath23 has a lower bound in l , will be automatically satisfied if @xmath24 has a countable cofinal subset , because @xmath421 is countably saturated ( see @xcite ) . so @xmath22 is a proper cut such that @xmath24 is closed under addition and every countable subset of @xmath23 has a lower bound in @xmath23 . by theorem @xmath423 there is a p - point ultrafilter @xmath8 and a rudin - keisler map @xmath5 such that @xmath424 and @xmath22 is the cut associated to @xmath5 and @xmath8 . since @xmath22 is not the standard cut theorem @xmath425 implies that @xmath8 is not ramsey for @xmath4 . since @xmath22 is not closed under exponentiation theorem @xmath426 implies that @xmath8 is not weakly - ramsey . recall that under ch , selective ultrafilters exist . let @xmath0 be a selective ultrafilter on @xmath1 , @xmath20 $ ] be a nonstandard element of @xmath421 and @xmath433 \in \omega^{\omega}/\mathcal{u } : [ g]\le { } ^{n}[f]\}.\ ] ] if @xmath434 $ ] and @xmath435 $ ] are in @xmath24 then there exists natural numbers @xmath108 and @xmath431 such that @xmath434\le { } ^{n}[f]$ ] and @xmath435\le { } ^{m}[f]$ ] . so @xmath434^{[h ] } \le { } ^{n+m}[f]$ ] . therefore @xmath24 is closed under exponentiation . additionally @xmath24 contains the standard part of @xmath421 . let @xmath23 be the complement of @xmath24 . the condition that @xmath23 is nonempty and every countable subset of @xmath23 has a lower bound in l , will be automatically satisfied if @xmath24 has a countable cofinal subset , because @xmath421 is countably saturated ( see @xcite ) . so @xmath22 is a proper cut such that @xmath24 is closed under exponentiation and every countable subset of @xmath23 has a lower bound in @xmath23 . by theorem @xmath426 there is a weakly ramsey ultrafilter @xmath8 and a rudin - keisler map @xmath5 such that @xmath424 and @xmath22 is the cut associated to @xmath5 and @xmath8 . since @xmath22 is not the standard cut theorem @xmath425 implies that @xmath8 is not ramsey for @xmath4 . in @xcite trujillo has shown that assuming ch there exists a selective for @xmath4 ultrafilter which is not ramsey for @xmath4 . using a similar proof to that in theorem @xmath425 it is possible to characterize the cuts obtainable from a selective for @xmath4 ultrafilter and the map @xmath287 ; they are exactly the standard cuts . however , it is unclear if there is a selective for @xmath4 ultrafilter @xmath0 and a rudin - keisler map @xmath5 such that the cut associated to @xmath5 and @xmath0 is not standard . this motivates the following question : if it is shown that the only cuts obtainable are standard then this method will not distinguish between the two notions . however if there exists a selective for @xmath4 ultrafilter and a rudin - keisler map @xmath5 such that @xmath22 is not standard then @xmath0 will not be ramsey for @xmath4 and the two notions will be distinguished . dobrinen and todorcevic in @xcite have defined generalizations of the space @xmath4 for @xmath436 . the spaces are built from trees @xmath437 in much the same way that @xmath4 is built from @xmath74 . trujillo @xcite has shown that for each positive integer @xmath108 , under ch , there are selective for @xmath438 ultrafilters on @xmath439 $ ] which are not ramsey for @xmath438 . however it is still unknown if , under ch , there are selective for @xmath440 ultrafilters on @xmath441 $ ] which are not ramsey for @xmath440 for @xmath442 . the methods used in @xcite fail for infinite @xmath312 as the trees @xmath437 have infinite height . assume that @xmath443 . can the notions of selective for @xmath440 and ramsey for @xmath440 be distinguished by characterizing the dedekind cuts obtainable from selective and ramsey for @xmath440 ultrafilters on @xmath441 $ ] ? assume ch holds and @xmath6 is a ramsey for @xmath4 ultrafilter on @xmath69 $ ] generated by a generic subset of @xmath73 . there exists a weakly ramsey ultrafilter @xmath8 on @xmath69 $ ] such that @xmath445 notice that @xmath446 is a selective ultrafilter by theorem [ selectthm ] . let @xmath20 $ ] be a nonstandard element of @xmath447 and @xmath448 \in \omega^{\omega}/\pi(\mathcal{u}_{1 } ) : [ g]\le { } ^{n}[f]\}.\ ] ] @xmath24 is closed under exponentiation . additionally @xmath24 contain the standard part of @xmath449 . let @xmath23 be the complement of @xmath24 . the condition that @xmath23 is nonempty and every countable subset of @xmath23 has a lower bound in @xmath23 , is automatically satisfied if @xmath24 has a countable cofinal subset , because @xmath447 is countably saturated ( see @xcite ) . so @xmath22 is a proper cut such that @xmath24 is closed under exponentiation and every countable subset of @xmath23 has a lower bound in @xmath23 . by theorem @xmath426 there is a weakly ramsey ultrafilter @xmath8 and a rudin - keisler map @xmath5 such that @xmath450 and @xmath22 is the cut associated to @xmath5 and @xmath8 . since @xmath22 is not the standard cut theorem @xmath444 implies that @xmath8 is not tukey - reducible to @xmath6 . assume ch holds and @xmath6 is a ramsey for @xmath4 ultrafilter on @xmath69 $ ] generated by a generic subset of @xmath73 . there exists a p - point ultrafilter @xmath55 on @xmath69 $ ] which is not weakly ramsey such that @xmath451 . notice that @xmath452 is a p - point ultrafilter by theorem [ selectthm ] . let @xmath20 $ ] be a nonstandard element of @xmath453 and @xmath433 \in \omega^{\omega}/\pi_{t(1)}(\mathcal{u}_{1 } ) : [ g]\le n\cdot[f]\ } \mbox { and } \ ] ] @xmath24 is under addition . additionally @xmath24 contain the standard part of @xmath454 . let @xmath23 be the complement of @xmath24 . the condition that @xmath23 is nonempty and every countable subset of @xmath23 has a lower bound in @xmath23 , is automatically satisfied if @xmath24 has a countable cofinal subset , because @xmath453 is countably saturated ( see @xcite ) . so @xmath22 is a proper cut such that @xmath24 is closed under addition and every countable subset of @xmath23 has a lower bound in @xmath23 . by theorem @xmath423 there is a p - point ultrafilter @xmath55 and a rudin - keisler map @xmath5 such that @xmath455 and @xmath22 is the cut associated to @xmath5 and @xmath55 . since @xmath22 is not the standard cut theorem @xmath444 implies that @xmath55 is not tukey - reducible to @xmath6 . on the other hand , @xmath452 is tukey equivalent to @xmath6 . therefore @xmath456 . notice that the previous corollary shows that the converse of theorem @xmath444 , under ch , does not hold . in particular there is a p - point @xmath55 such that there is no p - point in the tukey - type of @xmath6 which gives the standard cut in @xmath457 since otherwise @xmath55 would be tukey reducible to @xmath6 . the previous remark shows that at the very least one needs to assume that the not only is @xmath8 a p - point but it is also tukey - reducible to @xmath6 . we conclude this paper with a more general question of dobrinen from @xcite which has motivated much of the work in this paper . n. dobrinen and s. todorcevic , a new class of ramsey - classification theorems and their applications in the tukey theory of ultrafilters part 1 , transactions of the american mathematical society ( to appear ) . n. dobrinen and s. todorcevic , a new class of ramsey - classification theorems and their applications in the tukey theory of ultrafilters part 2 , transactions of the american mathematical society ( to appear ) .	associated to each ultrafilter @xmath0 on @xmath1 and each map @xmath2 is a dedekind cut in the ultrapower @xmath3 . blass has characterized , under ch , the cuts obtainable when @xmath0 is taken to be either a p - point ultrafilter , a weakly - ramsey ultrafilter or a ramsey ultrafilter . dobrinen and todorcevic have introduced the topological ramsey space @xmath4 . associated to the space @xmath4 is a notion of ramsey ultrafilter for @xmath4 generalizing the familiar notion of ramsey ultrafilter on @xmath1 . we characterize , under ch , the cuts obtainable when @xmath0 is taken to be a ramsey for @xmath4 ultrafilter and @xmath5 is taken to be any map . in particular , we show that the only cut obtainable is the standard cut , whose lower half consists of the collection of equivalence classes of constants maps . forcing with @xmath4 using almost - reduction adjoins an ultrafilter which is ramsey for @xmath4 . for such ultrafilters @xmath6 , dobrinen and todorcevic have shown that the rudin - keisler types of the p - points within the tukey type of @xmath6 consists of a strictly increasing chain of rapid p - points of order type @xmath1 . we show that for any rudin - keisler mapping between any two p - points within the tukey type of @xmath6 the only cut obtainable is the standard cut . these results imply existence theorems for special kinds of ultrafilters .
no - slip ( or rough ) billiards are a type of billiard dynamical system based upon a model in which linear and angular momentum of a hard spherical particle , moving in an n - dimensional euclidean domain , may be exchanged on collisions at the boundary with total energy conserved , as indicated in the diagram of figure [ fig : intro ] . gutkin and broomhead introduced no - slip collisions in two dimensions as a new model of a gas , in which `` the spheres interact with each other and with the container walls '' without slipping , in such a way that the impact `` conserves the total energy of the system , but mixes the tangential velocity components with the angular velocities of colliding spheres . '' @xcite they demonstrate that in two dimensions energy may be conserved only by the specular collisions of the traditional gas model , well - known from standard billiards , or the unique conservative alternative of no - slip collisions . very little ( none to our knowledge ) seems to have been done regarding the dynamics of no - slip billiards since @xcite . in particular , the broader implications of the main observation of that paper , concerning a striking boundedness property that they obtain in the very special case of an infinite strip billiard table , were not pursued . the main goal of this paper is to reintroduce this topic and begin a more systematic study of the dynamics and ergodic theory of no - slip billiards . in addition to extending the boundedness result of @xcite , we indicate by a combination of analytical and numerical observations that this property should hold in much greater generality than shown by gutkin and broomhead and that it may serve as the basis of a very general non - ergodicity result for this class of billiard systems . we have intentionally kept the technical level of this paper fairly low . a more systematic development will be taken up elsewhere . no - slip billiards in dimension @xmath0 arise naturally in @xcite , in which we obtain an exhaustive description of collisions of rigid bodies in @xmath1 satisfying natural physical requirements . we summarize here the more relevant facts from that paper . let @xmath2 denote the configuration manifold of two rigid bodies in @xmath1 having smooth boundaries . we endow @xmath2 with the riemannian metric whose quadratic form gives the system s kinetic energy function and assume that points @xmath3 represent configurations in which the bodies have a single contact point . at a boundary point @xmath4 of @xmath2 where @xmath5 is differentiable , we define a _ collision map _ as a linear map @xmath6 that sends vectors pointing out of @xmath2 into vectors pointing inward . we say that @xmath7 is a _ strict collision map _ at @xmath4 if 1 . energy is conserved . that is , @xmath7 is an orthogonal linear map ; 2 . linear and angular momentum are conserved in the unconstrained motion . ( this property , expressed in terms of invariance of a momentum map , may amount to no restriction at all when one of the bodies , representing the billiard table , is assumed fixed in place and the subgroup of euclidean symmetries of the whole system is trivial ) ; 3 . time reversibility . this amounts to @xmath7 being a linear involution ; 4 . impulse forces at collision are applied only at the single point of contact . ( see @xcite for an elaboration and geometric interpretation of this property , which may be regarded as a generalized momentum conservation law that is generally non - trivial and highly restrictive . ) the following result is shown in @xcite . at each boundary point of the configuration manifold of the system of two rigid bodies , assuming the boundary is differentiable at that point , the set of all strict collision maps can be expressed as the disjoint union of orthogonal grassmannian manifolds @xmath8 , @xmath9 , of all @xmath10 dimensional planes in @xmath11 . in particular , when @xmath12 , the set of strict collision maps is a two - point set consisting of the specular reflection and the no - slip collision . the geometric description of no - slip collisions is explained in figure [ fig : conemap ] . in addition to the standard reflection it involves a rotation by the special angle @xmath13 , which is the same for all table shapes . a billiard system is generally understood to consist of two rigid bodies , one of which is fixed in place and called the _ billiard table_. the focus is then on the motion of the second , referred to as the _ billiard particle_. for the system to be fully specified it is necessary to impose boundary conditions . from our perspective , this amounts to assigning a collision map @xmath14 to each boundary configuration @xmath3 from among those in the moduli of collision maps described by the above theorem . we offer a variety of examples in @xcite . on a general newtonian mechanical system without boundary the so - called liouville volume measure , obtained from the canonical ( symplectic ) form on the cotangent bundle of the configuration manifold @xmath2 , is invariant under the hamiltonian flow . it is also well - known that the liouville measure remains invariant for systems with boundary when the collision maps correspond to specular reflection , but it is not clear whether this still holds for more general boundary conditions . we show in @xcite that a sufficient condition for the invariance of the liouville measure is that the field of collision maps @xmath15 be parallel with respect to the levi - civita connection on @xmath5 associated to the kinetic energy riemannian metric . this condition is satisfied for no - slip billiards on the plane . the canonical billiard measure on the boundary of the phase space of a planar no - slip billiard is invariant under the billiard map . having established that no - slip billiard maps preserve the canonical billiard measure , we can venture into ergodic theory . we may ask , in particular , whether the known constructions giving rise to ergodic billiards of the standard kind also apply to the no - slip kind . the main result of @xcite about boundedness of trajectories in infinite strip billiard tables , suitably generalized , suggests that ergodicity for no - slip billiards is very difficult to come by . a feature of no - slip billiards that may preclude ergodicity is an _ axis of periodicity _ , a certain type of period two trajectory that will be seen below to be very ubiquitous . in contrast to periodic points that may exist in standard ergodic billiards , these axes appear to occur as the center of small invariant regions . the axes occur in a large class of billiards including polygons , as will be seen . furthermore , phase portraits suggest the dynamics persist under small smooth perturbations . indications of the invariant regions precluding ergodicity appear in curvilinear polygons in numerical experiments shown here in section [ ex ] . in this paper we provide evidence indicating that * gutkin and broomhead s observation that trajectories in infinite strip billiards are bounded holds in much greater generality ; * this boundedness property is a strong obstacle preventing ergodicity . this behavior demonstrates one marked difference between the dynamics of no - slip billiards and those of standard billiards . for example , generic polygons ( having angles of irrational multiples of @xmath16 ) are ergodic @xcite , but no - slip billiards on polygons appear never to be ergodic . note that no - slip billiards in dimension @xmath0 correspond to configuration spaces @xmath2 of dimension @xmath17 , whose points are parametrized by the center of the disk - particle and its angle of rotation . the boundary of @xmath2 is then a two - dimensional , piecewise smooth manifold , and the phase space is a @xmath18-dimensional manifold whose points are pairs @xmath19 in which @xmath3 and @xmath20 may be taken to lie in a hemisphere ( whose radius is determined by the conserved kinetic energy ) about the normal vector to @xmath5 at @xmath4 pointing into @xmath2 . in this paper we identify this hemisphere with the disk of same radius in the tangent space to @xmath5 at @xmath4 under the natural orthogonal projection . thus the phase space of our billiards , at least in the case of bounded ( and connected ) billiard tables , will typically be homeomorphic to the cartesian product of a @xmath0-torus and a disk . a simplification results by noting that the angle of rotation ( but certainly not the speed of rotation ) is typically immaterial . formally this means that we may for many purposes consider the _ reduced phase space _ , defined as the quotient of the @xmath18-dimensional phase space by the natural action of the rotation group @xmath21 . the resulting @xmath17-dimensional solid torus is indicated in the diagram of figure [ fig:3d ] . in many cases key feature may be gleaned from the projection to the velocity components , which we refer to as the velocity phase portrait . see figure [ fig : hexa ] for an example . in section [ perbd ] we consider several examples of no - slip billiards in dimension two , beginning by revisiting the `` no - slip billiard on a strip '' from @xcite . we give an alternate proof of boundedness which turns out to extend to dimension three and show that there are no non - trivial periodic orbits . turning to the open infinite wedge , we show that all orbits not eventually escaping are bounded and that the wedge angle may be chosen for any @xmath22 such that _ all _ bounded orbits are periodic with period @xmath23 . the kind of localization of orbits near what we have called _ axis of periodicity _ noted here for wedge billiards is observed for very general billiard shapes containing flat points , and perhaps many others . this suggests that a detailed analysis of wedge billiards and their smooth perturbations is fundamental for understanding the dynamics of general no - slip billiards . section [ ex ] comprises examples of polygons , circles , and other simple shapes . in standard billiards , two well - known techniques for generating chaotic dynamics are the defocusing and dispersing mechanisms . the stadium of bunimovich @xcite , the most well - known example of chaotic focusing , consists of two half circles connected by flat segments . the corresponding no - slip billiard , however , fails to be ergodic , as any arbitrarily small flat strip will have bounded orbits for all trajectories in a neighborhood of positive measure near the vertical trajectory ( figure [ fig : stad - disp ] , left ) . while apparently no such elementary argument can be used to show that the dispersing examples of standard billiards are not ergodic in the no - slip case , some insight is gleaned from the velocity phase portraits . sinai billiards with boundaries which are entirely dispersing generate strong chaotic behavior in standard billiards @xcite . while numerical analysis of phase portraits of corresponding no - slip examples tends to show potentially chaotic behavior for most orbits , exceptional orbits forming closed curves suggest the possibility of non - ergodic behavior near periodic points . figure [ fig : stad - disp ] ( right ) shows a bounded component for one of the simplest dispersing cases , a disk on a flat torus . the basic properties obtained from the strip and wedge examples of this section will be important tools in general for understanding no - slip dynamics . if the unconstrained model for two rigid bodies in @xcite is adapted to one fixed body and a non - intersecting but otherwise unconstrained disk of uniform mass distribution , then the post - collision rotational and linear velocity @xmath24 is the function of the pre - collision velocities @xmath25 given by @xmath26 r_{\frac{\pi}{2}}\nu - ( v^-\cdot \nu ) \nu , \end{aligned}\ ] ] where @xmath27 is the matrix of counterclockwise rotation by angle @xmath28 and @xmath29 is the inward normal vector at the point of contact . for dimension two we may use coordinates @xmath30 , with planar position @xmath31 and normalized rotational position @xmath32 , where @xmath33 is the radius of the disk . note that the rotation adds a third dimension to the two spacial dimensions . sometimes , if there is no loss of information , the given figures will be the planar projection . also , without any essential alteration of the theory we may use the orbit of the center of mass , with boundaries accordingly adjusted by distance @xmath33 . if the upward normal is in the direction @xmath34 , then by [ update2 ] the pre- and post - collision velocities are related by @xmath35 where @xmath36 and the transformation matrix @xmath37 is @xmath38 \frac{2\sqrt{2}}{3 } & \frac{1}{3 } & 0\\[6pt ] 0 & 0 & -1 \end{array}\right)}.\ ] ] as an application of no - slip collisions , broomhead and gutkin @xcite considered a disk between two parallel horizontal boundaries . we normalize the separation to give unit time between collisions . using complex coordinates for the phase space and finding a series for the horizontal displacement after @xmath39 collisions , it was noted in @xcite that the orbit is always bounded , except for the case of zero vertical velocity . the following strengthening of this main result in @xcite holds for unit velocity and fixed unit separation , letting @xmath40 be the parallel direction , so that the time between collisions will be @xmath41 , constant as @xmath42 only changes sign at collisions . [ pr : horbd ] orbits of a two dimensional system of no - slip collisions with parallel boundaries have horizontal displacement no more than @xmath43 . viewing the trajectories in three dimensions , we will show that for each boundary plane the set of collision points is contained in a line ( figure [ fig:3dview2dstrip ] ) . consider any two successive collision points @xmath44 and @xmath45 on the upper wall , choosing coordinates with the origin at the intermediate collision on the lower wall and letting @xmath46 be the velocity after the collision at @xmath44 . then @xmath47 and [ eq : rt ] gives @xmath48 and @xmath49 the slope in the @xmath50 ( horizontal - rotational ) plane is @xmath51 independent of the initial conditions . therefore , all of the collision points on the upper wall lie on a line of slope @xmath52 , and a similar argument on the lower wall shows that all collisions occur on a line of slope @xmath53 . as no velocity is exchanged between the horizontal - rotational component and the vertical component , the length of segments representing the projection of the orbits between successive collisions will be @xmath54 . geometrically , the orbits are contained in an astroid ( figure [ fig : parastr ] ) with the given bound , the maximal horizontal displacement being achieved when the ( projection of ) the trajectories is perpendicular to the contact line . the no - slip strip can never have more than three consecutive trajectory segments in the same horizontal direction : considering the trajectory segments in the upper , left , lower , and right quadrants of the @xmath55 plane as delineated by the contact lines ( as in figure [ fig : parastr ] ) , no two consecutive trajectories both lie in the upper ( or lower ) quadrants . at most two consecutive trajectories may lie in the left ( or right ) quadrants . in the case of three spacial dimensions , the argument of proposition [ pr : horbd ] may be generalized to two hyperplanes of the phase space in both of which the projected velocity is constant . the contact points occur on lines of slope @xmath56 and similar bounds can be obtained in terms of the constant velocity in ( or excluding ) the hyperplanes , yielding the following result . [ pr:3dbdd ] orbits of three dimensional no - slip collisions between parallel planes are bounded . next we consider the question of periodic orbits in the strip case . if the velocity is entirely vertical there will be a period two orbit , where the disk s velocity is vertical with no rotation . this is a simple example of an axis of periodicity . it is useful to consider the composition of transformations on the velocity through a complete cycle of two collisions . in the no - slip strip case , @xmath57 where @xmath58 is given by @xmath59 \frac{4}{9 } \ , \sqrt{2 } & -\frac{7}{9 } & 0 \\[6pt ] 0 & 0 & 1 \end{array}\right)}\ ] ] with @xmath60 as in equation [ eq : rt ] and @xmath61 the appropriate frame adjustment . the proof that there are no higher order periodic orbits will consider iterations of @xmath62 and use the following lemma of niven @xcite . [ lm : niv ] if @xmath28 is a rational multiple of @xmath16 , the only possible rational values of @xmath63 are @xmath64 and @xmath65 . the no - slip billiard system on the infinite strip has no periodic orbits besides the trivial period - two orbit . [ pr : parnonper ] notice that @xmath66 only if the velocity is vertical , giving the trivial period two orbit . suppose @xmath67 or @xmath68 . a necessary condition for higher order periodicity is then that @xmath69 for some @xmath70 . @xmath71 is a rotation matrix and will have finite order precisely when it is a rotation through an angle which is a rational multiple of @xmath16 . but here the angle is @xmath72 is not a rational multiple of @xmath16 by lemma [ lm : niv ] . it follows from proposition [ pr : parnonper ] that the bounds of proposition [ pr : horbd ] are optimal . if @xmath73 then the bound is @xmath74 which holds trivially for the simply periodic case otherwise , the orbit is not periodic and will come arbitrarily close to the geometric limit on the astroid by invariance of measure . notice that for a certain initial velocity it is possible to have a ray - like orbit which achieves the limit on one side and then reverses direction . however , by the non - existence of higher order periodic orbits it can not also achieve the limit on the other side . the matrix @xmath62 and corresponding rotation angle @xmath75 in the proof of proposition [ pr : parnonper ] will generalize . @xmath62 has an eigenvalue of @xmath76 corresponding to the eigenvector @xmath77 , the axis of periodicity . because the axis corresponds to a coordinate axis , @xmath62 is easily identified with an @xmath78 , which suggests a natural connection to the @xcite proof in terms of rotation in the complex plane . for the wedge example , though , the axis will have a component in the rotational direction ; however , the axis and rotation angle will still be fundamental . in this section we consider the case of no - slip billiards on an open wedge of angle @xmath28 , letting @xmath34 be the direction of the outward bisector , @xmath40 the perpendicular spatial direction , and @xmath79 the rotational position . an informal numerical survey of wedge systems reveals that for most angles @xmath80 the behavior is similar to the bounded , nonperiodic dynamics of the no - slip strip ( figure [ fig : wedgeexamples ] ) . however , certain isolated @xmath28 have periodic orbits which are stable in the sense of persisting when the initial conditions are altered . these observations will be made rigorous in proposition [ pr : bddwedge ] and corollary [ cor : anyper ] respectively . but first we show the elementary but useful fact that all wedges have an axis of periodicity . ( center left ) and @xmath81 ( center right ) give period four and period ten orbits which persist when initial velocities are changed.,height=105 ] [ pr : wedax ] an orbit of an angle @xmath28 wedge billiard with initial velocity @xmath82 will have period two if and only if @xmath83 and @xmath84 . informally , figure [ fg : wsp ] gives an intuition why such a direction must exist for the case @xmath85 . formally , consider @xmath86 giving the transformation of the velocity after two collisions for a wedge of angle @xmath28 . adjusting the normal for the angle , we have @xmath87 , where @xmath88 is the frame adjusted rotation . then @xmath89 is an eigenvector of @xmath62 corresponding to the eigenvalue @xmath76 . since the outward component @xmath42 is zero , the position as well as the velocity is unchanged . the no - slip strip case might be thought of as the limit as @xmath28 approaches zero . ( note , that the two spatial axes are reversed here : [ eq : spw ] would imply @xmath90 , giving the @xmath34 axis as the axis of periodicity . ) turning to the question of higher periodicity for no - slip wedges , it is again useful to approach the question in terms of the velocity transformation @xmath91 . in order to have an orbit of period @xmath92 it is necessary for the velocity to return to the initial velocity after @xmath39 iterations , hence a necessary condition is @xmath93 . as @xmath91 simply rotates the velocity vector around the axis of periodicity , this condition will depend on the rotation angle @xmath75 being a rational multiple of @xmath16 . notice that @xmath94 is orthogonal to the @xmath20 given in proposition [ pr : wedax ] giving the axis of periodicity . then @xmath95 . a direct calculation gives the following relation between @xmath28 and @xmath75 . [ pr : wedgeangle ] the wedge angle @xmath28 and the rotational angle @xmath75 of @xmath86 , the corresponding transformation of the velocity after one cycle of two no - slip collisions , are related by @xmath96 [ cor : anyper ] for any @xmath97 there is an angle @xmath80 such that all ( nondegenerate ) bounded orbits in an angle @xmath28 wedge have period @xmath92 . by the proposition @xmath28 can be chosen so that the velocities are transformed by a rotation of @xmath98 around the axis of periodicity in the half - sphere representation of the velocities , ensuring the necessary condition that the velocities are periodic . preservation of measure implies that this condition is also sufficient to ensure the orbit is periodic . ( see lemma [ lem : repeat ] below . ) degenerate cases may occur with period less than @xmath92 , but a small perturbation of initial conditions will yield an orbit of maximal period . the set of wedge angles @xmath28 giving periodic orbits is dense in @xmath99 . a wedge angle @xmath28 can be chosen such that all bounded orbits are periodic of period @xmath92 . for @xmath100 , @xmath101 gives a ( two - cycle ) rotation angle @xmath102 and period @xmath103 trajectories . _ , height=1 ] figure [ fig : perex ] shows the predicted period eight example . the degenerate cases occur where the spacial projection contains repeated segments ; these are resolved when the rotational dimension is incorporated . this may happen in any higher order periodic case with appropriate initial conditions . recall that the strip produced bounded orbits for all initial conditions except for the horizontal trajectory . with the wedge there are many trajectories that escape , but the analogous result below says that the velocity phase space may be similarly partitioned . by the assumption of unit velocity , the set of velocities may be identified with the sphere @xmath104 , excluding the two @xmath105 poles representing the cases in which the there is only rotational motion . by time reversibility antipodal points correspond to the same trajectories and may be identified . the existence of an axis of periodicity for all wedge angles @xmath28 implies the region of @xmath104 corresponding to bounded orbits is not empty . numerically , one finds two antipodal non - escape regions with area a decreasing function of @xmath28 , approaching @xmath74 as @xmath28 approaches @xmath106 and growing to @xmath107 , an entire hemisphere , as @xmath28 approaches @xmath74 . the latter limit is consistent with the fact that almost all initial conditions yield bounded orbits in the no - slip strip . to specify these regions formally , let @xmath108 be the sector of the sphere corresponding to directions of direct escape from the @xmath28 wedge , along with the antipodal sector . then for @xmath86 we may define the escape region for @xmath28 as @xmath109 and its complement @xmath110 will be the region of non - escape velocities . [ pr : bddwedge ] every orbit with a non - escape velocity remains in the non - escape region . first suppose @xmath28 is not periodic , that is , the corresponding angle of rotation for the transformation @xmath91 is not a rational multiple of @xmath16 . distance from the axis of periodicity is invariant under rotation by @xmath91 , so the closest point in @xmath108 is rotated around a fixed spherical circle whose center is the axis point . every point outside of it eventually is mapped to the escape sector , while no points inside are . hence the escape and non - escape regions partition the sphere , except for a circular boundary . see figure [ fig : escape ] . if @xmath28 is periodic of period @xmath92 , then the non - escape region @xmath110 is a curvilinear regular polygon , an n - gon if @xmath39 us odd and a 2n - gon in @xmath39 is even , which remains fixed with sides permuted under @xmath91 . ( left middle , right middle ) and caustic circle for non - periodic @xmath28 ( right ) . , title="fig:",width=5 ] we can now prove the following lemma which was used above in establishing periodicity . [ lem : repeat ] for a given orbit , if two collision points on the same edge of a no - slip wedge have identical velocities then their positions also correspond and the orbit is periodic . suppose an orbit has a velocity vector repeated at distinct positions , and consider the trajectory of @xmath10 collisions between the two points oriented in the direction from the outer point to the inner point . the velocity pattern will then repeat the same cycle of @xmath10 trajectories as the orbit continues inward indefinitely . consider the action of @xmath10 iterations of the billiard map on the subset of phase space consisting of the product of the non - escape velocities and a small rectangle in the rotational and spacial dimensions . the velocities are invariant while the rectangle contracts exponentially in the spacial dimensions and expands at most linearly in the rotational dimension , which violates preservation of measure . for a billiard system with circular table of radius @xmath111 and no - slip collisions , the projections of trajectories from the @xmath17-dimensional angle - position space to the disk in position plane have the property that the vertex angle at each collision is a constant of motion . moreover , for each projected trajectory @xmath112 , there exists a pair of concentric circles of radius less than @xmath111 that are touched tangentially and alternately by the sequence of line segments of @xmath112 at the middle point of these segments . implicit in the alternating caustics is the fact that the incoming angle relative to the tangent will alternate between angles @xmath113 and @xmath114 with successive collisions , or equivalently the internal angles between the trajectories and a radius will alternate between the complementary angles @xmath115 and @xmath116 . ( see figure [ fig : circlesetup ] . ) consider three successive collisions , @xmath117 , @xmath118 , and @xmath119 , and choose coordinates with @xmath118 at the origin , with rotational direction @xmath79 , @xmath40 in the tangent direction , and @xmath34 in the radial direction . let @xmath120 be the velocity after the collision at @xmath117 . all circular billiards up to a rotation or reflection may be considered by varying @xmath20 and changing @xmath117 and @xmath119 accordingly , leaving @xmath118 fixed , hence requiring @xmath121 . then the orbit is determined by the choice of @xmath20 , and in particular we can express the two caustic radii in terms of the components @xmath122 . immediately we have @xmath123 and @xmath124 and using the post - collision velocity @xmath125 we also have @xmath126 and @xmath127 hence @xmath128 and @xmath129 notice that in the above equations requiring the numerators to be equal guarantees the denominators are equal , thus we may ensure that @xmath130 ( and @xmath131 ) by choosing @xmath20 with @xmath132 if we further require @xmath133 the orbit will be a regular n - gon . these two requirements correspond respectively to segment @xmath134 and the numbered elliptic curves in figure [ fig : circles ] , with regular n - gons at the intersections . additionally , the family of orbits with coinciding caustics @xmath130 are those in which there is no change in rotational velocity , a family corresponding precisely to the orbits of standard billiards . for standard billiards on a circle , or more generally on a smooth convex boundary , there is a countably infinite family of n - periodic billiards for any @xmath97 . ( see @xcite . ) for no - slip billiards there is a continuum of initial conditions that give periodic orbits for a given @xmath39 . by proposition [ pr : wedgeangle ] , the wedge angle @xmath135 yields period four ( or degenerate period two ) orbits . as a result , the equilateral triangle has unique dynamics . immediately , any orbit which starts at one side and fails to contact both of the remaining sides by the third collision will be periodic . it turns out , in fact , that _ all _ orbits are periodic . first , we show that the velocies are periodic with the given possible values . labeling the sides , consider the possible combinations keeping track only of which side the collision occurs on and not the exact location , looking at equivalence classes up to relabeling sides . each possible orbit can be written as a sequence of ones , twos , and threes , without loss of generality beginning with @xmath136 . certain combinations are dynamically impossible : we can immediately rule out any sequence with repeated numbers , and @xmath137 can be ruled out because any @xmath138cycle between two edges will repeat . other combinations , like @xmath139 can be ruled out by comparing the dynamics with the wedge dynamics when side @xmath17 is removed . the possible parameters when leaving side @xmath76 for the second time are such that it must , in the wedge , return to side two , parameters which also ensure the orbit would next hit side @xmath0 if instead it collides with wall @xmath17 . figure [ fig : eqproof ] shows the combinatorial possibilities with dotted lines indicating equivalent forms and dynamically impossible combinations struck through . besides the four cycle @xmath140 , the only possibilities may be expressed @xmath141 and @xmath142 . fixing an orientation for the triangle , the transformation of the velocity vector after a collision , relative to the new frame , will either be @xmath143 or @xmath144 but the combined transformations for the two possibilities after the sixth collision are ( in the given representation ) @xmath145 or @xmath146 , and the velocity is preserved . [ 12 [ 121 [ 4-cycle , circle , draw ] [ 1213 , name = spec [ ] [ 12132 , name = spec2 [ , name = spec3 ] [ 121323 [ 6-cycle , circle , draw ] [ ] ] ] ] ] [ 123 [ 1231 [ 12312 [ 123121 , name = dup3 , draw ] [ 123123 [ 6-cycle , circle , draw ] [ ] ] ] [ 12313 , name = dup2 , draw ] ] [ 1232 , name = dup1,draw ] ] ] ( dup1 ) to[out = north west , in = north ] ( spec ) ; ( dup2 ) to[out = north west , in = north ] ( spec2 ) ; ( dup3 ) to[out = north west , in = north ] ( spec3 ) ; it remains to show that the orbits are actually periodic , that is , that the orbits return to the same boundary point and are not merely parallel trajectories . suppose that the orbit arrives back on the starting side but not at the same point , and consider subsequent returns after multiples of six collisions . the second set of trajectories will be parallel to the first all the way around and accordingly the sequence of walls will not change . now consider the finite partition of the original wall by degenerate points where an orbit in the fixed direction from that point eventually hits a vertex . all return points must be in the same interval , always at the same distance apart and with the order preserved . but this implies that the entire interval is fixed , and the orbit must return to the starting point . the projection of the velocity phase portrait of an isosceles triangle no - slip billiard is given in figure [ fig : isos ] , where overlapping orbits are the result of the projection . this pattern is typical of polygons , with most orbits appearing as closed curves around periodic points , with the number of components matching the period of the central point . numerical evidence suggests that for no - slip polygons there may be no positive measure ergodic components . standard billiards with entirely concave boundaries are dispersing , with wave fronts expanding at every collision . in contrast , the chaotic behavior of the bunimovich stadium depends upon a defocusing mechanism : convex boundaries create a focusing front , but under circumstances in which the front passes through a focusing point and sees a net expansion before colliding at the next boundary . looking at billiards with left and right edges identified and concave ( or convex ) upper and lower boundaries , both mechanisms appear to arise in the no - slip case as well . ( see figure [ fig : focus ] . ) to trigger a comparable level of defocusing after the same number of collisions for an equally spaced wave front , a greater curvature is required for the no - slip case . as the no - slip examples are projections , with independent expansion or contraction arising in the rotational dimension , the behavior may be more nuanced . in spite of the apparent defocusing , the stadium has a bounded positive measure component phase space and is not ergodic . similarly , other known standard billiard examples fail to be ergodic in the no - slip case : the ` mushroom ' billiard @xcite has a similar bounded component in the stem , the ` flower ' billiard fails to be ergodic by remark [ rem : circerg ] , while ` pocket ' billiards are not ergodic for both of the above reasons . in these cases the no - slip phase portraits show a mixture of structured orbits mixed surrounded by potential chaotic seas ( figure [ fig : combinederg ] ) . the ` moon ' billiard , recently shown to exhibit ergodic behavior for some parameters @xcite , does not appear to be ergodic in the no - slip case . along with the mushroom billiard , its phase portrait exhibits segments along which orbits linger before moving to apparent chaotic regions . this behavior is qualitatively similar to the known _ marginally unstable periodic orbits _ of standard billiards which both the moon @xcite and mushroom @xcite @xcite are known to exhibit . figure [ fig : hept ] gives the phase portraits for three small but increasing @xmath147 perturbations of the no - slip regular pentagon , demonstrating the persistence of the dynamics . numerical experiments using a range of polygons and a range of variation in curvature have not produced examples free from evidence of non - ergodic behavior .	abstract we investigate the dynamics of no - slip billiards , a model in which small rotating disks may exchange linear and angular momentum at collisions with the boundary . we give new results on periodicity and boundedness of orbits which suggest that a class of billiards ( including all polygons ) is not ergodic . computer generated phase portraits demonstrate non - ergodic features , suggesting chaotic no - slip billiards can not readily be constructed using the common techniques for generating chaos in standard billiards . _ dedicated to the memory of kolya chernov _
underwater exploration represents a very important economic , technologic and scientific challenge . this is closely related to arctic and antarctic offshore resources , pollution monitoring , general oceanographic data collection and , recently , to underwater actuation @xcite . due to very large underwater areas and high damping properties of water , application of multiple autonomous underwater vehicles ( auvs ) in cooperative missions seems very promising @xcite . for application of auvs in networked or swarm mode , there is a number of crucial issues : underwater sensing and communication ( s&c ) , cooperation and mission control , design of auv platforms , autonomous behavior and several collective aspects of running multiple auvs . in this work we concentrate on minimalistic local s&c @xcite , and related coordination strategies , being motivated by the following reasons @xcite . in several past and running projects devoted to underwater swarms , such as aquajelly @xcite , angels @xcite , cocoro @xcite , a number of auv platforms and sensing technologies has been developed . these works indicated two important issues : a successful auv platform needs a dedicated combination of different s&c technologies , moreover capabilities of underwater cooperation depends on the level of embodiment @xcite of on - board s&c systems . in several cases , even a simple multi - modal signal system leads to advanced cooperation @xcite ( see e.g. the case of multi - agent cooperation @xcite ) . since swarm approaches rely primarily on local interactions between auvs @xcite , the paper is devoted to local s&c systems ( unmodulated and modulated ir / blue light , rf and electric field ) , which can be used for robot - robot / robot - object detection and provide sub - modal information , such as direction and distances , as well as can be used for analog and digital communication @xcite . these systems , developed for each of the platforms is described in sec . [ sec : localcom ] , whereas sec . [ sec : conparison ] provides general overview over different s&c technologies . in sec . [ sec : experiments ] we shortly sketch a few behavioral experiments with these systems and finally in sec . [ sec : conclusion ] conclude about useful combinations of s&c and their embodiment for collective underwater systems . in this section we give a short overview over different state - of - the - art s&c systems in an underwater environment , see e.g. @xcite-@xcite . four systems will be compared : * _ sonar _ : sonic waves travel very well under water and the energy and build - space required for generating and receiving them is very low . this approach is used in so - called acoustic modems @xcite . drawbacks of this approach are , firstly , the relatively low sound travel speed of roughly 1500 m@xmath0s ( much slower than any other s&c system ) , and , secondly , multiple reflections causing essential distortions in the signal . * _ radio _ : electromagnetic waves are a standard communication method in air ; its application under water creates several problems @xcite . due to water connectivity , the attenuation of radio waves depends on the used frequency , which in turn results in the size of antenna . high frequencies ( @xmath1 100 mhz ) only need a small antenna ( @xmath2 m ) while their range is restricted to @xmath3 m. lower frequencies ( 100 khz ) have a long range ( 100 m ) , but need a large antenna ( 100 m ) . * _ optical _ : using light as a communication channel can provide a compact size of transmitting equipment and acceptable range @xcite . due to the color dependent attenuation of light in water , the communication range varies between a few centimeters in ir spectra and increases to over a meter by using blue or green light . * _ electric field _ : this is a new communication approach . it bases upon generating and measuring electric fields . the build - size and energy required for this system is very small . unfortunately the attenuation of electric fields is very high , liming the range of this communication channel to less than 1 m. we will discuss this approach in sec . [ sec : elect ] . table [ tab : comparisonofdifferentcommunicationchannels ] shows an overview of the discussed approaches . .comparison of different communication channels [ cols="<,>,>,>",options="header " , ] as a digital communication transceiver , the blue light system needs modulator , amplifier , signal conditioner , and protocol encoder / decoder . the one chip solution can be solved by using a cs 8130 irda chip from cirrus logic . since the blue light system has a directional s&c , two channels are not sufficient for the swarm robot to communicate in every direction . the half duplex behavior of each channel makes one channel unable to be applied for sensing , because the sensing mechanism requires to transmit and to receive the sensing signal in one time . therefore , the position of the transmitter and receiver are swapped with neighboring channels for sensing application . the system must be configured and calibrated for finding the best modulation type for underwater communication and sensing . according to the measurement results , both for communication and sensing , the quadrature amplitude modulation ( qam ) seems to be the best modulation for underwater application . [ fig : chat_blue_light](a ) shows the relation between current sensitivity and communication distance . by using this curve , an active sensing algorithm can be added to the inter - robot communication algorithm by varying the amplification and the sensitivity of the programmable amplifier via software , see fig . [ fig : chat_blue_light](b ) . the robot can approximate the distance with other robots by gradually decreasing the current sensitivity while communicating each other . the developed inter - robot communication algorithm has three phases . first , when two robots are in the communication range , they begin to establish the communication by sending their ids to each other . second , the communicating robots are approximating their distance by gradually decreasing the current sensitivity within the programmable gain amplifier . therefore , after knowing their own position , behavioral or cooperation phased can be performed . a robot will continuously iterate the first phase if there is no other robots in the communication range . hence an obstacle might reflect the transmitted signal and the robot would receive back the first phase communication packet that contains its own i d . after experimenting with optical s&c systems , we implemented another approach , which is inspired by weakly electric fish . these animals are capable of producing an electric field which they can use for localisation and communication @xcite . here we try to use this bio - inspired approach for analog communication and navigation in robot swarms . * electric fields . * electric charges generate electrical fields in their vicinity . electric fields are vector fields . for a point charge @xmath4 the field intensity @xmath5 can be calculated at each point @xmath6 as @xmath7 with the permittivities @xmath8 ( vacuum ) and @xmath9 ( relativ ) @xcite . the field vectors of multiple point charges follow the superposition principle . in our robot the electric field is generated by a dipole . the field intensity is proportional to the charge in the electrodes , which themselves are proportional to the applied voltage to the electrodes : @xmath10 with the voltage @xmath11 and the capacity @xmath12 . * * the intensity of an electric field can then be detected by measuring the differential potential between two electrodes in the field . this potential is proportional to the electric field intensity , which itself is proportional to the output voltage of the sender . by modulating the output voltage of the sender , information can be transmitted . * localization . * by using multiple pairs of electrodes in the receiver it is possible to calculate the bearing and distance to the sender . this is achieved by utilizing the drop in field intensity with relation to the distance . a sinus wave is impressed on the sender s electrodes which creates an oscillating electrical field . the field intensity depends mainly on the amplitude and frequency of the output voltage and some environmental conditions . the amplitude in the field intensity at a specific point @xmath6 is proportional to the amplitude of the output voltage : @xmath13 with the amplitude of the output signal @xmath14 and measured input @xmath15 , the frequency @xmath16 and the time @xmath17 . if sender and receiver are approximately in the same plane and the electrodes have the same orientation ( compare fig . [ fig : versuch ] left ) ( [ eq : amplitudesim ] ) can be simplified to : @xmath18 with the frequency dependent proportionality factor @xmath19 and the distance @xmath20 between sender and receiver . measuring the sinus amplitude ( @xmath21 , @xmath22 , @xmath23 , @xmath24 ) at four points with a specific geometrical pattern ( fig . [ fig : versuch ] right ) leads to the following equations : @xmath25 in setting up these equations it is assumed that @xmath26 so that the error in the angle @xmath27 and distance @xmath20 between the different sensors is minimal . in ( [ eq : amplitudessystem ] ) the proportional factor and output amplitude can be eliminated , under the condition of @xmath28 leading to : @xmath29 and @xmath30 * design limitations . * this approach has two design limitations : it requires the sender and receiver electrodes to have the same orientation ( i.e. vertical ) and to be roughly in the same plane ( horizontal to the orientation ) : * the first limitation holds no practical difficulties . our robot maintains a specific orientation , caused by its center of gravity . by placing one of the sender and receiver electrodes on top and one on the bottom of the robot the orientation is always vertical . * the derivation above is only correct if sender and receiver are on the same plane , which is horizontal to the orientation of their electrodes . in a three dimensional environment this is not always true , but usually the working space is wider than it is high , even in a 3d environment . even so , we are working on overcoming these limitations . we are confident that the second can be eliminated by rearranging the receiver electrodes . to overcome the first limitation additional electrodes may be needed . as described in the previous sections , different local s&c systems utilize the same hardware components for sensing and communication . moreover , they use modal and sub - modal approaches , which provide not only message transmission , but also deliver spatial information about position and distances of robots and objects . in this section we describe several behavioral experiments , performed with these systems . one of the implemented scenarios with aquajelly robots had the following form , see fig . [ fig : scenario ] . in water , a robot sends sequentially in all ir channels its own i d . listening and sending times are selected as approx . 95% listening and 5% sending , so that all robots most of time silently observe the environment . receiving another - than - own i d means meeting another robot , whereas non - id ir light means own reflection from passive objects . granularity of ir channels is enough for rough collision avoidance with objects , e.g. walls of the aquarium . collision avoidance based on digital channels are impossible for more than two robots ( or robot and object ) . in opposite , blue light channels emit almost all time . since light has additive properties , when two robots meet each other , intensity of light in the point of light - spheres - intersection creates a light gradient and can be locally sensed by both robots . especially interesting is the light gradient when several robots meet each other ; they create complex gradients , which can be used for precise multi - robot navigation . unfortunately , blue - light sensors continuously receive signals from their own light sphere so that no efficient communication is possible in this mode . initially all robots are fully charged . when a robot has a low energy value , it swims up and recharges . with the progress of experiment , more and more robots swim up for recharging . in this way , several robots meet in the upper part of the aquarium and compete for the docking station , see fig . [ fig : scenario](from left to right ) . since only a robot with lowest energy value should recharge , all robots bilaterally exchange values of their own energy level . the robot with the lowest energy value can swim up . this local behavior leads to the following interesting collective behavior . due to light gradient created by many robots , all robots exert optical pressure `` on each other , and collectively swim down , whereas only one most ' ' hungry " robot swims up and recharges , see fig . [ fig : scenario](right ) . in order to investigate the modulated blue light s&c system , we used underwater submarine toy as a mechanical platform with new electronic components for locomotion , computational and s&c capabilities . this submarine has three degrees of freedom and three actuators for moving forward / backward , turning left / right , and diving up to 1 meter . the necessary modifications of the submarine including the replacement of the original electronic parts with the new designed electronic boards , and drilling some new holes on the robot s body for communication / sensing transducers placements . cortex3 lm3s316 microcontroller with 25mhz of clock frequency , 16 kb of internal flash rom , and 16 kb of ram has been used in the platform . two motor drivers and two navigational sensors are placed on the main board of the electronic platform . the combination of the available pwm output from cortex3 and motor driver perform the ability to control the swimming velocity via software . a digital compass and pressure sensor are added as a three - dimensional orientation sensor ( an low - frequency rf part is foreseen for a backup communication with host ) . during the experiment , robot is deployed into the aquarium fulfilled with several obstacles . the available obstacles and aquarium s walls are used to examine the sensing capability , see fig . [ fig : blue_light_platform](left ) . both types of obstacles have different type of optical characteristic , which create different reflection behavior for the blue light . therefore , white papers can be put outside the aquarium walls to increase the reflection capability of the optical sensor . for testing the communication and active sensing capabilities , one robot is deployed underwater and a static encoded blue light transceiver is installed on the aquarium wall as a measurement reference point , see fig . [ fig : blue_light_platform](right ) . the static transceiver can illuminate several type of light signals if it receives a specific blue light packet data from the swimming robot . different types of blinking signals are used to examine the functionality of the active sensing capability . this approach underlies several other experiments , where a few passive robots are identified by one active auv as foraging targets . the circuit for electric field communication is very simple . it consists mainly of a digital - analog - converter ( dac ) for the sender and four amplifiers ( op ) with analog - digital - converters ( adc ) for the receiver ( fig . [ fig : esversuch ] ) . * _ sender : _ the output of the 14-bit dac is directly tied to one of the sender electrodes while the other is connected to vcc/2 . the electrodes are in direct contact with the surrounding water . the output of the dac can be set to a voltage between gnd and vcc . this setup allows control of the field intensity and polarity . * _ receiver : _ the receiver has four pairs of electrodes in the water to measure the difference in the potential of the electric field in four places . the electrodes use capacitors as highpass - filters to filter dc signals . the signals are amplified by differential ops ( magnitude 1000 ) and digitized by 14-bit adcs . for the experiments the sender and receiver are put under water ( fig . [ fig : versuch ] right ) . the receiver has a sampling rate of 10 khz . the measured data is transmitted to a pc , where the bearing and distance are calculated . in the experiment the sender was moved around the receiver and the received data was recorded together with a video tape of the experiment for comparison ( fig . [ fig : esversuch ] ) . the true bearing was extracted from the video stream and compared with the from the electrical sensor data calculated bearing . [ fig : vergleichcords ] shows the result . for 50% of the measuring points the error is less than 5@xmath31 and it exceeds never more than 15@xmath31 . this might be further improved by increasing the magnitude of the amplifiers and reducing the noise through optimized circuits and digital filters . in this work we considered several optical and electric - field - based approaches for sensing and communication within @xmath32 . together with s&c approaches for @xmath33 , such as acoustic and low - frequency rf , they represent the available spectra of s&c technologies for underwater networked and swarm robotics . as indicated in the sec . [ sec : experiments ] and from other performed experiments , these approaches combine communication with localization , distance measurement and object detection . in several cases , such a sub - modal information is available even during communication and can be used for very efficient behavioral strategies . each of the considered s&c system has its own benefits and weaknesses . it seems that no current single system is capable of achieving all the requirements on @xmath32/@xmath33 , sensing , minimal build space , energy consumption and complexity . therefore the best approach lies in combining several of the available systems , for example in the way shown in fig [ fig : scheme ] . the optical system provides split wave - length dependent channels . it can be used in analog and digital mode with existing control circuits for ir systems ( e.g. irda or different modulations e.g. pcm / qam ) , which with small modifications can be used for green , cyan and blue light . the range and bandwidth are sufficient for local communication . the channel is directional which can be of benefit for swarm - based coordination approaches . additionally the reflection in analog mode can be used for navigation and detection tasks . the electrical sensor is a good supplementary element to directional optics . electric fields - based channels are omni - directional ; hardware required for generation and detection of electric fields utilizes off - the - shelf components and is compact and energy efficient . it can be used to calculate the bearing between sender and receiver , i.e. for self - localization . the range is small but sufficient for @xmath32 . it is also necessary to supplement these s&c systems by acoustic or ultra - low - frequency rf to provide global communication . sonar requires a bit larger hardware equipment than the optical system . with additional components , it can be used for measuring distances to obstacles . rf systems represent a trade - off between the frequency ( i.e. communication distances ) and the size of integrated antennas ( i.e. the size of platform ) . the control circuits are more complex than those for optic or acoustic approaches . since bandwidth for low - frequency rf is not sufficient for application of standard protocols ( e.g. zigbee ) , global rf communication represents some open problems . if more than one receiver is used , the bearing between sender and receiver can be calculated . however this would make the hole system more complex and expensive . comparing acoustic and low - frequency rf approaches for @xmath33 , acoustic one is more favorable due to more less complex hardware . usage of global communication for networked and swarm systems should be reduced to absolute minimum ( see for instance minimalistic approaches for cooperation and decision making @xcite , @xcite ) . the angels and cocoro projects are funded by the european commission within the work programm future and emergent technologies `` and ' ' cognitive systems and robotics " under the grant agreements no . 231845 and 270382 . we want to thank all members of the project for fruitful discussions .	this paper is devoted to local sensing and communication for collective underwater systems used in networked and swarm modes . it is demonstrated that a specific combination of modal and sub - modal communication , used simultaneously for robot - robot and robot - object detection , can create a dedicated cooperation between multiple auvs . these technologies , platforms and experiments are shortly described , and allow us to make a conclusion about useful combinations of different signaling approaches for collective underwater systems .
the structurally simplest fe - based superconductor , fese , is one of the most mysterious at the present writing . it exhibits a tetragonal to orthorhombic structural phase transition at @xmath4 , and displays very strong electronic nematic behavior below this temperature ( @xmath5 ) , but never orders magnetically as do the more familiar fe pnictide systems . while its critical temperature @xmath6 is relatively low for this class of materials , a very rapid increase of @xmath7 to about @xmath8 is observed under modest pressure@xcite . various intercalates of fese have @xmath7 of roughly this magnitude at ambient pressure as well , and the highest @xmath7 values of the entire class of fe - based superconductors , @xmath9 , are found in monolayer films of fese grown on srtio@xmath10 substrates.@xcite thus the bulk fese material , of which excellent single crystals are now available@xcite , gives the impression of being poised to become a high - temperature superconductor , such that efforts to understand its properties have accelerated in the past few years . a starting point for these theoretical efforts is a reasonable band structure for fese . unfortunately , both density functional theory ( dft ) and dynamical mean field theory ( lda+dmft ) predict low - energy bands that deviate qualitatively from that observed by angle - resolved photoemission ( arpes)@xcite and quantum oscillation ( qo ) studies@xcite . at high @xmath5 , arpes finds splittings of the hole band associated with fe @xmath0 states that are consistent with a modest spin - orbit coupling of approximately @xmath11 . at low @xmath5 , an additional splitting that breaks the fourfold @xmath0 symmetry is also observed at the m - point and it increases significantly and smoothly in a manner consistent with orbital ordering . the fermi pocket sizes are generally much smaller than found in _ ab initio _ studies , and show a @xmath1 symmetry breaking significantly larger than one might anticipate given the @xmath12 ) orthorhombic splitting of the lattice constants . in ref . , we presented a tight - binding band structure `` engineered '' to give the correct arpes and qo results , both at high and low @xmath5 , including a @xmath5-dependent orbital order term added to the hamiltonian , and showed that the model gave reasonable results also for the knight shift and spin - lattice relaxation rate @xmath13 compared to experiment@xcite . using this band structure we also calculated the spin fluctuation exchange pairing interaction , and showed that it yielded a superconducting ground state with gap structure consistent with stm and penetration depth measurements.@xcite recently , the first inelastic neutron scattering ( ins ) results on this system became available@xcite , showing the dominance of `` conventional '' @xmath2 spin fluctuations at low energies and persisting up to @xmath14 , while previous _ ab initio _ approaches did not find low - energy excitations at that momentum@xcite . for @xmath5 greater than @xmath14 , both ins and nmr experiments have suggested that no significant low - energy spin fluctuation weight was present , in contrast to pnictide systems such as bafe@xmath15as@xmath15 , where such strong magnetic fluctuations were tied to the nematic transition@xcite . understanding the cause for suppression of @xmath2 spin fluctuations seen in neutron scattering and nmr at high @xmath5 is clearly important to understand the physical reasons for the suppression of magnetic order in these systems , and the relation , if any , between spin fluctuations , nematic order and orbital order . several novel proposals have been made recently about how unusual magnetic effects might prevent the occurrence of long range magnetic order@xcite or lead to a type of order that is difficult to observe@xcite . the @xmath2 spin fluctuations in fese have a broad maximum , according to ins , at very low energy , of order @xmath16 , which acquires significant amplitude immediately below @xmath14@xcite . this is in apparent conflict with nmr spin relaxation results , which probe low - energy spin fluctuations at all @xmath17 , but which are negligible until lower @xmath5 just above @xmath7 are reached@xcite . one proposal to reconcile the results of these two probes is the presence of an apparent spin gap in the ins measurements below about 2 mev ; there is no apparent explanation for such a spin gap , however . another possibility is that the spin fluctuations remain at intermediate @xmath5 and nmr spin relaxation is less sensitive to them in this material due to the presence of competing fluctuations that evolve differently with the evolution of @xmath5-dependent band structure . when the @xmath5 is lowered below @xmath7 , the @xmath2 fluctuations sharpen into a resonance similar to those observed in fe pnictide systems . it is remarkable , in fact , that at low @xmath5 and low energies , the magnetic response of the system is similar to canonical iron - based systems such as ba(fe@xmath18co@xmath19)@xmath15as@xmath15@xcite in this paper , we present results for the dynamical susceptibility derived from our band model for fese together with a random phase approximation ( rpa ) treatment of hubbard and hund - type interactions . we reproduce the low - energy @xmath2 spin fluctuations and the neutron resonance in the superconducting state , but show that at higher energies the spin fluctuations disperse to wave vectors along the @xmath20 line in momentum space . we continue to find a @xmath5-dependence of the knight shift and @xmath13 in agreement with experiment , despite the fact that our model does not have a low - energy normal state spin gap . we attribute the apparent insensitivity of the nmr to the @xmath2 fluctuations at high @xmath5 to the weak overall weight at this wave vector , due to the electronic structure of this material . the superconducting order parameter derived from spin - fluctuation pairing has @xmath21 character with accidental nodes , leading to a v - shaped density of states ( dos ) at the fermi surface . nodal lines occur and thus lead in our calculations to anisotropies in the low - temperature penetration depth , varying as @xmath5 and @xmath22 in the @xmath23 and @xmath24 directions , respectively . these predictions can be checked experimentally in a single - domain sample . the overall picture that emerges for this material is a system driven to the verge of a magnetic instability by orbital ordering . the hamiltonian for this system is given by @xmath25 where @xmath26 is the tight - binding hamiltonian , @xmath27 is a @xmath5-dependent orbital ordering term that breaks the crystal point - group symmetry , and @xmath28 is the standard hubbard - hund hamiltonian , see appendix a. the tight - binding term can be expressed as @xmath29 here ( @xmath30 ) are orbital indices , @xmath31 are the hopping integrals , and @xmath32 . a band structure that is consistent with the observed electronic structure in fese has been provided in ref . , where it was shown that the rather complex renormalizations of band structure relative to dft observed in fe - based superconducting systems@xcite , involving both fermi velocity changes and relative shifts of hole and electron bands , could be captured in terms of relatively simple modifications of a few short - range matrix elements ( details can be found in appendix a ) . the model assumes further that below the structural transition at @xmath33 the fourfold - symmetry - broken phase is described by a mixed bond- and site - centered @xmath5-dependent orbital ordering term in the hamiltonian given by @xmath34 here @xmath35 ( @xmath36 ) is the bond - centered ( site - centered ) orbital ordering which we assume has a mean - field @xmath5 dependence @xmath37 below @xmath38 with @xmath39,@xcite and leads to a maximum band splitting of @xmath40 . this combination of orbital ordering is motivated by a recent arpes measurement on detwinned fese@xcite , showing that the band splitting between the @xmath0 bands occurs in opposite directions at the @xmath41 and m points . we find a similar behavior in our calculated band energies @xmath42 as discussed in appendix a. band splittings consistent with the results in ref . are reproduced by @xmath43 , together with a spin - orbit coupling of @xmath44 determined by the high-@xmath5 bands at the @xmath41 point . and those in rpa approximation @xmath45 at @xmath46 and @xmath8 . the bare susceptibilities have been multiplied by a factor of 10 to improve readability of the plot . ] focusing first on the static , @xmath47 , magnetic properties , fig . [ fig : staticchi ] shows the bare and rpa spin susceptibilities @xmath48 and @xmath45 as a function of momentum . it is striking that in the high-@xmath5 ( @xmath1 symmetric ) magnetic response , the @xmath2 instability competes very closely with a state near @xmath3 [ @xmath49 , where @xmath50 is close to @xmath51 , even when interactions are included close to the rpa instability . this is reminiscent of the scenario of glasbrenner _ et al._,@xcite whereby fese is nonmagnetic due to competition of several nearly degenerate magnetic states with wave vector @xmath49 . as @xmath5 is lowered , this degeneracy is lifted by orbital ordering , as evident from fig . [ fig : staticchi ] , singling out the @xmath2 fluctuations as the dominant ones . this remains true also with the susceptibility enhanced by interactions in the rpa scheme ; spectral weight is moved downwards in energy near @xmath2 to dominate the spin spectrum up to @xmath52 as seen in fig . [ fig : highneut ] . from the fermi surface plots shown in fig . [ fig : bandfscut ] in appendix a it can be inferred that there is no obvious nesting at the fermi surface that can readily explain the larger noninteracting susceptibility at @xmath2 compared to @xmath53 . this property originates from scattering at higher energies . we will return to this point further below . in @xmath54 along high symmetry directions at @xmath55 . ( b ) susceptibility calculated at @xmath56 above the structural transition . ] inelastic neutron scattering ( ins ) experiments on fese have revealed the importance of conventional @xmath2 fluctuations@xcite . experiments performed on powder samples have found @xmath2 fluctuations extending to at least 80 mev . as can be seen in fig . [ fig : highneut ] we also find similar @xmath2 fluctuations extending up to @xmath57 when the interactions are tuned close to the stoner instability ( @xmath58 , @xmath59 , @xmath60 is the band renormalization factor ) such that there is no magnetic instability at least down to the temperature of the superconducting transition . note that at these high energies , the @xmath2 excitations are present both below and above the structural transition , as in experiment@xcite . we will come back to differences in weights of the stripe fluctuations close to @xmath2 and nel fluctuations close to @xmath3 as a function of @xmath5 later . additionally , the imaginary part of the susceptibility disperses and becomes more intense along the line from @xmath2 to @xmath3 as we move to higher energies . furthermore , there are weaker incommensurate branches which soften near @xmath3 , in contrast to observed spin excitations in fe pnictides@xcite . this feature of the high - energy excitations constitute a prediction of the present theory for ins experiments on single crystals extending to higher energies . dependence of the imaginary part of the dynamical susceptibility @xmath61 at @xmath62 calculated from our model with the interaction parameters as in fig . [ fig : highneut ] , shown for series of low energies @xmath63 . ] it is interesting to observe that above the structural transition , the low - energy @xmath2 fluctuations are strongly suppressed [ see fig . [ fig : highneut](b ) ] , in agreement with experiments@xcite . this suppression can be more clearly seen in fig . [ fig : lowneut ] where the @xmath5 dependence of the low - energy spin fluctuations at @xmath2 is plotted . the strength of the low - energy @xmath2 fluctuations dies off rapidly with increasing @xmath5 , and disappears completely above @xmath14 . note that the increase in @xmath64 at low @xmath5 with decreasing @xmath65 is due to the low - energy peak in @xmath64 since the system is close to the magnetic instability . this unusual @xmath5 dependence is primarily due to an enhancement of low - energy spin fluctuations induced by the orbital ordering . by introducing correlations that push the system close to the magnetic instability , the low-@xmath5 fluctuations are amplified more strongly compared to high @xmath5 . we argue therefore that the suppression of the high-@xmath5 spin fluctuations seen in nmr experiments is not necessarily due to the presence of a spin gap , but may arise simply due to @xmath5-dependent electronic structure effects . for example , we note that while the @xmath2 fluctuations strengthen at low @xmath5 , those near @xmath3 lose spectral weight . since both such fluctuations contribute to the nmr spin relaxation , they evidently compensate each other to some extent , leaving a flatter @xmath5 dependence in the intermediate-@xmath5 region ; thus @xmath13 does not rise until close to @xmath7 when the @xmath2 fluctuations completely dominate , as discussed in appendix b. at @xmath62 for the same interaction parameters as in fig . [ fig : highneut ] at three representative values of @xmath5 . for the lowest temperature ( @xmath66 ) , @xmath67 is calculated in the presence of the superconducting gap within the spin - fluctuation pairing approach as detailed in ref . . [ fig : nresonance ] ] in fig . [ fig : nresonance ] the energy dependence of the low - energy inelastic spectrum is shown at @xmath55 . it displays a peak at around @xmath68 , similar to ins experiments@xcite . note that such a peak occurs in our model primarily because the coulomb interaction is tuned close to the stoner instability . therefore within such an itinerant model the presence of the low - energy peak structure is a signature that fese is very close to a magnetic instability , consistent also with pressure experiments@xcite . recently , the magnetic excitations in fese have been studied in more detail via ins experiments.@xcite . for a closer comparison , we also present maps of the dynamical structure factor @xmath69 at different energies and two temperatures above and below the structural transition in . the dynamical structure factor has been calculated using at two representative temperatures above and below the structural transition obtained by integration over momentum space at @xmath70 . in ( a ) we show the contributions of nel fluctuations by integrating over one quarter of the brillouin zone centered around @xmath3 compare fig . [ fig : s_maps ] ( r ) . in ( b ) the area is centered around @xmath2 to estimate the stripe fluctuations , showing a transfer of spectral weight to stripe fluctuations due to orbital order , and a low - energy peak in the superconducting state . the total local susceptibility ( c ) is unaffected by lowering @xmath5 , but the weight transfer as a function of @xmath5 can be observed clearly at an energy @xmath71 ( d ) . [ fig : integrated_susc ] ] @xmath72 where @xmath73 is the magnetic form factor of @xmath74 , such that the result does not show periodicity with reciprocal vectors as the susceptibility does@xcite . in fig . [ fig : s_maps ] , we show plots of the dynamical structure factor as a function of momentum for various energies at high @xmath5 and low @xmath5 . for low energies @xmath75 it can be seen that the spot around @xmath2 is elongated perpendicular to the direction @xmath76-@xmath2 at all temperatures showing the dispersion of the fluctuations along the path @xmath2 to @xmath3 , which can also be seen in fig . [ fig : highneut ] . already from this series of plots one can see that the spectral weight around @xmath3 is present at all temperatures , but decreases when orbital order sets in and the fluctuations at @xmath2 sharpen and gain weight . to make this effect more visible , we also calculate the local susceptibility @xmath77 , where the integration has been performed for @xmath70 since the susceptibility shows only a very weak dependence on @xmath78 . in fig . [ fig : integrated_susc ] we present the results for partial integrals over square regions @xmath79 of one quarter of the brillouin zone around @xmath3 to estimate the weight of the nel fluctuations ( a ) and around @xmath2 to estimate the stripe fluctuations ( b ) . the regions of integration are indicated in fig . [ fig : s_maps ] ( r ) with dashed lines . restricting ourselves to the energy range where the two types of fluctuations are separated in momentum space , one can observe that the nel fluctuations decrease in weight as orbital order sets in , while the stripe fluctuations gain weight . the total local susceptibility is nearly unchanged over the whole bandwidth of the paramagnon excitation spectrum that extends up to energies of approximately @xmath80 with a maximum at around @xmath81 ; see fig . [ fig : integrated_susc ] ( c ) . in summary , our results agree extremely well with experimental results presented in ref . in the transfer of the spectral weight from nel fluctuations to stripe fluctuations on lowering @xmath5 , [ fig : integrated_susc ] ( d ) , while the former are already weaker above the structural transition . also the total local susceptibility is in agreement with the experimental result . note that the relative strength of nel fluctuations and stripe fluctuations changes slightly with the ratio @xmath82 and details of the electronic structure such that quantitative conclusions can not be drawn , but the trend of weight transfer is robust . to conclude this discussion , we mention that our result slightly deviates from experimental findings in ref . in the nature of the nel fluctuations where they seem to be gapped at low @xmath5 and do not deviate from a commensurate structure . we expect however that the incommensurate peaks near @xmath3 will coalesce in our model if self - energy effects from disorder and interactions are included in the model , but this calculation is beyond the scope of this paper . next , we calculate the superconducting order parameter using spin - fluctuation - mediated pairing interactions . within our previous approach of solving the linearized gap equation using a five - band model for computational simplicity@xcite , we obtain the gap symmetry function @xmath83 on the fermi surface as shown in fig . [ fig_sc](a ) . due to the reduction of the crystal symmetry by the orbital order parameter at low @xmath5 , this state has components of both @xmath84- and @xmath85-wave symmetry . in our model , the superconducting gap is very small on the @xmath41-centered pocket and obeys line nodes at the @xmath2 pocket . the sign change between the electron - like bands between @xmath2 and @xmath53 is induced primarily by pair scattering originating from the broad @xmath3 structure of the susceptibility . from the larger value of the susceptibility at momentum transfer @xmath2 compared to @xmath53 ( see fig . [ fig : staticchi ] ) one would naively expect a sign change between the @xmath41-centered pocket and the pocket at @xmath2 . instead , the sign change appears between the hole - like band and electron - like band with fermi surface centered at @xmath53 because of the effect of matrix elements in the pair - scattering vertex ; i.e. , it is driven by the orbital content at the fermi level . next , we solve the bogoliubov - de gennes ( bdg ) equations in momentum space following the method in ref . for the multiorbital case with symmetrized pairing interactions in order to calculate the dos which shows a v - shaped spectral dependence at low energies as seen in fig . [ fig_sc](b ) . this nodal v - shaped dos is in agreement with recent stm measurements.@xcite of the leading superconducting instability with @xmath86 as obtained from the solution of the linearized gap equation with @xmath87 and @xmath59 , where @xmath60 is the band renormalization factor . areas where @xmath88 is smaller than 1% of the maximal value are marked black . ( b ) dos calculated by solving the full bdg equation with the same interaction parameters in momentum space using a grid of @xmath89 @xmath90 points and interpolating the converged gap to calculate it with a reasonable spectral resolution . ] for the calculation of the imaginary part of the susceptibility in the superconducting state , we use the framework discussed in ref . . the gap symmetry function from the linearized gap equation is interpolated on a large momentum mesh , and we employ an exponential damping factor@xcite to specify its value away from the fermi surface with damping chosen one order of magnitude larger than the gap @xmath91 . we chose a smearing @xmath92 and use an integration grid of @xmath93 @xmath90 points to obtain reasonable results as shown in fig . [ fig : nresonance ] . at high @xmath5 above the structural transition , there is very little low - energy spectral weight at @xmath2 . it gradually increases when the @xmath5 is decreased and orbital order sets in , such that at lower @xmath5 a maximum at a few mev is obtained . deep in the superconducting state where the superconducting gap is fixed to yield coherence peaks at @xmath94 , see fig . [ fig_sc](b ) , spectral weight at low energies is suppressed due to the superconducting gap opening and the maximum is again at approximately @xmath95 , which also induces a low - energy peak in the local stripe fluctuations , see fig . [ fig : integrated_susc ] ( b ) , but not in the local nel fluctuations . the low-@xmath5 contributions near @xmath3 ( not shown ) are very small , similar to the paramagnetic state shown in figs . [ fig : highneut ] and [ fig : s_maps ] ( j ) . to conclude our discussion of the low - energy spin fluctuations , we emphasize that the increase of the peak upon lowering of @xmath5 into the superconducting state is somewhat different from the usual neutron resonance expected in the fe pnictides , where a mode corresponding to a pole in the rpa susceptibility is allowed to propagate due to the gapping of the fermi surface in the superconducting state , and the effect is weighted by a coherence factor @xmath96 $ ] that is maximized when the gap changes sign between hole and electron pockets . in the case of fese , at @xmath62 these coherence factors play a negligible role since the gap changes sign between the hole pocket and the electron pocket at @xmath2 only over a very small range of angles on the @xmath2 pocket . instead , the influence of the superconducting gap is to shift and sharpen the normal state peak such that spectral weight at low energies is removed and effectively a gapped paramagnon appears . the final peak position of the apparent `` neutron resonance '' as shown in fig . [ fig : nresonance ] is roughly given by the sum of the gaps on the pockets at @xmath41 and @xmath2 . at low @xmath5 showing power - law ( linear ) @xmath5 dependence in the @xmath24 ( @xmath23 ) direction where the nodes are not ( are ) captured . inset : corresponding components of the superfluid density tensor @xmath97 calculated over a wide range of @xmath5 assuming a mean - field - like superconducting order parameter as a function of @xmath5 . [ fig : pen ] ] a possible experimental probe that could confirm and elaborate on the nodal nature of the superconducting gap measured by stm is the london penetration depth . since the system under discussion shows a strong deviation from a @xmath1-symmetric electronic structure , this deviation should also imprint itself on the superconducting order parameter , yielding a gap that is a combination of @xmath84- and @xmath85-wave contributions as obtained in our calculation . indeed , the penetration depth @xmath98 should display an anisotropy that is directly related to the superconducting order parameter . to make predictions for our proposed gap state , we calculate the penetration depth from the current - current correlator as follows@xcite : assuming that contributions from individual bands @xmath99 at the fermi level simply sum up in the correlator@xcite , we obtain for the current direction @xmath100 @xmath101 the corresponding result from a single band calculation is given by @xmath102 evaluating eq . ( [ eq_pen ] ) for @xmath93 @xmath90 points as in the calculation of the susceptibility in the superconducting state , we set @xmath103 and use a mean - field - like @xmath5 dependence of the superconducting order parameter , setting @xmath104.@xcite in the inset of fig . [ fig : pen ] we show the result for the two eigenvalues of the superfluid density tensor , @xmath97 , corresponding to current in the @xmath24 and @xmath23 directions , respectively . for the choice @xmath105 the nodal region can be probed with finite @xmath106 yielding a linear penetration depth at low @xmath5 . on the other hand , the absence of nodes in the @xmath24 direction is reflected in the missing linear term in the penetration depth such that the finite but small gap yields a power law similar to @xmath22 as expected when the direction of the current is perpendicular to the nodes@xcite . note that the crossover energy scale is set not by the maximum of the gap , but by the smaller minimal gap on the @xmath41-centered fermi surface . as discussed previously in ref . ( and confirmed in our calculations ) , the gap velocity @xmath107 can be safely neglected in the calculation . our model , based on fits to arpes and qo data , provides an explanation of the observed splitting of the knight shift and @xmath13 as seen in nmr experiments , is consistent with the @xmath5 and @xmath65 evolution of the ins data , and predicts a superconducting order parameter within the spin - fluctuation pairing scenario consistent with stm . the striking result is that the @xmath5 evolution of all these properties can be explained by introduction of a simple , phenomenological orbital order with mean - field like temperature dependence.@xcite of course we have not specified the physical origin of this orbital ordering here , but it is interesting to note that the system found to fit experiments appears to be very close to two competing magnetic instabilities , at @xmath2 , and @xmath49 , consistent with the proposal of glasbrenner _ et al_.@xcite that long - range magnetism is suppressed in this system by closely competing magnetic states . if this is true , the orbital ordering effect could indeed be driven by spin fluctuations in the high-@xmath5 tetragonal state . we remind the reader , however , that rather small changes in band structure led to a different magnetic response in our earlier work , so for the moment this conclusion should be treated with caution . all current microscopic theories of orbital ordering rely on an interplay of orbital fluctuations and spin fluctuations , and generally link the structural and magnetic transitions . recently , a theory of the nematic transition in fese without long - range magnetic order was proposed in refs . , , but it is not clear whether it yields electronic structure at low @xmath5 consistent with arpes and qo data . we now ask why orbital ordering favors low - energy spin fluctuations at @xmath2 . in our model , it appears that differences in the magnetic responses in the @xmath24 and @xmath23 directions below @xmath14 due to the lowered symmetry drive the system to a state where a @xmath2 instability competes very closely with a state at @xmath49 . in ref . it is argued that this competition suppresses the magnetic ordering temperature . the transfer of spectral weight from @xmath53 to @xmath2 with onset of orbital order as seen in fig . [ fig : staticchi ] is one of the consequences of increased nesting , e.g. , that @xmath108 is fulfilled for a larger set of @xmath90 points . unlike in a single - band picture where fermi surface nesting ( at zero energy ) automatically gives rise to a logarithmic singularity in the lindhard function , however , in multiband systems significant contributions to the susceptibility can arise from finite energy scattering . in general , this complicates the identification of dominant nesting vectors for the multiband case , and can lead to a situation where the susceptibility at two different momenta exhibits an opposite trend than expected from the fermi surface . looking for example at fig . [ fig : bandfscut](a ) a better fermi surface nesting at @xmath109 than at @xmath110 can be seen , but the susceptibility is larger at @xmath111 . ): the partial integral down to energies of @xmath112 does not accumulate large differences ( red and blue areas ) , but at low energies there is a significant contribution that makes the susceptibility at @xmath2 larger than at @xmath53 . [ fig_kk ] ] one way to determine the origin of the larger weight at @xmath111 is to use the kramers - kronig relation that connects the real and imaginary parts of the susceptibility @xmath113 and identify the relevant energies where the integrand on the right - hand - side largely contributes . one key question is the origin of the splitting of the @xmath2 and @xmath53 response . a plot of the difference of the integrand in eq . ( [ eq_kk ] ) as a function of the lower boundary @xmath114 reveals that high - energy fluctuations are not affected by the orbital order , but scattering at an energy transfer below @xmath115 gives rise to a larger susceptibility at @xmath2 as shown in fig . [ fig_kk ] . the major contributions are made up by scattering of states dominantly of @xmath116 orbital character with momentum transfer @xmath2 which is significantly larger than the corresponding scattering of the states of @xmath117 orbital character with momentum transfer @xmath53 . as expected from the electronic structure , the superconducting gap does not display tetragonal symmetry and is a linear combination of @xmath84- and @xmath85-wave symmetries allowing for accidental nodes . the calculated state shown in fig . [ fig_sc ] indeed has vertical line nodes on the x - centered electron pocket where the gap reaches zero such that the dos becomes v - shaped at low energies ; see fig . [ fig_sc](b ) . we note that the gap function @xmath88 changes sign on the extremely 2d x - centered electron pocket , but only over a very small region in momentum space as it can be seen from the areas in fig . [ fig_sc](a ) marked in black where the gapfunction is very small . in the absence of a sign change in the gap function @xmath88 for momentum transfer @xmath2 , we argue that the corresponding susceptibility should not show a strong resonance of the usual @xmath118 type , but rather a peak due to the proximity to the magnetic instability in the normal state which is then modified in the superconducting state by gapping out states at the fermi level ; see fig . [ fig : nresonance ] . the gap state with accidental nodal lines can be probed by the anisotropy of the penetration depth . unlike symmetry - imposed nodes as , for example , in the @xmath85-wave state of cuprate superconductors , the nodes only occur in the @xmath23 direction , but not in the @xmath24 direction . this is a simple consequence of eq . ( [ eq_pen ] ) where , for @xmath119 , the derivative @xmath120 at a nodal line vanishes , and all the other terms in the @xmath90 sum stem from fully gapped quasiparticle states . the authors of ref . have studied the distortion from the tetragonal state by strain and resulting movements of nodal points in the gap structure and proposed observable @xmath121 penetration depth anisotropies in consequence . in our case , the distortion of the electronic structure is intrinsic , and we emphasize that also the @xmath5 dependence of the penetration depth , which should be easier to observe , exhibits qualitative anisotropy arising from the unusual nodal structure of the gap function . fese is one of the most challenging and intriguing of the fe - based superconductors . in this work we began with a tight - binding model which provided a good fit to both arpes and qo data on this system both at high and low @xmath5 , but assuming a phenomenological orbital ordering . however the dynamical magnetic susceptibility derived from this model proved inconsistent with subsequent ins measurements , so we posed the question of whether small local adjustments to the tight - binding hamiltonian could preserve the good fit to one - particle properties while simultaneously fitting two - particle susceptibility within an rpa approach . our slightly modified model provided an equally good fit to arpes and qo data and explains many aspects of the ins measurements semiquantititatively such as the transfer of spectral weight from nel to stripe fluctuations and a paramagnon dispersion along @xmath2 to @xmath3 . while our approach does not constitute a complete microscopic theory , we have provided an internally self - consistent model of the electronic structure and spin excitations which evolves with @xmath5 according to experiment . this has the principal virtue of confirming that the main physical phenomenon driving the subtle evolution of the electronic excitations is orbital ordering . we have proposed that both site - centered and bond - centered orbital order of similar order of magnitude are required to explain the existing experimental data . in the superconducting state , we find an order parameter with accidental nodes leading to a v - shaped dos and an anisotropy in the penetration depth as well as an enhanced magnetic response at momentum transfer @xmath2 . we acknowledge our useful discussions with a. boothroyd , j. zhao , b. bchner , m. h. christensen , , and h. kontani . a. k. and b. m. a. acknowledge financial support from a lundbeckfond fellowship ( grant no . a9318 ) . p. j. h. was partially supported by the department of energy under grant no . de - fg02 - 05er46236 . at @xmath122 showing the band splitting due to orbital ordering where the main orbital character is indicated by the colors red @xmath117 , green @xmath116 , blue @xmath123 , yellow @xmath124 , and purple @xmath125 , and the thickness represents the magnitude of the main orbital character . ( b ) fermi surface for the same bands in the folded brillouin zone that yields the frequencies for quantum oscillations as shown in fig . [ fig : qo ] . ] for our band structure at low @xmath5 including spin - orbit coupling obtained using a numerical method@xcite where the error bars indicate the numerical uncertainty in the determination of the extremal orbits . ] the band structure parameters utilized in the current paper are very similar to those presented in the supplemental material of ref . . in the present study we have included additional shifts @xmath126 and @xmath127 since this leads to a better agreement with recent neutron scattering experiments@xcite . the momenta @xmath128 and @xmath129 are measured in units of the inverse lattice constants along the two orthorhombic crystal directions and refer to the brillouin zone corresponding to the iron lattice , except for fig . [ fig : bandfs ] where the 2-fe brillouin zone has been used . the properties in terms of fermi surface and quantum oscillation frequencies are mostly unchanged compared to those presented earlier ; see figs . [ fig : bandfs ] and [ fig : qo ] . the main effect of these quite small changes is to move the leading magnetic instability from @xmath3 in ref . to @xmath2 in the present case . note that it is necessary to include a spin - orbit coupling for capturing a band splitting such that only one hole - like pocket is present at the @xmath41 point at high @xmath5@xcite ; see fig . [ fig : bandfscut ] . for the low-@xmath5 electronic structure where the bands are split by the orbital order , the spin - orbit coupling has little effect and is neglected for the calculation of physical quantities at low @xmath5 for simplicity . ( left ) and @xmath130 ( right ) for @xmath122 ( top ) and @xmath131 ( bottom ) in the presence of spin - orbit coupling . color code as in fig . [ fig : bandfs ] . in ( a ) it is visualized that the fermi surface nesting for momentum transfer @xmath132 is larger than for @xmath111 . ] the local interactions are included via the hubbard - hund hamiltonian @xmath133 where we set @xmath134 , @xmath135 to restrict ourselves to the spin - rotation - invariant case . dependence of the knight shift showing the splitting of the knight shift signal below @xmath14 . here @xmath136 and @xmath137 are the two orthorhombic domains and the @xmath138axis is perpendicular to the fese plane . ( b ) spin lattice relaxation rate @xmath139 calculated without the structure factor showing the strong upturn below @xmath140 . inset : @xmath139 calculated with the structure factor screening calculated with the correct expression for @xmath141 in eq . ( [ eq_nmr ] ) showing the effects of twin orthorhombic domains . ] nmr experiments in fese have probed the @xmath142se atom and mapped out both the @xmath5 dependence of the knight shift and the low - energy spin fluctuations through measurements of the spin lattice relaxation rate ( @xmath139)@xcite . both the knight shift and @xmath139 are split below the structural transition due to contributions from the two orthorhombic domains in twinned fese samples . it is found that spin fluctuations measured by @xmath139 are not affected by the structural transition itself but are enhanced at much lower @xmath5 . we have calculated the nmr knight shift as @xmath143 with a constant hyperfine form factor @xmath144 above the structural transition and an order - parameter - like @xmath5 dependence in the orbitally ordered state@xcite . as seen in fig [ fig : nmr](a ) by fitting the magnitude of the form factor we get a good agreement with knight shift measurements similar to the results presented in ref . . the @xmath5 dependence of the spin lattice relaxation rate @xmath139 can be evaluated from the expression , @xmath145 here @xmath141 is the @xmath50-dependent form factor , @xmath146 indices run over the cartesian coordinates . in the paramagnetic state the form factor matrix is diagonal implying that @xmath147 . in the orthorhombic phase , we have @xmath148 . assuming a nearest neighbor interaction between the nuclear spin of the @xmath142se atom and the iron conduction electrons , we express the form factor as @xmath149 . we obtain a good agreement with both the knight shift and @xmath139 experiments for @xmath150 $ ] , @xmath151 , where @xmath152 , @xmath153 , and @xmath37 is a mean - field @xmath5 dependence which is zero above the structural transition@xcite . as can be seen from fig . [ fig : nmr](b ) the spin lattice relaxation rate is enhanced at low @xmath5 compared to the structural transition . 49ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1038/nmat2491 [ * * , ( ) ] http://stacks.iop.org/0256-307x/29/i=3/a=037402 [ * * , ( ) ] http://dx.doi.org/10.1038/nmat3654 [ * * , ( ) ] http://dx.doi.org/10.1038/nmat4153 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.180505 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.237001 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.220506 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.121111 [ * * , ( ) ] link:\doibase 10.1103/physrevb.91.155106 [ * * , ( ) ] link:\doibase 10.1103/physrevb.91.214503 [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.205117 [ * * , ( ) ] @noop ( ) , link:\doibase 10.1103/physrevb.90.144517 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.026402 [ * * , ( ) ] http://dx.doi.org/10.1038/nmat4138 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.027001 [ * * , ( ) ] link:\doibase 10.1103/physrevb.91.180501 [ * * , ( ) ] http://dx.doi.org/10.1038/nmat4492 [ * * ( ) ] , @noop ( ) , link:\doibase 10.1103/physrevb.86.060412 [ * * , ( ) ] link:\doibase 10.1038/nphys2877 [ * * , ( ) ] http://dx.doi.org/10.1038/nphys3434 [ * * , ( ) ] , http://dx.doi.org/10.1038/nphys3456 [ * * , ( ) ] , link:\doibase 10.1103/physrevb.91.201105 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.116401 [ * * , ( ) ] link:\doibase 10.1016/j.crhy.2015.03.001 [ ( ) ] link:\doibase 10.1103/revmodphys.87.855 [ * * , ( ) ] http://dx.doi.org/10.1038/nphys3594 [ * * ( ) ] , link:\doibase 10.1103/physrevb.78.020514 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.177005 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.064517 [ * * , ( ) ] link:\doibase 10.7566/jpsj.84.063701 [ * * , ( ) ] , ed . , link:\doibase 10.1107/97809553602060000103 [ _ _ ] ( , ) link:\doibase 10.1103/physrevb.92.104505 [ * * , ( ) ] link:\doibase 10.1126/science.1202226 [ * * , ( ) ] link:\doibase 10.1073/pnas.1413477111 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.134520 [ * * , ( ) ] link:\doibase 10.1103/physrevb.70.054510 [ * * , ( ) ] http://stacks.iop.org/0953-8984/22/i=18/a=185704 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.014524 [ * * , ( ) ] link:\doibase 10.1007/bf01303700 [ * * , ( ) ] @noop ( ) , @noop ( ) , link:\doibase 10.1103/physrevlett.113.217001 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.cpc.2011.10.015 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.085110 [ * * , ( ) ] http://stacks.iop.org/1367-2630/16/i=8/a=083025 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.214514 [ * * , ( ) ]	we present a theoretical study of the dynamical spin susceptibility for the intriguing fe - based superconductor fese , based on a tight - binding model developed to account for the temperature - dependent band structure in this system . the model allows for orbital ordering in the @xmath0 channel below the structural transition and presents a strongly @xmath1 symmetry broken fermi surface at low temperatures which accounts for the nematic properties of this material . the calculated spin excitations are peaked at wave vector @xmath2 in the 1-fe brillouin zone , with a broad maximum at energies of order a few mev . in this range , the occurrence of superconductivity sharpens this peak in energy , creating a @xmath2 `` neutron resonance '' as seen in recent experiments . with the exception of the quite low energy scale of these fluctuations , these results are roughly similar to standard behavior in fe pnictide systems . at higher energies , however , intensity increases and shifts to wave vectors along the @xmath2 - @xmath3 line . we compare with existing inelastic neutron experiments and nmr data , and give predictions for further studies .
recently we proposed a new algorithm for computation of cohomology of a wide class of lie superalgebras . this algorithm reduces the computation for the whole cochain complex to a number of smaller tasks within smaller subcomplexes . one can demonstrate that if @xmath7 is the computation time for the whole complex , then partition of the complex into @xmath8 subcomplexes reduces the computation time roughly to the value @xmath9 . thus , the approach appeared to be efficient enough to cope with several difficult tasks in computing cohomology for particular lie ( super)algebras @xcite . more detailed experiments with the _ * c * _ implementation of the algorithm , including profiling , reveal that arithmetic operations over @xmath2 take the main part of computation time ( usually more than 90% for large tasks ) . the same is true if @xmath2 is replaced by @xmath10 ( though computation becomes somewhat faster ) . a standard way to reduce negative influence of this bottleneck is to compute several modular images of the problem with subsequent restoring the result over @xmath10 or @xmath2 by the chinese remaindering or an algorithm for recovering a rational number from its modular residue @xcite . though , as is clear , the sum of sizes of modules used for constructing images can not be less than the size of maximum integer in the final result , the modular approach allows the intermediate swelling of coefficients to be avoided . moreover the use of modular images is much more advantageous in the case of ( co)homology computation than in the traditional problems of linear algebra . as we demonstrate further , the overwhelming part of computation can be accomplished using only one modular image . recall that the gauss elimination , the basic constituent of algorithms for computing ( co)homology , over @xmath11 has only _ cubic _ computational complexity in contrast to the _ exponential _ one over @xmath2 or @xmath12 note that the approach presented here can be applied not only for the lie superalgebras , but in more general case of computation of homology or cohomology , especially when there is a practical method of splitting ( co)chain complex into smaller subcomplexes . to demonstrate the power of the new algorithm and program , we present the results of computation of cohomology in the trivial module for the algebra @xmath13 this is an example ( for @xmath14 ) of the lie superalgebra of special ( divergence free ) vector fields on the @xmath15-dimensional supermanifold preserving the odd version of symplectic ( periplectic , as a. weil called it ) structure @xcite . these superalgebras , being super counterparts of the lie algebras of hamiltonian vector fields , are vital in the batalin vilkovisky formalism , see @xcite . the @xmath16th cohomology is defined as the quotient group @xmath17 for the _ cochain complex _ @xmath18 here , the @xmath19 are abelian groups of cochains , graded by the integer @xmath16 ( called _ dimension _ or _ degree _ ) ; the @xmath20 are differentials ( @xmath21 ) ; the @xmath22 and @xmath23 are the subgroups of cocycles and coboundaries , respectively ( see @xcite for details ) . in order to apply without restrictions the linear algebra algorithms , we assume that the groups of cochains are additive groups of certain linear spaces or modules and we shall use the corresponding terms in the subsequent text . to compute the @xmath16th cohomology , it suffices to consider the following part of ( [ whole ] ) : @xmath24 first of all we split ( [ part ] ) using the @xmath10-grading in the cochain spaces induced by the gradings in the lie ( super)algebra ( and the module over this algebra ) involved in the construction of the cochain spaces : @xmath25 here , @xmath26 is a grading subset . it appears that , as a rule , any subcomplex in a given degree @xmath27 can be split , in turn , into smaller subcomplexes : @xmath28 here @xmath29 is a finite or infinite set of subcomplexes . equation ( [ spart ] ) means that the spaces @xmath30 split into the direct sum of subspaces @xmath31 and the matrices of the linear mappings @xmath32 can be represented in the block - diagonal form @xmath33 the construction of these subcomplexes is the _ central part _ of the splitting algorithm . thus , the whole task reduces to a collection of easier tasks of computing @xmath34 as a basis of the cochain space @xmath35 we choose the set of super skew - symmetric monomials of the form @xmath36 here , @xmath37 and @xmath38 are basis elements of the algebra and module , respectively , and @xmath39 is the dual element to @xmath40 that is , @xmath41 the degrees of factors in ( [ monom ] ) satisfy the relation @xmath42 notice that @xmath43 and this is a serious obstacle to extraction of finite - dimensional subcomplexes for infinite - dimensional lie ( super)algebras when computing cohomology in the adjoint module ( important in the deformation theory ) . we also assume that @xmath44 to construct a subcomplex @xmath45 from the sum in right hand side of ( [ spart ] ) , we begin with choosing somehow an arbitrary _ starting _ monomial @xmath46 of the form ( [ monom ] ) . there are various choices of the starting monomial and the time and space efficiency of computation depends on the choice . having no better idea , we use at present the following three strategies : choice of a lexicographically minimal , or lexicographically maximal , or random monomial . we call these strategies _ bottom _ , _ top _ and _ random _ , respectively . among these , the top strategy seems to be most efficient ( see experimental data in tables [ timesh ] and [ timessle ] ) and it is used by default . nevertheless , other strategies help sometimes to push through difficult tasks when the top strategy fails . then we construct the three sets @xmath47 @xmath48 and @xmath49 of basis monomials for @xmath50 @xmath51 and @xmath52 respectively , by the procedure _ * constructsubcomplex * _ presented on page . constructsubcomplex[constructsubcomplex ] * input : * @xmath46 , starting @xmath53-monomial + * output : * @xmath54 , @xmath55 , @xmath49 , monomial bases of cochain spaces in the current + subcomplex @xmath56 : @xmath57 such that @xmath58 + * local : * @xmath59 @xmath60 , @xmath61 , working subsets + of currently `` new '' ( not yet processed ) monomials ; + @xmath62 , @xmath63 , @xmath64 , working sets of monomials ; + @xmath65 , @xmath66 , @xmath67 , working monomials + * initial setting : * + 1 : @xmath68 + 2 : @xmath69 + 3 : @xmath70 + * loop over @xmath16-monomials : * + 4 : _ * while * _ @xmath71 _ * do * _ + 5 : @xmath72 + * supplement the set @xmath54 * : + 6 : @xmath73 + 7 : @xmath74 + 8 : @xmath75 + * supplement the set @xmath49 * : + 9 : @xmath76 + 10 : @xmath77 + 11 : @xmath78 + * exclude processed monomial @xmath66 * : + 12 : @xmath79 + * loop over @xmath80-monomials : * + 13 : _ * while * _ @xmath81 _ * do * _ + 14 : @xmath82 + * supplement the set @xmath83 * + 15 : @xmath84 + 16 : @xmath85 + 17 : @xmath86 + * exclude processed monomial @xmath87 * + 18 : @xmath88 + 19 : _ * od * _ + * loop over @xmath89-monomials : * + 20 : _ * while * _ @xmath90 _ * do * _ + 21 : @xmath91 + * supplement the set @xmath55 : * + 22 : @xmath92 + 23 : @xmath85 + 24 : @xmath93 + * exclude processed monomial @xmath65 : * + 25 : @xmath94 + 26 : _ * od * _ + 27 : _ * od * _ + 28 : _ * return * _ @xmath54 , @xmath55 , @xmath49 + the function _ * takemonomialfromset * _ called within _ * constructsubcomplex * _ takes the current monomial from a set of monomials . the function _ * inverseimagemonomials * _ generates the set of @xmath95-monomials whose images with respect to the mapping @xmath96 contain a given @xmath97-monomial . the function _ * imagemonomials * _ generates the set of @xmath98-monomials whose inverse images with respect to the @xmath99 contain a given @xmath97-monomial . in the finite - dimensional case , the loops in the procedure _ * constructsubcomplex * _ are finite and we obtain in the end a minimal subcomplex of the form ( [ subcomplex ] ) . this is the unique minimal subcomplex involving the starting monomial @xmath46 . let us consider in more detail the procedure of computation of cohomology within the subcomplex in accordance with formula ( [ insubcomplex ] ) . from now on we assume that @xmath100 and @xmath101 in ( [ subcomplex ] ) are finite - dimensional spaces over @xmath2 or @xmath11 or modules over @xmath12 since important in mathematics and physics fields @xmath102 and @xmath103 are , in principle , non - algorithmic objects , our main interest will be focused on the cohomology over the field @xmath2 ( or its algebraic extentions ) . in accordance with a general theorem in the homological algebra , called the _ universal coefficient theorem _ @xcite , ( co)homology with coefficients from an arbitrary abelian group @xmath104 can be expressed in terms of ( co)homology with coefficients in @xmath10 . thus , we can carry out the computation over @xmath10 and then go to the coefficient group we are interested in . let us consider now the connection between @xmath105 and @xmath106 in the finite - dimensional case , the group @xmath107 is a finitely generated abelian group having the following canonical representation @xmath108 here , @xmath109 , the number of copies of the integer group @xmath10 , is called the _ rank _ of the abelian group @xmath107 or the _ betti number_. the cyclic groups @xmath110 are called the _ torsion subgroups _ and their orders @xmath111 , having the property @xmath112 @xmath113 @xmath114 and so on , are called the _ torsion coefficients_. in the case of cohomology , the universal coefficient theorem is expressed by the following split short exact sequence @xmath115 where the operation @xmath116 is the _ periodic product _ of abelian groups . in our context @xmath117 the term `` split '' , in application to sequence ( [ exact ] ) , means the possibility to construct the isomorphism @xmath118 replacing @xmath104 by @xmath2 in ( [ isomorphism ] ) and taking into account that @xmath119 for any abelian group @xmath120 , we have @xmath121 . since @xmath122 for arbitrary @xmath123 and @xmath124 the dimension of @xmath125 interpreted as vector space over @xmath2 , coincides with the rank ( betti number ) @xmath109 of the group @xmath106 our modular approach is based on the following [ theorem ] @xmath126 * remarks : * 1 . inequality ( [ inequality ] ) means that non - trivial cohomology classes computed over the field of rational numbers @xmath2 can exist only in the subcomplexes with non - trivial cohomology classes computed over the finite field @xmath11 with arbitrary prime @xmath127 2 . h. khudaverdian turned author s attention to the fact that inequality ( [ inequality ] ) can be deduced immediately from the universal coefficient theorem : considering the product @xmath128 and taking into account representation ( [ abelgroup ] ) and isomorphism @xmath129 , we see that the dimension of @xmath130 , as a vector space over @xmath11 , can not be less than @xmath109 ( only additional dimensions may appear , if the torsions in @xmath107 or @xmath131 contain cyclic groups of the form @xmath132 ) . + nevertheless , we give here a direct constructive proof in order to demonstrate in parallel the main ideas of ( co)homology computation . _ proof . _ to prove inequality ( [ inequality ] ) , we have to compute ( [ insubcomplex ] ) in such a way as to avoid cancellations of integers and apply the modular homomorphism @xmath133 at the end of computation . thus , it is convenient to consider ( [ insubcomplex ] ) over @xmath10 instead of @xmath134 we assume that @xmath135 is odd and use a symmetric representation of @xmath11 , i. e. , @xmath136 we will also apply @xmath137 component - wise to multicomponent objects over @xmath138 like vectors and matrices . + we begin with the following setup : * @xmath139 and @xmath140 are represented as finite - dimensional modules @xmath141 respectively , i. e. , @xmath142 + * the differentials @xmath143 and @xmath144 are represented ( in the monomial bases of the form ( [ monom ] ) , in our case ) as integer @xmath145 and @xmath146 matrices @xmath147 d = d = , @xmath147 respectively . we write @xmath148 \end{tabular } } \hspace{-7pt}\stackrel{\mbox{\scriptsize{$j$}}}{a } } $ ] to indicate that matrix @xmath120 has @xmath149 rows and @xmath150 columns . the matrices @xmath151 and @xmath152 satisfy the relation @xmath153 the computation of cohomology , i.e. , construction of quotient module , can be reduced to the construction of so - called _ ( co)homology decomposition _ @xcite based on the computation of the smith normal forms @xcite of the matrices representing differentials . now we should consider the coboundary submodule @xmath168 combining the relation @xmath169 with the structure of the matrix @xmath29 ( see formula ( [ s0 ] ) ) , we can reduce the matrix @xmath152 determining coboundaries to the matrix @xmath170 acting in the submodule of cocycles : @xmath147 v^-1d = . @xmath147 computing the smith normal form @xmath171 for the reduced coboundary matrix we get @xmath147 s = = , @xmath147 where @xmath172 and @xmath173 \end{tabular } } \hspace{-8pt}\stackrel{\mbox{\scriptsize{$~r'$}}}{\widetilde{s ' } } } = \mathrm{diag}\left(s'_1,\ \ldots\ , \ s'_{r'}\right).$ ] we can extend the transformation matrix @xmath174 \end{tabular } } \hspace{-10pt}\stackrel{\mbox{\scriptsize{$m\!-\!r$}}}{\widetilde{u ' } } } $ ] acting in the submodule of cocycles to the transformation matrix acting in the whole module @xmath175 @xmath147 u = = . @xmath147 here , @xmath159 \end{tabular } } \hspace{-8pt}\stackrel{\mbox{\scriptsize{$\,r$}}}{i } } $ ] is the @xmath176 identity matrix . using the transformation matrices @xmath177 and @xmath178 we can transform the initial ( monomial in our case ) basis @xmath179 in the module @xmath180 into the basis @xmath181 making the cohomology decomposition explicit @xmath182 in this decomposition we have @xmath183 and @xmath184 the formula for the dimension of cohomology ( betti number ) follows from decomposition ( [ decomposition ] ) @xmath185 now let us consider how ( [ dimension ] ) changes under @xmath186 the image of ( [ dimension ] ) takes the form @xmath187 since @xmath137 is a ring homomorphism , we have for arbitrary unimodular matrix @xmath120 with integer entries @xmath188 that is @xmath137 maps the above transformation matrices into invertible matrices . hence the number of elements in the decomposition basis @xmath189 remains unchanged , @xmath190 on the other hand , the invariant factors @xmath191 and @xmath192 of the matrices @xmath193 and @xmath29 divisible by @xmath135 vanish , hence @xmath194 and inequality ( [ inequality ] ) is proved by comparing ( [ dimension ] ) and ( [ dimensionp ] ) . @xmath195 an algorithm based on the above ideas was implemented in the _ c * _ language . the program called _ * liecohomologymodular * _ has the following structure : 1 . input lie ( super)algebra @xmath120 , module @xmath196 over @xmath120 , cohomology degree ( dimension ) @xmath16 and grade @xmath27 . @xmath120 and @xmath196 should be defined over ( some algebraic extension of ) @xmath10 or @xmath134 2 . construct the full set @xmath197 of @xmath16-cochain monomials in grade @xmath27 . 3 . choose a prime @xmath135 for searching subcomplexes with non - trivial cocycles by computing over @xmath11 . [ choosemonom ] choose an element @xmath198 ( the starting monomial ) . 5 . construct a minimal subcomplex @xmath199 such that @xmath200 6 . compute @xmath201 . [ computeh ] if @xmath202 then compute @xmath203 over @xmath10 or @xmath2 ( or their extensions ) . we can use here the chinese remaindering or the rational recovery algorithm as more efficient procedures than direct computation over @xmath10 or @xmath2 . delete all basis monomials of @xmath51 from @xmath204 9 . if @xmath197 is empty , then stop computation , otherwise go to step [ choosemonom ] . in the current implementation we obtain the relations determining cocycles and coboundaries ( in fact , the rows of matrices of differentials ) within the procedure _ * constructsubcomplex_. these relations are generated one by one as by - product of the functions _ inverseimagemonomials * _ and _ * imagemonomials_. to prevent unnecessary memory consumption , every newly arising relation is reduced modulo the system of relations existing to the moment and , if the result is not zero , the new relation is added to the system . thus , we automatically have the matrices of differentials in the normal form just after completion of the procedure _ constructsubcomplex_. this process is obviously equivalent to the gauss elimination method , the most standard method for the computation of the smith normal form of a matrix . in recent years , a number of new fast algorithms for the determination of smith normal form have been elaborated @xcite . these algorithms appear to be well suited to the ( co)homology computation . it is worth to study the possibility to incorporate these algorithms in our implementation . we could , for example , remove the generation of relations from the functions _ inverseimagemonomials * _ and _ * imagemonomials * _ making them as fast as possible . then , after construction of subcomplex with the help of these modified functions , we should generate the matrices of differentials separately and apply the fast algorithms to these matrices . of course , this modification should be done if the total computation time decreases without substantial increase in the memory consumption . the works @xcite contain a detailed analysis of the properties of the sets of primes most appropriate for application of modular algorithms to a given matrix . here we give only a few comments concerning the choice of prime @xmath135 in our algorithm . these comments are based mainly on experiments with the program . from the practical point of view , we should use only primes @xmath135 for which all operations in @xmath11 can be done within one machine word . thus , for 32bit architecture we should choose @xmath135 from the set of 8951 primes @xmath205 a good choice should not produce excessive cocycles . of course , such cocycles will be removed at step [ computeh ] anyway , but at the expense of additional work . in our context , an `` unlucky '' prime is the one which divides the invariant factors of matrices of differentials ( or , in other words , the torsion coefficients of cohomology over @xmath10 ) and , as is clear , the probability for a given prime to be unlucky diminishes as the prime grows . on the other hand , there is an increase of time ( in the examples we have computed , up to factor 2 or 3 ) and space expenditures with increase of @xmath135 within the set @xmath206 so it makes sense not to use too large primes for searching subcomplexes with potentially non - trivial cohomology . in our practice , we use , as a rule , a compromise : a prime near the half of 32bit word , namely , @xmath207 i. e. , the 4th fermat number . however , quite satisfactory results can be obtained even with much smaller primes , as is illustrated in table [ tab : dimh ] . the symbols @xmath208 in the boxes of this table mean that ( non - zero ) @xmath209 , whereas @xmath210 means the cohomology with coefficients in the trivial module for the lie algebra @xmath211 of hamiltonian vector fields on the @xmath212-dimensional symplectic manifold . we performed computation over all modular fields @xmath213 through @xmath214 as is seen in table [ tab : dimh ] , the results for @xmath215 fully coincide with those for @xmath2 for all computed grades @xmath216 $ ] ( and for all cohomology degrees @xmath16 ) . the table also illustrates theorem [ theorem ] , i.e. , all boxes containing non - zero dimensions for the field @xmath2 contain also non - zero ( @xmath217 same for @xmath2 ) dimensions for all fields @xmath11 considered . in tables [ timesh ] and [ timessle ] we present ( considering both algebra and superalgebra cases ) the running times for computation over @xmath215 of cohomology @xmath218 ( for @xmath219 ) and @xmath220 ( for @xmath221 ) . the columns presented in these tables are : the dimensions of cochain spaces , i. e. , the sizes of matrices of differentials ; the running times in seconds for the _ top _ strategy of the choice of the starting monomial and comparison of the _ bottom _ and _ top _ strategies . the times in both tables were obtained on a 1133mhz pentium iii pc with 512 mb . note that the maximum memory consumption is near 46 mb and near 14 mb for the tasks in table [ timesh ] and in table [ timessle ] , respectively . .timing for @xmath222 , @xmath223 [ cols="^,^,^,^,^,^",options="header " , ] & @xmath224 + when computing cohomology we start with the construction of the full set of @xmath53-monomials . at the moment we do not see how to avoid this in deterministic algorithms . to represent the set of monomials , we need to allocate @xmath225 elements of memory representing basis elements of algebra and module ( @xmath226 for the trivial and @xmath227 for any non - trivial module ) . in our implementation we represent basis elements by two - byte integers . for the last box in column 9 ( @xmath228 ) of table [ tab : h1 ] we have @xmath229 and the set of monomials occupies near 12 mb . the dimensions grow very rapidly , so , in fact , we are working on the brink of abilities of 32bit architecture and , even theoretically , we can add only a few rows to table [ tab : h1 ] . but , as to arithmetical difficulties , we have some progress . i would like to thank h. khudaverdian , d. leites and i. shchepochkina for helpful comments on drafts of this text . 99 kornyak , v.v . : a new algorithm for computing cohomologies of lie superalgebras . in : _ computer algebra in scientific computing / casc01 _ , v.g.ganzha , e.w.mayr and e.v.vorozhtsov ( eds . ) , springer - verlag berlin heidelberg ( 2001 ) 391398 kornyak , v.v . : a method of splitting cochain complexes to compute cohomologies of lie ( super)algebras . _ russian journal for computer science ( `` programmirovanie '' ) . _ * 2 * ( 2002 ) 7680 ( in russian ) kornyak , v.v . : extraction of `` minimal '' cochain subcomplexes for computing cohomologies of lie algebras and superalgebras . in : _ computer algebra and its application to physics / caap-2001 _ , v.p . gerdt ( ed . ) , jinr dubna ( 2002 ) 186195 kornyak , v.v . : computation of cohomology of lie algebra of hamiltonian vector fields by splitting cochain complex into minimal subcomplexes . in : _ computer algebra in scientific computing / casc02 _ , v.g.ganzha , e.w.mayr and e.v.vorozhtsov ( eds . ) , springer - verlag berlin heidelberg ( 2002 ) 201206 kornyak , v.v . : a method of splitting cochain complexes for computing cohomology : lie algebra of hamiltonian vector fields @xmath230 _ russian journal for computer science ( `` programmirovanie '' ) . _ * 2 * ( 2003 ) 9499 ( in russian ) wang , p.s . , guy , m.j.t . , davenport , j.h . : @xmath135-adic reconstruction of rational numbers . _ sigsam bulletin , _ * 16 * ( 1982 ) 23 collins , g.e . , encarnacin , m.j . : efficient rational number reconstruction . _ j. symb . comp . , _ * 20 * ( 1995 ) 287297 leites , d. : lie superalgebras . in : _ modern problems of mathematics . recent developments _ , * 25 * , viniti , moscow ( 1984 ) p. 3 ( in russian ; english translation in josmar * 30(6 ) * ( 1985 ) p. 2481 ) leites , d. , shchepochkina , i. : _ classification of simple lie superalgebras of vector fields _ , preprint mpim-2003 - 28 ( www.mpim-bonn.mpg.de ) gomis j. , pars j. , samuel s. : antibracket , antifields and gauge - theory quantization . * 259 * ( 1995 ) , no . 12 , 145 pp fuks , d.b . : _ cohomology of infinite dimensional lie algebras . _ consultants bureau , new york ( 1987 ) fomenko , a.t . , fuks , d.b . : _ a course in homotopic topology . nauka ( 1989 ) . ( in russian ; an english translation of an earlier version was published by akadmiai kiad , budapest ( 1986 ) ) hilton , p. j. , wylie , s. : _ homology theory : an introduction to algebraic topology . _ cambridge university press , new york ( 1960 ) mac lane , s. , birkhoff , g. _ algebra . _ chelsea publishing co. , new york , third edition ( 1988 ) dumas , j - g . , saunders , b.d . , villard , g. : on efficient sparse integer matrix smith normal form computations . _ j. symb . comp . _ * 32(12 ) ( 2001 ) 7199 dumas , j - g . , heckenbach , f. , saunders , b.d . , welker , v. : computing simplicial homology based on efficient smith normal form algorithms . in _ algebra , geometry and software systems , _ m. joswig , n. takayama ( eds . ) springer ( 2003 ) 177206 batalin , i.a . , vilkovisky , g.a . : gauge algebra and quantization . lett . _ 102b * ( 1981 ) 2731 batalin , i.a . , vilkovisky , g.a . : closure of the gauge algebra , generalized lie equations and feynman rules . phys . _ * b234 * ( 1984 ) 106124 shander , v.n . : analogues of the frobenius and darboux theorems for supermanifolds . _ comptes rendus de l academie bulgare des sciences , _ * 36 * , no . 3 ( 1983 ) 309311 khudaverdian , h. m. : laplacians in odd symplectic geometry . in _ quantization , poisson brackets and beyond _ , th . voronov , ( ed . ) , contemp . , v.15 , amer . math . soc . , providence , ri ( 2002 ) 199212	we describe an essential improvement of our recent algorithm for computing cohomology of lie ( super)algebra based on partition of the whole cochain complex into minimal subcomplexes . we replace the arithmetic of rational numbers or integers by a much cheaper arithmetic of a modular field and use the inequality between the dimensions of cohomology @xmath0 over any modular field @xmath1 and over @xmath2 : @xmath3 . with this inequality we can , by computing over arbitrary @xmath4 quickly find the ( usually , rare ) subcomplexes for which @xmath5 and then carry out the full computation over @xmath2 within these subcomplexes . we also present the results of application of the corresponding _ * c * _ program to the lie superalgebra of special vector fields preserving an `` odd - symplectic '' structure on the @xmath6-dimensional supermanifold . for this algebra , we found some new basis elements of the cohomology in the trivial module .
a _ ripplet _ is a function @xmath2 whose integer translate are _ totally positive _ @xcite , i.e. for any ordered real numbers @xmath3 , and any ordered integers @xmath4 , @xmath5 , it holds @xmath6 total positivity implies that the integer translates of @xmath2 are _ variation diminishing _ , i.e. for any finite sequence @xmath7 @xmath8 where @xmath9 denotes the strict sign changes of its argument . the disequality ( [ var.dim.1 ] ) in turn implies that the system @xmath10 has shape - preserving properties , which are known to play a crucial role in several applications , from approximation of data to cagd @xcite . the concept of a ripplet was first introduced by goodman and micchelli in + @xcite , where the authors focused their interest on _ two - scale refinable ripplets _ , i.e. ripplets that are solution of a _ two - scale refinable equation _ @xmath11 where the _ scaling mask _ @xmath12 is a suitable real sequence . well known examples of refinable ripplets are the cardinal b - splines , i.e. the polynomial b - splines on integer nodes . starting from the seminal paper of goodman and micchelli , many families of two - scale refinable ripplets were constructed ( see , for instance , @xcite , @xcite ) . more recently , @xmath13-scale refinable ripplets , with dilation @xmath13 greater than 2 , were addressed in @xcite and refinable ripplets with dilation 3 were constructed in @xcite ( see also @xcite ) . the interest in refinable ripplets lies in the fact that they give rise to convergent subdivision algorithms for the reconstruction of curves and the limit curves they generate preserve the shape of the initial data @xcite . refinable ripplets can also be proved to solve the cardinal interpolation problem @xcite . for instance , the construction of cardinal interpolants by by means of the refinable ripplets in @xcite was addressed in @xcite . refinable ripplets have good properties not only in the context of geometric modeling and function approximation but they also enjoy some optimality properties useful in signal processing . in fact , refinable ripplets induces a multiresolution analysis in @xmath14 that allows us to generate a nested sequence of wavelet spaces . actually , it is always possible to construct compactly supported semiorthogonal wavelet bases starting from a refinable ripplet @xcite . moreover , refinable ripplets have asymptotically , i.e. when their smoothness tends to infinity , the same optimal time - frequency window achieved by the gaussian function @xcite . since the rate of convergence can be proved to be very fast for a large class of ripplets , including , for instance , the refinable ripplets introduced in @xcite , refinable ripplets can approximate the gaussian with high accuracy giving rise to efficient algorithms for signal analysis @xcite . finally , a ripplet can be seen as a discrete kernel satisfying a causality property so making the refinable ripplets particularly attractive in the scale - time analysis of signals @xcite . all the refinable ripplets quoted above are _ stationary _ in the sense that they satisfy the functional equation ( [ stazref ] ) with the same mask sequence at each dyadic scale . for this reason usually ( [ stazref ] ) is referred to as a _ stationary two - scale equation_. from the point of view of signal processing this means that the same analysis and synthesis filters are used at all dyadic scales @xcite . nevertheless , the use of the same set of filters at each scale does not give great flexibility in applications , especially when some preprocessing steps with different filters are required . from the functional point of view , the use of different filters at different scales gives rise to a _ nonstationary multiresolution analysis _ ( see , for instance , @xcite , @xcite , @xcite , @xcite ) . a nonstationary multiresolution analysis can be generated by a set of _ nonstationary refinable functions _ , i.e. an infinite set of functions @xmath15 satisfying an infinite set of _ nonstationary two - scale equations _ @xmath16 for some infinite sequence of masks @xmath17 , each mask @xmath18 being different at each dyadic scale . the use of different scaling masks at different scales allows us to construct refinable functions endowed with properties that can not be achieved in the stationary setting . for instance , the nonstationary process generated by the b - spline masks of increasing support gives rise to a refinable function that is compactly supported while belongs to @xmath19 @xcite . other families of @xmath19 nonstationary refinable functions based on pseudo - spline masks were constructed in @xcite . exponential splines too can be associated to a nonstationary process @xcite . they reproduce exponential polynomials and the associated wavelet bases can be successfully used in the analysis of signals with exponential behavior @xcite . families of nonstationary refinable ripplets were introduced in @xcite and @xcite . in particular , the latter are highly localized in the scale - time plane , a property that is crucial in many applications . for this reason , in the present paper we focus our interest in this family of nonstationary refinable ripplets and prove that these ripplets enjoy several properties which are relevant in the context of both geometric modeling and signal processing . then , we construct the associated wavelet bases . we notice that , since the refinable ripplets we are considering are non orthogonal , compactly supported orthogonal wavelets belonging to the space generated by their translates do not exists . this motivates us to construct nonstationary _ semiorthogonal prewavelets_. in fact , giving up the orthogonality condition we can build wavelet bases with compact support . moreover , we give a procedure to construct the nonstationary compactly supported biorthogonal bases associated with the nonstationary refinable ripplets we are considering . since we are interested in implementing efficient nonstationary decomposition and reconstruction algorithms for signal processing , we construct also the corresponding pairs of nonstationary filters to be used in the analysis and synthesis of a given discrete signal . the paper is organized as follows . in section 2 we give some basic definition concerning nonstationary multiresolution analysis and wavelet spaces and recall some results about the existence of solutions of the nonstationary refinable equations ( [ nseq_int ] ) . the class of nonstationary refinable functions we are interested in is described in section 3 , while in section 4 we proved some approximation properties that were not addressed in @xcite . in section 5 we construct the nonstationary prewavelet bases associated with the nonstationary refinable ripplets in the class and give an efficient algorithm to evaluate nonstationary prewavelet masks . the construction of compactly supported biorthogonal bases and filters , which give rise to efficient decomposition and reconstruction formulas for discrete signals , is addressed in section 6 . finally , in section 7 some examples of nonstationary masks and refinable bases are given and the corresponding nonstationary filters are constructed . a simple test on the analysis of a spike - like signal is also shown . wavelet spaces are usually constructed starting from a multiresolution analysis that is a sequence @xmath20 of nested subspaces which are dense in @xmath14 and enjoy the separation property . thus , the corresponding wavelet space @xmath21 is defined as the orthogonal complement of @xmath22 in @xmath23 @xcite . in the stationary case all the spaces @xmath24 are generated by the dilates and translates of a unique refinable function @xmath25 , solution of the stationary two - scale equation ( [ stazref ] ) , while the spaces @xmath26 are generated by the dilates and translates of a unique _ wavelet _ @xmath27 . in contrast , in the nonstationary setting any space @xmath22 ( resp . @xmath26 ) is generated by the @xmath28-shifts of a different refinable function @xmath29 ( resp . wavelet @xmath30 ) , which are not dilates of one another . as a consequence , the spaces @xmath20 and @xmath31 are not scaled versions of the spaces @xmath32 and @xmath33 , respectively . nonstationary multiresolution analysis are addressed in several papers in the literature ( see , for instance , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite and references therein ) . here , following @xcite and @xcite , we say that a space sequence @xmath34 forms a nonstationary multiresolution analysis of @xmath14 if l ( _ i _ ) @xmath35 , @xmath36 ; ( _ ii _ ) @xmath37 ; ( _ iii _ ) @xmath38 ; + + ( _ iv _ ) for any @xmath36 , there exists a @xmath14-stable basis @xmath39 in @xmath22 such that + @xmath40 property ( _ i _ ) implies that the _ refinable functions _ @xmath41 satisfy a set of _ nonstationary refinable equations _ , i.e. @xmath42 for some sequence of _ scaling masks _ @xmath17 , where @xmath43 and @xmath44 + properties ( _ ii _ ) and ( _ iii _ ) are always true if any @xmath29 is compactly supported . as for _ ( iv ) _ , the function system @xmath45 , @xmath36 , is @xmath14-stable if and only if the fourier transform of @xmath29 has no @xmath46-periodic real zeros ( cf . @xcite ) . the existence of a unique set of functions @xmath47 solution to ( [ nsre ] ) , as well as their properties , are related to the properties of the mask sequence @xmath48 . since we are interested in the the case of compactly supported masks , we assume that any mask @xmath49 is compactly supported with @xmath50 and there exists a _ fundamental mask _ @xmath51 satisfying the sum rules @xmath52 such that @xmath53 ( cf . even if the assumption above excludes the mask sequence associated with the up function and the nonstationary mask sequence considered in @xcite , nevertheless it covers many nonstationary scaling masks , such us those ones associated with the exponential splines @xcite , @xcite and the mask families associated with the ripplets introduced in @xcite , @xcite . for the case of mask sequences with growing support we refer the reader to @xcite , @xcite . the nonstationary refinable equations ( [ nsre ] ) can be associated to a _ nonstationary cascade algorithm _ @xcite , which generates at any iteration @xmath54 the sequence of functions @xmath55 by @xmath56 without loss of generality , we assume that the starting function @xmath57 is the same for all @xmath58 , i.e. @xmath59 , where @xmath60 is a given @xmath14-stable function with @xmath61 , so that @xmath62 for any @xmath63 . the convergence of the cascade algorithm is related to the spectral properties of the _ fundamental transition operator _ @xmath64 , defined as @xmath65 where @xmath66 is the autocorrelation of the fundamental mask @xmath67 . the cascade sequence @xmath68 converges strongly to @xmath29 in @xmath14 as @xmath69 , uniformly in @xmath70 , if and only if the fundamental transition operator has unit spectral radius , 1 is the unique eigenvalue on the unit circle and is simple ( * ? ? ? 1.3 ) . under these hypotheses the sequence @xmath41 converges strongly to the solution of the stationary refinable equation @xmath71 we notice that any nonstationary mask sequence @xmath72 having a b - spline mask as fundamental mask , gives rise to a convergent cascade algorithm @xcite . the nonstationary refinable equations in the fourier space read @xmath73 where the laurent polynomials @xmath74 are the _ mask symbols_. we notice that any refinable function @xmath29 is normalized so that @xmath75 . in case the nonstationary cascade algorithm converges , the fourier transform of @xmath29 is given by @xmath76 given a nonstationary multiresolution analysis @xmath77 , at any level @xmath36 the _ wavelet space _ @xmath26 is defined as the orthogonal complement of @xmath22 in @xmath23 , i.e. @xmath78 any wavelet space is generated by the @xmath28-shifts of a wavelet function @xmath30 , i.e. @xmath79 any @xmath30 being different at any scale @xmath70 . the existence of a set of generating wavelets @xmath80 is always assured when all the refinable functions @xmath81 are compactly supported . moreover , it is always possible to construct compactly supported @xmath14-stable wave - lets associated with compactly supported @xmath14-stable refinable functions @xcite . the class of univariate compactly supported nonstationary masks and refinable functions we are interested in was introduced in @xcite . let us denoted the masks in the class by @xmath82 where @xmath83 is an integer @xmath84 , related to the support of the mask @xmath85 . + the explicit expression of @xmath85 is as follows . for any @xmath83 , the starting mask @xmath86 has entries @xmath87 while the entries of the higher level masks @xmath85 , @xmath88 , have expression @xmath89 } , \quad 0 \le\alpha \le n+1\,,\\ \\ a^{(n , m)}_\alpha = 0\ , , \qquad \hbox{otherwise}\ , . \end{array } \right.\ ] ] ( we assume @xmath90 when @xmath91 or @xmath92 . ) + we notice that @xmath93 is a real parameter that acts as a _ tension parameter_. in fact , the larger @xmath94 the faster the factor @xmath95 in the nonstationary mask coefficients goes to 1 when @xmath96 . this means that the nonstationary process behaves in practice as a stationary one if we choose large values of @xmath94 . thus , the more interesting cases are obtained for values of @xmath94 that are close to 1 . + for any @xmath83 and @xmath94 , the mask @xmath85 is compactly supported on @xmath97 $ ] and is _ bell - shaped _ , i.e. its entries are positive , centrally symmetric and strictly increasing on @xmath98\bigr]$ ] . + we notice that for any @xmath99 the 0-level scaling mask @xmath86 is the mask of the characteristic function of the interval [ 0,1 ] , while the @xmath70-level scaling masks @xmath85 , @xmath100 , are related to the class of stationary masks introduced in @xcite . + any mask sequence @xmath101 can be associated with the set of nonstationary refinable equations @xmath102 where @xmath103 , & m=0\ , , \\ \left [ 0,n+1\right ] , & m>0\ , . \end{array } \right.\ ] ] from @xcite it follows that for any @xmath104 and @xmath93 , @xmath105 , @xmath36 , is compactly supported with @xmath106= \left \ { \begin{array}{ll } \left [ 0,\frac n 2 + 1 \right ] , & m=0\ , , \\ \\ \left [ 0 , 2^{-m}(n+1 ) \right ] , & m>0\ , , \end{array } \right.\ ] ] and belongs to @xmath107 . moreover , any system @xmath108 is linearly independent and @xmath14-stable , forms a partition of unity , is _ totally positive _ , and enjoys the variation diminishing property . for any @xmath104 and @xmath109 , the mask sequence @xmath101 has the sequence @xmath110 as fundamental mask @xcite . since @xmath111 is the mask of the b - spline of degree @xmath83 having integer knots on @xmath97 $ ] , the sequence of functions @xmath112 , generated by the cascade algorithm @xmath113 converges strongly to @xmath114 . the convergence of the cascade algorithm implies that the fourier transform of @xmath105 is @xmath115 where @xmath116 a straightforward computation gives @xmath117 it is worthwhile to observe that any symbol @xmath118 is a hurwitz polynomial @xcite , i.e. it has only zeros with negative real part . moreover , the fundamental symbol of @xmath118 is the b - spline symbol @xmath119 in this section we analyze some properties of the refinable ripplets @xmath105 that are useful in both geometric modeling and signal processing applications . first of all , let us denote by @xmath120 the @xmath121-dilates of the b - spline of degree @xmath83 with knots on @xmath122 , normalized so that @xmath123 , @xmath124 , @xmath36 . in the fourier space @xmath120 satisfies the refinable equation @xmath125 its fourier transform is given by @xmath126 the refinable functions @xmath105 are generated by a convolution low involving the b - spline @xmath120 . [ th_conv ] for @xmath127 , @xmath105 , @xmath58 , satisfies the _ convolution _ property @xmath128 from ( [ infprodnm ] ) , ( [ fattor_anm ] ) and ( [ bn_fou ] ) it follows @xmath129 for @xmath100 , and @xmath130 , for @xmath131 . the claim follows by applying the inverse fourier transform to the relations above . as a first consequence of the theorem above , we can prove that @xmath105 satisfies suitable strang - fix conditions . [ corsf ] for any @xmath132 and @xmath100 , @xmath133 has a zero of order @xmath134 for @xmath135 , @xmath136 . @xmath137 has a simple zero for @xmath138 , @xmath136 . by repeated application of th . [ th_conv ] we get @xmath139 where @xmath140 since the symbols @xmath141 have the hat function symbol @xmath142 as fundamental symbol , the infinite product converges @xcite . moreover , for @xmath100 the symbol @xmath141 has no zeros on the unit circle , so that @xmath143 has no zeros , too . thus , @xmath144 , @xmath145 , has zeros of the same order as @xmath146 does , and the claim follows . + for @xmath131 , @xmath147@xmath148 has a simple zero for @xmath149 , @xmath150 , while @xmath151 has a zero of order @xmath134 for @xmath152 , @xmath153 . thus , @xmath154 has a simple zero for @xmath155 , @xmath153 . the strang - fix conditions allow us to conclude that polynomials of suitable degree are contained in the space generated by the @xmath28-translates of @xmath105 , i.e. the space @xmath156 [ th_poly ] let @xmath157 be the space of polynomials up to degree @xmath158 . for any @xmath104 and @xmath145 , the space @xmath159 contains @xmath160 , i.e. for any polynomial @xmath161 there exists a sequence of real numbers @xmath162 such that @xmath163 the space @xmath164 contains the space of the constants @xmath165 . _ note_. the roots of the nonstationary symbols ( [ fattor_anm ] ) can not fulfill the hypotheses of th . 1 in @xcite . thus , the refinable functions @xmath105 can not generate exponential polynomials . actually , we are interest in the construction of efficient decomposition and reconstruction formulas for general ( possibly non exponential ) signals . to this end high algebraic polynomial generation is more effective since it induces a high number of vanishing moments in the analyzing wavelet . from the results above it follows the approximation order of the system @xmath166 @xcite . for any @xmath104 and @xmath145 , the system @xmath166 has _ approximation order _ @xmath134 , i.e. for any @xmath167 there exists a constant @xmath168 , independent from @xmath70 , such that @xmath169 the system @xmath170 has approximation order 1 . _ note_. even if the system @xmath170 reproduces just the constants , polynomials of higher degree can be represented by suitable integer translates of @xmath171 . in fact , @xmath172 and its fourier transform has a zero of order @xmath134 for @xmath155 , @xmath153 . thus , polynomials of degree @xmath173 can be represented by the integer shifts of the dilate @xmath171 and the system @xmath174 has approximation order @xmath134 ( cf . @xcite ) . interestingly enough , the convolution property ( [ convm ] ) allows us to prove a _ differentiation rule _ for the functions @xmath105 . [ derphinm ] let @xmath175 be the backward finite difference operator defined recursively as @xmath176 for @xmath132 , the derivatives of @xmath105 are given by @xmath177 we will prove ( [ dervm ] ) just for @xmath100 ; the case @xmath131 can be proved in a similar way . + from ( [ bn_fou ] ) and the first of ( [ conv_tfphi ] ) it follows @xmath178 the inverse fourier transform gives @xmath179 which is the derivation rule for @xmath180 . repeated applications of this rule give the derivation rules of higher order . from ( [ dervm ] ) and the variation diminishing property , we can infer that the derivative @xmath181 , @xmath182 , has the same behavior of the finite difference @xmath183 with @xmath184 for @xmath131 and @xmath185 for @xmath145 . in particular , the number of strict sign changes of @xmath181 is not greater than the number of strict sign changes of the sequence @xmath186 so that we can infer the shape of @xmath105 from @xmath187 . for any @xmath188 and @xmath58 , @xmath105 is _ bell - shaped _ , i.e. @xmath105 is centrally symmetric , strictly increasing on @xmath189 $ ] , and its second derivative has just 2 sign changes . since any mask @xmath85 is centrally symmetric , any @xmath105 is centrally symmetric , too . as a consequence @xmath190 is centrally antisymmetric and @xmath191 is centrally symmetric . for @xmath145 ( [ dervm ] ) and ( [ derdiffvm ] ) give @xmath192 where @xmath185 , while for @xmath131 ( [ dervm])-([derdiffvm ] ) together with ( [ rephinm ] ) yield @xmath193 where @xmath184 . since @xmath194 is positive in @xmath195 and vanishes for @xmath196 , @xmath197 and @xmath198 are positive for @xmath199 . now , by the variation diminishing property from the first relations in ( [ der12 m ] ) and ( [ der120 ] ) it follows that @xmath197 has just one sign change which has to be in @xmath200 ; thus , @xmath201 in @xmath202 and @xmath203 in @xmath204 . as for the second derivative , a direct computation shows that @xmath198 is negative in @xmath200 so that @xmath198 has at least two sign changes ; but from the variation diminishing property and the second relations in ( [ der12 m ] ) and ( [ der120 ] ) it follows that @xmath198 can not have more than two sign changes , so proving the claim . from the results in the previous sections it follows that , for any @xmath104 held fix , the spaces @xmath159 , @xmath36 , generate a nonstationary multiresolution analysis as defined in section 2 . we note that the @xmath14-stability of the basis @xmath166 implies that the symbol of the autocorrelation of @xmath205 , i.e. the polynomial @xmath206 is non vanishing for any @xmath207 . the vector @xmath208^t$ ] , where @xmath209 is the eigenvector corresponding to the eigenvalue 1 of the transition operator @xmath210 where @xmath211 , is the autocorrelation of the mask @xmath212 . the sequence @xmath213 is positive , compactly supported with @xmath214 , \left [ \dfrac n2 + 1\right]\bigr]\ , , & m=0\ , , \\ \\ { [ -n-1,n+1]}\ , , & m>0\ , , \end{array } \right.\ ] ] and centrally symmetric . as a consequence @xmath215 with @xmath216 since @xmath217 at any level @xmath70 the basis @xmath166 is a non orthogonal basis . the nonstationary multiresolution analysis @xmath218 allows us to define a wavelet space sequence @xmath219 , where each space @xmath220 is the orthogonal complement of @xmath159 in @xmath221 . since @xmath105 is non orthogonal , orthogonal wavelets with compact support do not exists . on the other hand , due to the @xmath14-stability and the compact support of each @xmath105 , it is always possible to construct compactly supported semiorthogonal wavelets ( * ? ? ? * th . 3.12 ) . the explicit expression of the prewavelet of minimal support can be obtained generalizing to the nonstationary case the results in @xcite ( see also @xcite ) . [ th_prewave ] for any @xmath222 , the functions @xmath223 where @xmath224 are the prewavelets generating the wavelet spaces @xmath225 the functions @xmath226 are compactly supported . moreover , any system @xmath227 is @xmath14-stable and linearly independent . finally , @xmath226 has @xmath134 vanishing moments . a straightforward computation yields @xmath228 for any @xmath229 , i.e. @xmath226 is orthogonal to the space @xmath159 . due to the support of @xmath105 it follows @xmath230 , & m=0\ , , \\ \\ \left [ -2n-1 , n \right ] , & m>0\ , , \end{array } \right.\ ] ] so that @xmath226 is compactly supported with @xmath231 , & m=0 , \\ \\ \left [ -2^{-m}n , 2^{-m}(n+1 ) \right ] , & m>0 . \end{array } \right.\ ] ] now , the wavelet equations ( [ prewave ] ) in the fourier space reads @xmath232 where @xmath233 , @xmath234 an explicit calculation gives @xmath235 as a consequence , @xmath236 is non - vanishing . it follows that at any level @xmath237 the system @xmath238 , generating the wavelet space @xmath220 , is linearly independent and @xmath14-stable . moreover , @xmath226 is non orthogonal with @xmath239 where @xmath240 is the autocorrelation of the sequence @xmath241 . + finally , since @xmath242 , from th . [ th_poly ] it follows that @xmath243 , @xmath244 , so concluding the proof . [ th_suppmin ] for any @xmath99 and @xmath58 , @xmath226 is the unique minimally supported wavelet generating the wavelet space @xmath245 . the orthogonality conditions @xmath246 , @xmath247 can be proved to be equivalent to the conditions @xmath248 let @xmath249 , @xmath250 , and , similarly , for the polynomial @xmath251 , let @xmath252 , @xmath253 , so that @xmath254 , and @xmath255 . it follows that conditions ( [ dg_orth ] ) are equivalent to @xmath256 thus , @xmath226 is the minimally supported prewavelet if and only if @xmath257 is the minimally supported polynomial satisfying ( [ orth_prew ] ) , i.e. if and only if @xmath258 and @xmath259 have no common zeros or , equivalently , @xmath260 and @xmath261 have no common zeros for @xmath262 . but this easily follows from ( [ dnm ] ) since any symbol @xmath118 has only zeros with negative real part and @xmath263 is positive . we note that @xmath226 is the @xmath121-dilate of the prewavelet constructed in @xcite and th . [ th_prewave ] and th . [ th_suppmin ] generalize to the nonstationary case the stationary prewavelet constructed in @xcite . even if the prewavelet coefficients @xmath264 do not have an explicit expression , they can be evaluated efficiently by an iterative algorithm . [ th_4.3 ] let @xmath265^t$ ] be the nonstationary prewavelet coefficients , and let @xmath266^t$ ] be the prewavelet coefficients corresponding to the stationary fundamental mask @xmath111 . + for any @xmath132 and @xmath58 held fixed , consider the iterative procedure @xmath267 where @xmath268 with @xmath269 and @xmath270\ , , & m=0\ , , \\ \left [ -2(n+1),n+1\right]\ , , & m>0\ , . \end{array } \right.\ ] ] the sequence @xmath271 converges strongly to @xmath272 when @xmath273 . moreover , the following error estimate holds @xmath274 where @xmath275 is a positive constant independent from @xmath276 . using the refinable equation ( [ rephinm ] ) in ( [ munm ] ) we get @xmath277 , which in matrix form can be written as @xmath278 . repeated applications of both algorithm ( [ algprew ] ) and the relation above give @xmath279 now , let @xmath280 , where @xmath281 . we note that @xmath282 . since @xmath111 is the fundamental mask of the mask sequence @xmath283 , it follows @xmath284 , and @xmath285 ( cf . moreover , @xmath286 thus , @xmath287 and the claim follows with @xmath288 . the refinable functions @xmath105 and the prewavelets @xmath226 are linearly independent and @xmath14-stable , but they are not orthogonal . as a consequence , the dual bases of @xmath105 and @xmath226 in @xmath159 and @xmath220 , respectively , have infinite support . nevertheless , compactly supported bases giving rise to efficient reconstruction and decomposition formula of a given discrete signal , can be obtained introducing biorthogonal bases . the theory of biorthogonal bases in the stationary case @xcite can be generalized to the nonstationary framework ( cf . @xcite ) . for any multiresolution analysis @xmath289 we can introduce a biorthogonal multiresolution analysis @xmath290 with @xmath291 and biorthogonal wavelet spaces @xmath292 and @xmath293 , such that for any @xmath58 @xmath294 at each level @xmath70 , the spaces @xmath159 , @xmath295 , @xmath220 and @xmath296 are generated by the @xmath28-integer translates of the biorthogonal functions @xmath105 , @xmath297 , @xmath226 and @xmath298 , respectively , satisfying the following biorthogonality conditions : @xmath299 we stress that all the biorthogonal functions @xmath105 , @xmath297 , @xmath226 and @xmath298 , @xmath36 , can not be obtained each other by dilation . thus , none of the biorthogonal spaces at level @xmath70 is a scaled versions of the spaces at level @xmath300 . biorthogonal bases for the exponential splines were constructed in @xcite . here , we want to construct the biorthogonal bases associated with the nonstationary refinable functions @xmath105 , @xmath36 . as a consequence of ( [ biorthmra ] ) and ( [ biorthwave ] ) , the wavelet @xmath226 belongs to @xmath221 , while the biorthogonal functions @xmath298 and @xmath297 belong to @xmath301 , so that @xmath302 moreover , perfect reconstruction at each level @xmath70 is guaranteed if for any @xmath58 the biorthogonal symbols @xmath303 satisfy @xmath304\ ] ] with @xmath305\ ] ] identity ( [ eqbezout ] ) is a bezout s equation which has a unique polynomial solution @xmath306 of a given degree @xcite . under mild conditions on the symbols @xmath307 and @xmath306 , biorthogonality conditions ( [ biorthnm ] ) guarantee that the biorthogonal bases @xmath308 are @xmath14-stable , so that , for any @xmath167 , the following decomposition formula holds @xmath309 where @xmath310 is the @xmath311-level approximation . some examples of nonstationary biorthogonal bases will be given in the next section . the sequences @xmath312 , @xmath313 , @xmath314 , @xmath315 are pairs of biorthogonal fir filters that give rise to the decomposition and reconstruction algorithms @xmath316 @xmath317 \,,\ ] ] which can be efficiently used for the analysis and synthesis of a given data sequence @xmath318 . in this section we give some examples of both nonstationary prewavelets and biorthogonal bases in the case when @xmath319 . in this case the nonstationary refinable functions @xmath320 , @xmath321 , belong to @xmath322 , i.e. they have the same smoothness as the cubic b - spline . interestingly enough , any @xmath320 with @xmath323 has the same support as @xmath324 , i.e. @xmath325 $ ] , while @xmath326 is more localized in the scale - time plane having @xmath327 $ ] , a property that appears very useful in applications ( see the example below ) . in order to obtain refinable beses and nonstationary filters significantly different from those ones generated by the cubic b - spline , we choose @xmath328 as a value for the tension parameter . + .numerical values ( rounded to the forth digit ) of the mask coefficients @xmath329 , @xmath330 , @xmath331 for @xmath332 . here @xmath328 [ cols="^,^,^,^,^,^,^,^,^,^",options="header " , ] we studied the properties of a class of refinable ripplets associated with sequences of nonstationary scaling masks . one of the most interesting property of these functions is in that they have a smaller support than the stationary refinable ripplets with the same smoothness . this localization property is crucial in several applications , from geometric modeling to signal processing . + after proving some approximation properties , such as strang fix conditions , polynomial reproduction and approximation order , we proved also that any refinable function in the family is bell - shaped , so that they can efficiently approximate a gaussian . moreover , since these refinable functions generate nonstationary multiresolution analyses , we constructed the minimally supported nonstationary prewavelets and proved that their @xmath121-shifts form @xmath333-stable bases . we note that this construction can be generalized to other class of nonstationary refinable functions , like exponential splines . moreover , we constructed nonstationary biorthogonal bases and filters to be used in efficient decomposition and reconstruction algorithms . + the localization property of the refinable ripplets we studied implies that the corresponding nonstationary wavelets have a small support too , a property which is very desirable in the case when the relevant information of a function to be approximated or of a signal to be analyzed are focused in small regions of the scale - time plane . the preliminary test in section 6 shows the good performances of the constructed nonstationary wavelets in a simple compression test . more tests will be the subject of a forthcoming paper . l. gori and f. pitolli . nonstationary subdivision schemes and totally positive refinable functions . in m. neamtu and schumaker l. l. , editors , _ approximation theory xii : san antonio 2007 _ , mod . methods math . , pages 169180 . nashboro press , brentwood , tn , 2008 .	we study the approximation properties of the class of nonstationary refinable ripplets introduced in @xcite . these functions are solution of an infinite set of nonstationary refinable equations and are defined through sequences of scaling masks that have an explicit expression . moreover , they are variation - diminishing and highly localized in the scale - time plane , properties that make them particularly attractive in applications . here , we prove that they enjoy strang - fix conditions and convolution and differentiation rules and that they are bell - shaped . then , we construct the corresponding minimally supported nonstationary prewavelets and give an iterative algorithm to evaluate the prewavelet masks . finally , we give a procedure to construct the associated nonstationary biorthogonal bases and filters to be used in efficient decomposition and reconstruction algorithms . as an example , we calculate the prewavelet masks and the nonstationary biorthogonal filter pairs corresponding to the @xmath0 nonstationary scaling functions in the class and construct the corresponding prewavelets and biorthogonal bases . a simple test showing their good performances in the analysis of a spike - like signal is also presented . + * keywords * : total positivity , variation - dimishing , refinable ripplet , bell - shaped function , nonstationary prewavelet , nonstationary biorthogonal basis + * msc * : 41a30 @xmath1 42c40 @xmath1 65t60
on 2005 july 4 at 05:52:03 ut ( as seen from earth ) a 370 kg nasa spacecraft was maneuvered into the path of comet 9p / tempel 1 @xcite . the results of the impact were recorded by a host of ground and space based observatories with wavelength coverage spanning the sub - mm to the ultraviolet @xcite . here we report on the results of low resolution slitless spectroscopy acquired with the _ galaxy evolution explorer _ ( _ galex _ ) in the near ultraviolet nuv channel 2 days before , immediately following , and 1 day after the impact event . the primary spectral features in this range include a solar continuum scattered from cometary dust and emissions from oh and cs molecular bands centered near 3085 and 2575 , respectively . _ galex _ _ ( galaxy evolution explorer ) _ is a nasa small explorer whose primary mission is to map the history of star formation using two modes : two - band photometry ( fuv , 13501750 ; nuv , 17503100 ) and integrated field grism spectroscopy with 1020 spectral resolution . because of its large field - of - view , 12 , _ galex _ is also well suited to cometary coma studies as demonstrated by the 2005 march observations of c/2004 q2 machholz @xcite . it has a limiting spatial resolution in the nuv channel of @xmath7 5 sampled with 15 pixels . for _ deep impact _ only nuv data was obtained as the fuv side was not operating at time of the event . in all there were 7 contiguous pre - impact orbits devoted to grism observations , with @xmath8 s visibility per orbit . two orbits of direct imaging were taken before and after the grism orbits . for the impact event there were 6 orbits of grism observations beginning a few minutes after impact , with direct imaging before ( 1 orbit ) and after ( 2 orbits ) . one day post - impact there were 4 orbits of grism observations with one direct imaging orbit before the grism observations . a log of the spectroscopic observations is given in table [ obslog ] . here we concentrate on the results derived from the low resolution slitless spectroscopy in the nuv channel . at the time of impact comet 9p / tempel 1 was at a distance of @xmath9 = 1.51 au from the sun and @xmath10 = 0.89 au from the earth . a 15 pixel subtended 975 km at the comet . emission from oh , cs , and solar scattered light were detected prior to impact . just after impact the brightness and spatial extent of these emissions increased with time . the following day saw the return of these emissions to near the pre - impact levels . the _ galex _ project supplied a time - tagged photon list of the observations with photon coordinates referenced to ra and dec ( j2000 ) . the project also supplied the output of emmanuel bertin s sextractor program run on the direct - mode ( imaging ) observations centered on the same fields . sextractor provides accurate astrometry for sources found in the direct images . we converted these positions into rectangular strips using the galex dispersion relation for box length and the magnitude estimates from sextractor for box widths in order to mask sources in the grism images . all of our astrometric corrections were done on a spherical projection with the idlastro routines . co - alignment between the strips and the grism images of the stars across the entire image confirmed the accuracy of these calculations . we used the jpl horizons ephemeris generator with _ galex _ as the observatory ( `` @galex '' ) to calculate the astrometrically correct ephemeris for comet tempel 1 . the ephemeris was used to reconstruct the comet image from the photon list in the reference frame of the comet . counts and exposure times were accumulated into separate images and the source mask and flat field ( supplied by the _ galex _ project ) were applied every time the stars moved by one pixel relative to the comet . as a result , emission from well - characterized sources was effectively erased and very little exposure time was lost . the final rate images , constructed by dividing the total count images by the total exposure time images for the corresponding observations , were used in the analysis of the temporal evolution of the coma following the impact . because of the low total counts in each individual grism image , individual exposures were summed to produce a composite image from each day of the program . in this case the source mask was not applied and the stellar spectra appear as multiple images separated by the comet s motion between _ galex _ orbits . an example is shown in figure [ image ] for the sum of the six grism images obtained following the impact event . note that the stellar images ( each corresponding to a different orbit ) are broadened by their apparent relative motion nearly normal to the dispersion direction during the exposure . analysis of the spectrum is begun with the composite images ( such as in figure [ image ] ) corresponding to the _ galex _ orbits from the three dates of observation . the spectra are extracted over 13 rows of the rotated image array , corresponding to an effective slit height of @xmath11 to ensure that only pixels with statistically significant counts are included in the aperture photometry . the detector background is evaluated over the same rows in the spectral range 1700 to 2000 where the nuv throughput for a solar - like spectrum is effectively zero . this background is subject to variation due to the presence of faint stars or galaxies in the field . the three composite spectra , after subtraction of background , are shown in the top panels of figure [ spec ] . the extraction of the discrete emissions from cs and oh is complicated by the blended continuous solar spectrum produced by the scattering of sunlight by cometary dust grains in the coma . an additional complication arises from the different spatial distributions of the dust and gas components that result from the quiescent outflow and the subsequent time evolution of these distributions after the impact , as the spatial variation is superimposed on the actual spectrum in the dispersion direction . between 2000 and 3400 the solar spectrum rises steeply and at the long wavelength end can contribute to the extended oh(0,0 ) band near 3085 . since the effective area , @xmath12 , supplied by the _ galex _ project did not extend beyond 3100 , we identified a field f8 v star and an _ iue _ calibration standard to extend @xmath12 to 3400 . a model of the dust scattered solar continuum is required to account for the contamination of the oh(0,0 ) and cs(0,0 ) bands . we used a solar spectrum from uars / solstice , described by @xcite , multiplied by this @xmath12 , and convolved with a superposition of @xmath1380 and @xmath13250 gaussian instrument functions to simulate the radially outflowing dust in the dispersion direction , in order to generate a template for the solar scattered light . this template , normalized to regions of the spectrum where strong molecular emission is absent , and reddened by 4% per 100 @xcite , is also shown in the figure . the difference between the observed spectra and the template is shown in the lower panels of the same figure . the three identified emission features are the cs(0,0 ) band at 2576 , and the oh(1,0 ) and ( 0,0 ) bands at 2820 and 3085 , respectively . for near zero heliocentric velocity , the ratio of fluorescence efficiencies of the oh(1,0 ) and ( 0,0 ) bands is 0.10 @xcite , but the _ galex _ effective area is 4.2 times higher at 2820 than at 3085 , so the ( 1,0 ) band should give a count rate @xmath130.4 that of the ( 0,0 ) band . this is consistent with the lower panels of figure [ spec ] , and serves to validate the continuum subtraction . in principle , we could use the template scaling factor to derive the temporal variation of the dust ejected by the impact , but this is much better done with visible observations from the ground or from space ( e.g. , * ? ? ? * ; * ? ? ? * ) because of uncertainties in the background subtraction . _ galex _ observations began only a few minutes after the impact and grism spectra were recorded on six successive orbits , or for about 8.5 hours following the impact . the temporal evolution of the cs and oh emissions can be seen in the individual spectra shown in figure [ deep ] . to construct a light curve , we used a rectangular aperture wide enough to capture almost all of the band emission , keeping the aperture height at @xmath11 , as above . for cs , the width was @xmath14 , corresponding to a spectral bandwidth of 100 , while for oh the width was @xmath15 , or @xmath13200 . the projected sizes on the sky are @xmath16 km and @xmath17 km , respectively for cs and oh . similar photometry was performed on the pre- and post - impact data sets , and the results are shown in figure [ time ] , in which the plotted times are the mid - points of the exposure . the observations beginning about 24 hours after impact show the gaseous emissions returning to near quiescent levels . the quiescent production rates were determined using the vectorial model of @xcite . the parent lifetimes at 1 au were 71,000 s for , with 85% of the dissociations producing oh @xcite , and 1,000 s for @xcite . the lifetimes of oh and cs were taken to be 140,000 s and 100,000 s , respectively , but are not critical for the aperture size used . the parent outflow velocity in the model is @xmath18 km s@xmath3 , while the daughters are given 1.05 and 0.90 km s@xmath3 , for oh and cs , respectively . the fluorescence efficiencies at 1 au are @xmath19 and @xmath20 photons s@xmath3 molecule@xmath3 for the oh(0,0 ) band @xcite and the cs(0,0 ) band ( r. yelle , private communication ) , respectively . the derived production rates are given in table [ prates ] , together with other recently reported values . our value for the water production rate fits squarely within the range reported previously . we also note that the / production rate ratio , @xmath21 , is typical of values found from _ iue _ observations of comets @xcite . the linear increase with time following impact in the number of dissociation products in the aperture can be modeled , using known photochemical rates @xcite , to determine the abundance of the parent molecular species produced in the collision . this approach was used by @xcite for the analysis of wide - field camera images of oh emission . however , for our aperture height of 12,600 km , molecules flowing radially outward at a mean velocity of 0.5 km s@xmath3 will begin to exit the aperture 3.5 hours after impact . this problem was considered by @xcite , and we adopt their formalism . in addition to the parent and daughter species , whose photochemical lifetimes are given above , @xcite also include a `` grandparent '' molecule to account for gas excavation over an extended period ( several minutes ) or evaporation from excavated icy grains . the `` lifetime '' of the grandparent is an adjustable parameter . the model results are shown with the data in figure [ time ] . for a baseline , we adopt the mean value of pre - impact count rates , which correspond to an observation almost exactly one rotation period ( @xmath1341 hours ) earlier . the cs light curve is well fit by a mean velocity of 0.4 km s@xmath3 and a velocity dispersion of 0.1 km s@xmath3 , similar to the values derived by @xcite for various radicals , and a `` grandparent '' lifetime of 2000 s. for oh , models with 0.4 km s@xmath3 and 0.5 km s@xmath3 are shown in the figure . what matters most in deriving the number of parent molecules produced is the linear increase immediately following the impact before any of the dissociation products has left the field - of - view . we note that @xcite found a maximum velocity of co produced by the impact to be 0.7 km s@xmath3 . the total number of molecules found is @xmath22 molecules of and @xmath23 molecules of ( corresponding to @xmath24 kg and @xmath25 kg , respectively ) . the / ratio is @xmath26 , comparable to the quiescent value within the given uncertainties . a comparison with other reported values is given in table [ mols ] , and shows generally good agreement . our numbers are lower , by 25% for and a factor of 3 for , than those presented in 2006 in a preliminary report on these results @xcite . the spectral images used in that analysis did not have the comet s motion removed and so a large extraction box was used to ensure that all of the cometary photons were counted . this made it difficult to properly subtract the dust continuum , which is more important for isolating the cs emission than it is for the oh emission . we note that the only uncertainties in the derivation of the total number of molecules produced ( other than due to photon statistics and absolute calibration ) are those in the photodissociation lifetimes and fluorescence efficiencies , which are known at the 10% level . the combined uncertainties are @xmath27% . after 25 hours all gas and dust emissions within the summation aperture have returned to nearly quiescent values . however , as the lower right panel of fig . [ spec ] shows , the oh bands one day after impact appear much broader in wavelength than prior to impact , likely due to outflow along the dispersion direction . this is not surprising as at 1.5 au the photodissociation lifetimes of and oh are 2 and 3 days , respectively . two other orbiting spacecraft made nuv observations of the _ deep impact _ event : both are x - ray observatories with uv / optical imagers that were used with wide - band nuv filters . the filter passband of the _ swift _ imager included both the cs and oh bands in addition to dust continuum @xcite but the photometric images are dominated by the oh emission . the number of water molecules produced by the impact derived from their photometry ( see table [ mols ] ) is in excellent agreement with our result . the _ xmm - newton _ observatory imager had two nuv filters , one of which included only cs emission while the other included both cs and oh @xcite . the light curves derived from both filters show behavior similar to that of the two light curves in figure [ time ] . @xcite attribute the difference in shape of the light curves to the presence of small `` icy grains '' in the ejecta . however , as the _ galex _ spectra of figure [ spec ] demonstrate , their data can be reasonably accounted for by cs emission . _ galex _ observations of comet 9p / tempel 1 at the time of _ deep impact _ provide a unique measurement of the cs parent produced by the excavation of cometary material , and as with other daughter molecules , appears to have the same abundance relative to as in quiescent outgassing of the comet . our derived value for the amount of excavated is in excellent agreement with the results of several other investigations . this reinforces the idea that nearly all the volatile species are uniformly mixed with the water ice from the near - surface layer at which sublimation normally occurs to the tens of meters to which _ deep impact _ excavated material . this lack of differentiation from near - surface to tens of meters implies either that the surface erodes ( due to hydrodynamic drag on the dust by the gas ) at about the same rate as the differentiation front proceeds deeper or that the differentiation has reached depths of many tens of meters while somehow still leaving the same relative fraction of all the volatiles throughout the upper tens of meters @xcite . notwithstanding the relative faintness of the comet , the _ galex _ results demonstrate the unique capability of wide - field imaging spectroscopy of comets in the ultraviolet at wavelengths below the atmospheric cutoff . we thank the _ galex _ project , and karl forster in particular , for their efforts in planning and executing these observations . we would like to acknowledge tim conrow and the _ galex _ pipeline processing team for helping us to fully utilize the moving target capability of _ galex _ , jon giorgini and the jpl horizons team for calculating high precision comet ephemerides from the perspective of _ galex _ , and wayne landsman and the many contributors to the idlastro library . this work was supported by the _ galex _ guest investigator program under nasa grant nnc06gd35 g to the johns hopkins university . lccc ` gi1_024001_deep_impact_pre_0001-ng - int.fits ` & 820.0 & 2005 - 07 - 02 & 06:22:17 + ` gi1_024001_deep_impact_pre_0002-ng - int.fits ` & 940.0 & 2005 - 07 - 02 & 08:00:53 + ` gi1_024001_deep_impact_pre_0003-ng - int.fits ` & 940.0 & 2005 - 07 - 02 & 09:39:29 + ` gi1_024001_deep_impact_pre_0004-ng - int.fits ` & 960.0 & 2005 - 07 - 02 & 11:18:04 + ` gi1_024001_deep_impact_pre_0005-ng - int.fits ` & 940.0 & 2005 - 07 - 02 & 12:56:41 + ` gi1_024001_deep_impact_pre_0006-ng - int.fits ` & 940.0 & 2005 - 07 - 02 & 14:35:17 + ` gi1_024001_deep_impact_pre_0007-ng - int.fits ` & 940.0 & 2005 - 07 - 02 & 16:13:53 + ` gi1_024002_deep_impact_0001-ng - int.fits ` & 759.0 & 2005 - 07 - 04 & 06:01:47 + ` gi1_024002_deep_impact_0002-ng - int.fits ` & 959.0 & 2005 - 07 - 04 & 07:40:23 + ` gi1_024002_deep_impact_0003-ng - int.fits ` & 939.0 & 2005 - 07 - 04 & 09:18:59 + ` gi1_024002_deep_impact_0004-ng - int.fits ` & 938.1 & 2005 - 07 - 04 & 10:57:36 + ` gi1_024002_deep_impact_0005-ng - int.fits ` & 940.0 & 2005 - 07 - 04 & 12:36:11 + ` gi1_024002_deep_impact_0006-ng - int.fits ` & 939.0 & 2005 - 07 - 04 & 14:14:48 + ` gi1_024003_deep_impact_post_0001-ng - int.fits ` & 299.0 & 2005 - 07 - 05 & 05:05:56 + ` gi1_024003_deep_impact_post_0002-ng - int.fits ` & 919.0 & 2005 - 07 - 05 & 06:40:52 + ` gi1_024003_deep_impact_post_0003-ng - int.fits ` & 931.2 & 2005 - 07 - 05 & 08:19:29 + ` gi1_024003_deep_impact_post_0004-ng - int.fits ` & 939.0 & 2005 - 07 - 05 & 09:58:05 +	_ galex _ observations of comet 9p / tempel 1 using the near ultraviolet ( nuv ) objective grism were made before , during and after the deep impact event that occurred on 2005 july 4 at 05:52:03 ut when a 370 kg nasa spacecraft was maneuvered into the path of the comet . the nuv channel provides usable spectral information in a bandpass covering 2000 3400 with a point source spectral resolving power of @xmath0 . the primary spectral features in this range include solar continuum scattered from cometary dust and emissions from oh and cs molecular bands centered near 3085 and 2575 , respectively . in particular , we report the only cometary cs emission detected during this event . the observations allow the evolution of these spectral features to be tracked over the period of the encounter . in general , the nuv emissions observed from tempel 1 are much fainter than those that have been observed by _ galex _ from other comets . however , it is possible to derive production rates for the parent molecules of the species detected by _ galex _ in tempel 1 and to determine the number of these molecules liberated by the impact . the derived quiescent production rates are @xmath1 ( ) @xmath2 molecules s@xmath3 and @xmath1 ( ) @xmath4 molecules s@xmath3 , while the impact produced an additional @xmath5 molecules and @xmath6 molecules , a similar ratio as in quiescent outgassing .
the properties of materials are strongly influenced by electronic interaction and randomness . in particular , coulomb correlations and disorder are both driving forces behind metal - insulator transitions ( mit ) connected with the localization and delocalization of particles . the mott - hubbard mit is caused by coulomb correlations ( electronic repulsion ) in the pure system @xcite . the anderson mit , also referred to as anderson localization , was established in @xcite and is the basis of the theory of localization of electrons in disordered systems . it is due to coherent backscattering from randomly distributed impurities in a system without interaction @xcite . it is therefore a challenge to investigate quantum models where both aspects are simultaneously present . the mott - hubbard mit is characterized by opening a gap in the density of states at the fermi level . at the anderson localization , the character of the spectrum at the fermi level changes from a continuous one to a dense discrete one . it is plausible that both mits could be detected by knowing the local density of states ( ldos ) , as it discriminates between a metal and an insulator , which is driven by correlations and disorder @xcite . in many recent works @xcite , the change of the whole ldos distribution function at the anderson transition were explored for the bethe and for the simple cubic lattices , where many other effects , like interaction , percolation and binary alloy disorder , were included . the theoretical investigation of disordered systems requires the use of probability distribution functions ( pdfs ) for the random quantities of interest and one is usually interested in typical values of these quantities which are mathematically given by the most probable value of the pdf . in many cases the complete pdf is not known , i.e. , only limited information about the system provided by certain averages ( moments or cumulants ) is available . in this situation it is of great importance to choose the most informative average of a random variable . for examples , when a disordered system is near the anderson mit , most of the electronic quantities fluctuate strongly and the corresponding pdfs possess long tails @xcite and at the anderson mit the corresponding moments might not even exist . this is well illustrated by the ldos of the system . the arithmetic mean of this random one - particle quantity does not resemble its typical value at all . in particular , it is noncritical at the anderson transition and hence can not help to detect the localization transition . by contrast , it was shown @xcite that the geometric mean gives a better approximation of the averaged value of the ldos , as it vanishes at a critical strength of the disorder and hence provides an explicit criterion for anderson localization @xcite . besides that , the work @xcite suggests that by considering the anderson localization , the moments of the distribution of the resolvent kernel are of essential importance . theoretical descriptions of the mit have to be nonperturbative if no long - range order exists on either side of the transition . a nonperturbative framework to investigate the mott - hubbard mit in lattice electrons with a local interaction and disorder is given by the dynamical mean - field theory ( dmft ) @xcite . in this paper , we investigate the anderson - falicov - kimball model . the pure falicov - kimball model describes two species of particles , mobile and immobile , which interact with each other when both are on the same lattice site . the falicov - kimball model captures some aspects of the mott - hubbard mit , i.e. , upon increasing the interaction the ldos for mobile particles splits into two subbands opening a correlation gap at the fermi level . in the anderson - falicov - kimball model the mobile particles are disturbed by a local random potential . the model is solved within the dmft framework . anyway , there is no reason to believe that just the geometric mean can offer a good approximation for the averaged ldos . actually , if we consider the generalized mean we can show that the averaged ldos can vanish in the band center at a critical strength of the disorder for a wide variety of averages . here , we use the hlder mean in order to analyze how the averaged ldos depends on each hlder parameter that is used and whether , in particular , the geometric mean can offer a better approximation than the arithmetic one . a generalized mean , also known as hlder mean , is an abstraction of the pythagorean means including arithmetic , geometric and harmonic means . if @xmath0 is a non - zero real number , we can define the generalized mean with exponent @xmath0 of the positive real numbers @xmath1 as @xmath2 as special cases , we can find , for example , the minimum ( @xmath3 ) , the geometric mean ( @xmath4 ) , the arithmetic mean ( @xmath5 ) and the maximum ( @xmath6 ) . a power mean emphasizes small ( big ) values for small ( big ) @xmath0 . in section [ model ] we present the anderson - falikov - kimball model and its solution using the dmft @xcite . we also present the relation between the disorder strength and the coulombian repulsion for the linearized dmft with @xmath0-averaging . in section [ results ] we present the numerical results concerning the ground - state phase diagram obtained using the hlder mean and compare them with the results of previous works @xcite . finally in section [ conclusions ] we present our conclusions and final remarks . the falicov - kimball model is , presently , the simplest model to study metal - insulator transitions in mixed valence compounds of rare earth and transition metal oxides , ordering in mixed valence systems , order - disorder transitions in binary alloys , itinerant magnetism , and crystallization . recently , it was also applied to study the possibility of electronic ferroelectricity in mixed - valence compounds , and also of the phase diagram of metal ammonia solutions @xcite . in its most simplified version , namely the static model , it consists in assuming that in the system exist two species of spinless fermions : one of them possess infinite mass and hence does not move while the other one is free to move @xcite . in the anderson - falicov - kimball model @xcite the mobile particles are disturbed by a local random potential , giving rise to a competition between interaction and disorder yielding stabilization of metalicity . this model is defined by the following hamiltonian : @xmath7 where @xmath8 ( @xmath9 ) and @xmath10 ( @xmath11 ) are , respectively , the creation ( annihilation ) operators for the mobile and immobile fermions ( electrons and ions , respectively ) at a lattice site @xmath12 , @xmath13 is the electron transfer integral connecting states localized on nearest neighbor sites and @xmath14 is the coulombian repulsion that operates when one ion and one electron occupy the same site . the average number of electrons ( ions ) on site @xmath12 is denoted as @xmath15 ( @xmath16).the energy @xmath17 is a random , independent variable , describing the local disorder disturbing the motion of electrons . for simplicity , it was assumed that just mobile particles are subjected to the structural disorder @xcite . the number of ions and electrons is considered independent of each other and fixed . if there is no long - range order , the position of the ions on a lattice is random , what introduces additional disorder apart from that given by the @xmath17 term in the hamiltonian ( [ hamil ] ) . we consider that the occupation @xmath18 on the @xmath12th site has probability @xmath19 ( @xmath20 ) . in the @xmath14 term in the hamiltonian ( [ hamil ] ) one has to take the quantum - mechanical average over a given quantum state of the @xmath21 particles , what does not transform the extended states into the localized ones . in the @xmath17 term one has to average the quantum - mechanical expectation values over different realizations of @xmath17 , what can lead to the anderson localization @xcite . in the pure falicov - kimball model , if @xmath22 , the fermi energy for electrons is inside of the correlation ( mott ) gap opened by increasing the interaction @xcite . how the disorder changes this gap and how the @xmath0-averaging influences the presence of localized states are the subjects of the present study . the anderson - falicov - kimball model , where the interaction and disorder are local , is solved within the dmft . the formalism that is used is the same from @xcite , where @xmath23 denotes the energy , the number of the immobile particles is conserved being zero or one and the chemical potential @xmath24 is introduced only for the mobile subsystem . using the hybridization function @xmath25 , which is a dynamical mean field ( molecular field ) describing the coupling of a selected lattice site with the rest of the system @xcite , the @xmath17-dependent ldos is @xcite @xmath26 where @xmath27 is the local @xmath17-dependent green function described in @xcite . from the @xmath17-dependent ldos , we introduce the @xmath0-averaged ldos @xmath28^{q}\right\}^{1/q}.\label{eq3}\]][aq ] where the subscript @xmath0 stands for the chosen average . the lattice , translationally invariant , green function is given by the corresponding hilbert transform @xmath29 -average . dots are determined from the numerical solution of the dmft equations . solid lines are obtained by numerical integration from the linearized dmft.,width=85,height=60 ] the self - consistent dmft equations are closed through the hilbert transform @xmath30 $ ] , where @xmath31 is the non - interacting density of states . using the semi - elliptic density of states for the bethe lattice , that is @xmath32 @xcite , the anderson - falicov - kimball model can be exactly solved . in this case , we can find @xmath33 @xcite . we considered @xmath17 an independent random variables characterized by a probability function @xmath34 . we assume here a box model , i.e. , @xmath35 , with @xmath36 as the step function . the parameter @xmath37 is a measure for the disorder strength . if we consider the real and imaginary parts of @xmath25 we can use @xmath38 , what leads to @xmath39 ^ 2-u^2\alpha_{i}^2}\label{eq2}\ ] ] where @xmath40 . the chemical potential @xmath41 , corresponding to a half - filled band ( i.e. , @xmath42 ) , and @xmath43 are assumed in this paper . @xmath44 sets the energy units . at the mit , the ldos vanishes in the band center and is arbitrarily small in the vicinity of the mit ( but on the metallic side ) . the ground - state properties in the half - filled band case are solely determined by the quantum states in the band center ( @xmath45 ) . in this sense , by linearizing the dmft equations we can determine the transition points on the phase diagram . due to the symmetry of @xmath46 , in the band center we have that @xmath47 and is purely imaginary @xcite , what leads to the recursive relation @xmath48 . the left hand side in the @xmath49th iteration step is given by the result from the @xmath50th iteration step . ( [ eq2 ] ) and expanding it with respect to small @xmath51 , we find that the recursive relations within the linearized dmft with @xmath0-averaging are @xmath52^{1/q},\ ] ] where @xmath53 ^ 2}.\ ] ] at the boundary curves between metallic and insulating solutions the recursions are constant , @xmath54 , because in a metallic phase the recursions are increasing , whereas in the insulating phase they are decreasing . this observation leads directly to the exact ( within dmft ) equations determining the curves @xmath55 , i.e. , @xmath56 for the linearized dmft with @xmath0-averaging . as a function of @xmath0 for @xmath57 the point represents the exact result for the cubic lattice corresponding to @xmath58 and @xmath59.,width=85,height=60 ] in the limit @xmath57 , the equation ( [ du ] ) can be analytically solved and we can find the critical @xmath37 as a function of @xmath0 @xmath60 we plot this function in fig . [ deltaq ] . as @xmath61 is real , finite and positive , we must have @xmath62 . this shows that the anderson localizations can be detected when we consider a generalized mean with @xmath62 . in this context , one can easily note that the geometrical and the arithmetical averages ( @xmath63 and @xmath64 , respectively ) are just particular means , where , respectively , the anderson localization can either be detected or not . for @xmath65 ( upper panel ) and @xmath66 ( lower panel ) for different disorder strengths @xmath37 ( @xmath67 and @xmath68).,width=80,height=120 ] the numerical integration of equation ( [ du ] ) is shown as solid curves in fig . [ cur ] for different @xmath0 s . we can clearly see that for values of @xmath0 greater than @xmath69 , @xmath37 always increases with @xmath14 , what characterizes the mott - hubbard mit . for values of @xmath0 smaller than @xmath69 we can always identify three regions with respect to the states at the fermi level : extended gapless phase , localized gapless phase and gap phase @xcite . as already pointed out in @xcite , depending on the spectral properties of the anderson - falicov - kimball model , we distinguish three different regimes : weak , intermediate and strong interaction regime . in the weak interaction regime the mott gap is not open ; in the intermediate interaction regime , the mott gap is open at @xmath70 and in the strong interaction regime the mott gap , determined within the dmft framework with @xmath0 less than @xmath69 , is always open even at large disorder . the mott and anderson insulators are continuously connected . hence , by changing @xmath14 and @xmath37 it is possible to move from one type of insulator to the other without crossing the metallic phase . this occurs because the anderson mit(@xmath57 ) and the mott - hubbard mit ( @xmath70 ) are not associated with a symmetry breaking @xcite . when we look at the curves obtained for @xmath0 smaller than @xmath69 , where the anderson localization can be detected , we see ( fig . [ cur ] ) that the curves obtained for greater values of @xmath0 enclose those ones for smaller values of @xmath0 . if we look at the curve corresponding to the smallest value of @xmath0 we see that in order to obtain the anderson localization , one needs to consider a small range of values of @xmath14 and @xmath37 ; in the other extreme , for values of @xmath0 close to @xmath69 the range of values of @xmath14 that can be considered is large , but one needs extremely high values of @xmath37 to detect it . , for @xmath71 and @xmath72 , @xmath73 , @xmath74 , @xmath69 , @xmath75 , @xmath76 and @xmath77 . as @xmath0 decreases @xmath78 is smaller.,width=85,height=60 ] in order to understand how the numerical solutions of the curves of fig . [ cur ] are obtained one needs to observe at first the behavior of the average local density of states for different values of @xmath14 as the disorder strength @xmath37 is varied . this is made considering equations ( [ eq1])-([eq4 ] ) , from which we can obtain the stable value of @xmath46 as a function of @xmath23 , as can be seen in fig . [ graf1 ] . for @xmath72 , @xmath73 , @xmath74 , @xmath69 , @xmath75 , @xmath76 e @xmath77 and for ( a ) @xmath79 , ( b ) @xmath80 and ( c ) @xmath81 . as @xmath0 decreases @xmath78 is smaller.,width=80,height=180 ] in all simulations we considered that the initial value of @xmath82 is a uniform distribution with bandwidth @xmath83 and then we determined @xmath84 in order to obtain @xmath25 and finally the new values of @xmath82 . this loop is performed until we find the stable configuration for @xmath46 . if we consider fixed values of @xmath14 and @xmath37 we can observe the dependence of the averaged ldos on the value of @xmath0 that is used . some of these results are presented in fig . [ var1 ] and fig . [ var2 ] . in the first one , the simulations are performed in a weak interaction regime ( @xmath65 ) . in the second one , we consider an intermediate interaction regime ( @xmath66 ) . although the behavior of the ldos is not the same in both cases , we can see that the smaller the value of @xmath0 that we consider , the smaller is the value of @xmath85 . we can also see that the shape of the curves change when we compare the curves for @xmath0 smaller than @xmath69 and greater than @xmath69 . a signature of the anderson localization is the vanishing of @xmath85 as we increase @xmath37 . the way to obtain the results in fig . [ cur ] consists in considering a range of values of @xmath14 for a fixed @xmath0 and varying the disorder strength @xmath37 for each value of @xmath14 , in order to determine the values of @xmath37 when @xmath86 . in the simulations , we used values of @xmath37 varying of @xmath87 and determined the stable values of @xmath85 for each of them . as we use an iterative process , the values of @xmath85 that we obtain converge always to the stable one . in this sense , for some values of @xmath37 the stable value of @xmath85 is exactly null . in fig . [ ult ] we show the dependence of @xmath85 on @xmath37 for two different values of @xmath14 ( @xmath88 and @xmath89 ) that , again refers to the weak and intermediate interaction regime , respectively . for the first one , increasing @xmath37 at fixed @xmath14 and @xmath0 leads to further decreasing of the ldos ; just for values of @xmath0 smaller than @xmath69 , @xmath85 can vanish for a finite value of @xmath37 ; for @xmath0 greater than @xmath69 this would just happen at infinity . for @xmath0 smaller than @xmath69 we found that the critical exponent of the curves @xmath90 is always @xmath91 @xcite , which guarantees that they all belong to the same universality class . for the second plot , @xmath85 is zero until a certain value is reached , which corresponds to the transition from the gap phase to the extended gapless phase . here again just for values of @xmath0 smaller than @xmath69 , @xmath85 turns to vanish for a finite value of @xmath37 . performing the same process for other values of @xmath14 we obtain easily all the simulation points of fig . [ cur ] for a variety of different values of @xmath0 . in the present paper , we generalized the solutions of the anderson - falicov - kimball model using a hlder mean for the averaged local density of states . in this context , where each mean is characterized by the hlder parameter @xmath0 , we found that the averaged ldos can be calculated for arbitrary parameters in order to detect the anderson - localization . as we vary the parameter in the hlder mean , we emphasize different signal - values of the ldos . besides that , there is a critical value for this parameter , that is @xmath92 . for values of @xmath0 greater and smaller than this , the dependence of the critical strength of disorder @xmath37 on the coulombian repulsion @xmath14 is , respectively , similar to the one found in @xcite using the arithmetic mean ( @xmath64 ) and the geometric mean ( @xmath63 ) . we showed that not only the geometric mean can be used to detect the anderson mean , but also an infinity of other averages . one example occurs for the cubic lattice @xcite , where the exact solution @xmath59 corresponds to @xmath58 in our model ( fig . [ deltaq ] ) , which is not exactly the geometric mean . in this way , using the hlder mean , we were able to understand why in many previous works the geometric mean has offered a better approximation for the averaged ldos than the arithmetic one . the work was supported by conselho nacional de pesquisas cientificas ( cnpq ) and deutsche akademische austauschdienst ( daad ) . h. j herrmann thanks the max - planck prize . 99 n. f. mott , proc . london , sect . a * 62 * , 416 ( 1949 ) r. abou - chacra , p. w. anderson and d. j. thouless , j. phys . c. * 6 * , 1734 ( 1973 ) p. w. anderson , phys . rev . * 109 * , 1492 ( 1958 ) k. byczuk , w. hofstetter and d. vollhardt , phys . rev . lett . * 94 * , 056404 ( 2005 ) f. x. bronold and a. alvermann and h. fehske , phil . mag . * 84 * , 673 ( 2004 ) a. alvermann and h. fehske , eur . phys . j b * 48 * , 295 ( 2005 ) g. schubert , a. weisse and h. fehske1 , phys . rev . b * 71 * , 045126 ( 2005 ) e. w. montroll and m. f. schlesinger , j. stat . phys . * 32 * , 209 ( 1983 ) k. byczuk , phys . rev . b * 71 * , 205105 ( 2005 ) m. romeo , v. da costa and f. bardou , eur . phys . j b * 32 * , 513 ( 2003 ) v. dobrosavljevic , a. a. pastor and b. k. nikolic , europhys . lett . * 62 * , 76 ( 2003 ) v. dobrosavljevic and g. kotliar , phys . * 78 * , 3943 ( 1997 ) m. aizenmann and s. molchanov , commun . phys . * 157 * , 245 ( 1993 ) k. leung and f. s. csajkall , j. chem . phys . * 108 * , 9050 ( 1998 ) c. a. macedo , l. g. azevedo and a. m. c. de souza , phys . rev . b * 64 * , 184441 ( 2001 ) m. eckstein , m. kollar , k. byczuk and d. vollhardt , phys . rev . b * 71 * , 235119 ( 2005 ) k. slevin and t. ohtsuki , phys . rev . lett . * 82 * , 382 ( 1999 )	the phase diagram of correlated , disordered electron systems is calculated within dynamical mean - field theory using the hlder mean local density of states . a critical disorder strength is determined in the anderson - falicov - kimball model and the arithmetically and the geometrically averages are found to be just particular means used respectively to detect or not the anderson localization . correlated metal , mott insulator and anderson insulator phases , as well as coexistence and crossover regimes are analyzed in this new perspective .
detecting and characterising periodic variations in the brightness ( and other properties ) of astrophysical sources is a cornerstone of observational astronomy . examples include the discovery of extrasolar planetary systems , using stellar pulsations to establish the cosmological distance scale , and the study of pulsars and interacting binary star systems . nearly sinusoidal modulations are usually the result of orbital motion or rotation . however , many other astrophysical sources notably accreting sources such as interacting binary stars , young stellar objects , and active galactic nuclei ( agn ) show persistent , random ( aperiodic , stochastic , noise ) variations in their brightness driven by the complex and turbulent accretion process . see @xcite for a brief review of random time series in astronomy . the random variations in agn can be described as ` red noise ' meaning a random process with a broad power spectrum increasing smoothly in power density to low frequencies ( often with an approximately power law shape : @xmath1 , with @xmath2 ) . * henceforth ) reported the detection of periodic modulations ( with a period of @xmath3 years ) in the optical brightness of the quasar pg 1302@xmath0102 ( @xmath4 , @xmath5 , virial mass estimate @xmath6 ) based on @xmath7 years of photometric data . this was found during a search of light curves from @xmath8 spectroscopically - confirmed quasars observed with the catalina real - time transient survey ( crts ; @xcite ) . further details are discussed in ( * ? ? ? * henceforth ) . they interpreted their discovery in terms of a short - period binary supermassive black hole system ( e.g. * ? ? ? * ) . further claims for periodic optical variability in agn have been made by @xcite and @xcite . the discovery of short period binary black holes in quasars is of great importance to a number of research areas including accretion physics @xcite , hierarchical structure formation @xcite , and gravitational physics @xcite . over the years there have been many reports of periodic or quasi - periodic variations from agn , spanning the range of agn types , from radio to gamma - rays , and on timescales from minutes to years . however , this field has a chequered history . many reports of periodic variations are based on very few observed cycles of the claimed period , and a failure to properly account for the random ( red noise ) variations which can produce intervals of seemingly periodic behaviour . see @xcite for a general discussion of this point , and @xcite for some specific examples of periodicity claims drawn just from x - ray observations of nearby agn @xcite , which showed @xmath9 ` cycles ' in a single , continuous x - ray observation . ] . further observations of the same targets usually fail to show the strictly repeating , coherent oscillations expected from a truly periodic process . as we enter the era of massive time - domain surveys capable of studying @xmath10 targets , it is becoming more important to carefully assess detection procedures in order to understand and control the number of false detections . in this paper we re - examine the case of pg 1302@xmath0102 , and we consider the broader problem of how different stochastic models can make it difficult to distinguish periodic modulation among light curves selected from large time - domain surveys . figure [ fig : examples ] ( top panel ) shows the eight years of crts photometric data for pg 1302@xmath0102 fitted with a sinusoidal model . the data comprise @xmath11 v - band magnitude estimates with a mean of @xmath12 mag . the data were taken with two very similar telescopes ( css and mls ; these provided @xmath13 and @xmath14 photometric points , respectively ) . the sampling pattern is irregular , comprising nine ` seasons ' each spanning @xmath15 months with gaps of @xmath16 months . within each season there are @xmath17 nights of data , each containing four closely spaced ( @xmath18 few minutes ) photometric measurements . the error bars provided by the crts pipeline processing are in this case overestimated by a factor of @xmath19 . this effect can be seen by examining the short timescale variations in the data : the rms variation of the magnitude estimates within groups of nearby data ( each group spanning @xmath20 days , where intrinsic variability is expected to be weak , and only including groups with @xmath21 points ) is a factor @xmath22 smaller than the error bars . the data clearly show significant variations , with an rms @xmath23 mag . we fitted the data ( using weighted least squares ) with a model comprising a sinusoid plus a constant offset : @xmath24 ( this is equivalent to a model @xmath25 with amplitude given by @xmath26 and phase @xmath27 . ) the best - fitting amplitude is @xmath28 mag and the best - fitting ( observer frame ) period is @xmath29 yr , slightly different from the @xmath30 yr found by . for fitting their sinusoidal model g15a included additional archival data notably linear data @xcite extending the observational baseline . the overall fit statistic is @xmath31 for @xmath32 degrees of freedom , again indicating that the error bars are too large . comparing this model to a constant gives @xmath33 . years of @xmath34-band catalina real - time transient survey ( crts ) data for pg 1302@xmath0102 . panels ( b ) and ( c ) show example simulations of red noise with the same sampling pattern as the crts data ( black points ) plus additional data to simulate three seasons of linear data ( blue points ) . panels ( b ) and ( c ) were generated by random processes with no periodicity present ( a bending power law power spectrum , and a damped random walk , respectively ) . in each case , the continuous , error - free simulation is shown as a pink curve and the sampled data are shown as circles . the red curve shows the best - fitting sinusoid . examples ( b ) and ( c ) were randomly selected from the @xmath35 best candidates in runs of @xmath36 simulations of each process . ] it is also possible to fit the data using a stochastic model . however , is not meaningful to simply compare the @xmath37 values for these fits . when fitting stochastic models to individual time series , the @xmath37 fit statistic loses its simple meaning as a diagnostic of the ` goodness of fit ' . ( this is because the variance of the process is itself a parameter to be fitted ; the standard @xmath37 statistic only makes sense as a likelihood proxy when the variance is fixed . in fact , @xmath38 is possible for any sufficiently flexible stochastic process . see also @xcite ) . in order to compare a periodic model to a stochastic model , we have performed a bayesian model comparison between the sinusoidal model and a simple stochastic process , the damped random walk model . we first computed the posterior densities for the parameters of each model using markov chain monte carlo ( mcmc ) method . we used a method based on the ensemble sampler proposed by @xcite with @xmath39 draws based on @xmath35 ` walkers ' after removing a ` burn - in ' period . we then computed the full marginal likelihoods ( often called ` evidence ' ) of each model using the methods described in @xcite and used by @xcite . the damped random walk ( drw ) is a stochastic process often used to describe quasar variability from survey data . see @xcite , @xcite , @xcite , @xcite . it is among the very simplest continuous - time stochastic processes . the drw has an auto - covariance function @xmath40 , specified by two parameters , @xmath41 and @xmath42 , which determine the ( total ) variance and characteristic timescale , respectively . equivalently , the drw has a power spectrum ( the fourier transform of the @xmath43 ) that is flat ( power law index @xmath44 ) below @xmath45 , and smoothly bends to a power law with index @xmath46 at higher frequencies . we have fitted this model , with three parameters @xmath47 , to the mean - subtracted pg 1302@xmath0102 data . the parameter @xmath48 is a scale factor applied to the photometric error bars ( see also @xcite ) . we assigned the following prior densities on these parameters . for @xmath42 and @xmath41 ( which are positive - valued ) we assigned lognormal priors , in both cases with @xmath49 ( corresponding to @xmath50 dex ) . the means of each lognormal prior were set based on the geometric mean @xmath42 of @xmath51 d from quasar samples in @xcite , @xcite and @xcite , and to give @xmath52 mag . for @xmath48 we assigned a uniform prior . the posterior and prior densities of these parameters are shown in figure [ fig : mcmc ] ( left ) and summarised in table [ tab ] . the posterior shows @xmath48 is low , consistent with the conclusion above ( section [ sect : obs ] ) that the error bars are a factor @xmath53 too large . .summary of prior and posterior distributions . [ cols= " < , < , < , < " , ] [ tab ] we also fitted a sine model with four parameters @xmath54 to the mean - subtracted pg 1302@xmath0102 data . the model is based on equation [ eqn : sin ] with an additional , @xmath48 , to rescale the error bars . the two amplitude parameters , @xmath55 and @xmath56 , were given zero - mean normal priors ( and @xmath57 chosen to give a prior mean @xmath58 mag ( the same mean prior variance as the drw model ) . the period @xmath59 was assigned a uniform distribution ( equivalent to using a pareto prior distribution for the frequency , @xmath60 ) . the @xmath48 parameter was given the same uniform prior as for the drw model . the posterior and prior densities of these parameters are shown in figure [ fig : mcmc ] ( right ) and table [ tab ] . we then estimated the marginal likelihood of each of these models , @xmath61 where @xmath62 represents all the parameters of model @xmath63 , @xmath64 is their combined prior density , and @xmath65 is the usual likelihood function given data @xmath66 . the ratio , @xmath67 , called the ` bayes factor ' , provides a way to weigh the probabilities of two models , @xmath68 and @xmath69 . here , the two model are the drw and sine models . marginal likelihoods are usually difficult to compute . we therefore used three methods to calculate @xmath70 draws , the laplace approximation of the posterior distribution , and the @xmath71-fold cross - validation method described in @xcite ] . using all the three methods we found @xmath72 , indicating a very strong preference for the drw over the sine model . this does not mean the drw provides an adequate description of the data , only that it is strongly favoured over the sine model . this may at first seem surprising , given the smooth nature of the light curve ( top panel of figure [ fig : examples ] ) . however , close inspection reveals that the sine model fails to capture structure in the light curve ( such as the different peak magnitudes of the two maxima ) that can be modelled by the drw . this is despite the fact that pg 1302@xmath0102 was chosen to be among the most periodic from @xmath73 light curves . in the above analysis we found the simple stochastic model to be strongly preferred over the sinusoidal model , despite the undulating appearance of the light curve of pg 1302@xmath0102 . we next study how often simple stochastic processes produce nearly sinusoidal light curves by producing a number of fake time series with the same sampling pattern as the pg 1302@xmath0102 crts + linear data but generated using 1 . a gaussian noise process with a steep , bending power law ( bpl ) power spectrum , 2 . a damped random walk ( drw ) , 3 . a sinusoidal process . we refer to the three types of process as the bpl , drw , and sine models , respectively . the bpl and drw are intended to simulate observations of normal ( single bh ) quasars , but with different assumptions about the typical quasar power spectrum . we initially set the drw timescale parameter @xmath74 d , the geometric mean of the quasar samples in @xcite and @xcite . the bpl model has a steep high frequency power spectrum . this choice of model is motived by the analysis of the high - quality _ kepler _ light curves of nearby agn which showed much steeper power spectra than the standard drw model allows @xcite , with power law indices @xmath75 . steep high - frequency power spectra are also common in rapid x - ray variability of nearby agn ( e.g. * ? ? ? it has the form given in section 3.5 of @xcite , with two bend frequencies , @xmath76 and @xmath77 . we use power law indices of @xmath44 ( below @xmath76 ) , @xmath46 ( between @xmath76 and @xmath77 ) , and @xmath78 ( above @xmath77 ) . the frequencies were initially set to @xmath79 yr@xmath80 ( timescale @xmath81 yr ) and @xmath82 yr@xmath80 ( timescale @xmath83 d ) . this is based on the high quality optical power spectrum of zw 229 - 15 from _ kepler _ data @xcite with frequencies scaled down by a factor @xmath7 as expected for a @xmath84 quasar . the sine model simulations are intended to fake data as if from short - period binary quasars , with each simulation having a single modulation period randomly drawn from a uniform distribution over the range @xmath85 yr ( corresponding to a frequency distribution @xmath86 over the range @xmath87 yr@xmath80 ) . more details of the simulation procedure are given in the appendix ( section [ sect : simdets ] ) . we simulated @xmath36 time series using each of the three processes . we then fitted each with the sinusoidal model . we identified as ` periodic candidates ' all simulations for which ( a ) the fit is good ( @xmath88 ) and the improvement in the fit compared to a constant is large ( @xmath89 ) , and ( b ) the period is in the range @xmath90 yr . the first criterion is used to select data with significant variability that resembles a sinusoidal modulation ( @xmath89 is comparable to that found for pg 1302@xmath0102 above ) . the longest allowed period was chosen to match that of and who selected only periodic candidates with @xmath91 cycles in the @xmath92-year crts data . we found it necessary to impose a limit on the shortest allowed periods that is slightly longer than the typical spacing of the crts seasons . allowing shorter periods results in a large number of good fits with periods @xmath93 yr or shorter , where the quasi - periodic sampling pattern of the crts data occasionally aligns with local maxima or minima of the simulation . this is an aliasing effect also discussed in the appendix of @xcite . figure [ fig : examples ] ( b)-(c ) show examples of candidate periodicities drawn from simulations of the bpl and drw processes . apart from our choice of the range of accepted periods , the above criteria are not intended to reproduce the period detection methods of and , or any other paper . we are simply selecting time series that have a sinusoidal shape ( those that give a good match , in a least squares sense , to a sinusoid , and a poor match to a constant model ) . any reasonable period detection algorithm should be able to identify the same time series as appearing to be periodic over at least @xmath94 cycles . a more general selection and fitting procedure that allowed for non - sinusoidal periodicities and for additional trends in the data will mostly likely identify additional false periods that were not selected by our fitting , and so our method is conservative . furthermore , we selected candidates based on fits to @xmath95 data points spanning @xmath96 yr ( appropriate for a combined crts + linear dataset ) this included more points and a longer baseline than most of the crts data used by and so our criteria for selecting periodic variability are in this sense more strict . simulated data meeting our selection criteria were produced with a rate of @xmath97 per @xmath98 simulations for the drw and bpl processes , with parameters defined as above . the periods of the fitted sinusoids are long , most are in the range @xmath99 yr ( i.e. @xmath100 cycles over the simulated data ) , and the strongest cases have periods of @xmath101 yr , always near the lower limit of the allowed range . the distribution of periods is shown in figure [ fig : periods ] for the @xmath102 candidates identified by from the crts data , and from the @xmath35 strongest period detections in simulations of the bpl , drw , and sine processes . the steep spectrum ( red noise ) random processes produce nearly - sinusoidal variations when sampled intermittently , and most of these ` phantom periodicities ' show only a few cycles , typically less than three . we have repeated the simulation experiments above with different choices for the bpl bend frequencies and the drw timescale . the rate of phantom periods is highest when @xmath103 d ( drw ) or @xmath104 yr@xmath80 ( bpl ) . the power spectra that show @xmath105 peaks ( @xmath106 for the drw model ) near the observable frequency range ( @xmath107 yr@xmath80 ) produce time series with strong , smooth variations on the timescales sampled , and are mostly likely to produce phantom periods . @xcite and @xcite found a geometric mean of @xmath108 d ( @xmath109 yr ) from their drw model fitting to large samples of quasars . this is the right order of magnitude for phantom periods to be most easily produced in data spread over @xmath110few years . if the drw spectrum is modified to have a high frequency slope of @xmath111 ( rather than @xmath112 ) , the rate of phantom periods is increased by a factor of @xmath110few , to @xmath113 in @xmath51 simulations . the reduction in the rate of phantom periods with higher / lower drw characteristic timescale can be understood as follows . if the timescale above which the power spectrum flattens to @xmath114 ( @xmath115 in the drw model , @xmath116 in the bpl model ) is shorter than the @xmath117 yr inter - season spacing of the data , phantom periods are rare . in such cases the inter - season variability is essentially white noise and this is unlikely to produce smooth undulations between seasons . on the other hand , if this timescale is considerably longer than the @xmath118 yr span of the observations , and the power spectrum remains steep far below the lowest observable frequencies ( longest timescales ) , the variations will be dominated by smooth , quasi - linear trends that get weaker as @xmath76 moves lower ( in our models the total power in the power spectrum is constant , so power moves out of the observed band as @xmath76 decreases ) . the chance of the variations being dominated by a succession of roughly equally spaced peaks ( and/or troughs ) is therefore reduced ( unless one applies ` detrending ' to the data , which then increases the rate of phantom periods ) . period candidates identified by . panels ( b)-(d ) show the frequencies of the @xmath35 best candidates from @xmath36 simulations of ( b ) red noise with bpl spectrum , ( c ) red noise with a drw spectrum , ( d ) sinusoids with uniformly distributed periods . the histograms show the density of periods , i.e. each is normalised such that its total area is unity . there is a clear tendency to find phantom periodicities with long periods ( @xmath81 yr ) while truly sinusoidal signals are most easily recovered at shorter periods ( a preference for periods of @xmath119 years is expected from sampling theory given the roughly annual spacing of the observing seasons ) . ] these simulation experiments demonstrate that when trying to detect periodic signals from a large pool of red noise time series sampled like crts data , ` phantom ' periodicities will be found , and their periods tend to be near the longest allowed period ( @xmath120 cycles over the available data , assuming obvious aliasing periods are ignored ) . this effect was previously discussed by @xcite and @xcite from large surveys of quasar light curves . by contrast , genuinely periodic variations are most easily detected with periods @xmath119 yr due to the seasonal sampling of the data . we stress that the precise number of phantom periods we find should not be directly compared to the survey of , , or of @xcite , and @xcite our detection criteria are different , and our simulation experiment focusses on above - average data quality ( a bright , variable quasar ) . the key point is that light curves with a sinusoidal appearance are produced at a rate that is not insignificant , and this rate depends strongly on the power spectrum . our lack of knowledge ( or poor assumptions ) about the true range of power spectral shapes for normal quasars translates to uncertainty about ( or poor calibration of ) the rate of phantom periods in quasar surveys . that apparently sinusoidal time series are generated is not due to a problem with the analysis procedures , it is an intrinsic property of random processes with steep power spectra . time series from red noise processes , which span timescales over which the power spectrum is steep ( @xmath121 ) , will usually be dominated by smooth variations showing modulations occurring on timescales of order the length of the time series . intermittent time sampling and low signal / noise of the data make it easier to mistake a phantom periodicity for a real one . fundamentally they can only be distinguished by much longer time series spanning many cycles of the putative period ; truly periodic processes will continue to oscillate while red noise processes are progressively less likely to show further oscillations . ( an intermediate possibility , beyond the scope of this paper , is that of quasi - periodic oscillations ( qpos ) which show drifts in period , phase or amplitude . ) when searching for rare events in a very large survey of sources , it is particularly important to understand the false positive rate . once the false positive probability per source is higher than the true incidence of the event in the survey population there will ( on average ) be more false than true detections . the periodic candidate identified by , and the @xmath102 candidates identified by , were selected from @xmath122 quasars . the characteristic timescales and other properties of quasar power spectra are still only poorly understood , but are likely to depend on the mass and other properties of the agn @xcite . the quasars in any large sample spanning a range of @xmath123 and @xmath124 will likely include a range of power spectral shapes . those quasars with a high power density and steep spectra over the observed timescale range will be most likely to produce phantom periods , and a survey containing a few thousand such quasars should be expected to produce many phantom periods . used drw simulations to assess the significance of the pg 1302@xmath0102 detection . they performed two different tests , one was a simulation for each quasar in their survey , the other was an analysis of @xmath98 simulations of data like that of pg 1032@xmath0102 . the apparent significance in the latter test only demonstrates that their simulations are not good at reproducing particular properties of the data , they do not demonstrate that a period has been detected . if quasar power spectra are steeper than the drw model ( as indicated by e.g. the _ kepler _ studies cited above ) the simulation test based on drw simulations could underestimate the number of false positives . by the same argument , it is meaningless to quote as and @xcite do the detection significance of periodicities using periodogram statistics ( including the lomb - scargle periodogram ) that are calibrated against a white noise null hypothesis , when the alternative hypothesis is non - white noise . a small @xmath125-value in such cases simply rejects the white noise null hypothesis ( already known to be false ! ) , it does not necessarily support a periodic alternative . there are several other potential problems with the simulation test of . one is that the pg 1302@xmath0102 was selected to be among the most periodic of @xmath126 quasars , hence there is a large ` look elsewhere ' effect for the number of quasars searched . further , the wavelet method they used decomposes the data by time and frequency , increasing still more the ` look elsewhere ' effect . precise calibration of the survey - wide false positive rate would require one to simulate the entire distribution of quasars many times over , accounting for the plausible range of aperiodic power spectra for each quasar , and with sufficient statistics to determine the probability of a false positive to an accuracy of @xmath127 per object . another issue is the treatment of measurement errors in simulations . the crts photometric errors for pg 1302@xmath0102 are significantly overestimated ; simulating random measurement errors that are larger than the measurement errors of the real data ( as in @xcite ) will lead to unrealistic simulations . in this case , that means too much ` white noise ' in the simulations , which then reduces the probability of the simulations producing strong , smooth modulations . if we repeat our simulations tests adding random measurement errors with a standard deviation equal to the crts pipeline error bars , the number of phantom periodicities drops by more than an order of magnitude . ( our selection procedure relies on obtaining a good fit , i.e. @xmath128 , which is much harder to achieve in the presence of increased white noise . ) fortunately , in particular cases such as pg 1302@xmath0102 or pso j334.2028 + 01.4075 @xcite , the issue of stochastic or nearly - periodic variations can be resolved by further observations . if more ` cycles ' of data ideally with a higher sampling rate and improved precision match the extrapolation of the current sinusoidal model , that would strongly support the periodic model . the more future cycles that remain coherent with the model ( based on current data ) , the stronger the evidence for a true periodicity . if the light curve diverges from the model , and in an apparently random manner , that will be evidence against the periodic model . on the other hand a smooth and systematic period or phase drift could indicate the presence of an optical quasi - periodic oscillation ( qpo ) , e.g. a strong resonance in the accretion flow not related to the orbit of a binary smbh . in the short term , however , it is often more practical to ` go wide ' ( light curves from many more targets ) than to ` go deep ' ( longer , better quality time series of individual targets ) , so it is particularly important to calibrate the false positive rate of few - cycle periodicities in irregular and noisy data . if true periodicities are rare as one might expect given than most quasars are powered by accretion onto a single bh and show stochastic variations , with only a small minority in every @xmath129 field quasars ( at @xmath130 ) may be expected to harbour a short period smbh binary system . ] of spectroscopic quasars expected to be short period binary smbhs a slight underestimate in the adopted value of the false positive rate could mean the false period detections outnumber the true detections . improving our understanding of the range of quasar noise power spectra would improve any search for outliers in samples of quasar variability . any true binary accreting black hole system likely undergoes stochastic variability in addition to periodic variations , so it is not yet clear that searches for the purely periodic signals are the best approach to finding these systems . we encourage further development of methods to identify periodic and mixed periodic / stochastic processes hidden among a range of stochastic processes , and to better identify the limitations of such methods when applied to sparsely sampled photometric data . in a preliminary investigation ( applying the bayesian analysis of section [ sect : bayes ] on sinusoidal simulations from section [ sect : sims ] ) we found that even strictly sinusoidal variations were difficult to distinguish from a simple stochastic process when the number of cycles was @xmath131 , but relatively straightforward to distinguish with @xmath81 cycles . further work needs to be done to uncover the fundamental limitations of distinguishing periodic and stochastic signals , given a particular sampling strategy . in 1978 , bill press closed his article on _ flicker noises in astronomy and elsewhere _ with a note of caution about how easy it is for the eye - brain system to select ` three - cycle ' periodicities in random time series @xcite . we might note here that , as we enter the era of ` big - data ' time domain surveys , one might do well to also regard with caution few - cycle periodicities selected by machine methods when they come from large samples of noisy time series . one should not be too pessimistic , however . as our understanding of the variability of different source populations improves , we will be better able to calibrate detection procedures and realise the potential of machine learning methods for mining the large time series compendia for rare , exotic behaviour . we thank an anonymous referee for a prompt and thoughtful report . sv acknowledges support from stfc consolidated grant st / k001000/1 . wna acknowledges support from the european union seventh framework programme ( fp7/2013 - 2017 ) under grant agreement n.312789 , stronggravity . mjm acknowledges support from an stfc ernest rutherford fellowship . dh acknowledges support by the moore - sloan data science environment at nyu . this research made use of nasa s astrophysics data system . the css survey is funded by the national aeronautics and space administration under grant no . nng05gf22 g issued through the science mission directorate near - earth objects observations program . the crts survey is supported by the u.s . national science foundation under grants ast-0909182 and ast-1313422 . sv thanks tom maccarone for early discussions of some ideas in this paper , and c. bailer - jones for advice about his r code . 99 algeri , s. , van dyk , d. a. , conrad , j. , anderson , b. , 2016 , eur . j. phys . , submitted ( arxiv:1602.03765 ) andrae , r. , kim , d .- w . , bailer - jones , c. a. l. , 2013 , a&a , 554 , a137 bailer - jones , c. a. l. , 2012 , a&a , 546 , 89 begelman m. c. , blandford , r. d. , rees , m. j. , 1980 , nature , 287 , 307 charisi , m. , bartos , i. , haiman , z. , price - whelan , a. m. , graham , m. j. , bellm , e. c. , laher , r. r. , marka , s. , 2016 , mnras , submitted ( arxiv:1604.01020 ) davies , r. b. , harte , d. s. , 1987 , biometrika , 74 , 95101 drake , a. j. et al . , 2009 , apj , 696 , 870 . edelson , r. , mushotzky , r. , vaughan , s. , scargle , j. , gandhi , p. , malkan , m. , baumgartner , w. , 2013 , apj , 766 , 16 edelson , r. , vaughan , s. , malkan , m. , kelly , b. c. , smith , k. l. , boyd , p. t. , mushotzky , r. , 2014 , apj , 795 , 2 frank j. , king a. , raine d. j. , 2002 , accretion power in astrophysics : third edition . cambridge university press gierlinski , m. , middleton , m. , ward , m. , done , c. , 2008 , nature , 455 , 369 graham m. j. , et al . 2015a , nature , 518 , 74 graham m. j. , et al . , 2015b , mnras , 453 , 1562 gonzalez - martin , o , vaughan s. , 2012 , a&a , 544 , 80 goodman , j. , weare , j. , 2010 , commun . , 5 , 65 haiman , z. , kocsis , b. , menou , k. , 2009a , apj , 700 , 1952 haiman , z. , kocsis , b. , menou , k. , lippai , z. , frei , z. 2009b , cqgra , 26 , 094032 kalamkar , m. , casella , p. , uttley , p. , obrien , k. , russell , d. , maccarone , t. , van der klis , m. , vincentelli , f. , 2016 , mnras , submitted ( arxiv:1510.08907 ) kelly , b. c. , bechtold , j. , siemiginowska , a. , 2009 , apj , 698 , 895 kelly , b. c. , sobolewska , m. , siemiginowska , a. , 2011 , apj , 730 , 52 kelly , b. c. , becker , a. c. , sobolewska , m. , siemiginowska , a. , uttley , p. , 2014 , apj , 788 , 33 kozowski , s. , et al . , 2010 , apj , 708 , 927 kozowski , s. , et al . , 2016 , mnras , submitted ( arxiv:1604.01773 ) liu , t. , et al . , 2015 , apj , 803 , l16 macleod , c. l. , et al . , 2010 , apj , 721 , 1014 mchardy , i. m. , koerding , e. , knigge , c. , uttley , p. , fender , r. p. , 2006 , nature , 444 , 730 mushotzky , r. f. , edelson , r. , baumgartner , w. , gandhi , p. , 2011 , apj , 743 , l12 press w. h. 1978 , comment . astrophys . , 7 , 103 sesar b. , et al . , 2011 , aj , 142 , 190 summons , d. p. , arvalo , p. , mchardy , i. m. , uttley , p. , bhaskar , a. , 2007 , mnras , 387 , 649 timmer , j. , knig , m. , 1995 , a&a . 300 , 707710 van der klis , m. , 2006 , in compact stellar x - ray sources ( w.lewin & m. van der klis eds . ) , cambridge university press ( cambridge , uk ) , p. 39 uttley , p. , mchardy , i. m. , vaughan , s. , 2005 , mnras , 359 , 345 vaughan , s. , uttley , p. , 2006 , adspr , 38 , 1405 vaughan , s. , 2013 , phil . trans . r. soc . a , 371 , 1984 volonteri , m. , haardt , f. , madau , p. , 2003 , apj , 582 , 559 zheng , z .- y . butler , n. r. , shen , y. , jiang , l. , wang , j.x . , chen , x. , cuadra , j. , 2016 , apj , submitted ( arxiv:1512.08730 ) zu , y. , kochanek , c. s. , kozowski , s. , udalski , a. , 2013 , apj , 765 , 106 for each stochastic simulation we generated a time series with @xmath132 points and @xmath133 points per year , spanning @xmath134 years . the simulations are of linear , stationary , gaussian processes with zero mean and a smooth ( ` broad - band noise ' ) power spectrum , generated with the fast method of ( * ? ? ? * see also @xcite ) . strictly , we simulate variations in the v - band magnitude , so the simulations naturally account for the lognormal distribution of ( linear ) flux ( see * ? ? ? the power spectrum is normalised such that the expected total power ( integral over all positive frequencies ) is @xmath135 mag@xmath136 ( which translates to a fractional rms of @xmath137 for the linear fluxes ) . this is equivalent to @xmath138 mag in the notation of @xcite . to the simulated @xmath139 yr light curves we add an offset of @xmath140 , ignore the first @xmath141 years of fake data ( to mitigate ` edge effects ' in the simulation ) and use linear interpolation between the regularly spaced ( @xmath142 hr ) fake data points to recover fluxes at ` observation ' times with the same sampling pattern as the crts data for pg 1302@xmath0102 . we then add independent random , gaussian noise ( @xmath143 , @xmath144 mag ) to simulate observational noise . finally , we truncate the magnitudes to two decimal places to represent the discretisation of the crts photometric data . error bars are then assigned with @xmath145 mag , i.e. @xmath53 times larger than the random error ( as in the real data ; see section [ sect : obs ] ) . include archival photometric data in addition to crts to extend the length of their light curve . the best sampled of these is the linear data @xcite which adds another two seasons of data prior to the start of the crts data . we have simulated this by taking the sampling pattern of the first three crts seasons and reversing it around the mid - point of the first crts season , to create new time points with realistic sampling extending @xmath146 seasons before the first crts season . the fake crts + linear data contain @xmath95 data points . for each periodic simulation we employ a similar procedure except the stationary gaussian process is replaced by a sinusoid with a random phase ( in the range @xmath147 ) , an amplitude of @xmath148 mag ( obtained by fitting the pg 1302@xmath0102 data ; section [ sect : obs ] ) , and a period drawn uniformly from the range @xmath85 yr , i.e. @xmath149 cycles over the @xmath150-year duration of the crts+linear data . we then add the mean magnitude , apply the same time sampling pattern , include random errors , and discretize the magnitudes as above . , the right shows @xmath105 which better illustrates the power per decade in frequency . frequency is in units of yr@xmath80 . bpl is the ( doubly ) bending power law model with bend frequencies at @xmath151 yr@xmath152 and @xmath153 d@xmath152 . the mbpl is a modified model with the low frequency bend moved down to @xmath154 yr@xmath152 . the drw is the ` damped random walk ' model which bends to an index @xmath46 above @xmath155 . in this case @xmath74 d ( the mean from @xcite ) . mdrw is a modified drw model with a high frequency index of @xmath156 . ]	there have recently been several reports of apparently periodic variations in the light curves of quasars , e.g. pg 1302@xmath0102 by @xcite . any quasar showing periodic oscillations in brightness would be a strong candidate to be a close binary supermassive black hole and , in turn , a candidate for gravitational wave studies . however , normal quasars powered by accretion onto a single , supermassive black hole usually show stochastic variability over a wide range of timescales . it is therefore important to carefully assess the methods for identifying periodic candidates from among a population dominated by stochastic variability . using a bayesian analysis of the light curve of pg 1302@xmath0102 , we find that a simple stochastic process is preferred over a sinusoidal variations . we then discuss some of the problems one encounters when searching for rare , strictly periodic signals among a large number of irregularly sampled , stochastic time series , and use simulations of quasar light curves to illustrate these points . from a few thousand simulations of steep spectrum ( ` red noise ' ) stochastic processes , we find many simulations that display few - cycle periodicity like that seen in pg 1302@xmath0102 . we emphasise the importance of calibrating the false positive rate when the number of targets in a search is very large . [ firstpage ] methods : data analysis methods : statistical surveys , quasars : general quasars : supermassive black holes
the quantum heisenberg antiferromagnet ( qhafm ) is a prototype model to describe the magnetic properties in the parent compounds of high temperature superconducting cuprates.@xcite non - linear sigma model is shown that its ground state has long - range order and its low energy excitations are in a renormalized classical state.chakravartyprl,chakravartyprb,chubukov these results are confirmed by the quantum monte carlo methods@xcite and are consistent with experimental data.@xcite moreover , the schwinger boson mean field theory@xcite and the variational spin - wave theorytakahashi have also been developed for the two - dimensional qhafm , leading to similar conclusions . previous theoretical studies have concentrated in the low temperature region . however , the low - energy interesting physics of the two - dimensional qhafm can persist up to the temperature region @xmath0 . moreover , the anomalous normal state in the charged doped qhafm in the cuprates@xcite relies on a theory for the magnetic fluctuations in the intermediate temperature range . therefore , a theory is needed to be well defined in the full temperature range . although the schwinger boson mean field theory@xcite and takahashi s variational spin - wave theory@xcite for a qhafm in square lattices can be extended to the finite temperature range , the mean field ansatz results in a finite temperature phase transition at @xmath1 , which explicitly violates the mermin - wagner theorem@xcite . to overcome this artificial mean filed divergence of the magnetic susceptibility , we will propose a perturbation spin - wave theory . in the takahashi s variational spin - wave theory , the dyson - maleev spin - boson representation is used for the qhafm on a square lattice with a nel ordered ground state : @xmath2for the spins in the spin - up sublattice , and @xmath3for the spin - down sublattice . the dyson - maleev or the holstein - primakoff spin - wave theory works well for the ground state . the corresponding results , such as the spin - wave velocity , the spin stiffness constant , the sublattice magnetization , and the perpendicular susceptibility , igarashi , canali , zheng agree well to the quantum monte carlo simulations.@xcite to develop a finite temperature spin - wave theory , takahashi introduced a constraint for the vanishing local magnetization , i.e. , @xmath4at the @xmath5th lattice site , which automatically fulfills the mermin - wagner theorem . takahashi also introduced a mean field order parameter @xmath6 to account for the spin - wave interaction in mean field level . this mean field ansatz can describe reliably the low temperature physics , but fails at high temperatures , where it produces an artificial phase transition . in this paper , we present a perturbation spin - wave theory with the constraint eq . ( [ eqn1.3 ] ) . the spin - wave interactions are considered within a many - body perturbation method . the mean field divergence in the previous variational spin - wave theory is removed . to verify our perturbation spin - wave theory , we calculate the uniform static susceptibility . its temperature dependence in the zeroth order can catch most features given by the quantum monte carlo simulations and high temperature series expansion . it shows a linear temperature dependence at low temperatures , a smooth crossover in the intermediate range and the curie law at high temperatures . by including the first order corrections from the spin wave interactions , our perturbation spin - wave theory agrees quantitatively well with the numerical simulations . this paper is organized as follows . in sec . [ sec2 ] , we present the linearized spin wave results up to the zeroth order . in sec . [ sec3 ] , a perturbation theory with the first order corrections from the spin - wave interactions are developed . we also give the comparison of the static magnetic susceptibility with the results of quantum monte carlo simulations and high - temperature series expansions . in sec . [ sec4 ] , we give the discussion and summary . in this paper we mainly discuss a qhafm on a quare lattice with hamiltonian defined by @xmath7where @xmath8 and @xmath9 denotes nearest - neighbor sites . since the ground state of this qhafm hamiltonian is an nel ordered state , we separate the square lattice with @xmath10 sites into sublattice @xmath11 for the up spins and sublattice @xmath12 for the down spins . the dyson - maleev spin representation given by eq.([eqn1.2-a ] ) and ( eqn1.2-b ) are defined upon the nel ordered state . in terms of boson operators , the model can be divided into @xmath13where @xmath14 contains the classical energy and the quadratic terms , @xmath15and @xmath16 contains the quartic terms , describing the spin - wave interactions , @xmath17 in this section , we neglect the spin - wave interactions and focus on the hamiltonian @xmath14 . we refer this approximate to a linearized spin - wave theory ( lsw ) . the spin - wave interaction in the hamiltonian @xmath16 will be studied in sec . . the constraint eq . ( [ eqn1.3 ] ) , from the dyson - maleev representation , can be described by introducing a lagrangian hamiltonian as@xcite @xmath18where the lagrangian multipliers @xmath19 and @xmath20 are assumed to be site independent , i.e. @xmath21 . physically the lagrangian multiplier @xmath22 plays a role of an effective chemical potential for the spin waves and leads to a finite gap , thus an effective mass , for these bosonic excitations at finite temperatures . for convenience in our following discussion , we define @xmath23 with the coordinate number @xmath24 . by using the fourier transformations and performing the bogoliubov transformations , the hamiltonian @xmath25 can be expressed as @xmath26where @xmath27 and @xmath28 . here @xmath29 and @xmath30 are operators describing two branches of bogoliubov excitations . the free energy per site is given by @xmath31minimizing the free energy with respect to @xmath32 leads to a self - consistent equation @xmath33as a consequence of the constraint eq . ( [ eqn1.3 ] ) . the uniform static susceptibility per site can be calculated from the static spin - spin correlation function , @xmath34here the thermal average is defined in the zeroth order linearized spin - wave approximation as @xmath35 . it can be easily shown that the transverse spin - spin correlation function @xmath36 . therefore , only the longitudinal spin - spin correlation function @xmath37 has a finite value . the vanishing transverse spin - spin correlation function can also be shown in the takahashi s variational spin - wave theory.@xcite physically it stems from the local constraint given in eq . ( [ eqn1.3 ] ) . the longitudinal spin - spin correlation function can be calculated by introducing a weak external magnetic field @xmath38 along @xmath39-axis which introduces an additional term @xmath40 to the model hamiltonian . the longitudinal static susceptibility thus reads @xmath41 , where the free energy @xmath42 includes the contribution from the external field @xmath43 . then the uniform static susceptibility within the linearized spin - wave theory has the form : @xmath44a similar expression has also been obtained by takahashi in his variational spin - wave theory.@xcite the ground state of the qhafm on a square lattice has a long - range ordertakahashi , hirsch due to the presence of bose - einstein condensation ( bec ) at the momentum @xmath45 . since @xmath46 at zero temperature , the spin - wave spectrum is given by @xmath47comparing to the takahashi s spin wave spectrum@xcite @xmath48 with @xmath49 , we find that it can be obtained by rescaling our spin - wave result as @xmath50the additional factor @xmath51 is the main difference at low temperatures between our linearized spin - wave theory and takahashi s variational theory . the spin - spin correlation function in the long distant limit can be derived as @xmath52where @xmath53 if @xmath54 are in the same sublattice , and @xmath55 if @xmath54 belong to different sublattices . @xmath56 is the magnetization at zero temperature , which is the same as that in the schwinger boson mean field theory@xcite . both the spin - spin correlation function and the local spontaneous magnetization @xmath57 in our linearized spin - wave theory agree with those of the variational spin - wave theory.@xcite these results imply that , to the order of @xmath58 , the spin - wave interactions have no contributions to the spontaneous magnetization and static spin - spin correlation at @xmath59k , consistent to the previous results.@xcite in the linearized spin - wave theory for the qhafm on a square lattice with different spin magnitude @xmath60 and @xmath61 . ] for @xmath60 and @xmath61 in our linearized spin wave theory of the qhafm on a square lattice . ] at finite temperatures , according to the mermin - wagner theorem , the long - range order is destroyed in the qhafm model by quantum fluctuations.@xcite fig . [ fig2.1 ] shows the temperature dependence of the lagrangian multiplier @xmath62 with different quantum spins @xmath63 . at low temperatures @xmath64 , we follow the takahashi s method and show analytically that @xmath32 approaches exponentially to its bec value @xmath65 as @xmath66the uniform static spin susceptibility in this low temperature region can also be derived analytically , leading to a linear - temperature dependence , @xmath67the first term gives a finite spin susceptibility at zero temperature . the linear term results from the linear density of states of the spin - wave excitations in the low energy limit . the low - energy spin - wave excitations at low temperatures have an approximate dispersion @xmath68where a dimensionless bosonic gap is defined by @xmath69 . at low temperatures , @xmath32 decays exponentially to @xmath65 as eq.([eqn2b.1 ] ) , therefore the energy spectrum eq.([eqn2b.3 ] ) is approximately as @xmath70 . the density of states @xmath71 of the spin - wave excitations is @xmath72 . it is this linear density of states that leads to the linear - temperature dependence of the static magnetic susceptibility at low temperatures . note that by scaling the linear temperature susceptibility @xmath73 by the factor @xmath51 , we can obtain the low - temperature susceptibility given by takahashi,@xcite @xmath74 . at low temperatures , the spin - spin correlation function follows @xmath75where @xmath76 is defined in eq.([eqn2a.3 ] ) and @xmath77 is the correlation length defined by @xmath78 . from eq.([eqn2b.1 ] ) , it can be easily shown that @xmath79takahashi s result for the spin - spin correlation length @xmath80 can be obtained by rescaling with the factor @xmath51 defined in eq.([eqn2a.2 ] ) , @xmath81 . this exponential temperature dependence of the spin correlation length agrees with that obtained from one - loop renormalization group approach of the non - linear sigma model.kopietz,manousakis compared to the experimental data and the numerical simulations , the two - loop approximation is necessary to convert the @xmath82 prefactor of the exponential into a constant.chakravartyprl,chakravartyprb when increasing the temperature , the lagrangian multiplier @xmath62 in fig . [ fig2.1 ] evolves from the low - temperature exponential behavior to the linear temperature dependence at high temperatures . at the same time , the uniform susceptibility as shown in fig . [ fig2.2 ] increases firstly to a broad peak at the intermediate temperature @xmath83 , and then decreases at high temperatures . at high temperatures @xmath84 , the dispersion of the energy spectrum becomes relatively weak and @xmath85 . the self - consistent equation eq.([eqn2.7 ] ) gives rise to @xmath86 in the high - temperature region , the uniform static magnetic susceptibility obeys the curie law , @xmath87such a high - temperature curie - law behavior in the linearized spin - wave theory is exactly the same as that in the takahashi s theory.@xcite why our linearized spin - wave theory can give rise to a curie - law behavior , in contrast to the general spin - wave theory is a very interesting question . the basic reason is that there is an effective chemical potential introduced for our spin waves by the local constraint eq.([eqn1.3 ] ) . this gives rise to a temperature dependent finite gap for the spin waves which behave as bosons with effective mass at finite temperatures . it should be noted that in our numerical solution to the self - consistent equation eq.([eqn2.7 ] ) and in calculating the uniform magnetic susceptibility eq.([eqn2.9 ] ) , there is a trick to include the contribution of the low - energy spin - wave excitations at low temperatures . although no exact bec occurs at finite temperatures , in the actual numerical calculation eq.([eqn2b.1 ] ) shows us that the exponential decay of the bosonic gap @xmath88 at @xmath45 will lead to a difficulty in numerically calculating the momentum integral from the low - energy region @xmath89 . we call this difficulty as a bec - like pseudo - singularity . a trick to deal with the bec - like pseudo - singularity is to separate the momentum space into two regions : region ( i ) @xmath90 and region ( ii ) is the rest part in the first brillouin zone . here the momentum @xmath91 satisfies @xmath92 . @xmath93 is an irrelevant parameter and we choose @xmath94 in our calculations . the momentum integral in region ( i ) can be firstly calculated with an approximate spin - wave energy dispersion eq.([eqn2b.3 ] ) . linearized spin - wave theory ( lsw ) and the takahashi s variational theory@xcite for the qhafm on a square lattice . ] for comparison , the uniform static magnetic susceptibility from our linearized spin - wave theory and the takahashi s variational results are shown in fig . [ fig2.3 ] . the sharp peak value at @xmath95 in the takahashi s theory comes from an artificial phase transition due to vanishing of the mean - field ansatz . at low temperatures , the static magnetic susceptibility from our linearized spin - wave theory is slightly larger than the takahashi s result . at high temperatures , the curie law occurs exactly in both theories . clearly , the artificial phase transition has been removed in the linearized spin - wave theory . one special feature in the uniform magnetic susceptibility shown in fig . fig2.2 is a kink structure around the temperature @xmath96 , where the slop of the susceptibility changes from a smaller value into a larger one . this special feature is also observed in the quantum monte carlo data . @xcite it was argued to originate from a crossover from a renormalized classical regime into a quantum critical regime . however , the monte carlo simulation shows that a similar abrupt change also appears in a classical heisenberg model.@xcite therefore , whether there is a crossover is still controversial . however , this kink structure implies obviously different scenarios occurring in these two temperature regions . in this section , we will go beyond the linearized approximation by including the first order corrections from the spin - wave interactions of the qhafm model . the method is based on the many - body perturbation theory . after the fourier and bogoliubov transformations , the total hamiltonian @xmath97 can be approximated by@xcite @xmath98 , \notag \\ e_{c } & = & \sum_{\mathbf{k}}\varepsilon _ { \mathbf{k}}+njzs\left [ s-\lambda \left ( 1 + 2s\right ) \right ] -\frac{njz}{4}c_{1}^{2 } , \label{eqn4.2}\end{aligned}\]]where @xmath99 , @xmath100 , @xmath101 , @xmath102 , and @xmath103 represents the spin - wave energy spectrum with corrections from the normal ordering of @xmath16 . at zero temperature , @xmath104 and the spin - wave energy dispersion is given by @xmath105 with @xmath106 for @xmath107 . the parameters @xmath108 and @xmath109 are defined by @xmath110 in the above approximation , we have ignored the off - diagonal parts in the fock space and picked up all the terms of the two spin - wave interactions . this is enough when we only consider the first order corrections from the spin - wave interactions for the free energy , since the linked cluster theorem tells that @xmath111 , where the thermal average is defined by @xmath112 and @xmath113 is defined for the free bosons with energy spectrum @xmath114 . the free energy per lattice site reads @xmath115 -\frac{jz}{2}c_{2}^{2}+\frac{1}{2n}e_{c } , \label{eqn3.2}\]]where @xmath116 and @xmath117 . minimizing this free energy , we can obtain the following self - consistent equation for the lagrangian multiplier @xmath32 , @xmath118where @xmath119 and @xmath120it should be noted that the self - consistent equation eq.([eqn3.2 ] ) is physically equivalent to the first order perturbation expansion for the constraint eq.([eqn1.3 ] ) , @xmath121 the uniform static susceptibility can also be calculated from the spin - spin correlation function given in eq.([eqn2.8 ] ) . when the first order corrections from the spin - wave interactions are included , the transverse static susceptibility has a small but finite contribution , @xmath122 in the derivation for @xmath123 , we have used the perturbation constraint equation eq.([eqn3.3 ] ) and ignored the higher order terms of the spin - wave interactions . the longitudinal static susceptibility can also be calculated from the free energy by introducing a weak external field hamiltonian @xmath43 , yielding @xmath124 , \label{eqn3.5}\]]where the second term comes from the two spin - wave interaction compared to the linearized spin - wave expression eq.([eqn2.9 ] ) , and @xmath125the uniform static susceptibility is thus given by @xmath126 . \label{eqn3.6}\ ] ] the self - consistent equation eq.([eqn3.2 ] ) and the static magnetic susceptibility eq.([eqn3.6 ] ) have been numerically calculated . when the first order corrections of the spin - wave interactions are included , the temperature dependence of both the lagrangian multiplier @xmath32 and the uniform static magnetic susceptibility have similar behaviors to the linearized spin - wave theory . the lagrangian multiplier @xmath32 is shown in fig . [ fig3.1 ] for different spin @xmath60 and @xmath61 . at low temperatures , it decays exponentially to its zero - temperature value @xmath65 , indicating the gapless spin wave excitations . at high temperatures , @xmath32 has a linear - temperature variation @xmath127 , where @xmath128 and @xmath129 have weak temperature dependence . for @xmath107 , @xmath130 and @xmath131 in the temperature region @xmath132 . the static magnetic susceptibility is displayed in fig . [ fig3.2 ] . it has three main features : ( i ) the linear - temperature behavior at low temperatures , ( ii ) a broad peak at an intermediate characteristic temperature @xmath83 , ( iii ) the curie - weiss law at high temperatures . for @xmath107 , a low - temperature linear behavior can be fitted by @xmath133 , with @xmath134 and @xmath135 . compared to the takahashi s theory where a similar linear behavior with @xmath136 and @xmath137,@xcite our zero temperature susceptibility @xmath138 is slightly smaller than that of takahashi s . also the slope @xmath139 is also slightly larger than that of the takahashi s . and @xmath61 in our perturbation spin - wave theory with the first order corrections from the spin - wave interactions . ] at high temperatures , the spin - wave interactions modify the exact curie law of the static magnetic susceptibility eq.([eqn2b.7 ] ) into a curie - weiss behavior . as shown in fig . [ fig3.3 ] , it evolves into the behavior obtained from the numerical high - temperature series expansions . physical origin of the high - temperature curie - weiss law also comes from the linear - temperature dependence of lagrangian multiplier @xmath32 in the energy spectrum of the spin waves . it should be noted that the kink structure in the static magnetic susceptibility in the linearized spin - wave theory shown in fig . [ fig2.2 ] also occurs in the perturbation theory . this implies that the kink feature is a characteristic property in the qhafm model . comes from ref . and qmc@xmath140 comes from ref . . htse comes from eq.(@xmath141 ) of ref . . for comparison , qmc@xmath140 have been multiplied by a factor @xmath142 . ] in fig . [ fig3.3 ] , we compare the static magnetic susceptibility of our spin - wave theory to the quantum monte carlo data@xcite and the high - temperature series expansions . it shows that our linearized spin - wave theory can reproduce all the main features in the numerical simulations : the low temperature linear behavior , the smooth crossover in the intermediate temperature region and the curie - weiss law behavior at high temperature . this good agreement encourages us to study the frustrated @xmath143 model proposed to describe the magnetic fluctuation in the newly discovered iron - pnictide superconductors.@xcite the first order correction from the spin - wave interactions makes the uniform static susceptibility both at low and high temperatures agrees quantitatively with the numerical quantum monte carlo and high temperature series expansions . this good fitting in fig . [ fig3.3 ] indicates that the physics of qhafm model , at least to the static magnetic response , has been captured by our perturbation spin - wave theory in the full temperature region . the good agreement of our perturbation spin - wave theory with the numerical simulations shows the validity of our theory at finite temperatures . it seems surprising because the spin - wave theory developed in the pioneering papers@xcite is generally thought to work well only at low temperatures in the ordered phase . the main difference of our finite temperature spin - wave theory is the requirement of the vanishing local magnetization . it is this local constraint that restricts the number of the spin wave excitations which thus behaves as bosons with an effective mass at finite temperatures . the extension of the spin - wave theory to the low - dimensional spin systems have been shown to work well . in one - dimensional quantum magnets , no magnetic order can survive at zero temperature due to the strong quantum fluctuations . although there is no long range order in the ground state , the free energy and the static magnetic susceptibility of the one - dimensional ferromagnet from a modified spin - wave theory show same behaviors to bethe - ansatz solutions.@xcite the haldane gap for the one - dimensional quantum antiferromagnet with @xmath144 from a spin - wave theory is @xmath145 ( ref . ) which well agrees with the numerical density - matrix renormalization group ( @xmath146 from ref . ) . the spin - wave theory for the one - dimensional quantum antiferromagnet with integral spin can reproduce the static magnetic susceptibility which agrees well with the quantum monte carlo simulations.@xcite these previous results , together with our study , indicate that the formulation of the spin - wave theory is a valid formalism to apply to the low - dimensional quantum magnets . in summary , we have studied the qhafm on a square lattice by using a perturbation spin - wave theory . the lagrangian constraint for the absence of the local magnetization plays an important role and leads to an effective mass for the spin wave excitations at finite temperatures . in linearized spin - wave theory , the calculation of the uniform static magnetic susceptibility can reproduce almost all features obtained from the quantum monte carlo and the high - temperature series expansions . it has a linear - temperature dependence at low temperatures and a broad peak feature at an intermediate characteristic temperature @xmath83 . at high temperatures , the uniform magnetic susceptibility follows a curie - weiss law . the static magnetic susceptibility from our perturbation theory with the first corrections from the spin - wave interactions agrees well with the numerical simulations .	a perturbation spin - wave theory for the quantum heisenberg antiferromagnets on a square lattice is proposed to calculate the uniform static magnetic susceptibility at finite temperatures , where a divergence in the previous theories due to an artificial phase transition has been removed . to the zeroth order , the main features of the uniform static susceptibility are produced : a linear temperature dependence at low temperatures and a smooth crossover in the intermediate range and the curie law at high temperatures . when the leading corrections from the spin - wave interactions are included , the resulting spin susceptibility in the full temperature range is in agreement with the numerical quantum monte carlo simulations and high - temperature series expansions .
studies of the clustering of cosmological dark matter have progressed enormously in the past twenty years . the dynamical evolution of the dark matter is driven by gravity and fully specified initial conditions are provided in current cosmological models . this problem can therefore be attacked quite cleanly using n - body simulations ( see jenkins et al . 1998 , gross et al . 1998 and references therein . ) studies of the clustering properties of galaxies , on the other hand , are much more complicated because galaxy formation includes messy astrophysical processes such as gas cooling , star formation and feedback from supernovae . these processes couple with the gravitational evolution of the dark matter to produce the clustering pattern of galaxies . because of this complexity , progress in understanding galaxy clustering has been slow . yet , theoretical modelling of galaxy clustering is essential if we are to make the most of the new generation of galaxy redshift surveys , the two - degree field ( 2df , colless 1996 ) and sloan digital sky survey ( sdss , gunn & weinberg 1995 ) , and of the new data on galaxy clustering at high redshift that has been accumulating recently ( e.g. adelberger et al . 1998 , governato et al . 1998 , baugh et al . 1999 ) . two kinds of simulation techniques are being used to approach galaxy clustering from a theoretical standpoint . the first of these attempts to follow galaxy formation by simulating directly dark matter and gas physics in cosmological volumes ( eg . katz et al . 1992 , evrard et al 1994 , weinberg et al . 1998 , blanton et al 1999 , pearce et al . 1999 ) . because the resolution of such simulations is limited , phenomenological models are required to decide when and where stars and galaxies form and to include the associated feedback effects . the advantage of this approach is that the dynamics of cooling gas are calculated correctly without the need for simplifying assumptions . the disadvantage is that even with the best codes and fastest computers available , the attainable resolution is still some orders of magnitude below what is required to resolve the formation and internal structure of individual galaxies in cosmological volumes . for example , the gas resolution element in the large eulerian simulations of blanton et al . ( 1999 ) is around half a megaparsec . lagrangian hydrodynamic methods offer better resolution , but even in this case , this is poorer than the galactic scales on which much of the relevant astrophysical processes occur . a different and complementary approach to studying galaxy clustering is to use semi - analytic models of galaxy formation . in this case , resolution is generally not a major issue . the disadvantage of this technique , compared to hydrodynamic simulations , is that , in calculating the dynamics of cooling gas , a number of simplifying assumptions , such as spherical symmetry or a particular flow structure , need to be made ( some of these assumptions are tested against smoothed particle hydrodynamics simulations by benson et al . 1999 ) . as in the direct simulation approach , a model for star formation and feedback is required . in addition to adequate resolution , semi - analytic modelling offers a number of advantages for studying galaxy clustering . firstly , it is a much more flexible approach than full hydrodynamic simulation and so the effects of varying assumptions or parameter choices can be readily investigated . secondly , with detailed semi - analytic modelling it is possible to calculate a wide range of galaxy properties such as luminosities in any particular waveband , sizes , bulge - to - disk ratios , masses , circular velocities , etc . this makes it possible to construct mock catalogues of galaxies that mimic the selection criteria of real surveys and to investigate clustering properties as a function of magnitude , colour , morphological type or any other property determined by the model . semi - analytic modelling has been used in two different modalities to study galaxy clustering . in the first , an analytic model for the clustering of dark matter halos developed by mo & white ( 1996 ) is assumed and the semi - analytic machinery is used to populate halos , generated using monte - carlo techniques , with galaxies . this technique has been extensively applied by baugh et al . ( 1998 , 1999 ) . in the second , more direct , approach , the semi - analytic modelling is applied to dark matter halos grown in a cosmological n - body simulation . the advantages of this latter strategy are that it allows a proper treatment of the small scale regime where the mo & white model breaks down and it bypasses any inaccuracies in the analytic ( press - schechter ) model used to compute the mass function of dark halos in the pure semi - analytic approach . this technique has been implemented in two ways . in the simplest case ( kauffmann , nusser and steinmetz , 1997 , roukema et al . 1997 , governato et al . 1998 ) , a statistical merger tree for each halo identified in the n - body simulation is generated in a monte - carlo manner . in the second implementation ( kauffmann et al . 1999a , 199b , diaferio et al . 1999 ) , the halo merger trees are extracted directly from the n - body simulation . in this paper , we adopt the first approach to the combined use of semi - analytic and n - body techniques ( i.e. with monte - carlo merger trees ) to study galaxy clustering . we focus on the specific question of how the process of galaxy formation couples with the large scale dynamics of the dark matter to establish the clustering properties of the galaxy population . we investigate in detail processes that bias galaxies to form preferentially in certain regions of space . previous cosmological dark matter simulations have established that the dark matter in popular cdm models tends to be more strongly clustered on small scales than the observed galaxy population ( jenkins et al . 1998 , gross et al . 1998 ) . we investigate whether the required antibias arises naturally in these cosmologies . more generally , we compare the predictions of these models with observations over a range of scales . the techniques that we use are described in [ sec : desc ] . the clustering properties of galaxies in our model are presented in [ sec:2pt ] . the various processes that play a role in determining how the galaxy distribution is biased relative to the mass are discussed in [ sec : nature ] . in [ sec : norms ] , we show that our results are robust to changes in model parameters and finally in [ sec : discuss ] we discuss our main conclusions . the two techniques that we employ in this paper , n - body simulations and semi - analytic modelling , are both well established and powerful theoretical tools . we do not intend to describe them in detail here , but instead refer the reader to the appropriate sources . we use the semi - analytic galaxy formation model of cole et al . ( 1999 ) to populate dark matter halos with galaxies . the merger history of dark matter halos is followed using a monte - carlo approach based on the extended press - schechter formalism ( press & schechter 1974 ; bond et al . 1991 ; bower 1991 ; lacey & cole 1993 ) . within each halo , galaxy formation is followed using a set of simple , physically - motivated rules that model the processes of gas cooling , star formation , feedback from supernovae and stellar evolution . the result is a fully specified model of galaxy formation with a relatively small number of free parameters which can be fixed by constraining the model to match the observed properties of the local galaxy population ( e.g. the luminosity function or the tully - fisher relation ) . once constrained in this way , the model makes predictions for a whole range of galaxy properties ( e.g. colours , sizes , bulge - to - disk ratios , rotation speeds etc . ) at both the present day and at high redshift . in this work we extend this list of predictions to include the spatial clustering of galaxies , in particular the two - point correlation function . whilst our models are similar in principle to those of kauffmann et al . ( 1999a ) they use different rules for star formation and feedback and also include the effects of chemical enrichment due to star formation . we also choose to constrain the model parameters in a rather different way , as discussed below . [ cols="^,^ " , ] we have shown that matching the local galaxy luminosity function our preferred method for constraining the parameters of our semi - analytic model leads to model predictions for galaxy clustering that are robust to reasonable changes in these parameters . kauffmann et al . ( 1999a ) adopted a different philosophy : they chose to constrain their models by matching the i - band tully - fisher relation , rather than the luminosity function . we explore the effect of this choice by constraining our own models in a similar way . specifically , we require the median magnitude of central spiral galaxies with halo circular velocities in the range 215.0 to 225.0 km s@xmath3 to be @xmath4 ( where @xmath5 is the dust - extincted i - band magnitude of each galaxy corrected to the face - on value ) . this can be achieved by a suitable choice of the luminosity normalisation parameter , @xmath6 , and leads to a model tully - fisher relation that agrees well with data from mathewson , ford & buchhorn ( 1992 ) . since our original @xmath7cdm models ( that is , the reference model and its variants ) already agreed quite well with the tully - fisher relation , ( see fig . [ fig : reftf ] ) , this different choice of constraint has only a minor effect on the correlation function . the only noticeable change is an increase in the scatter of the asymptotic bias in the variant models . in the @xmath8cdm models , on the other hand , the new constraint has an important effect because the original models that agreed well with the luminosity function , missed the tully - fisher relation by about 1 magnitude . forcing a fit to the tully - fisher relation destroys the good agreement of the reference model with the luminosity function , as may be seen in fig . [ fig : nofeedb ] . this figure also shows a @xmath8cdm model in which we have attempted to obtain a better luminosity function by dramatically reducing the amount of supernovae feedback into the interstellar gas . [ fig : nofeedb ] shows that our two @xmath8cdm models that match the tully - fisher relation have different correlation functions . in other words , this exercise demonstrates that when models are constrained in this way , the resulting correlation functions are rather sensitive to the choice of model parameters . this explains why kauffmann et al . ( 1999a ) concluded that their clustering predictions depended strongly on the way they parametrised star formation , feedback and the fate of reheated gas in their model . by contrast , we have found that our predictions for the correlation function are robust to changes in model parameters , _ so long as the models match the bright end of the galaxy luminosity function_. in this study , we have considered some of the physical and statistical processes that determine the distribution of galaxies and its relation to the distribution of mass in cold dark matter universes . the approach that we have adopted exploits two of the most successful techniques currently used in theoretical cosmological studies : n - body simulations to follow the clustering evolution of dark matter and semi - analytic modelling to follow the physics of galaxy formation . our main conclusion is that the efficiency of galaxy formation depends in a non - trivial fashion on the mass of the host dark matter halo and , as a result , galaxies , in general , have a markedly different distribution from the mass . this result had been anticipated in early cosmological studies ( eg . frenk , white & davis 1983 , davis et al . 1985 , bardeen et al . 1986 ) , but it is only with the development of techniques such as semi - analytic modelling that realistic calculations have become possible . the statistics of the spatial distribution of galaxies reflect the interplay between processes that determine the location where dark matter halos form and the manner in which halos are lit up " by galaxy formation . if the resulting mass - to - light ratio of halos were independent of halo mass , then the distribution of galaxies would be related in a simple manner to the distribution of dark matter in halos . in current theories of galaxy formation , however , the mass - to - light ratio has a complicated dependence on halo mass . on small mass scales , galaxy formation is inhibited by the reheating of cooled gas through feedback processes , whereas in large mass halos it is inhibited by the long cooling times of hot gas . as a result , the mass - to - light ratio has a deep minimum at the halo mass , @xmath9 , associated with @xmath1 galaxies , where galaxy formation is most efficient . although our calculations assume a specific model of galaxy formation , the dependence of mass - to - light ratio on halo mass displayed in fig . [ fig : mllf ] is likely to be generic to this type of cosmological model . the consequence of such a complex behaviour is a scale dependent bias in the distribution of galaxies relative to the distribution of mass . on scales larger than the typical size of the halos that harbour bright galaxies , the bias in the galaxy distribution is related in a simple way to the bias in the distribution of massive halos . in our @xmath10 @xmath8cdm model , galaxies end up positively biased on large scales , but in our flat , @xmath11 @xmath7cdm model , they end up essentially unbiased . on small scales , the situation is more complicated and the correlation function depends on effects such as the spatial exclusion of dark matter halos , dynamical friction , and the number of galaxies per halo . in particular , our simulations show how the statistics of the halo occupation probability influence the amplitude of the galaxy correlation function on sub - megaparsec scales . in our models , the occupation of halos by galaxies is not a poisson process . since the amount of gas available for star formation is limited , the mean number of pairs per halo is less than that of a poisson distribution with the same mean . this property plays an important role in determining the amplitude of small scale correlations . remarkably , the correlation function of galaxies in our @xmath7cdm model closely approximates a power - law over nearly four orders of magnitude in amplitude . this is in spite of the fact that the correlation function of the underlying mass distribution is not a power - law , but has two inflection points in the relevant range of scales . somehow , the various effects just discussed conspire to compensate for these features in the mass distribution . in particular , on scales smaller than @xmath12 mpc , the galaxy distribution in the @xmath7cdm model is _ antibiased _ relative to the mass distribution . the apparently scale - free nature of the galaxy correlation function in this model seems to be largely a coincidence ( although whether this is also true of the real universe remains to be seen ) . our @xmath8cdm model which has similar physics although a different initial mass fluctuation spectrum , does not end up with a power - law galaxy correlation function . coln et al . ( 1999 ) have carried out a very high resolution n - body simulations ( of a @xmath7cdm cosmological model similar to ours ) which resolves some substructure within dark matter halos . the correlation function of these sub - halos is remarkably similar to the correlation function of the galaxies in our @xmath7cdm reference model . in a sense , the merger trees of our semi - analytic models keep track of sub - halos within dark matter halos since they follow the galaxies that form within them . ( unlike the simulations , however , the semi - analytic model does not follow the spatial distribution of sub - halos . ) coln et al . select sub - halos by circular velocity , whereas we select galaxies by luminosity . since , in our models , the luminosity of a galaxy is correlated with the circular velocity of the halo in which it formed , there is some correspondence between the type of objects studied by coln et al . and us . however , the connection could be complicated by effects such as tidal disruption or stripping of halos within halos but which are not included in our model . nevertheless , the abundance of sub - halos considered by coln et al is similar to the abundance of galaxies in our @xmath7cdm model and this might account for the similarity between the two correlation functions . another noteworthy outcome of our simulations is the close match of the galaxy correlation function in our @xmath7cdm model to the observed galaxy correlation function , itself also a power - law over a large range of scales ( groth & peebles 1977 ; baugh 1996 ) . this match is particularly interesting because the parameters that specify our semi - analytic galaxy formation model were fixed beforehand by considerations that are completely separate from galaxy clustering ( see cole et al . 1999 ) . our procedure for fixing these parameters places special emphasis on obtaining a good match to the observed galaxy luminosity function ( c.f . fig . [ fig : reflf ] ) , but makes no reference whatsoever to the spatial distribution of the galaxies . to summarize , the combination of high resolution n - body simulations with semi - analytic modelling of galaxy formation provides a useful means for understanding how the process of galaxy formation interacts with the process of cosmological gravitational evolution to determine the clustering pattern of galaxies . in general , we expect galaxies to be clustered somewhat differently from the dark matter , and the relation between the two can be quite complex . a flat cdm model with @xmath2 gives an acceptable match to the observed galaxy correlation function over about four orders of magnitude in amplitude ( as does an open model with the same value of @xmath13 . ) the @xmath7cdm model is also in reasonable agreement with a number of other known properties of the galaxy distribution . ajb , smc and csf acknowledge receipt of a pparc studentship , advanced fellowship and senior fellowship respectively . csf also acknowledges a leverhume fellowship . cgl was supported by the danish national research foundation through its establishment of the theoretical astrophysics center . this work was supported in part by a pparc rolling grant , by a computer equipment grant from durham university and by the european community s tmr network for galaxy formation and evolution . we acknowledge the virgo consortium and the gif collaboration for making available the gif simulations for this study , and the anonymous referee for insightful and helpful comments . adelberger k. l. , steidel c. c. , giavalisco m. , pettini m. , kellogg m. , 1998 , apj , 505 , 18 bagla j. s. , 1998 , mnras , 299 , 417 bardeen j.m . , bond j.r . , kaiser n. , szalay a.s . , 1986 , apj , 304 , 15 baugh c. m. , 1996 , mnras , 280 , 267 baugh c. m. , cole s , frenk c. s. , lacey c. g. , 1998 , apj , 498 , 504 baugh c. m. , benson a. j. , cole s. , frenk c. s. , lacey c. g. , 1999 , mnras , 305 , l21 blanton m. , cen r. , ostriker j. p. , strauss m. a. , tegmark m. , 1999 , astro - ph/9903165 bond j. r. , kaiser n. , cole s. , efstathiou g. , 1991 , apj , 379 , 440 bower r.g . , 1991 , mnras , 248 , 332 cole s. , lacey c. g. , baugh c. m. , frenk c. s. , 1999 , mnras submitted coln p. , klypin a. , kravtsov a. , khokhlov a. , 1999 , apj , 523 , 32 colless m. , 1996 , in proceedings of the heron island conference , http://msowww.anu.edu.au/ heron / colless / colless.html davis m. , huchra j. , latham d. w. , tonry j. , 1982 , apj , 253 , 423 davis m. , peebles p. j. e. , 1983 , apj , 267 , 465 davis m. , efstathiou g. , frenk c. s. , white s. d. m. , 1985 , apj , 292 , 371 diaferio a. , kauffmann g. , colberg j. m. , white s. d. m. , 1999 , mnras , 307 , 537 efstathiou g. , frenk c.s . , white s.d.m . , davis m. , 1988 , mnras , 235 , 715 efstathiou g. , bond j. r. , white s.d.m . , 1992 , mnras , 258 , 1 evrard a. e. , summers f. j. , davis m. , 1994 , apj , 422 , 11 ferrara a. , bianchi s. , cimatti a. , giovanardi c. , 1999 , apjs , 123 , 437 frenk c. s. , white s. d. m. , davis m. , 1983 , apj , 271 , 417 gardner j. p. , sharples r. m. , frenk c. s. , carrasco b. e. , 1997 , apj , 480 , l99 glazebrook k. , ellis r. , colless m. , broadhurst t. , allington - smith j. , tanvir n. , 1995 , mnras , 273 , 157 governato f. , baugh c. m. , frenk c. s. , cole s. , lacey c. g. , quinn t. , stadel j. , 1998 , nat , 392,359 gross m. a. k. , somerville r. s. , primack j. r. , holtzman j. , klypin a. , 1998 , mnras , 301 , 81 gunn j. e. , weinberg d. h. , 1995 , wide - field spectroscopy and the distant universe , proceedings of the 35th herstmonceux workshop , cambridge university press , cambridge groth e. j. , peebles p. j. e. , 1977 , apj , 217 , 38 ella m. , scaramella r. , stirpe g. m. , vettolani g. , 1999 , astro - ph/9901378 hamilton a. j. s. , kumar p. , lu e. , matthews a. , 1991 , apj , 374 , l1 huchra j. p. , davis m. , latham d. w. , tonry j. , 1983 , apjs , 53 , 89 jenkins a. , frenk c. s. , pearce f. r. , thomas p. a. , colberg j. m. , white s. d. m. , couchman h. m. p. , peacock j. p. , efstathiou g. , nelson a. h. , ( the virgo consortium ) , 1998 , apj , 499 , 20 jing y. p. , 1998 , apj , 503 , l9 katz n. , hernquist l. , weinberg d. h. , 1992 , apj , 399 , 109 kauffmann g. , nusser a. , steinmetz m. , 1997 , mnras , 286 , 795 kauffmann g. , colberg j. m. , diaferio a. , white s. d. m. , 1999a , mnras , 303 , 188 kauffmann g. , colberg j. m. , diaferio a. , white s. d. m. , 1999b , mnras , 307 , 529 kennicutt , r. 1983 , apj , 272 , 54 lacey , c.g . , cole , s. , 1993 , mnras , 262 , 627 lacey c. g. , cole s. , 1994 , mnras , 271 , 676 lemson g. , kauffmann g. , 1999 , mnras , 302 , 111 loveday j. , peterson b. a. , efstathiou g. , maddox s. j. , 1992 , apj , 390 , 338 mathewson d. s. , ford v. l. , buchhorn m. , 1992 , apjs , 81 , 413 mo h. j. , white s. d. m. , 1996 , mnras , 282 , 347 mobasher b. , sharples r. m. , ellis r. s. , 1993 , mnras , 263 , 560 moore b. , frenk c.s . , white s.d.m . , 1993 , mnras , 261 , 827 navarro j. f. , frenk c. s. , white s. d. m. , 1996 , apj , 462 , 563 peacock j. a. , dodds s. j. , 1996 , mnras , 280 , l19 pearce f. r. , jenkins a. , frenk c. s. , colberg j. m. , white s. d. m. , thomas p. a. , couchman h. m. p. , peacock j. a. , efstathiou g. ( the virgo consortium ) , apj , 512 , l99 press w. h. , schechter p. , 1974 , apj , 187 , 425 ratcliffe a. , shanks t. , parker q. a. , fong r. , 1998 , mnras , 296 , 173 roukema b. f. , peterson b. a. , quinn p. j. , rocca - volmerange b. , 1997 , mnras , 292 , 835 salpeter , e.e . , 1955 , apj , 121 , 61 somerville r.s . , lemson g. , kolatt t.s . , dekel a. , 1998 , astro - ph/9807277 ( submitted to mnras ) tegmark m. , peebles p. j. e. , 1998 , apj , 500 , 79 owen j. m. , weinberg d. h. , evrard a. e. , hernquist l. , katz n. , 1998 , apj , 503 , 160 white s. d. m. , frenk c. s. , 1991 , apj , 379 , 52 zucca e. , zamorani g. , vettolani g. , cappi a. , merighi r. , mignoli m. , stirpe g. m. , macgillivray h. , collins c. , balkowski c. , cayatte v. , maurogordato s. , proust d. , chincarini g. , guzzo l. , maccagni d. , scaramella r. , blanchard a. , ramella m. , 1997 , a&a , 326 , 477	we have used a combination of high resolution cosmological n - body simulations and semi - analytic modelling of galaxy formation to investigate the processes that determine the spatial distribution of galaxies in cold dark matter ( cdm ) models and its relation to the spatial distribution of dark matter . the galaxy distribution depends sensitively on the efficiency with which galaxies form in halos of different mass . in small mass halos , galaxy formation is inhibited by the reheating of cooled gas by feedback processes , whereas in large mass halos , it is inhibited by the long cooling time of the gas . as a result , the mass - to - light ratio of halos has a deep minimum at the halo mass , @xmath0 , associated with @xmath1 galaxies , where galaxy formation is most efficient . this dependence of galaxy formation efficiency on halo mass leads to a scale - dependent bias in the distribution of galaxies relative to the distribution of mass . on large scales , the bias in the galaxy distribution is related in a simple way to the bias in the distribution of massive halos . on small scales , the correlation function is determined by the interplay between various effects including the spatial exclusion of dark matter halos , the distribution function of the number of galaxies occupying a single dark matter halo and , to a lesser extent , dynamical friction . remarkably , these processes conspire to produce a correlation function in a flat , @xmath2 , cdm model that is close to a power - law over nearly four orders of magnitude in amplitude . this model agrees well with the correlation function of galaxies measured in the apm survey . on small scales , the model galaxies are less strongly clustered than the dark matter , whereas on large scales they trace the occupied halos . our clustering predictions are robust to changes in the parameters of the galaxy formation model , provided only those models that match the bright end of the galaxy luminosity function are considered . galaxies : formation , galaxies : statistics , large - scale structure of the universe
recent proper motion measurements of the lmc and the small magellanic cloud ( smc ) have suggested that the lmc is now moving around the galaxy with a velocity ( @xmath6 ) of @xmath7 km s@xmath8 with respect to the galaxy ( e.g. , kallivayalil et al . 2006 , k06 ; piatek et al . the latest observational studies on the proper motions of the clouds ( vieira et al . 2010 , v10 ) have however derived significantly different @xmath6 ( @xmath9 km s@xmath8 ) , which is significantly smaller than @xmath6 derived by the above earlier studies ( see also costa et al . 2009 , c09 , which derived @xmath10 km s@xmath8 ) . using n - body models of the lmc , bekki ( 2011 ) has recently shown that if the total number of the lmc fields used for the derivation of the center - of - mass proper motion of the lmc , then the derived center - of - mass proper motion of the lmc can significantly deviate from the true one . these observational and theoretical studies imply that the proper motion of the lmc observationally derived so far does not allow us to construct a model for the very precise 3d orbit of the lmc . [ cols="^,^,^,^,^,^,^,^,^,^ " , ] the past 3d motions of the lmc and the smc are the most important parameters in the construction of the formation model for ms and las ( murai & fujimoto 1980 ) . a recent `` first passage '' scenario has shown that ms was formed owing to lmc - smc interaction well before the accretion of the clouds onto the galaxy for the first time ( besla et al . although this model has failed to explain the observed bifurcated structure of the ms and the elongated leading arm , it is consistent with some of the observational results of the above - mentioned hst proper motions of the clouds . on the other hand , classical tidal models can better explain physical properties of ms and la ( e.g. , gardiner & noguchi 1996 , gn96 ; yoshizawa & noughci 2003 , yn03 ; connors et al . 2006 , c06 ) , their models appear to be inconsistent with the above - mentioned hst proper motion results . given that there are significant differences in orbital evolution of the clouds between these previous models , it is observationally important to give more stringent constraints on the 3d motions of the clouds , in particular , the lmc which gravitationally dominates the magellanic system . recently photometric and spectroscopic observations of stars in the surrounding regions of the lmc have revealed possible evidence for the presence of an extended stellar halo of the lmc ( e.g. , minniti et al . 2003 ; muoz et al . 2006 ; majewski et al . 2009 ; saha et al . for example , minniti et al . ( 2003 ) found that the velocity dispersion of the 43 rr lyrae stars in the lmc region is @xmath11 km 2@xmath8 and suggested that the lmc has a dynamically hot stellar halo . muoz et al . ( 2008 ) and majewski et al . ( 2009 ) have revealed a possible presence of the lmc stars out to @xmath12 from the lmc center ( corresponding roughly 10 times the scale radius of the lmc stellar disk ) , which can possibly consist of the outer halo of the lmc . subramaniam & subramaniam ( 2009 ) found that the distribution of rr lyrae stars in the inner region of the lmc has a disk - like distribution and suggested that the old rr lyrae stars do not trace the extended metal - poor stellar halo of the lmc . the possible extended stellar halo of the lmc can be much more strongly influenced by the galaxy than the stellar disk during the lmc - galaxy tidal interaction . therefore the stellar halo can be stripped much more efficiently by the galaxy to become a part of the galactic stellar halo . thus , the outer part of the galactic stellar halo can contain stars initially from the stellar halo of the lmc and therefore can have fossil records on the past orbital history of the lmc . however , the tidal stripping of the _ original metal - poor _ stellar halo of the lmc has not been investigated so far in previous numerical simulations . it is thus unclear what structural and kinematical properties of the stripped halo stars of the lmc have in the outer stellar halo of the galaxy . the purpose of this paper is thus to propose that if the lmc has an extended stellar halo before its commencement of the lmc - galaxy tidal interaction , then the stripped halo stars in the outer galactic halo can have fossil information as to when and where the lmc was first accreted onto the galaxy . using self - consistent n - body simulations of the lmc - galaxy tidal interaction , we investigate how the tidal field of the galaxy can influence the dynamical evolution of the stellar halo of the lmc . we show how the 3d spatial distribution and kinematics of stars stripped from the lmc stellar halo depend on the past 3d orbits of the lmc with respect to the galaxy . we investigate how the lmc dynamically evolves after it is first accreted onto the galaxy from outside the virial radius of the dark matter halo of the galaxy . in the present model , both the lmc and the galaxy are represented by n - body particles so that dynamical friction of the lmc against the galactic dark matter halo can be self - consistently modeled . the present model is therefore more sophisticated than those used in recent models for the magellanic system by besla et al . ( 2010 ) and diaz & bekki ( 2011 , db11 ) which treated the galaxy as a fixed gravitational field . owing to the smaller mass of the smc in comparison with the lmc mass , the smc does not significantly influence the orbital evolution of the lmc . also we focus exclusively on how the stellar halo initially in the lmc dynamically evolves during the lmc - galaxy tidal interaction . thus we do not investigate extensively the models with the smc in the present study for clarity . we however briefly discuss whether the smc can influence the evolution of the lmc stellar halo using two representative models with the smc . the lmc is modeled as a bulge - less disk galaxy embedded in a massive dark matter halo . since the lmc is assumed to fall onto the galaxy from outside the dark matter halo of the galaxy , it is highly likely that the lmc initially has an extended dark matter halo ( i.e. , not truncated at the tidal radius of the lmc ) . the total mass and the virial radius of the dark matter halo of the lmc are @xmath13 and @xmath14 , respectively . we adopted an nfw halo density distribution ( navarro , frenk & white 1996 ) suggested from cdm simulations : @xmath15 where @xmath16 , @xmath17 , and @xmath18 are the spherical radius , the characteristic density of a dark halo , and the scale length of the halo , respectively . we adopted @xmath19 ( @xmath20 ) and the mass ratio of @xmath13 to @xmath21 was fixed at 16.7 . the stellar disk with the size @xmath22 and mass @xmath21 is assumed to have an exponential profile with the scale length ( @xmath23 ) of 1.5 kpc and the truncation radius of 7.5 kpc ( @xmath24 ) . in order to investigate the models which have the maximum rotation speed of @xmath25 km s@xmath8 as suggested by recent observations by piatek et al . 2008 , we assume that the total mass of the lmc ( @xmath26 ) is @xmath27 and @xmath28 . the lmc has a stellar halo with the mass @xmath29 and the size @xmath30 and the radial density profile is assumed to follow the power - law one with the power - law index of @xmath31 ( @xmath32 ) . we mainly investigate the models with @xmath33 and @xmath34 . gas dynamics , star formation , and chemical evolution modeled in our previous models ( bekki & chiba 2005 , 2009 ) are not included in the present model , because we focus exclusively on the dynamical evolution of the stellar halo of the lmc . observational studies for 1047 edge - on disk galaxies in the sloan digital sky survey ( sdss ) revealed almost ubiquitous presence of stellar halos around disk galaxies ( zibetti et al . their fig . 2 clearly shows that the stellar halos extend at least @xmath35 , where @xmath36 is the scale length of a stellar disk . the galaxy is observed to have the stellar halo which extends at least 30 kpc corresponding to @xmath37 ( @xmath38 ) for the stellar scale length of 3.5 ( 2.5 ) kpc ( e.g. , chiba & beers 2000 ) and has a mass that is about @xmath39% of the galactic stellar disk ( freeman & bland - hawthorn 2002 ) . thus the present models with @xmath33 and @xmath34 ( @xmath40 ) are quite reasonable . we however investigate models with different @xmath30 in order to make more robust conclusions of the present numerical simulations . the galaxy is assumed to consist of the nfw dark matter halo , exponential stellar disk , and hernquest bulge . the dark halo has a mass @xmath41 , a virial radius @xmath42 , and a @xmath43 parameter . we investigate two dark halo models in which the maximum circular velocity can be @xmath44 km s@xmath8 that is observed for the galaxy . one has @xmath45 , @xmath46245 kpc , and @xmath47 and the other has @xmath48 , @xmath46347kpc , and @xmath49 . the stellar disk has a mass of @xmath50 , a size of 17.5 kpc ( scale - length of 3.5 kpc ) , and the q - parameter of 1.5 . the bulge has a mass of @xmath51 and a size of 3.5 kpc ( scale - length of 700pc ) . like the lmc model , the galaxy is assumed to have no gas and no star formation . the disk plane of the galaxy is set to be the @xmath52-@xmath53 plane in the present study ( i.e. , the @xmath54-axis is the polar direction of the galaxy ) . the lmc is initially located at @xmath55 from the center of the galaxy and has a velocity ( @xmath56 ) of @xmath57 , where @xmath58 is the circular velocity at @xmath55 and @xmath59 is a parameter controlling the 3d orbit of the lmc . the initial orbital plane of the lmc is inclined by @xmath60 degrees with respect to the galactic plane ( i.e. , the @xmath52-@xmath54 plane ) . @xmath55 is fixed at @xmath61 for all models so that the lmc can be accreted from outside the virial radius of the galaxy . previous best models for the ms formation predicted that the orbital plane of the lmc is almost coincident with the @xmath53-@xmath54 plane ( e.g. , yn03 ) . we thus assume that @xmath62 : the spin vector of the initial orbital angular momentum of the lmc is the same as the @xmath52-axis . the initial velocity components of the lmc are given as ( 0 , @xmath63 , -@xmath64 ) , where @xmath65 . the @xmath53- and @xmath54-components of the velocity correspond to initial tangential and radial velocities , respectively . the spin vector of the stellar disk of the lmc is inclined by 45 degrees with respect to the orbital plane of the lmc for all models . by changing @xmath59 ( also @xmath66 and @xmath67 ) , we can determine the time ( referred to as @xmath68 from now on ) when ( i ) the lmc - galaxy distance can become the observed 50 kpc and ( ii ) the lmc is moving away from the galaxy . we mainly investigate the dark matter model of the galaxy with @xmath69 . if we adopt @xmath70 for the above dark matter model , then @xmath68 is just after the first pericenter passage : this model corresponds to the first passage scenario ( referred to as `` fps '' and the model m1 ) . if we adopt @xmath71 , then @xmath68 is just after two pericenter passages : this corresponds to the second pericenter passage scenario ( `` sps '' and m2 ) . 1 shows the orbital evolution of the lmc in the standard fps ( m1 ) and sps ( m2 ) . we mainly discuss the results of these standard fps and sps in order to more clearly show whether stars stripped from the lmc stellar halo have fossil information of the past orbital history of the lmc : we however investigate different models and discuss the results briefly in 3.2 . the parameter values for the representative 11 models investigated in the present study are shown in the table 1 . if we adopt models with larger @xmath72 ( @xmath73 ) , then @xmath68 can be after the third or fourth pericenter passage of the lmc . the simulated distribution of the stripped lmc halo stars in these models are essentially similar to that in the sps . the aim of this paper is to more clearly show how the spatial distribution of the stripped halo stars depends on the past orbit of the lmc . accordingly we focus on the differences in physical properties the stripped halo stars between the fps and sps in the present paper . also it should be stressed that the present models can not reproduce exact locations of the lmc in the galactic coordinate ( i.e. , ( @xmath52 , @xmath53 , @xmath54 ) for the lmc ) : we adopt the present models in order to point out the possible differences in physical properties of the stripped lmc stellar halo between the two scenarios . fully self - consistent 3d models which can explain the present 3d locations and velocities of the lmc and the smc will be discussed in forthcoming papers . we estimate the velocity of the lmc when ( i ) the lmc passes the pericenter around the galaxy and ( ii ) the pericenter distance ( @xmath74 ) is close to @xmath75kpc that corresponds to the present distance of the lmc from the center of the galaxy . we confirm that the velocity at @xmath74 ( @xmath76 ) is in good agreement with the observed one ( @xmath77 km s@xmath8 ; k06 ; c09 ; v10 ) for all models except the model m11 with @xmath78 kpc ( which is therefore unreasonable for the lmc ) . for example , @xmath76 is 323 km s@xmath8 , 300 km s@xmath8 , 342 km s@xmath8 , 347 km s@xmath8 , and 355 km s@xmath8 for models m1 , m2 , m3 , m4 , and m9 , respectively . for all models , the lmc can not merge with the galaxy within @xmath5 gyr after the first passage of the galactic virial radius . these results demonstrate that the present lmc model is reasonable and realistic in discussing dynamical evolution of the lmc . numerical computations were carried out both on the latest version of grape ( gravity pipe , grape - dr ) which is the special - purpose computer for gravitational dynamics ( sugimoto et al . 1990 ) at university of western australia ( uwa ) and on the hp 390s blade machine with three gpu cards ( nvidia tesla m2070 ) with cuda g5/g6 software package being installed for calculations of gravitational dynamics at ivec in uwa . the total number of particles used for dark matter halo , stellar disk , and bulge of the galaxy in a simulation are 800000 , 200000 , and 33400 , respectively . the total number of particles used for dark matter , stellar halo , and stellar disk in the lmc are 100000 , 100000 , and 50000 , respectively . the adopted gravitational softening length is fixed at 200pc for all models . the time @xmath79 represents the time that has elapsed since the simulation starts . the polar axis of the galactic disk does not change during a simulation in the present study owing to the adopted lmc orbit . the lmc disk particles are not stripped owing to the adopted large initial mass of the lmc ( @xmath80 ) . fig . 2 shows how the lmc stellar halo is stripped by the strong tidal field of the galaxy to form a disk - like or ring - like stellar structure with a long diffuse stellar stream in the outer galactic halo ( @xmath81 kpc ) in the standard sps ( m2 ) . after the first pericenter passage ( @xmath82 gyr ) , the outer part of the lmc stellar halo is partially stripped to form diffuse leaning and trailing stellar stream ( @xmath83 gyr ) . these stripped stars can be then trapped by the galactic potential to form a smaller ring - like structure with the radius of @xmath75 kpc ( @xmath84 gyr ) , while the lmc stellar halo continues to be stripped . finally an irregular ring - like stellar structure connected to the lmc and the outer diffuse stellar stream can be formed after the second pericenter passage ( @xmath85 gyr ) . owing to the initially extended structure of the lmc stellar halo , the stripped stars do not form narrow tidal streams . 3 shows how the 3d distributions of the stripped lmc stellar halo particles depend on the past orbits of the lmc . here stripped stars are those outside the original stellar halo size ( @xmath86 kpc ) at @xmath87 . clearly , a thick disk structure along the polar axis ( i.e. , @xmath54-axis ) can be seen for @xmath88 kpc in the sps whereas such a disky structure can not be seen in the fps . furthermore , diffuse leading and trailing stellar streams can be seen in the @xmath53-@xmath54 projection for the fps whereas a closed irregular ring can be seen in the sps . since it is almost impossible for any models based on the fps to show a closed ring just after the last pericenter passage , the derived clear differences in distributions of stripped lmc halo stars projected onto the @xmath53-@xmath54 plane between the two scenarios can be used for distinguishing the two scenarios . 4 demonstrates that the tidally stripped lmc stellar halo can show a characteristic distribution in the @xmath89 plane , where @xmath90 ( @xmath91 , @xmath92 ) and @xmath93 are the angular momentum component in the @xmath54 ( @xmath52 , @xmath53 ) direction and @xmath94 , respectively . distributions of halo stars on this @xmath89 plane have been discussed in the context of the formation of the galactic stellar halo due to accretion of dwarf galaxies ( e.g. , chiba & beers 2000 ) . it should be stressed that only the stripped stars ( @xmath95 ) shown in fig . 3 ( i.e. , not all stars ) are plotted in this fig . 4 . owing to the diffuse and wide stellar distributions in the tidal streams and the polar ring , the stellar distribution in phase space does not show any strong concentration in the fps and sps . however , there is a clear difference between the two in the sense that the fps shows a much larger number fraction of stars with @xmath93 larger than @xmath96 kpc km s@xmath8 in comparison with the sps . the presence of stars with larger @xmath93 in the fps is due to the fact that the lmc is moving faster in the fps . thus this appreciable difference in phase space can be used to distinguish the two scenarios . the present study predicts that the stripped dark matter halo of the lmc can form a polar thick disk with a central hole ( `` dark polar ring '' ) in the sps . a `` ripple '' ( of `` shell '' ) structure can be clearly seen at @xmath97kpc@xmath98kpc and @xmath97kpc@xmath99kpc in the @xmath52-@xmath54 plane . if dark matter particles outside @xmath95 are regarded as those consisting of the dark polar ring , then the total mass of the ring is @xmath100 ( i.e. , about 61% of the original dark halo mass ) which corresponds to 5.5% of the total mass of the galactic dark matter in the present model . although this dark polar ring would be currently impossible to be detected by observational studies , it could dynamically influence orbital evolution of dwarf satellite galaxies around the galaxy , in particular , those in polar orbits ( e.g. , carina dwarf galaxy ) . 6 shows the orbital evolution of the lmc in the four representative models with different @xmath59 and @xmath41 . clearly , the lmc can not merge within the last 6.8 gyr in these four models , which demonstrates a possibility that the lmc was accreted onto the galaxy long time ago but has not merged with the galaxy yet . although the lmc initially has a larger total mass ( @xmath80 ) , it can lose a significant fraction of the mass owing to tidal stripping by the galaxy . this mass loss can make the time scale of dynamical friction significantly longer for the lmc . as a result of this , the lmc can not so quickly merge with the galaxy as it otherwise could if it keeps the large initial mass . the lmc in the model m4 with @xmath101 can finally show @xmath102 kpc corresponding to the present distance of the lmc from the center of the galaxy . the lmc can not merge with the galaxy within @xmath103 gyr in this model . given that the adopted @xmath101 in the model m4 is predicted to be typical in recent @xmath4cdm cosmological simulations ( e.g. , wetzel 2011 ) , this means that the lmc orbit needs not be unusual one in a @xmath4cdm universe . only the model with lower @xmath59 for @xmath104 can show @xmath102 kpc within the last 6 gyr such a lower @xmath59 is not typical in the @xmath4cdm simulations , which implies that the lmc could be an unusual object if the galaxy is as massive as @xmath104 . 7 shows the distributions of the stripped lmc stellar halo particles around the galaxy projected onto the @xmath53-@xmath54 plane for six representative models . since the distributions of the stars projected onto the @xmath105 plane are less significantly different between fps and sps , we here focus on the distributions projected onto the @xmath53-@xmath54 plane . clearly , the model m3 , which is regarded as fps , does not show a ring - like distribution for the stripped lmc stellar halo stars . this means that irrespective of orbital models , fps can not show a ring - like distribution of the stripped stars along the polar axis of the galaxy . the model m9 ( which is regarded as sps ) shows a ring - like distribution of the stripped stars whereas the model m10 ( fps ) does not . this clear difference between fps and sps in models with larger masses of the galaxy ( @xmath104 ) strongly suggests that the stripped lmc stellar halo stars can be used for distinguishing fps and sps . the model m6 has a smaller size of the lmc stellar halo ( @xmath106 kpc ) so that the halo can not be so strongly influenced by the tidal field of the galaxy in comparison with the model m2 . shows that there is no significant difference in the distribution of the stripped lmc stellar halo stars between m2 and m6 ( i.e. , the two models with the same orbit of the lmc ) . this result combined with that derived for the model m2 suggests that as long as the lmc has an extended stellar halo ( @xmath107 ) , the stripped halo stars can have fossil record of the past orbit of the lmc . if the lmc stellar halo has @xmath108 , then the stripping of the halo stars become less efficient . however , such models with @xmath108 are inconsistent with observations of galactic stellar halos by zibetti et al . ( 2004 ) . in order to discuss how the lmc - smc interaction can influence the final distribution of the stripped lmc halo stars in the outer halo of the galaxy , we construct a n - body model for the smc . the smc has an initial total mass of @xmath109 and is modeled as a bulge - less disk galaxy embedded in a massive dark matter halo . the structural and kinematical properties of stellar disks and dark matter halos are assumed to be self - similar between the lmc and the smc . the smc has an exponential stellar disk with the mass of @xmath110 and the size of 2.0 kpc . the initial distance of the smc from the center of the lmc is a free parameter and denoted as @xmath111 . the initial velocity of the smc with respect to that of the lmc is given as @xmath112 , where @xmath113 is the circular velocity of the lmc at @xmath111 . the orbital evolution of the smc depends not only on the orbital parameters of the lmc but also on @xmath111 . we here describe the results for the models m7 ( @xmath114 ) , because the lmc and the smc do not merge so quickly with each other in this model . although we have investigated a number of models with different @xmath111 , the models with smaller @xmath111 like m8 with @xmath115 show a short time scale of lmc - smc merging ( within @xmath0 gyr ) . they are not so useful for discussing physical properties of the present lmc - smc system which has not completely merged yet . 8 shows the orbital evolution of the smc with respect to the galaxy and the lmc in the model m7 . this is the first orbital model of the smc based on a fully self - consistent n - body simulation in which all of the three galaxies are represented by n - body particles . dynamical friction between the galaxy , the lmc , and the smc is self - consistently included so that the orbit of the smc can be more precisely predicted in this model . clearly , the clouds can keep their binary status for the last 5 gyr , if they were initially strongly bound . owing to dynamical friction between the lmc and the smc , the clouds appear to be almost merged at @xmath116 gyr . 7 shows that the stripped lmc halo stars in this model has a ring - like distribution along the polar axis of the galaxy and the distribution is quite similar to that in the model m2 . thus the smc can not change significantly the spatial distribution of the stripped lmc halo stars in the outer galactic halo owing to its low mass . recent theoretical studies based on high - resolution collisionless n - body simulations have investigated orbital properties ( e.g. , eccentricities and pericenter distances ) of satellite galaxies around luminous ones like the galaxy ( e.g. , benson 2005 ; lux et al . 2010 ; boylan - kolchin et al . 2011 , b11 ; wetzel 2011 ) . one of their results relevant to the present study is distributions of normalized total velocities ( @xmath117 in fig.2 by wetzel 2011 ) which correspond to @xmath59 in the present study . the predicted @xmath59 distribution in wetzel ( 2011 ) has a peak value of 1.15 and shows only a small fraction of satellites with @xmath118 . this means that the orbits used in models with low @xmath59 @xmath119 are less common in the @xmath4cdm model : the present models with @xmath120 would be more appropriate for describing the orbit of the lmc . given that both low and high @xmath59 models show clear differences in 3d distributions of stripped lmc halo stars between fps and sps , the above less consistent models do not become a problem in the present study . the lmc in the models with @xmath121 and @xmath122 kpc can finally have @xmath102 kpc ( at @xmath123 gyr ) owing to the orbital decay of the lmc caused by dynamical friction against the galaxy . given that @xmath121 is typical in recent cosmological simulations ( wetzel 2011 ) , this result means that if the lmc has a typical orbit for satellites in the galaxy , then it should have been accreted onto the galaxy more than @xmath5 gyr ago . this also suggests that successful formation models of ms and las will be able to be constructed by using results of cosmological n - body simulations on orbital properties of lmc - type satellites in galaxy - type halos . the lmc even in the models with lower @xmath59 ( @xmath124 ) can not merge with the galaxy within the last 6 gyr . this demonstrates a possibility that the lmc was accreted onto the galaxy more than @xmath5 gyr ago and has been interacting with the galaxy since then . it is well known that the observed locations , kinematics , fine - structures ( e.g. , bifurcation ) in ms and las can give very strong constraints on the orbital evolution of the lmc and the smc for a given mass model of the galaxy ( e.g. , gn96 ; yn03 ; c06 ; db11 ) . previous models that reproduced very well the observed bifurcated ms and long la predict that the lmc and smc have passed their pericenter about the galaxy at least twice for the last 2@xmath1253 gyr ( c06 ; db11 ) . previous dynamical models for the formation of ms based on fps have reproduced neither the observed location of ms ( fig . 21 in bekki & chiba 2009 ) nor the bifurcated structure of ms ( besla et al . 2010 ) . these models also failed to reproduce the observed properties of the las so far : their models can not be regarded as successful ms formation models . these results imply that the lmc arrived within the virial radius of the galaxy ( for the first time ) more than 4@xmath1255 gyr ago . the question is whether this earlier arrival of the lmc is consistent with the latest results of dissipationless cosmological n - body simulations on accretion processes of lmc - type satellite galaxies onto galaxy - type halos . using the results of the millennium ii simulation , b11 investigated the epochs ( @xmath126 ) of first accretion onto galaxy - mass galaxies and orbital properties for the `` lmc analogs '' that have masses ranging from @xmath127 to @xmath128 at the accretion epochs . they showed that ( i ) only @xmath12930% of lmc analogs have @xmath130 gyr ( their fig.1 ) , ( ii ) lmc analogs with @xmath131 gyr have lower angular momentum ( fig . 2 ) , ( iii ) lmc analogs with @xmath131 gyr are more likely to be strongly bound ( figs . 3 and 4 ) , and ( iv ) about 50% of lmc analogs with @xmath132 gyr have orbital eccentricities ( @xmath133 ) smaller than 0.6 ( fig . 5 ) . these results are consistent with the previous successful ms models ( e.g. , gn96 ; yn03 , and db11 ) that predict bound orbit of the lmc with @xmath134 and at least two pericenter passages of the lmc in the galaxy ( i.e. , @xmath132 gyr ) . if we estimate the specific angular momentum of the lmc ( _ j _ ) in gn96 by adopting the same method as used in b11 , then _ j _ is 0.26 for @xmath135 kpc . the results in fig . 2 in b11 show that about 60% of early - accreted lmc analogs with @xmath132 gyr have _ j _ @xmath136 . this means that the predicted orbit of the lmc in gn96 is a common one in b11 . we thus conclude that the orbital properties in the above successful ms formation models based on the early accretion of the lmc are consistent with those of lmc analogs predicted from recent cosmological simulations by b11 . our conclusion above therefore appears to be at odd with b11 which concluded that the lmc was accreted within the past four gyr and is currently making its first pericenter passage about the galaxy . clearly , their fig . 4 shows that about 60% of lmc analogs are accreted onto the galaxy more than 4 gyr ago ( i.e. , @xmath132 gyr ) . this result appears to be inconsistent with the above conclusion by b11 . this apparent inconsistency is due to the fact that b11 used only observations by k06 and thereby estimated orbital energy and angular momentum of the lmc in order to find lmc analogs that have orbital properties similar to the observed ones : their conclusions depend on the adopted assumptions on the orbital angular momentum and energy of the lmc . now new observational results on the proper motion of the lmc ( e.g. , c09 ; v10 ) have shown that the orbital energy and angular momentum of the lmc are significantly different from those by k06 . these significant differences between different observations imply that it would be worthwhile for cosmological n - body simulations to find the best lmc analogs separately for each of observations by k06 , c09 , and v10 . the present study has first shown that the outer regions of the galactic stellar halo ( @xmath137 kpc ) , in particular , those along the galactic polar axis , can have fossil information on the past 3d orbit of the lmc . if the lmc has already experienced the pericenter passage at least two times , then the stars stripped from the lmc stellar halo can have a thick disk or ring - like distribution along the @xmath54-axis in the galactic coordinate . this polar disk ( or ring ) has a mass of @xmath138 ( @xmath139 ) and thus can consist of @xmath124% of the entire galactic stellar halo mass . the rotating kinematics of the polar disk and the spin vector of the polar disk ( i.e. , the @xmath52-direction ) would be distinguished from those of other possible stellar substructures formed from tidal destruction of dwarf galaxies in the galactic stellar halo . therefore , if the polar disk really exists in the outer stellar halo , then it would be feasible for ongoing and future observations to detect them along the galactic polar axis . previous models showed that the ms along the polar axis of the galaxy can be formed from efficient stripping of gas from the smc due to tidal interaction between the lmc , smc , and the galaxy ( yn03 ) . their models also predicted that there should be no / little stars along the ms , because tidal stripping is efficient only for gas that is initially located in the outer part of the smc in comparison with stars ( see fig . 8 in yn03 ) . therefore , if observations discover stars in the ms region , then the stars are unlikely to originate from the stellar disk of the smc . however , if the smc also has an extended stellar halo , as suggested by recent observations by nol & gallart ( 2007 ) , then the stellar halo can be stripped by the galaxy during the lmc - smc - galaxy tidal interaction to form a stellar stream / disk along the ms . if this is the case , the stellar structure along the ms would consist of stars from metal - poor stellar halos of the lmc and the smc . guhathakurta & reitzel ( 1998 ) searched for stars in the @xmath140 field of the ms iv region but could not find possible evidence for the present of stars in the region . they therefore suggested that the total mass ratio of stars and neutral hydrogen gas ( @xmath141 ) is less than 0.1 . given that the mass of the ms is @xmath142 ( e.g. , brns et al . 2005 ) , @xmath141 , which is the mass ratio of the stripped lmc stellar halo to the ms here , can be well less than 0.05 . therefore , the presence of the polar disk / ring composed of the stripped lmc halo stars would not be inconsistent with the above observation . the narrow observational field in guhathakurta & reitzel ( 1998 ) would possibly have prevented them to detect the high - density regions of the possible polar disk / ring . brck and hawkins ( 1983 ) did not find stars brighter than 20.5 mag in the b - band for the ms and thus suggested that the ms is unlikely to contain intermediate - age stars ( @xmath143 yr ) . given that this observation was difficult to detect old faint stars , the non - detection of stars does not necessarily mean the absence of any stars in the ms . thus , although observational studies have not revealed evidence for the presence of old stars in the ms so far , this non - detection could be due largely to observational constraints ( e.g. , narrow field and shallow photometric survey etc ) . the present study predicts that the ms and the possible polar disk / ring can overlap to some extent in the projected distributions on the sky , their distributions in the 3d space are yet quite different . thus a very wide and deep photometric survey _ along and nearby the ms _ is necessary to map the entire stellar distribution of the possible stellar polar disk / ring . future observational surveys such as the southern sky survey by the skymapper telescope ( keller et al . 2007 ) would be ideal for the detection of the polar disk / ring . it should be stressed that the present results depend strongly on whether the lmc has an extended stellar halo well before its commencement of tidal interaction with the galaxy : if the lmc initially does not have the halo , then the outer stellar halo of the galaxy would not have valuable information on the past 3d orbit of the lmc . extensive observational studies on physical properties of the outer part of the lmc based on photometric properties of stars ( e.g. , color - magnitude diagrams ) are currently ongoing ( e.g. , saha et al . 2010 ) . the physical extension and projected 2d distribution of the possible lmc stellar halo revealed by these studies will assess the viability of the present new idea to give constraints on the past orbit of the lmc . using the results of numerical simulations on dynamical properties of substructures in a milky way - like halo in a @xmath4cdm model , li & helmi ( 2008 ) showed that accretion of a small group of galaxies on to the galaxy about 8 gyr ago can explain the polar distribution of satellite galaxies in the galaxy . bekki ( 2008 ) suggested that if the lmc was a central galaxy of a group being accreted onto the galaxy ( i.e. , with other small galaxies , like the smc , carina , and fornax ) , then the lmc now can have a `` common halo '' that surrounds both the lmc and the smc.if this is the case , the stripped low - mass dwarfs from the lmc group might well have a polar distribution like the stripped lmc stellar halo . the observed thin disky distribution of the galactic satellite galaxies along the galactic polar axis ( e.g. , metz et al . 2009 ) could be thus possible evidence of early accretion of the lmc onto the galaxy , if the lmc was accreted onto the galaxy as a group . yoon & lee ( 2002 ) revealed an intriguing distribution of seven metal - poor ( [ fe / h]@xmath144 ) globular clusters ( gcs ) along the polar - axis of the galaxy and suggested that the gc distribution is consistent with accretion of a satellite galaxy which lost their gcs by tidal stripping of the galaxy . the gcs along the polar - axis have ring - like or disky spatial distribution and rotational kinematics , which appear to be similar to dynamical properties of the stripped lmc stellar halo stars in sps of the present study . these results imply that some of original gcs in the lmc halo were tidally stripped during the accretion of the lmc more than @xmath0 gyr ago to become the halo gcs with a disky spatial distribution along the polar axis of the galaxy . this picture would be consistent with the observed similarities in physical properties of gcs ( e.g. , locations on [ fe / h]-@xmath145 relation , where @xmath145 is the mean period of type ab rr lyraes ) between the seven gcs and the lmc gcs . we have investigated dynamical evolution of the stellar halo of the lmc orbiting the galaxy using fully self - consistent n - body simulations in which the lmc , the smc , and the galaxy are represented by collisionless n - body particles . we also have investigated orbital evolution of the lmc for different initial velocities of the lmc at its first passage of the virial radius of the galaxy . this is for the purpose of discussing whether orbital properties of the lmc in the previous successful formation models for ms and las are consistent with those predicted from recent cosmological simulations based on a @xmath4cdm model . the main results are summarized as follows . \(1 ) the tidal interaction between the lmc and the galaxy can strip stars initially in the extended stellar halo of the lmc and the stripping processes depend strongly on the past orbit of the lmc . for example , if the epoch of the first lmc accretion onto the galaxy from outside its viriral radius is more than @xmath0 gyr ago ( i.e. , at least two pericenter passages ) , the stars stripped from the stellar halo of the lmc can form an irregular polar ring or a thick disk with a size of @xmath1 kpc and rotational kinematics . on the other hand , if the lmc was first accreted onto the galaxy quite recently ( @xmath2 gyr ago ) , the stripped stars form shorter leading and trailing stellar stream at @xmath3 kpc . the distributions of the stripped stars in phase space between the two cases can be significantly different . \(2 ) differences in dynamical properties of the stripped lmc halo stars between fps and sps do not depend on model parameters such as @xmath41 , @xmath30 , and @xmath59 . also the smc is found to be unable to significantly change the final distribution of the stripped lmc stellar halo around the galaxy . therefore the derived differences between fps and sps suggest that if we compare the observed three - dimensional ( 3d ) distribution and kinematics of the outer galactic stellar halo along the polar - axis , then we can give strong constraints on the past orbit of the lmc . future observational surveys such as the southern sky survey by the skymapper telescope would be ideal for the detection of the polar stellar disk / ring possibly formed from the past lmc - galaxy interaction . the galaxy can have a massive polar dark matter ring formed from the lmc - galaxy interaction , if the lmc was accreted onto the galaxy more than several gyr ago . the observed disky distributions of the galactic satellite galaxies and seven galactic halo gcs along the polar axis of the galaxy could be possible evidence of the early accretion of the lmc . \(3 ) the lmc can lose its mass , energy , and orbital angular momentum during its orbital decay caused by dynamical friction against the dark matter halo of the galaxy . the significant mass loss of the lmc during its orbital evolution can make the timescale of dynamical friction significantly longer . as a result of this , the lmc can not merge with the galaxy within @xmath129 6 gyr after its first passage of the galactic virial radius , even if the initial total mass of the lmc is about 10% of that of the galaxy . the lmc with @xmath121 ( @xmath146 and @xmath147 ) that is predicted to be typical in recent @xmath4cdm cosmological simulations can not merge within @xmath103 gyr for @xmath80 and @xmath148 . these results allow the previous successful ms formation models which assume that the lmc and the galaxy have been dynamically interacting with each other without merging at least a few gyr . \(4 ) the lmc in the models with @xmath121 can finally have @xmath74 ( @xmath75 kpc ) and @xmath76 ( @xmath77 km s@xmath8 ) that are observed in the present lmc . this result implies that the orbit of the lmc needs not be unusual one in the @xmath4cdm model . this also suggests that it is promising for future formation models of ms and las based on cosmological simulations to reproduce both the present location / velocity of the lmc and the physical properties of ms and las in a self - consistent manner . \(5 ) the previous successful formation models of ms and las predict at least two pericenter passages of the lmc about the galaxy and thus accretion of the lmc onto the galaxy ( i.e. , the first passage of the galactic virial radius ) more than 4 gyr ago . the models also predict @xmath134 , _ j_@xmath149 , and strongly bound orbit for the lmc . orbital properties of early - accreted lmc analogs ( @xmath132 gyr ) predicted from recent cosmological simulations are therefore consistent with those in the successful formation models of ms and las . this consistency strongly suggests that the lmc was accreted onto the galaxy more than 4 gyr ago so that the bifurcated ms and elongated las could be formed as a result of tidal interaction between the lmc , the smc , and the galaxy . we thus conclude that the lmc was accreted onto the galaxy more than 4 gyr ago ( i.e. , the first pericenter passage more than @xmath150 gyr ago ) . this early accretion suggests that the global star formation of the lmc could have been enhanced by its dynamical interaction ( bekki & chiba 2005 ) and hydrodyanmical one ( mastropietro et al . 2009 ) with the galaxy since more than @xmath150 gyr ago ( i.e. , after the first pericenter passage ) . this lmc - galaxy interaction which can trigger star formation can provide physical explanations both for the observed enhancement of star formation in the lmc about a few gyr ago ( e.g. , gallagher et al . 1996 ) and for the unusually blue colors of the lmc as a massive satellite in galaxy - scale halos ( tollerud et al . furthermore , the long - term lmc - smc - galaxy interaction , which is possible in early accretion of the lmc , can repetitively enhance star formation in the smc for the last few gyr ( yn03 ) , which is indeed observed ( e.g. , harris & zaritsky 2004 ) . thus , the observed star formation histories of the lmc and the smc , bifurcated ms , and elongated las strongly suggest that the lmc first passed the galactic virial radius more than 4 gyr ago . i am grateful to the anonymous referee for constructive and useful comments that improved this paper . kb acknowledge the financial support of the australian research council throughout the course of this work . the work was supported by ivec through the use of advanced computing resources located at the university of western australia . majewski , s. r. , nidever , d. l. , muoz , r. r. , patterson , r. j. , kunkel , w. e. , carlin , j. l. , 2009 , the magellanic system : stars , gas , and galaxies , proceedings of the international astronomical union , iau symposium , volume 256 , p. 51	using fully self - consistent n - body models for the dynamical evolution of the large magellanic cloud ( lmc ) in the galaxy , we show that if the lmc initially has an extended old stellar halo before its commencement of tidal interaction with the galaxy , physical properties of the stars stripped from the lmc stellar halo can have fossil information as to when and where the lmc was accreted onto the galaxy for the first time . if the epoch of the first lmc accretion onto the galaxy from outside its viriral radius is more than @xmath0 gyr ago ( i.e. , at least two pericenter passages ) , the stars stripped from the stellar halo of the lmc can form an irregular polar ring or a thick disk with a size of @xmath1 kpc and rotational kinematics . on the other hand , if the lmc was first accreted onto the galaxy quite recently ( @xmath2 gyr ago ) , the stripped stars form shorter leading and trailing stellar stream at @xmath3 kpc . also distributions of the stripped stars in phase space between the two cases can be significantly different . the derived differences in structure and kinematics of the stripped stars therefore suggest that if we compare the observed three - dimensional ( 3d ) distribution and kinematics of the outer galactic stellar halo along the polar - axis , then we can give strong constraints on the past orbit of the lmc . we also discuss whether the orbital properties of the lmc in successful formation models for the magellanic stream ( ms ) can be consistent with orbital properties of the lmc - type systems in galaxy - type halos predicted from recent high - resolution cosmological simulations in a @xmath4cdm universe . we find that the orbital properties of the lmc in the successful formation models are consistent with those predicted from the cosmological simulations . we also find that the lmc can not merge with the galaxy within the last @xmath5 gyr in models consistent with predictions from the @xmath4cdm simulations . given that the successful ms formation models predict at least two pericenter passages of the lmc in the galaxy , we conclude that the lmc was accreted onto the galaxy more than @xmath0 gyr ago so that interaction between the lmc , the small magellanic cloud ( smc ) , and the galaxy could form the ms and its leading arms ( las ) . [ firstpage ] magellanic clouds galaxies : structure galaxies : kinematics and dynamics galaxies : halos galaxies : star clusters
a major result in ergodic theory in the late 1980 s was the proof of the return times theorem by j. bourgain @xcite ( which was later simplified by j. bourgain , h. furstenberg , y. katznelson , d. ornstein in @xcite ) . this theorem created a key strengthening of the birkhoff s pointwise ergodic theorem @xcite . [ rtt1 ] let @xmath1 be an ergodic dynamical system of finite measure and @xmath2 . then there exists a set @xmath3 of full measure such that for any other ergodic dynamical system @xmath4 with @xmath5 and any @xmath6 : @xmath7 converges @xmath8-a.e . for all @xmath9 . note that the set @xmath10 depends not only on the function @xmath11 chosen , but on the transformation @xmath12 in our dynamical system . in the bfko proof @xcite of the return times theorem , one of the keys to the argument was to decompose the given function using the kronecker factor in order to prove the result independently for both the eigenfunctions and those functions in the orthocomplement of the kronecker factor . using factors in convergence proofs in ergodic theory has long been a very useful tool . the notion of a characteristic factor is originally due to h. furstenberg and is explicitly defined by h. furstenberg and b. weiss in @xcite . when the limiting behavior of a non - conventional ergodic average for @xmath13 can be reduced to that of a factor system @xmath14 , we shall say that the latter is a * characteristic factor * of the former . for each type of average under consideration , one will have to specify what is meant by reduced in the given case . in the case of h. furstenberg and b. weiss @xcite , they define the notion of characteristic factor for averages of the type @xmath15 therefore their specific definition of characteristic factor is as follows . if @xmath16 are @xmath17 integer - valued sequences , and @xmath18 is a factor of a system @xmath19 , we say that @xmath20 is a * characteristic factor for the scheme * @xmath16 , if for any @xmath21 we have @xmath22\ ] ] converges to @xmath23 in @xmath24 . in 1998 , d. rudolph @xcite extended the return times theorem to averages with more than two terms with his proof of the multiterm return times theorem . his proof answered one of the questions on the return times raised by i. assani who proved the same result for weakly mixing systems in @xcite . [ mrtt ] let @xmath17 be any positive integer . for any ergodic dynamical system @xmath25 and any @xmath2 , there exists a set of full measure @xmath26 in @xmath27 such that if @xmath28 for any other dynamical system @xmath29 and any @xmath30 there exists a set of full measure @xmath31 in @xmath32 such that if @xmath33then for any other dynamical system @xmath34 and any @xmath35 there exists a set of full measure @xmath36 in @xmath37 such that if @xmath38 for any other dynamical system @xmath39 the average : @xmath40 converges @xmath41-a.e .. as we are interested in finding characteristic factors for ergodic dynamical systems this theorem is quoted here with the extra assumption of ergodicity for the dynamical system @xmath25 . d. rudolph s proof of the multiterm return times theorem utilized the method of joinings and fully generic sequences . this led to an elegant proof of the theorem which avoided the study of the factor of the @xmath42-algebra which was characteristic for the averages . so the higher order version of the kronecker factor @xmath43 which had been key to the bfko @xcite proof was not needed in d. rudolph s argument . this paper seeks to determine what factors serve a role similar to the kronecker factor @xmath43 in this multiterm setting . for our purposes we define the notion of pointwise characteristic factors for the multiterm return times averages as follows . consider @xmath19 a measure preserving system . the factor @xmath44 is * pointwise characteristic for the @xmath17-th return times averages * if for each @xmath2 we can find a set of full measure @xmath10 such that for each @xmath9 , for any other dynamical system @xmath45 and any @xmath30 , there exists a set of full measure @xmath31 such that for each @xmath46 in @xmath31 then @xmath47 for any other dynamical system @xmath34 and any @xmath35 , there exist a set of full measure @xmath36 in @xmath37 such that if @xmath38 for any other dynamical system @xmath48 for @xmath49-a.e . @xmath50 the average @xmath51g_{1}(s_{1}^{n}y_{1})g_{2}(s_{2}^{n}y_{2})\cdots g_{k}(s_{k}^{n}y_{k})\ ] ] converges to 0 . in looking for potential characteristic factors for the general multiterm return times averages we consider the factors first used by h. furstenberg to prove szemerdi s theorem @xcite . these factors are called @xmath17 step distal factors in @xcite . we denote these factors ( which will be further defined in section 2 ) as @xmath52 using the notation from @xcite where these factors were shown to be @xmath0-characteristic for the averages @xmath53 while the norm convergence of averages for @xmath0-characteristic factors can sometimes lead to pointwise characteristic properties , this is not always guaranteed to be the case . thus it is of consequence to look at pointwise convergence in addition to investigating factors with respect to the norm convergence . we will show that these @xmath52 factors can be characterized in an inductive way by seminorms using lemma 1.3 of @xcite . using these seminorms we will prove our first result . characteristic for the furstenberg averages . ] [ main ] the factors @xmath52 are pointwise characteristic for the multiterm return times averages . more precisely , using the seminorms defining the @xmath52 we can find pointwise uniform upper bounds of the multiterm return times averages . the study of the nonconventional furstenberg averages has seen important progress being made in the last seven years . in @xcite and @xcite the host - kra - ziegler factors @xmath54 were created independently by b. host , b. kra and t. ziegler and were shown to be characteristic in @xmath0 norm for the furstenberg averages . using these factors we prove our second result . [ hkz ] let @xmath19 be an ergodic measure preserving system . the host - kra - ziegler factors @xmath54 are pointwise characteristic for the multiterm return times averages . as the @xmath54 factors are smaller than the factors @xmath52 , and thus @xmath55 , the fact that the @xmath52 factors are pointwise characteristic for the multiterm return times averages is a consequence of theorem [ hkz ] . but in our proof of theorem [ main ] using the seminorm defining the factors @xmath52 we obtain pointwise uniform upper bounds of the multiterm return times averages . with the @xmath54 factors we do not have such pointwise estimates . the uniform upper bounds are derived after integration combined with a @xmath56 argument . an unusual feature of our proof is that we use the previously established convergence result of d. rudolph and use that to prove that the factors we are studying are characteristic . let @xmath19 be an ergodic dynamical system on a probability measure space . the factors @xmath52 are defined in the following inductive way . * the factor @xmath57 is equal to the trivial @xmath42-algebra @xmath58 * for @xmath59 the factor @xmath60 is characterized by the following . a function @xmath61 if and only if @xmath62 note that the factor @xmath63 is the kronecker factor of our ergodic transformation @xmath12 because @xmath64 in lemma [ seminorms ] we will verify that the @xmath52 as defined above do actually form well - defined factors . we will want to verify that these @xmath52 are maximal isometric extensions . there are several equivalent ways of expressing this feature . we will be using the terminology found on pages 373 - 374 of @xcite to specify how these factors form maximal isometric extensions . furstenberg has shown in @xcite that for any ergodic dynamical system @xmath19 and any @xmath12-invariant factor @xmath65 there is a unique maximal factor action @xmath66 which contains @xmath65 so that in the rohlin representation of @xmath67 , the space @xmath68 can be taken to be a compact metric space with isometric factor maps . this factor @xmath69 arises from the invariant algebras of the relatively independent joinings of @xmath1 over the factor @xmath65 . this is precisely the notion of maximal isometric extension referred to in the next lemma . [ seminorms ] let @xmath19 be an ergodic dynamical system on a probability measure space . for @xmath70 for each function @xmath71 the quantities @xmath72 are well defined . furthermore , they characterize factors of @xmath12 which are successive maximal isometric extensions . let us consider a general factor @xmath44 of @xmath12 and @xmath73 the projection onto this factor . the relatively independent joining of @xmath74 over the factor @xmath44 is the measure @xmath75 defined for @xmath76 @xmath77 bounded functions as @xmath78 by birkhoff s ergodic theorem applied to @xmath74 and the invariant measure @xmath75 we have @xmath79 where @xmath80 is the @xmath74-@xmath75 invariant @xmath42-algebra . if we denote by @xmath81 the quantity @xmath82 then lemma 1.3 in @xcite tells us that @xmath83 if and only if @xmath84 where @xmath85 is the maximal isometric extension of @xmath86 using these observations one can characterize the successive maximal isometric extensions . the trivial @xmath42-algebra is @xmath57 . then we define @xmath87 , @xmath88 and more generally @xmath89 the seminorms characterizing these factors are well defined as @xmath72 where @xmath90 in order to simplify the inductive parts of our argument , we first clarify the techniques that we will use in a series of small lemmas . this next lemma relies on an application of the spectral theorem which allows us to alternate between wiener - wintner and return times averages in our inductive argument . [ wwtype ] let @xmath91 be a sequence of complex numbers . if @xmath92 then @xmath93 in @xmath94 for all measure - preserving systems @xmath95 . this follows immediately from the proof of theorem 3.1 in @xcite . next , we will use the following lemma which is an easy consequence of the van der corput lemma @xcite . it will help us simplify the wiener - wintner averages which will appear in the inductive argument . [ vdctype ] there exists an absolute constant @xmath96 such that for any @xmath91 bounded sequence of complex numbers and any positive integer @xmath97 we have @xmath98 for @xmath99 . the following lemma will be useful in establishing the basis step of our next theorem . it gives a pointwise upper bound for the return times averages for two terms , case studied in the bfko @xcite paper . [ upbound0 ] let @xmath19 be an ergodic dynamical system and @xmath100 let us denote by @xmath101 its kronecker factor . then there exists a universal set @xmath10 depending only on @xmath11 and the system @xmath13 such that for any dynamical system @xmath45 and any @xmath102 we have @xmath103 by the bfko return times theorem , we know that the kronecker factor is pointwise characteristic . so we have on a universal set @xmath10 of full measure @xmath104 @xmath105 a.e . , for any dynamical system @xmath45 and any @xmath102 . therefore for any dynamical system @xmath45 and any @xmath102 , @xmath106 using the cauchy schwartz inequality we have then @xmath107 then by applying birkhoff s pointwise ergodic theorem we have @xmath108 which is the upper bound announced in this lemma . we will prove our first main result , theorem [ main ] , in the course of proving the following more detailed statement . [ a_k ] let @xmath17 be any positive integer . for any ergodic dynamical system @xmath13 and for each @xmath2 we can find a set of full measure @xmath10 such that for each @xmath9 , for any other dynamical system @xmath45 and any @xmath30 with @xmath109 , there exists a set of full measure @xmath31 such that for each @xmath46 in @xmath31 then @xmath47 for any other dynamical system @xmath34 and any @xmath35 with @xmath110 there exist a set of full measure @xmath36 in @xmath37 such that if @xmath38 for any other dynamical system @xmath48 for @xmath49-a.e . @xmath50 * the average @xmath111g_{1}(s_{1}^{n}y_{1})g_{2}(s_{2}^{n}y_{2})\cdots g_{k}(s_{k}^{n}y_{k})\ ] ] converges to 0 . * thus for @xmath112 the average @xmath40 converges to 0 @xmath49-a.e .. * also we have the following pointwise upper bound for our limit @xmath113 the basis step for the induction of the statement in ( [ rtt ] ) was done in the bfko @xcite proof of bourgain s return times theorem . here it was shown that @xmath114 was pointwise characteristic for averages of the type @xmath115 in lemma [ upbound0 ] , we showed that the quantity @xmath116 is a pointwise upper bound for the @xmath56 of the absolute value of the averages where @xmath117 is the kronecker factor for @xmath118 this last term is itself less than @xmath119 because @xmath120 thus we have reached the inequality @xmath121 we want to get a better upper bound namely @xmath122 where @xmath96 is an absolute constant . to this end we apply the van der corput lemma to obtain @xmath123 @xmath124 \leq\ ] ] @xmath125.\ ] ] applying the inequality ( [ bound ] ) to each of the functions @xmath126 yields @xmath127.\ ] ] by taking the limit with @xmath128 we get the better estimate @xmath129 which shows clearly that @xmath63 satisfies ( [ upbound ] ) in the basis step . assume that for any @xmath2 and @xmath130 we can find sets @xmath10 of full measure such that if @xmath9 , then for any other dynamical system @xmath45 and any @xmath30 with @xmath109 , there exists a set of full measure @xmath31 such that for each @xmath46 in @xmath31 then @xmath47 for any other dynamical system @xmath131 and any @xmath132 with @xmath133 there exist a set of full measure @xmath134 in @xmath135 such that if @xmath136 for any other dynamical system @xmath137 and any @xmath138 with @xmath139 for @xmath140-a.e . @xmath141 we have * the average @xmath142g_{1}(s_{1}^{n}y_{1})\cdots g_{j}(s_{j}^{n}y_{j})\ ] ] converges to 0 . * also we have the upper bound @xmath143 [ bn ] let @xmath11 be an element of @xmath144 and let @xmath145 , @xmath146 and @xmath147 be as defined in the preceding paragraph . if @xmath148 then @xmath149 for some absolute constant @xmath96 . here the constant @xmath96 is independent of the @xmath11 , @xmath145 , @xmath146 and @xmath147 . by lemma [ vdctype ] , there exists a constant @xmath96 such that for @xmath150 @xmath151 from our inductive hypothesis , we know that for each @xmath152 there is a set of full measure @xmath153 on which @xmath154 & & \\ \cdot(g_1\cdot \overline{g_1 \circ s_1^h})(s_{1}^{n}y_{1})\cdots ( g_{k-1}\cdot \overline{g_{k-1 } \circ s_{k-1}^h})(s_{k-1}^{n}y_{k-1})\bigg\| & \rightarrow & 0.\end{aligned}\ ] ] therefore , the intersection of these sets @xmath153 over @xmath152 gives a set of full measure @xmath155 on which @xmath156 for all @xmath128 . the cauchy - schwartz inequality gives us @xmath157 by birkhoff s pointwise ergodic theorem we know that there is a set of full measure @xmath158 on which for each @xmath152 the average over @xmath159 in the above inequality converges to @xmath160 therefore on the set of full measure @xmath161 @xmath162 as functions @xmath11 in @xmath163 are characterized by the property that @xmath164 , lemma [ bn ] implies that when @xmath11 is an element of @xmath165 we have @xmath166 on the set of full measure @xmath167 . therefore @xmath168 converges to 0 @xmath169-a.e .. hence by an application of lemma [ wwtype ] , we know that for any other dynamical system @xmath170 and any @xmath171 @xmath172 converges to 0 in @xmath173 . as pointwise convergence of the average in equation ( [ norm ] ) follows from theorem [ mrtt ] , we have @xmath174 converges to 0 @xmath49-a.e . , when @xmath11 is in @xmath165 . therefore for all @xmath2 we have @xmath175g_{1}(s_{1}^{n}y_{1})g_{2}(s_{2}^{n}y_{2})\cdots g_{k}(s_{k}^{n}y_{k})\ ] ] converges to 0 @xmath49-a.e .. thus , we have shown that the factors @xmath52 are pointwise characteristic for the multiple term return times averages . to finish the proof of the theorem it remains to show that @xmath176 we use the property just established that the factors @xmath52 are pointwise characteristic for the multiple term return times averages of @xmath177 functions including the arbitrary function @xmath11 and the van der corput lemma . we apply this characteristic property to each of the functions @xmath178 and apply the cauchy schwartz inequality to obtain our result . we have @xmath179 & = & \\ c \bigg [ \frac{1}{h } + \frac{1}{h}\sum_{h=1}^h \limsup_n\big\| \frac{1}{n}\sum_{n=1}^{n - h}\mathbb{e}(f \cdot f(t^h)\|\mathcal{a}_k)(t^nx)\cdot & & \\ g_{1}(s_{1}^{n}y_{1})g_{1}(s_{1}^{n+h}y_{1})\cdots g_{k}(s_{k}^{n}y_{k})g_{k}(s_{k}^{n+h}y_{k})\big\|\bigg ] & & \end{aligned}\ ] ] applying the characteristic property to each of the functions @xmath178 the above inequality is @xmath180\ ] ] because @xmath181 . by birkhoff s pointwise ergodic theorem and the ergodicity of @xmath12 , the inequality in ( [ last2 ] ) is @xmath182.\ ] ] using the cauchy schwartz inequality we obtain that the inequality in ( [ last3 ] ) is less than or equal to @xmath183.\ ] ] taking the limit with @xmath128 gives us the upper bound @xmath184 as noted above , the factors @xmath54 are smaller than the @xmath52 factors and thus their orthogonal complements @xmath185 are bigger . therefore theorem [ hkz ] , which we are proving in this section , is an extension of theorem [ main ] . we will prove theorem [ hkz ] directly from the properties of the factors @xmath54 . the host - kra - ziegler factors , @xmath54 , were defined in @xcite by seminorms as follows . * the factor @xmath186 is equal to the trivial @xmath42-algebra . * the factor @xmath187 can be characterized by the seminorms @xmath188 where @xmath189 * the factor @xmath68 is the conze - lesigne factor , @xmath190 . functions in this factor are characterized by the seminorm @xmath191 such that @xmath192 a function @xmath193 if and only @xmath194 * more generally b. host and b. kra showed in @xcite that for each positive integer @xmath17 we have @xmath195 with the condition that @xmath196 if and only if @xmath197 * the factors @xmath57 and @xmath186 are equal to the trivial @xmath42-algebra . * the factors @xmath63 and @xmath187 are also identical . the seminorm @xmath188 and @xmath199 are equal indeed @xmath200 * the difference starts with the factors @xmath201 and @xmath202 it is not difficult to find examples where @xmath203 . on the two torus the transformation @xmath204 where @xmath205 denotes the fractional part of @xmath206 is an example for which the two factors differ . more generally it can be shown that if the transformation on the two torus is given by @xmath207 , where @xmath208 is measurable , then @xmath201 always coincides with the full algebra ( _ i.e. _ , the system is 2-step distal ) , and @xmath209 only when @xmath210 is cohomologous to the affine co - cycle . [ vdc2 ] let @xmath19 be an ergodic dynamical system and @xmath100 then for all positive integers @xmath128 we have @xmath211 where @xmath96 is an absolute constant derived from the application of the van der corput lemma . in particular we have for @xmath169-a.e . @xmath212 @xmath213 [ induct ] given @xmath19 an ergodic measure preserving system on a probability measure space and @xmath214 . then we can find a set of full measure @xmath10 such that for every @xmath215 for each measure preserving dynamical system @xmath216 and each @xmath102 with @xmath109 we have @xmath217 where @xmath96 is an absolute constant derived from the application of the van der corput lemma and @xmath218 by the bfko @xcite proof of the return times theorem we have pointwise convergence of the above averages , therefore the @xmath56 on the left hand side of the above expression becomes a limit . therefore , we have @xmath219 where @xmath220 is the spectral measure associated to @xmath221 with respect to the dynamical system @xmath222 . thus @xmath223 as @xmath109 , using lemma [ vdc2 ] we derive the inequality @xmath224 [ alr ] let @xmath19 be an ergodic measure preserving system on a probability measure space and @xmath71 . then for @xmath169-a.e . @xmath225 for every measure preserving system @xmath226 and each @xmath102 with @xmath109 we have @xmath227 where @xmath96 is the absolute constant from the application of the van der corput lemma . in particular , for @xmath193 ( or equivalently @xmath228 ) we have for @xmath105-a.e . @xmath46 @xmath229 similarly to the proof of lemma [ induct ] as pointwise convergence of the averages holds by theorem [ mrtt ] we can rewrite the above @xmath232 as a @xmath233 and use the spectral theorem to rewrite the integral in ( [ ineq2 ] ) as @xmath234 thus by lemma [ vdc2 ] the expression in [ ineq2 ] is @xmath235 on a set of full measure depending only on @xmath11 as @xmath236 . this set of full measure is , in fact , the intersection of the sets of full measure obtained by the bfko @xcite proof of the return times theorem for each function @xmath237 [ cl ] let @xmath19 be an ergodic measure preserving system on a probability measure space . the factor @xmath68 , the conze lesigne factor , is pointwise characteristic for the three term return times theorem . we denote by @xmath242 the three term averages with our original function @xmath11 , the fixed system @xmath216 and the variable one @xmath243 more precisely we have @xmath244 by theorem [ mrtt ] we have a set of full measure @xmath245 on which the pointwise convergence of the return times averages with three terms holds for any choice of measure preserving dynamical system @xmath246 and @xmath247 with @xmath248 . therefore for @xmath33 we have using the spectral measure as before , we can continue from ( [ lim ] ) @xmath250 as @xmath251 . note that this upper bound is now independent of the choice of @xmath252 and @xmath253 so in fact we have @xmath254 note that the left hand side of this last inequality is not necessarily measurable . however one can conclude by making the following observation . by lemma [ alr ] we have @xmath255 therefore , for @xmath256 ( i.e. @xmath257 ) there exists a set of full measure in @xmath32 on which we have @xmath258 for @xmath259 we have @xmath260 for any choice of dynamical system @xmath252 and @xmath247 with @xmath248 . therefore @xmath68 is pointwise characteristic for the three term return times averages . it remains to finish the induction argument which we have started in the above lemmas . suppose that we know that the @xmath261 are pointwise characteristic for the @xmath262-th return times averages and that the bound in ( [ lemmaalr ] ) from lemma [ alr ] holds for @xmath262-th averages for @xmath263 . specifically for @xmath169-a.e . @xmath225 , for every measure preserving system @xmath264 and each @xmath265 with @xmath266 with @xmath267 we have for @xmath268 @xmath269 where @xmath270 and @xmath271 is any measure preserving dynamical system and @xmath35 with @xmath272 . by theorem [ mrtt ] , we know that for any @xmath2 and @xmath130 we can find sets @xmath10 of full measure such that if @xmath9 , then for any other dynamical system @xmath273 and any @xmath30 with @xmath109 , there exists a set of full measure @xmath31 such that for each @xmath46 in @xmath31 then @xmath47 for any other dynamical system @xmath274 and any @xmath35 with @xmath272 there exist a set of full measure @xmath36 in @xmath37 such that if @xmath38 we have the pointwise convergence of the return times averages with @xmath17 terms for any other dynamical system @xmath275 and any @xmath138 with @xmath139 . we would like to show that @xmath279 is pointwise characteristic for the @xmath17-return times averages and that we have for @xmath280 @xmath281 where the @xmath282 @xmath283 and @xmath284 are as defined above . this will complete our proof by induction . to make the reading of the induction proof easier we show how one can prove that @xmath289 is pointwise characteristic for the @xmath290-th return times averages . the reader will check that the arguments extend without difficulty to arbitrary @xmath17 . so we want to show that if @xmath291 then @xmath292 one can conclude that for the appropriate universal sets for @xmath206 @xmath297 @xmath298 and @xmath299 if @xmath300 then @xmath301 for all measure preserving system @xmath302 and each @xmath303 with @xmath304 1 . in theorem [ a_k ] ( see equation ( [ upbound ] ) ) , we proved that we have a pointwise upper bound on the average of multiple terms as follows @xmath319 we asked in a previous version of this paper whether for @xmath320 one can replace in these inequalities the @xmath321 seminorms for the @xmath52 with those defining the @xmath54 factors . * it was shown in @xcite that @xmath322 but we can not find an absolute constant @xmath96 for which @xmath323 * as pointed out by the referee one can not replace the @xmath321 seminorms in equation ( [ upbound ] ) with those defining the @xmath54 factors . for @xmath12 an irrational rotation on the circle the kronecker factor is @xmath324 and @xmath325 . then @xmath326 . direct computations show that @xmath327 therefore one can not find an absolute constant @xmath96 for which @xmath328 consider the average @xmath329 then we have @xmath330 the right hand side of the above inequality is less than @xmath331 however one can not find an absolute constant @xmath96 for which @xmath332 2 . the authors of this paper are writing a survey of the return times theorem @xcite which will include more details of the historical developments of theorem [ rtt1 ] and [ mrtt ] and related questions such as the ones noted above . 17 i. assani . `` multiterm return time theorem for weakly mixing systems '' _ annales de linstitut henri poincare_,vol 36,2,153 - 165,(2000 ) . i. assani . _ wiener wintner ergodic theorems_. world scientific publishing co. , new jersey , london , singapore , hong kong , 2003 . i. assani . `` characteristic factors for some non conventional ergodic averages . '' _ preprint _ , 2003 . i. assani . `` pointwise convergence of ergodic averages along cubes . '' _ j. analyse math . _ 110 ( 2010 ) , 241 - 269 . i. assani , e. lesigne and d. rudolph . `` wiener wintner return - times ergodic theorem . '' _ isr . j. math . _ 92 ( 1995 ) 375 - 395 . i. assani and k. presser . `` survey of the return times theorem . '' _ in progress _ , 2010 . g. d. birkhoff . `` proof of the ergodic theorem . '' _ proc . _ 17 ( 1931 ) , 656 - 660 . j. bourgain . `` return time sequences of dynamical systems . '' _ unpublished preprint _ , 1988 . j. bourgain , h. furstenberg , y. katznelson , and d. ornstein . appendix to j. bourgain : `` pointwise ergodic theorems for arithmetic sets . '' _ inst . hautes tudes sci . _ 69 ( 1989 ) , 5 - 45 . h. furstenberg . `` the structure of distal flows . '' _ american journal of mathematics _ 85 no . 3 ( 1963 ) , 477 - 515 . h. furstenberg . `` ergodic behavior of diagonal measures and a theorem of szemerdi on arithmetic progressions . '' _ j. analyse math . _ 31 ( 1977 ) , 204 - 256 . h. furstenberg and b. weiss . `` a mean ergodic theorem for @xmath333 . '' _ convergence in ergodic theory and probability _ 5 ( 1996 ) , 193 - 227 . b. host and b. kra . `` nonconventional ergodic averages and nilmanifolds . '' _ ann . of math . ( 2 ) _ 161 ( 2005 ) , 397 - 488 . l. kuipers and h. niederreiter . _ uniform distribution of sequences . _ wiley - interscience , john wiley & sons , new york , 1974 . d. rudolph . `` a joinings proof of bourgain s return time theorem . '' _ _ ergodic theory dynam . systems__14 no . 1 ( 1994 ) , 197 - 203 . d. rudolph . `` eigenfunctions of @xmath334 and the conze - lesigne algebra . '' ergodic theory and its connections with harmonic analysis . _ london math . lecture note ser . _ 205 ( 1995 ) , 369 - 432 . d. rudolph . `` fully generic sequences and a multiple - term return times theorem . '' _ 131 no . 1 ( 1998 ) , 199 - 228 . t. ziegler . `` universal characteristic factors and furstenberg averages . '' _ j. amer . _ , 20 , ( 2007 ) , 53 - 97 .	this paper is an update and extension of a result the authors first proved in 2003 . the goal of this paper is to study factors which are known to be @xmath0-characteristic for certain nonconventional averages and prove that these factors are pointwise characteristic for the multiterm return times averages . in memory of dan rudolph .
the classical problem of water waves over a variable seabed has attracted much attention ( see @xcite and references therein ) . there are some significant differences in this interesting and practically important problem , as compared to the theory of waves on a deep water or in canals with a flat horizontal bottom . in situations where the fluid depth is less or of the same order as a typical length of surface wave , inhomogeneity of the bottom is a reason for linear and nonlinear wave scattering and transformation , and it strongly affects wave propagation . these phenomena occur so widely that one can meet them almost everywhere , although with different scales . examples of strongly nonlinear dynamics are ocean waves running on a beach , or motion of disturbed water in a puddle after a car . among linear effects due to bottom topography is existence of special edge - localized waves discovered by stokes @xcite , that propagate along the shore line of a beach . over an axially symmetric underwater hill , quasi - localized wave modes with non - zero angular momentum can exist , similar to long - life - time states of a quantum particle confined by a potential barrier of a finite width @xcite . it is necessary to say that underwater obstacles of definite shapes and sizes can serve as waveguides ( a narrow and long underwater crest ) or as lenses ( an oblong underwater hill oriented crosswise to the wave propagation ) . a qualitative explanation for all the linear effects is simple . indeed , let @xmath0 be the coordinate in the horizontal plane , @xmath1 the depth corresponding to quiet surface . then , looking at the well known dispersion relation for small - amplitude gravitational surface waves , @xmath2 ( where @xmath3 is the frequency , @xmath4 is the absolute value of the wave vector , @xmath5 is the gravitational acceleration ) , one can see that the local refraction index @xmath6 increases as the depth @xmath1 decreases , in accordance with the formulas @xmath7 @xmath8 where the function @xmath9 is determined by eq . ( [ omega_k_h ] ) . therefore , as in the conventional light optics , here an oblique wave changes its direction of propagation when meets gradient of @xmath10 . also , the total internal reflection is possible in propagation from smaller depth to larger depth . besides observing such natural phenomena , a set of laboratory experiments has been carried out to investigate various aspects of the given problem in more idealized and controlled conditions than are achieved in nature @xcite . in particular , waves over locally periodic bottoms were studied experimentally @xcite , and such a general for periodic media effect was observed as the bragg resonances and the corresponding band structure with gaps in wave spectrum . it is worth to say that in natural conditions quasi - periodic sand bars occur quite often . in general , a qualitative picture of the mentioned phenomena is clear . as concerning the quantitative side of the mathematical theory of waves over a variable bottom , here not everything that necessary has been done , because practically all developed up to now analytical models and methods are related to the limit cases where the fluid is considered as ideal , and the slope of the bottom is small ( or amplitude of the bottom undulations is small ) . for the general three - dimensional ( 3d ) hamiltonian theory of water waves , such restriction seems to be unavoidable even in considering the most simple , irrotational flows when the state of the system is described by a minimal set of functions , namely by a pair of canonically conjugated quantities as the deviation @xmath11 of the free surface from the horizontal plane and the boundary value @xmath12 of the velocity potential @xcite . a technical difficulty exists here that , when working in 3d space , it is impossible to represent in convenient and compact form the kinetic energy functional @xmath13 which is part of the hamiltonian of the system . small values of the bottom slope and of the free surface slope make possible expansion of the hamiltonian to asymptotic series and subsequent application of various variants of the perturbation theory . in such traditional approach , an inhomogeneous bottom does not allow to write in exact form even linearized equations , not speaking about nonlinear corrections . there are more favorable conditions for progress in theory of 2d potential ideal flows with a free boundary , and the reason for this is the possibility to employ such powerful mathematical tools as analytical functions and the corresponding conformal mappings . time - dependent conformal mappings were successfully used for studying strongly nonlinear 2d wave dynamics on deep water and over straight horizontal bottom @xcite . in the cited works the region occupied by resting fluid ( the lower half - plane or a horizontal stripe ) was mapped onto the region with disturbed free boundary , and the real axis was transformed into moving boundary . such a conformal `` straightening '' of free surface has provided a compact representation for the hamiltonian , derivation of exact equations of motion , and possibility for precise numerical simulations of the system evolution . the purpose of this work is to study the effect of a strongly undulating bottom on 2d ideal potential flows with a free surface . here conformal mappings are used as well , and this is done in two variants . in the first , `` moderate '' variant ( sec . 2 ) , a fixed conformal mapping `` straightens '' the bottom , but not the free boundary . more exactly : instead of the cartesian coordinates @xmath14 and @xmath15 ( with @xmath15-axis up - directed ) , curvilinear coordinates @xmath16 and @xmath17 are introduced , and the change of coordinates is performed with the help of an analytical function @xmath18 which maps the stripe @xmath19 onto the region between the horizontal line @xmath20 and the inhomogeneous bottom @xmath21 . in this case @xmath22 , the horizontal line @xmath20 corresponds to @xmath23 , and on the bottom @xmath24 . the bottom may have arbitrary large slope and even impending pieces where the dependence @xmath25 is multi - valued , as shown in fig.[bandstructure-1 ] . the shape of free surface will be described by a function @xmath26 . the lagrangian for weakly - nonlinear waves is represented as an integral series in powers of the dynamical variables @xmath27 and @xmath28 , with coefficients explicitly depending on the spatial coordinate @xmath16 . in the small - amplitude limit , the wave dynamics is governed by linear integral - differential equations . it is using the conformal variables @xmath16 and @xmath17 , that allows us to obtain these equations in exact form , contrary to the traditional approach where even linearized equations can be obtained only approximately by expansion in the small parameter , the slope of the bottom . the definition `` moderate '' for this variant emphasizes that straightening of the bottom without straightening the free boundary is able to provide not more than a weakly - nonlinear theory . nevertheless , such a theory seems to be helpful and applicable in many practical cases when wave amplitude is small . the results of this part of the work are the derivation of the hamiltonian functional for weakly - nonlinear potential surface waves in canals having arbitrary bottom shape , as well as calculations for band structure of spectrum for a number of periodic bottom profiles . as an example how to treat the linearized equations , also the problem is considered of wave reflection on a smooth `` step '' changing the depth from @xmath29 to @xmath30 . the other variant of using the conformal mappings may be called `` radical '' in the sense it is valid for arbitrary shape of the bottom and for arbitrary shape of the free surface . it is an exact combined theory where a time - dependent conformal mapping straightens both the bottom and the free boundary ( sec . such a mapping can be represented as the result of two mappings : @xmath31 , where the first function @xmath32 maps the horizontal stripe @xmath33 onto the region @xmath34 with the straight lower boundary ( @xmath35 ) and with a perturbed upper boundary , after that the time - independent function @xmath36 maps the half - plane @xmath37 onto the region @xmath38 in the physical plane bounded from below by the bottom . the shape of the free surface will be described by the formula @xmath39 . however , it appears that exact nonlinear equations for @xmath40 in the inhomogeneous case have the same form as the known equations for waves over a horizontal bottom @xcite , but with different analyticity requirements imposed on the solutions . numerical solutions obtained by the spectral method are presented that describe a running and breaking wave ( sec . so , suppose we know the analytical function @xmath41 which maps the horizontal stripe @xmath19 onto the region occupied by the fluid at rest , and this function takes real values on the real axis : @xmath42 . the velocity field is irrotational , and the velocity potential @xmath43 satisfies the laplace equation @xmath44 in the flow region @xmath45 , with the boundary conditions @xmath46 , @xmath47 . due to conformal invariance of the laplace equation in 2d - space , hence equation @xmath48 is satisfied as well , with no - penetration boundary condition on the bottom : @xmath49 . let us now take into account the fact that the lagrangian functional for potential surface waves has the following structure @xcite : @xmath50 where @xmath51 , @xmath52 , and the hamiltonian functional @xmath53 is the total energy of the system sum of the kinetic energy and the potential energy in gravitational field ( in this paper we neglect surface tension effects , though they can be easily incorporated by adding to the hamiltonian the surface energy ) . in our variables @xmath54 this system has the obvious stable equilibrium @xmath55 , @xmath56 , hence one may consider weak oscillations near this equilibrium state . in a standard way ( see , for instance @xcite ) , let us expand the lagrangian ( [ l_general_uv ] ) in powers of the dynamical variables @xmath57 and @xmath58 . it is clear that due to the symmetry principle the expansion for @xmath59 contains only the odd powers of @xmath17 , while the expansion for @xmath60 contains only the even powers of @xmath17 . therefore up to the third order in powers of @xmath57 and @xmath58 the lagrangian ( [ l_general_uv ] ) is equal to @xmath61 where @xmath62 , and the equality @xmath63 has been taken into account in the first integral in r.h.s . the expansion for the kinetic energy ( calculation of the functionals @xmath64 and @xmath65 ) is performed in a standard manner @xcite and gives @xmath66 \psi du,\ ] ] @xmath67 \psi)^2]v du.\ ] ] here the linear hermitian operator @xmath68 $ ] has been introduced , acting as @xmath69 \psi(u)=-\mbox{p.v.}\int\limits_{-\infty}^{+\infty } \frac{\psi_{\tilde u}(\tilde u ) d\tilde u}{2\sinh[(\pi/2)(\tilde u - u)]}.\ ] ] in fourier - representation this operator simply multiplies the fourier - harmonics @xmath70 by @xmath71 . quadratic on @xmath58 part of the potential energy is @xmath72 it is convenient to deal with the function @xmath73 canonically conjugated to @xmath28 , and write the corresponding up - to - third - order hamiltonian in terms of @xmath74 and @xmath57 : @xmath75 \psi du+ \frac{g}{2}\int \frac{\xi^2 } { x'(u ) } du\nonumber\\ \label{tilde_h } & + & \frac{1}{2}\int \frac{[\psi_u^2-([\hat k\tanh \hat k ] \psi)^2]\xi}{x'^2(u ) } du.\end{aligned}\ ] ] physically , this asymptotic expansion of the hamiltonian is on a small parameter the slope of the free surface ( see @xcite for more comments and references ) . the weakly - nonlinear equations of motion have the standard hamiltonian structure @xmath76 \psi -\frac{\partial}{\partial u}\left ( \frac{\xi\psi_u}{x'^2(u)}\right)\nonumber\\ \label{xi_t_1 } & & \qquad-[\hat k\tanh \hat k]\left ( \frac{\xi [ \hat k\tanh \hat k]\psi}{x'^2(u ) } \right ) , \end{aligned}\ ] ] @xmath77 \psi)^2]}{2x'^2(u)}.\ ] ] if @xmath78 , then @xmath79 is approximately equal to the equilibrium depth @xmath80 . for long waves over a such slowly varying bottom , only fourier - harmonics @xmath81 and @xmath82 with small @xmath83 are excited , so in this case the hamiltonian ( [ tilde_h ] ) can be simplified to the local form @xmath84}{2x'^2(u)}\right ] du,\ ] ] which is suitable for consideration of such phenomena as interaction of solitons with the bottom topography . now let us consider the linearized system @xmath85\psi , \qquad -\psi_t = g \frac{\xi}{x'(u)}.\ ] ] for a monochromatic wave ( @xmath86 ) eqs . ( [ lin_xi_psi ] ) are reduced to the single integral equation @xmath87 in the low - frequency limit this equation can be considerably simplified . a variant of simplification is to introduce a new function @xmath88 by the equality @xmath89 . then we obtain the equation @xmath90 the low - frequency limit corresponds to long wave - lengths , when @xmath91 , so we have to deal with the second - order differential equation @xmath92 where @xmath93 should be small ( only in this case the wave length is indeed effectively long ; remember that @xmath79 is of the same order as the depth ) . higher - order approximations to equation ( [ monochrom ] ) can be derived in a similar manner , for instance by change @xmath94 f(u)$ ] and subsequent expanding @xmath95 $ ] and @xmath96 $ ] in powers of @xmath97 . as an explicit example of using eq.([secondorderappr ] ) , we consider reflection of a long wave from a step - shaped bottom inhomogeneity described by the function @xmath98 where @xmath99 , @xmath100 . if frequency of the wave is small , @xmath101 , then equation ( [ secondorderappr ] ) may be applied . calculating the derivative @xmath102 we have for @xmath103 the equation @xmath104}{[c e^{-\alpha u}+1]}f(u)=0,\ ] ] where @xmath105 a general solution for equation ( [ eq_for_f_step ] ) is known @xcite . in particular , the reflection coefficient is given by the expression @xmath106 } { \sinh[\frac{\pi\omega } { \alpha\sqrt{g}}(\sqrt{\tilde h_1}+\sqrt{\tilde h_2})]}\right)^2.\ ] ] interesting phenomena occur if shape of the bottom is periodic : @xmath107 here @xmath108 is a dimensional parameter , @xmath109 are some complex fourier - coefficients . obviously , @xmath110 and @xmath111 decay rapidly at large @xmath112 , since @xmath113 does not have any singularities at @xmath114 . the equation ( [ monochrom ] ) for eigen - functions @xmath115 ( where @xmath116 ) now has the form @xmath117\psi(u)=0,\ ] ] or in fourier - representation @xmath118 for convenience let us denote @xmath119 now we have the infinite chain of linear equations @xmath120 where @xmath121 and @xmath122 interact if the difference between @xmath123 and @xmath124 is an integer number . let us fix some @xmath125 . nontrivial solutions of the system ( [ psi_nu ] ) exist only at definite values @xmath126 , where @xmath127 . it is necessary to note that the functions @xmath128 are periodic : @xmath129 , and even : @xmath130 . this determines the band structure of the spectrum with frequency gaps ( see figs.[bandstructure-1]-[anotherbandstructure ] ) . for numerical computing @xmath128 it is necessary to cut the infinite chain ( [ psi_nu ] ) at some large but finite length , thus considering only @xmath125 between @xmath131 and @xmath132 . practically @xmath132 should be several times larger than the index @xmath133 of @xmath134 . numerical results for @xmath135 shown in the figures [ bandstructure-1]-[anotherbandstructure ] were obtained with the help of the mathematical package maple 8 taking @xmath136 . fig.[anotherbandstructure ] shows that in some cases even for strongly undulating bottom the coefficients @xmath109 with @xmath137 can be still small ( @xmath138 ) . in these cases it is easy to calculate analytically in the main approximation positions of the gaps . for example , let us consider the bottom profile as in fig.[anotherbandstructure ] , where @xmath139 . the gaps in spectrum correspond to integer or half - integer @xmath125 s . it is important that at these values of @xmath125 , solutions of the linear chain ( [ psi_nu ] ) possess definite parity , in the sense that @xmath140 . this allows us in gap calculation consider only positive @xmath125 . let us first consider half - integer @xmath125 s and the corresponding half - infinite chain @xmath141 obviously , the even and odd cases result in different @xmath142 s , and it is this difference that determines the gaps in spectrum . for main - order calculation of the first and third gaps , we cut this chain : @xmath143 , @xmath144 , and so on . now we have to solve the equation for zeros of the determinant @xmath145 @xmath146 ( \lambda - f_{5/2})&&\nonumber\\ -\lambda^2\epsilon^2[\lambda(1\pm\epsilon ) -f_{1/2}]=0.&&\end{aligned}\ ] ] first we take @xmath147 , where @xmath148 is a small quantity of the order @xmath149 . in the main order @xmath150 , and this gives us the first gap : @xmath151 . for the third gap we write @xmath152 , where @xmath153 is of order @xmath154 . the equation for @xmath153 with the third order accuracy is @xmath155(f_{3/2}-f_{5/2 } ) & & \nonumber\\ -\epsilon^2f^2_{3/2}(f_{3/2}(1\pm\epsilon)-f_{1/2})=0.&&\end{aligned}\ ] ] from here we find @xmath156,\ ] ] where we may keep only the second- and third - order terms . this gives us the position of the third gap @xmath157 : @xmath158\nonumber\\ & & \pm \frac{\epsilon^3f^3_{3/2}}{(f_{3/2}-f_{1/2})^2}.\end{aligned}\ ] ] analogously , the gaps at integer @xmath125 s can be considered . these are determined by the system @xmath159 for instance , position of the second gap in second order is given by the formulas @xmath160 in exact nonlinear theory , the shape of free boundary is given in parametric form by a compound function @xmath161 , where @xmath36 is a known function completely determined by the bottom shape [ for example , @xmath162 corresponds to a narrow vertical barrier of the height @xmath163 at @xmath164 on the straight horizontal bottom with the depth @xmath165 . the unknown function @xmath32 should be analytical in the stripe @xmath19 and the combination @xmath166 $ ] should take real values . these conditions relate the real and the imaginary parts of @xmath167 at the real axis @xcite : @xmath168 where @xmath169 is a real function , and the linear anti - hermitian operator @xmath170 is @xmath171 in fourier - representation . in @xmath16-representation @xmath172}.\ ] ] the inverse operator @xmath173 acts as @xmath174}.\ ] ] note that the previously considered operator @xmath68 $ ] is @xmath175 . the kinetic energy functional is now exactly equal to the expression at the r. h. s. of eq.([k_2 ] ) . the lagrangian for @xmath28 and @xmath167 is given by the formula @xmath176du,\end{aligned}\ ] ] where the ( real ) lagrangian indefinite multiplier @xmath177 has been introduced in order to take into account the analytical properties of the function @xmath178 given by eq.([zeta_rho ] ) . from the above lagrangian one can obtain the equations of motion . variation of the action @xmath179 by @xmath180 gives the equation @xmath181 which can be easily transformed to the form ( compare with @xcite ) @xmath182.\ ] ] the variation of the action by @xmath183 results after simplifying in the equation @xmath184 since the product @xmath185 has the same analytical properties as both @xmath186 and @xmath187 , we can multiply eq.([d_l_exact_d_zeta ] ) by @xmath186 and write @xmath188 where @xmath189 is another real function . the imaginary part of the above equation together with eq.([var_l_exact_psi ] ) result in @xmath190.\ ] ] using this equality , we can reduce the real part of eq.([d_l_exact_d_zeta_tilde ] ) to the form @xmath191 -\frac{\hat t[\psi_u\hat r\psi_u]}{\|z'(\zeta)\|^2\|\zeta_u\|^2},\ ] ] which is the bernoulli equation in the conformal variables . exact equations ( [ kinematic_exact ] ) and ( [ bernoulli_exact ] ) [ with given analytical function @xmath36 and with the condition ( [ zeta_rho ] ) ] completely determine the evolution of gravitational surface waves over the undulating bottom parameterized by a real parameter @xmath192 as @xmath193 . equations ( [ kinematic_exact ] ) and ( [ bernoulli_exact ] ) can be represented in another form by using the identity @xmath194=\psi_u^2-(\hat r\psi_u)^2 $ ] @xcite and introducing the complex potential @xmath195 ( which is analytically continued to the stripe @xmath196 ) : @xmath197,\\ \phi_t&=&-\phi_u(\hat t+i)\left [ \frac{\mbox{im\,}\phi_u}{\|z'(\zeta)\|^2\|\zeta_u\|^2}\right]\nonumber\\ \label{bernoulli_exact_phi_zeta } & & -(1+i\hat r)\left[\frac{\|\phi_u\|^2}{2 \|z'(\zeta)\|^2\|\zeta_u\|^2 } + g\mbox{im\,}z(\zeta)\right].\end{aligned}\ ] ] a very interesting point is that one can re - write equations ( [ kinematic_exact ] ) and ( [ bernoulli_exact ] ) without the intermediate function @xmath167 , but directly for @xmath198 . indeed , after multiplying eq.([kinematic_exact ] ) by @xmath199 we obtain the equations @xmath200,\ ] ] @xmath201 -\frac{\hat t[\psi_u\hat r\psi_u]}{\|z_u\|^2},\ ] ] that is exactly the same system as was derived in @xcite for a straight horizontal bottom . however , in our case analytical properties of the function @xmath202 are different : @xmath203.\ ] ] the only requirements for @xmath202 now are that it should be analytical in the stripe @xmath196 and the corresponding mapping should have a physical sense ( no self - intersections are allowed ) . the question may arise : where is the bottom shape in eqs . ( [ kinematic_exact_z])-([bernoulli_exact_z ] ) ? the answer is simple : the shape of the bottom is an integral of motion for this system . roughly speaking , each particular solution of eqs.([kinematic_exact_z])-([bernoulli_exact_z ] ) corresponds to a flow over a definite topography determined by the initial condition @xmath204 . analogously , eqs.([kinematic_exact_phi_zeta])-([bernoulli_exact_phi_zeta ] ) can be represented as @xmath205,\ ] ] @xmath206 -(1+i\hat r)\left[\frac{\|\phi_u\|^2}{2\|z_u\|^2}+g\mbox{im\,}z\right].\ ] ] for numerical simulations , still other equivalent forms of exact equations may be useful , since numerical stability depends dramatically on the choice of dynamical variables @xcite . two alternative sets of equations were used in computations presented below . first , as it was pointed in @xcite for the case of deep water , a good practical choice for the dynamical variables is @xmath207 and @xmath208 . it is easy to derive the equations of motion for @xmath209 and @xmath210 from eqs.([kinematic_exact_phi_z])-([bernoulli_exact_phi_z ] ) , and they are very elegant ( compare with @xcite ) : @xmath211 @xmath212.\ ] ] the variables @xmath213 and @xmath214 do allow stable numerical simulations for waves over varying seabed . however , analytical properties of @xmath213 and @xmath214 are not restricted by conditions similar to eq.([zeta_rho ] ) , and therefore the shape of the bottom is preserved in this case only approximately . the second set of variables , that were used in numerical experiment , consists of two complex functions : @xmath167 and @xmath215 both having effectively controlled analytical properties . with this choice , the bottom shape is preserved exactly , but the corresponding equations of motion are slightly less compact : @xmath216.\end{aligned}\ ] ] it is necessary to explain here some important details about space - periodic solutions of the system ( [ kinematic_exact_zeta_beta])-([bernoulli_exact_zeta_beta ] ) , since spectral numerical methods deal with periodic functions . such solutions exist if the function @xmath199 is periodic with a fixed real period @xmath217 , so that @xmath218 . however , this does not imply that the functions @xmath219 and @xmath220 have a fixed @xmath16-period . it would be so , but the linear operator @xmath221 is singular at small @xmath83 , and its action on a constant function is not periodic in @xmath16 : @xmath222 . thus in the right - hand - sides of the eqs.([kinematic_exact_zeta_beta])-([bernoulli_exact_zeta_beta ] ) we have non - periodic terms . therefore @xmath223 and @xmath224 can not retain a constant @xmath16-period . however , @xmath223 and @xmath224 can be space - periodic with a * time - dependent @xmath16-period*. so , at arbitrary moment of time we will have the equality @xmath225 , where @xmath226 with an unknown real function @xmath227 . the unknown complex fourier coefficients @xmath228 correspond to a real ( @xmath229-periodic on the variable @xmath230 ) function @xmath231 : @xmath232 the linear operator @xmath233 is diagonal in the discrete fourier representation : @xmath234 . analogously , @xmath220 can be represented as @xmath235 , where @xmath236 equations of motion for the real functions @xmath227 , @xmath231 , and @xmath237 follow from eqs.([kinematic_exact_zeta_beta])-([bernoulli_exact_zeta_beta ] ) : @xmath238 \big),\end{aligned}\ ] ] where @xmath239 , @xmath240 and so on . the linear operator @xmath241 is regular . in the discrete fourier representation it is defined as follows : @xmath242 two numerical experiments are briefly reported below , first of them employing eqs.([kinematic_exact_ab])-([bernoulli_exact_ab ] ) , and the second one employing eqs.([kinematic_exact_zeta_beta])-([bernoulli_exact_zeta_beta ] ) . both the systems ( [ kinematic_exact_ab])-([bernoulli_exact_ab ] ) and ( [ kinematic_exact_zeta_beta])-([bernoulli_exact_zeta_beta ] ) are equally convenient for numerical solution by spectral methods , inasmuch as the multiplications can be performed in @xmath16-representation while the linear operators @xmath170 and @xmath221 ( also the @xmath16-differentiation ) are simple in the fourier representation . efficient subroutine libraries for the fast fourier transform are now available . the integration schemes in both cases were based on the runge - kutta 4-th order method , similarly to work @xcite . for computing the discrete fourier transform , the fftw library was used @xcite . the length scale was normalized by a factor @xmath243 and the velocity scale by @xmath244 . the above remark about time - dependent @xmath16-period equally concerns the system ( [ kinematic_exact_ab])-([bernoulli_exact_ab ] ) as ( [ kinematic_exact_zeta_beta])-([bernoulli_exact_zeta_beta ] ) . however , for localized disturbances the relative change of @xmath227 remains small . therefore solitary waves are possible to simulate with a constant @xmath245 and with the regularized @xmath221 . it was done so the first numerical experiment [ referred as ( i ) ] , where periodic boundary conditions were applied with the fixed @xmath16-period @xmath246 . the initial conditions were taken @xmath247 and @xmath248 where @xmath249,\nonumber\\ c(u)&=&\cosh[(l/\pi)\sin\{\pi(u+i)/l\}].\nonumber\end{aligned}\ ] ] this initial configuration is symmetric ( even ) . it results after some time in two oppositely propagating , nearly solitary waves . the waves are created in the region where the depth is maximal . in the course of motion each wave approaches a shallow region where the surface profile @xmath250 becomes steeper and finally multi - valued . some of the results of this numerical experiment are presented in figs.[t0-a]-[crest7-a ] , where the velocity distribution on the surface is shown , as well as shapes of free surface and of the bottom for several moments of time . in general , the computed wave profiles look quite realistic , though the present theory does not take into account viscous effects . steeping of the wave profile is clearly seen . it should be noted , however , in these simulations the bottom shape is preserved only approximately due to discretization errors , and the same concerns the total energy . the computation was stopped well before the moment of formation of a singularity on the crests of the waves , when the numerical scheme becomes invalid . in real world this moment corresponds to development of a three - dimensional instability resulting in vortices , splash and foam . in the second numerical experiment [ referred as ( ii ) ] , the shape of the bottom was fixed by analytical function @xmath251 with the dimensionless parameters @xmath252 , @xmath253 . the initial velocity field was taken zero : @xmath254 , while @xmath255 had the form @xmath256 where @xmath257 qualitatively , these initial conditions are similar to those in the first experiment , however now two oppositely propagating waves are created over inclined region of the bottom , so there is no left - right symmetry in the evolution . the full system ( [ dot_alpha])-([dot_chi ] ) was solved with a high accuracy for @xmath133 in the limits @xmath258 ( the energy conservation was up to 12 digits for a long `` smooth '' initial stage of the evolution , and it was up to 5 digits for the final stage just before the breaking ) . this more accurate numerical solution was compared to a less accurate solution obtained with fixed @xmath259 . the difference was found very small . the corresponding numerical results are presented in figs.[t0]-[t7 ] . here we again observe steeping of the wave profile with the tendency towards finite time singularity formation on the crest . such behavior indeed takes place in natural conditions when the flows are almost two - dimensional . in this paper we have derived approximate weakly - nonlinear , as well as exact nonlinear equations of motion for potential water waves over a strongly inhomogeneous bottom . the consideration was based on using the conformal mappings . for linear waves over periodic seabed , the band structure of the spectrum has been calculated . though the obtained exact equations can be written in formally the same form as those derived in @xcite for a straight horizontal bottom , but admissible solutions have different analytical properties if the bottom is inhomogeneous . when the equations are written in this form , the bottom shape is preserved as an integral of motion . of course , the above `` inviscid theory '' works only on large enough spatial scales and only until the singularity moment , as it was clear from the very beginning . practically , this theory is good for description sea and ocean waves before their breaking . these investigations were supported by the intas under grant number 00292 , by rfbr , by the russian state program for support of the leading scientific schools , and by the russian science support foundation .	two - dimensional free - surface potential flows of an ideal fluid over a strongly inhomogeneous bottom are investigated with the help of conformal mappings . weakly - nonlinear and exact nonlinear equations of motion are derived by the variational method for arbitrary seabed shape parameterized by an analytical function . as applications of this theory , band structure of linear waves over periodic bottoms is calculated and evolution of a strong solitary wave running from a deep region to a shallow region is numerically simulated .
the knowledge of the qcd dynamics at high energies is essential in understanding the hadronic interactions studied at current ( hera , tevatron ) and future ( lhc ) accelerators . in the physics of saturation @xcite the relevant observable is the forward scattering amplitude of a quark - antiquark dipole off a target . it enters several processes at high energy like deep inelastic scattering ( dis ) , photoproduction and hadron - hadron scattering . its quantum evolution has recently been the object of many studies @xcite . the balitsky - kovchegov ( bk ) equation @xcite valid in the large @xmath3 limit and in the mean field approximation provides a tool to study the rapidity behavior of the gluon distribution in a large momentum range including the saturation region where the effects of gluon recombination , taken into account in the non - linear term , become important . the main features of analytical solutions of bk have been used to parameterize @xcite the unintegrated gluon distribution , @xmath4 , which is the fourier - transform of dipole - target amplitude . the knowledge of the corresponding nuclear unintegrated gluon distribution is indeed crucial . it is the basic object required to calculate physical observables in the @xmath5-factorized picture employed by the gluon saturation models . in principle , @xmath4 possesses a bjorken-@xmath2 dependence determined ( here , @xmath6 ) by nonlinear evolution equations of the cgc theory @xcite and its dependence on transverse momentum @xmath7 is fixed by a characteristic saturation momentum , @xmath8 . in most of cgc approaches , the gluon distribution is suppressed below the saturation scale @xmath9 @xcite compared to the perturbative form @xmath10 . the physics of dense partonic systems and their non - linear perturbative evolution to higher energy has motivated several attempts at understanding bulk properties of ultra - relativistic heavy ion collisions such as the multiplicity , rapidity distribution e centrality dependence of particle production @xcite . notice that saturation effects has been claimed to account for the suppression of the high-@xmath11 hadronic spectra in gold - gold and deuterium - gold collisions at rhic . for the evolved gluon distribution , the yield of produced gluons in @xmath0 and @xmath1 collisions at central rapidity can be calculated in the factorized way @xcite , @xmath12 where @xmath13 . under the assumption of local parton - hadron duality , the multiplicity in @xmath1 collisions at central rapidity rises proportional to the nuclear saturation scale , @xmath14 @xcite . for collision energies reaching to 5.5 tev , the upcoming program in lead - lead collisions at the cern large hadron collider ( lhc ) , the confirmation of the conclusions obtained from rhic is expected to take place . it is also expected the discrimination between the different physical mechanisms proposed to explain particle production in high energy nuclear reactions . as noticed in ref . @xcite , purely empirical parameterizations of multiplicity data of a large variety of colliding systems allow a logarithmic dependence on collision energy . however , rhic data by themselves do not differentiate between this and other functional forms like power - laws , negating any possibility to usefully constraint the expectations for lhc energies without further theoretical guidance . in this work we propose a parameterization for the nuclear gluon distribution based on the numerical / analytical solutions of the bk evolution equation . starting with a a proton target , we present a parameterization which encodes the main features of analytical solution of bk equation as geometric scaling property and inclusion of sub - leading terms . it describes all data on dis in the region of @xmath15 in a wide interval of photon virtualities . relying on geometric scaling arguments , we propose a simple model for the nuclear unintegrated gluon distribution . such a model is then compared to data on nuclear ratio and its relation to saturation models is investigated . lets start by considering the collision between a virtual photon and a proton at high energy . in a frame where the photon travels fast , one can consider that it fluctuates into a @xmath16 dipole @xcite . the lifetime of this dipole being much longer than the time of interaction with the proton , one can write the cross section as a product of the wavefunction for a photon to go into a dipole times the dipole - proton cross section , which leads to usual formula @xcite @xmath17 where @xmath18 is the dipole - proton cross section . the transverse and longitudinal photon wavefuctions in this expression are computable in perturbative qed . the proton structure function @xmath19 can be obtained from the @xmath20 cross section through the relation @xmath21 $ ] . if one treats the proton as a homogeneous disk of radius @xmath22 , the dipole - proton cross section in eq . ( [ cross_section ] ) is usually taken to be proportional to the dipole - proton forward scattering amplitude @xmath23 through the relation @xmath24 . here , @xmath6 is the rapidity variable . the bk evolution equation describes the high - energy evolution of the dipole - proton scattering amplitude @xmath25 , where the asymptotic behavior of its solutions is naturally expressed in momentum space . therefore , the photo - absorption cross section can be written in terms of @xmath4 , the fourier transform of @xmath23 : @xmath26 accordingly , the proton structure function can be expressed in momentum space : @xmath27 where the wavefunction is now expressed in momentum space as well ( see ref . @xcite for details ) . the scattering amplitude @xmath28 , the forward scattering amplitude in momentum space , can be obtained as a numerical result of the bk evolution equation . in case of neglecting its dependence on the impact parameter , the bk equation can be expressed as @xmath29 where @xmath30 and @xmath31 is the characteristic function of the balitsky - fadin - kuraev - lipatov ( bfkl ) kernel @xcite and we used @xmath32 with @xmath33 some fixed soft scale . in ref . @xcite , the properties of analytical solutions of bk equation were used to build an analytical expression for @xmath4 . for instance , the bk equation reduces to the fisher - kolmogorov - petrovsky - piscounov ( f - kpp ) equation @xcite ( after changing of variables ) when its kernel is approximated in the saddle point approximation _ i.e. _ to second order in the derivative @xmath34 , the so - called diffusive approximation . the f - kpp equation is a well - known equation in non - equilibrium statistical physics , whose dynamics is used to describe many reaction - diffusion systems in the mean - field approximation . this equation has been extensively studied recently and , in particular , it is known to admit traveling waves as asymptotic solutions @xcite . at asymptotic rapidities , the amplitude @xmath4 , instead of depending separately on @xmath7 and @xmath35 , depends only on the scaling variable @xmath36 , where we have introduced the saturation scale @xmath37 , measuring the position of the wavefront . as discussed in ref . @xcite , a more detailed calculation allows also for the extraction of two additional sub - leading corrections , resulting into the following expression for the tail of the scattering amplitude : @xmath38,\ ] ] where @xmath39 denotes the second derivative of the bfkl kernel with respect to @xmath40 and where the saturation scale contains sub - leading terms . the critical parameters @xmath41 and @xmath42 are obtained from the knowledge of the bfkl kernel alone and correspond to the selection of the slowest possible wave , i.e. @xmath43 . for the lo bfkl kernel , one finds @xmath44 and @xmath45 . concerning the property of geometric scaling @xcite , when one moves along the saturation line , the behavior of the scattering amplitudes remains unchanged . in addition , the sub - leading corrections in also play an important role as the last term introduces an explicit dependence on the rapidity @xmath35 and hence violates geometric scaling . however , this term in negligible when @xmath46 . this means that geometric scaling is obtained for @xmath47 , i.e. , in a window which extends @xmath48 above the saturation scale . it is a remarkable property that , at high - energy , the consequences of saturation are observed arbitrarily far in the tail , where @xmath4 is much smaller than 1 . notice that eq . only gives a description of the tail of the wavefront @xmath49 ( @xmath50 ) . in the infrared domain , one can show that the amplitude behaves like @xmath51 , with @xmath52 being a constant . in ref . @xcite , an analytic interpolation between both asymptotic behavior of the amplitude was proposed . the final expression for the unintegrated gluon function is given by : @xmath53 \ , \left[1-\exp \left(-t_{\text{dil}}\right)\right].\ ] ] this equation is our master formula for the further extrapolation to the nuclear case . in expression above , @xmath54 is an expression which is monotonically decreasing with @xmath55 and which reproduces ( up to the logarithmic factor ) , the amplitude for geometric scaling : @xmath56,\ ] ] with @xmath57 . the parameters of model have been determined from a fit of all the last hera measurements of the proton structure function from h1 @xcite , zeus @xcite , with the analysis being restricted to the following kinematic range @xmath58 and @xmath59 . in our numerical analysis we use the fit including charm @xcite ( @xmath60 gev and @xmath61 gev ) , where @xmath62 , @xmath63 , @xmath64 gev@xmath65 and @xmath66 gev . having presented the unintegrated gluon distribution for the nucleon case , in what follows we investigate its extension for nuclear targets relying on geometric scaling properties . based on geometric scaling arguments and on the main features of numerical solutions of bk equation , in ref . @xcite it is assumed that both the energy and the nuclear size ( or centrality ) dependence on the scattering amplitude @xmath67 can be encoded in the saturation scale @xmath68 for any hadron @xmath69 ( proton or nucleus ) . then the cross section in eq . ( [ cross_section ] ) can be written as @xmath70 as the wavefunction is proportional to @xmath71 times a function of @xmath72 , the cross section is only a function of @xmath73 . this geometric scaling was found to describe all lepton - proton data with @xmath15 . in order to compare protons and different nuclei , in ref . @xcite the following assumption about the impact parameter dependence was considered : it can be scaled by the nuclear radius of the hadronic target @xmath69 , @xmath74 , with @xmath75 fm . then , the condition for geometric scaling in lepton nucleus data is @xmath76 for the dependence on atomic number , the saturation scale in the nucleus grows with the ratio of the transverse partons densities to some power @xmath77 , which was taken as a free parameter : @xmath78 where @xmath8 is the saturation scale for a proton target . in ref . @xcite , it was found @xmath79 , with @xmath80 , which translate into a growth of the nuclear saturation scale faster than @xmath81 for large nuclei . now , we are ready to construct a model for the unintegrated nuclear gluon distribution . using the geometric scaling arguments , we can rewrite eq . ( [ eq : tmodel ] ) , by replacing @xmath82 and @xmath83 : @xmath84 \ , \left\ { 1-\exp \left[- t_{\text{dil}}(\tau_a , y)\right]\right\},\ ] ] where we have defined @xmath85 . we think this procedure is suitable once it has been shown in ref . @xcite that the data for nuclear structure functions lie on the region where @xmath86 . therefore , the condition @xmath87 is roughly valid and geometric scaling violation should be small . in next section , we investigate the numerical calculation of the nuclear structure function using eq . ( [ eq : tmodela ] ) and compare them to the experimental measurements of the nuclear ratios . as a function of bjorken-@xmath2 for several nuclear targets . solid line represents our results ( bm model ) and dot - dashed curves the result of asw model . ] let us now compare our model for the nuclear unintegrated gluon distribution to the experimental data on nuclear ratios . the shadowing in nuclei is usually studied through the ratios of cross sections per nucleon for different nuclei . in fig . [ fig:1 ] , the results of the model are compared with experimental data from refs . @xcite . in these figures , the quantity @xmath88 $ ] is presented . we have selected data where @xmath89 and only statistic errors are shown . the solid lines ( here labeled as bm model ) represents the numerical results using @xmath90 . this means that we are using the conservative value of @xmath77 which reproduces the same growth for the nuclear saturation scale as in ref . it is verified that we fairly reproduce the behavior on @xmath2 for the nuclear ratio , predicting a larger nuclear shadowing for the carbon case . the model is able to describe data even in the region @xmath91 , taking part of the anti - shadowing region . considering both the absence of any free parameter in the nuclear unintegrated gluon distribution we find the agreement quite reasonable . for sake of comparison , we show the numerical results using the armesto - salgado - wiedemann ( asw ) model @xcite ( dashed lines ) for the nuclear cross section . it is verified a saturation of nuclear ration for @xmath92 and the model underestimate the data for lead target . for consistency , we have checked the results for nmc data ( @xmath93 ) and a smaller @xmath77 values seems to be preferred . this fact has been discussed in ref . @xcite , where the nuclear enhancement in the saturation scale has been addressed in distinct saturation model ( with and without geometric scaling violation ) . as pointed out in that study , a precise extraction on the @xmath94 dependence of @xmath95 will play an important role in distinguishing between `` classical '' and `` quantum '' evolution in the cgc . similar discussion can be found in ref . @xcite . it is timely discuss the qualitative aspects of the nuclear ratios using the present parameterization for the unintegrated gluon distribution . assuming geometric scaling in the region covered by the experimental data , we can investigate the asymptotic behaviors of the nuclear ratio . at intermediate @xmath2 , around @xmath96 the unintegrated gluon distribution has the power - like behavior ( modulo logarithmic corrections ) @xmath97 and @xmath98 ( in fact , for the proton case we have additional scaling violations ) . this leads to the following result for the nuclear ratio , taking into account geometric scaling arguments : @xmath99^{\gamma_c}= \left(\frac{r_a^2}{a\,r_p^2}\right)^{1-\gamma_c\delta},\end{aligned}\ ] ] where the asymptotic behavior seems to be dependent on the product @xmath100 . notice that in our case @xmath101 against @xmath102 for the asw parameterization . therefore , the expression above help us to understand the deviations between asw model and the current model . we have checked that eq . ( [ assymp1 ] ) exactly gives the plateau in @xmath103 seen in the asw results in the region @xmath104 . the enhancement observed in our case is a consequence of the anomalous dimension to be dependent on @xmath2 and @xmath71 in that kinematic region , where scaling violations start to take place . therefore , the discussion is more involved in that region , which resembles the discussion on the cronin peak in @xmath1 collisions . on the other hand , for very small-@xmath2 ( and low @xmath71 ) , where @xmath105 , the unintegrated gluon distribution takes a logarithmic behavior . thus , one has @xmath106 $ ] and @xmath107 $ ] . this leads to the following result for the nuclear ratio , @xmath108=\frac{r_a^2}{a\,r_p^2 } \left [ 1+\delta\frac{\log(\frac{r_a^2}{ar_p^2})}{\log(\frac{q^2}{q_{s}^2})}\right]\nonumber \\\end{aligned}\ ] ] which is clearly dependent on the saturation scale and on photon virtuality . in particular , the ratio seems to be closely independent of anomalous dimension @xmath41 . as a final discussion , we discuss the relation of present model with recent determinations of the nuclear unintegrated gluon distribution . notice that the usual notation for the unintegrated gluon density , @xmath109 is related to function @xmath4 by the following relation ( see for instance ref . @xcite for a derivation ) : @xmath110 there is an comprehensive research program in calculating the numerical solution for the nuclear gluon distribution from the bk equation including nlo effects such as the running coupling corrections @xcite . from the phenomenological point of view , our results are similar to the numerical solution as on average the energy behavior for the saturation scale is consistent with @xmath111 and the main features of solutions are present . recently , a new modified bfkl equation incorporating the shadowing and anti - shadowing corrections of the gluon recombination to the unintegrated gluon function has been proposed through the rsyz equation @xcite . the numerical results from such an equation is not identical to the bk results in the transverse momentum space but is similar to a mean field result . of course , our model is less general than gluon function obtained from rsyz equation as we focus on the shadowing region . finally , in ref . @xcite the authors have tested several parameterizations of unintegrated gluon function in nucleus inspired on asymptotic solution of bk . most of them fit our expression for the @xmath4 , which gave satisfactory description of the data for ratio central / peripheral at rhic . this work was supported by the funding agencies cnpq and fapergs , brazil . one of us ( mvtm ) acknowledges the organizers of the _ xvii international workshop on deep - inelastic scattering and related subjects dis2009 _ ( madrid 26 - 30 april 2009 ) , for their invitation and hospitality at universidad autonoma de madrid , where this work has started . we thank raju venugopalan for very useful discussions . j. jalilian - marian , a. kovner , a. leonidov and h. weigert , nucl . b * 504 * ( 1997 ) 415 [ arxiv : hep - ph/9701284 ] ; e. iancu , a. leonidov and l. d. mclerran , nucl . a * 692 * ( 2001 ) 583 [ arxiv : hep - ph/0011241 ] ; h. weigert , nucl . a * 703 * ( 2002 ) 823 [ arxiv : hep - ph/0004044 ] . a. l. ayala , m. b. gay ducati and e. m. levin , nucl . * b493 * , 305 ( 1997 ) ; nucl . phys . * b551 * , 335 ( 1998 ) . l. n. lipatov , sov . . phys . * 23 * , 338 ( 1976 ) ; e. a. kuraev , l. n. lipatov and v. s. fadin , sov . jetp * 45 * , 199 ( 1977 ) ; i. i. balitsky and l. n. lipatov , sov . j. nucl . phys . * 28 * , 822 ( 1978 ) . j. l. albacete , n. armesto , j. g. milhano , c. a. salgado and u. a. wiedemann , phys . d * 71 * , 014003 ( 2005 ) ; j. l. albacete and y. v. kovchegov , phys . d * 75 * , 125021 ( 2007 ) ; j. l. albacete , n. armesto , j. g. milhano and c. a. salgado , phys . d * 80 * , 034031 ( 2009 ) .	the solution of non - linear evolution equations for dense nuclear gluon density has been suggested as one of the relevant mechanisms of @xmath0 and @xmath1 collisions at collider energies . here we study a simple parameterization for the unintegrated gluon distribution using the knowledge of asymptotic solutions of the balitsky - kovchegov equation , describing high - energy qcd in the presence of saturation effects . a satisfactory description of nuclear shadowing at small-@xmath2 is obtained and it allows us to understand the qualitative behavior shown by data .
radio - quiet quasars ( rqq ) represent more than half the known agn at high redshift ( @xmath3 ) , yet information on the galaxies that host these active nuclei is currently extremely limited . at lower redshift , the luminous rqq are found primarily in giant ellipticals with luminosities of several times a @xmath4 galaxy at @xmath5 ( e.g. mclure et al . 1999 ) . in addition , the luminosity of the host seems to correlate roughly with that of the nucleus ( mcleod & rieke 1995 ; mclure et al . 1999 , mcleod , rieke , & storrie - lombardi 1999 ) . similarly , almost every nearby bulge - dominated galaxy contains a supermassive black hole candidate whose mass is roughly proportional to the mass of its bulge ( magorrian et al . 1998 ; van der marel 1999 ) , implying a strong link between the formation and evolution of galaxies and those of the quasars and their hosts . semi - analytic hierarchical clustering models of galaxy formation have been applied by kauffman & haehnelt ( 1999 ) to this question , and they have made some specific predictions about the evolution of the nuclear magnitude host relation for radio - quiet quasars . to test such theories we have made hst nicmos observations of radio - quiet quasar hosts near the epoch of the peak quasar density ( @xmath02 3 ) . we have chosen a sample of quasars whose nuclei are faint enough to provide a good comparison sample to those of the well - studied low-@xmath6 quasars . throughout this paper we use @xmath7 = 50 km s@xmath8 mpc@xmath8 and @xmath9 = 1 . we selected 5 quasars from the faint quasar survey of zitelli et al . ( 1992 ) with 21.6@[email protected] at @xmath0 23 . the exact redshift range was constrained to avoid emission lines falling in the nicmos h filter passbands ; this resulted in a sample of 3 objects at @xmath01.8 and 2 objects at @xmath02.7 . their nuclear @xmath11 are in the range @xmath1222 to @xmath13 , making them comparable ( and somewhat fainter ) in absolute magnitude to many low-@xmath6 quasar samples ( e.g. bahcall et al . 1997 , mcclure et al . 1999 , mcleod et al . 1999 ) the observations were made using the nic2 aperture of hst s nicmos camera , which has a field size of 192 @xmath14 192 , at a scale of 0075 pixel@xmath8 . to achieve emission - line - free imaging , the @xmath15 objects were imaged in the f165 m filter , resulting in a restframe wavelength of @xmath2v , and the @xmath16 objects in the f160w filter , corresponding to rest - frame @xmath2b . we chose a nearby star for each of the 5 quasars , and observed this star in the same visit as the quasar using the identical dither pattern in order to characterize the point spread function ( psf ) . the dither pattern included half - pixel offsets to improve resolution by adequately sampling the hst psf at these wavelengths . we observed each of the 5 quasars ( and its corresponding psf star ) with two visits separated by several months , resulting in observations of each field at significantly different position angles on the sky relative to the psf pattern this allows an independent check on the reality of any residual emission seen . the final fwhms achieved were 014 016 . observing times per frame were on the order of 1500 s per quasar . total exposures for the quasars were on the order of 6000 s per visit . however , due to variable saa cr persistence problems , sky noise levels in the final images differ . we recalibrated the raw nicmos data using a modified version of the standard pipeline process . we have combined each of the two visits to each object separately , by using a simple method that determines the locations of the bad pixels in the initial frames and creates bad pixel masks using crrej , resamples the corrected frames to double the linear dimensions , and combines these using standard rejection . we achieved the cleanest subtraction with a psf star observed during the same visit as the quasar . we have made a simple , direct psf subtraction through an iterative method that varies the centering and scaling of the psf versus the quasar , subtracts the two images , and measures the chi - square of the residuals . finding an unambiguous relative centering is simple ; determining the best scaling is more subjective . here , we require flatness of the residual across some central region ( generally within a radius of @xmath2015 ) . this will still likely result in an oversubtraction , depending on how peaked the real host galaxy is within this inner radius . in these images , the psf residuals seem to dominate within a radius of @xmath201 . within this region , we do not expect to recover morphological information ; we can however find real excess flux . we have clearly resolved excess flux around 4 out of the 5 quasars relative to the psf stars , and also relative to a star with an apparent magnitude similar to those of the quasars found within the field of mzz 9592 . this field star provides a good check that our results are not the product of some difference in the way we have observed and reduced the psf stars versus the quasars . we have thus also treated this star as if it were a quasar and applied the same psf subtraction techniques to it ; we find no significant flux in the residual . in figure 1 , we show the results of these analyses for both visits of mzz 9592 , separately . we have rotated the results of the second visit to match the orientation of the first visit . in figure 2 , we give the psf - subtracted combined results for 4 of the quasars , gaussian - smoothed . this shows the range from the brightest to faintest of the residual host galaxies . to provide the simplest basis for cross - comparison between samples , we have first calculated simple aperture magnitudes . we have used an aperture of 2 ( corresponding to @xmath216 kpc ) to include most of the flux expected from a host galaxy , while excluding most nearby discrete companions . to estimate errors , we have also calculated the host magnitudes for the two visits separately . in most cases the differences were @xmath170.3 mag . we give in table 1 the results of our magnitude analysis ; we find the detected hosts vary from @xmath2l@xmath1 to 4 l@xmath1 , using @xmath18 and @xmath19 at @xmath5 from the field galaxy luminosity function of loveday et al . the faintest residual host ( around mzz 4935 ) we may barely detect at a flux @xmath20.4l@xmath1 . these fluxes will generally need a correction for flux lost from the psf subtraction process which will vary depending on how compact the intrinsic , underlying host is . in the next section we discuss simple models to estimate the amount of this correction . .mzz quasar nuclear and host properties [ cols="<,^,^,^,^,^,^ " , ] the morphologies of these hosts are quite compact , generally less than @xmath205 ( @xmath24 kpc ) effective radius . we also find in 2 cases galaxies close to the quasar in projection ( @xmath17 10 kpc ) , and in mzz 9592 an off - center host residual . ( 350,400 ) ( -80,-170 ) ( -25,360 ) ( 194,360 ) ( -25,130 ) ( 194,130 ) * figure 1 . * mzz 9592 , psf subtractions and tests . _ a. _ visit one , psf - subtracted , using the observed bright psf star from the same visit . the same image with a different scaling is shown in the lower left inset , demonstrating that the residual host is not centered on the nucleus . the upper right inset shows the unsubtracted quasar . _ b. _ visit two , psf - subtracted , rotated to match the orientation of visit one . extra sky noise comes from unremoved low - level cr residuals . _ c. _ visit one , psf - subtracted with the star that falls within the field , demonstrating that the residual is not an artifact of the mismatch between the observational strategies applied to the quasar fields and the bright psf stars . _ d. _ the field star minus the psf star : no residual flux . ( 350,500 ) ( -80,-100 ) ( -25,430 ) ( 194,430 ) ( -25,200 ) ( 194,200 ) * figure 2 . * psf - subtracted mzz quasar hosts , gaussian - smoothed with a kernel of 006 . each panel is 57 square ( or roughly 45 kpc ) , n up , e left . _ a. _ mzz 9592 , @xmath20 2.7 , with an inset of the central region ( unsmoothed ) . there is an off - center residual host component . _ b. _ mzz 9744 , @xmath0 1.8 _ c. _ mzz 1558 , @xmath21 2.7 _ d. _ mzz 4935 , @xmath201.8 , host not detected . note the apparent close ( @xmath2 10 kpc ) companion galaxies in panels _ b _ and _ c_. to compare the results of our study with studies of lyman break galaxies and radio galaxies at similar redshifts , we have made some simple simulations of the effect the presence of the quasar nucleus and the psf - subtraction process have on the derived magnitudes and morphologies . we have used for the first set of these simulations hst nicmos observations of a number of spectroscopically identified @xmath0 2 3 lyman break galaxies from the hdf north ( dickinson et al . 1999 ) , which were observed in the f160w filter to a much greater depth than our observations . for the observed galaxies that were at the same redshift as our quasars , we first resampled these imaging data to match our final pixel scale , added `` quasar nuclei '' of varying magnitudes at the center of the galaxy by scaling and adding the mzz9592 field star , then added poissonian noise to the appropriate level . we then used one of the observed psf stars to run the same automatic psf - subtraction process that we used on the quasars . we have found that the subtraction process in most cases on these lyman break galaxies resulted in a loss of flux of @xmath20.3 magnitude , giving us an indication of the amount to correct our derived hosts . even after such a correction , most of our quasar hosts have relatively moderate total magnitudes of @xmath2l@xmath1 , and these magnitudes and compact sizes are basically consistent with the magnitudes and compactness of star - forming galaxies at similar epochs ( dickinson et al . 1999 ) . to enable a comparison with previous ground - based infrared imaging work on radio - loud galaxies , we adjust our fluxes to match those of eales et al . 1997 , who studied 6c radio galaxies up to @xmath03 . extrapolating our h band fluxes to the k band gives k band magnitudes of 22.0 , 20.3 , and 20.5 for the @xmath201.8 quasar hosts , and 20.1 and 20.8 for the @xmath02.7 quasar hosts ( after correction to the larger aperture used by eales et al . ) . they found that @xmath22 6c galaxies have a median @xmath23 magnitude of @xmath218.5 , while our 3 quasar hosts are at least a magnitude fainter than this , even allowing for a half magnitude correction to our fluxes lost in the subtraction process . our @xmath02.7 quasar hosts are also a magnitude fainter at k than the 6c objects . as the quasars in our sample have nuclear magnitudes that are comparable to those of the well - studied samples of low redshift quasars , we can make a direct comparison , and find that these @xmath22 3 rqq hosts are of similar or fainter m@xmath24 as those of the sample of bahcall et al . ( 1997 ) and mclure et al . ( 1997 ) , for example . however , if the host galaxies of luminous radio - quiet quasars evolve passively into giant ellipticals today , then at high redshift they should have had similar host magnitudes to the radio galaxies , rather than the moderately faint hosts we have found . our results are consistent with other recent results on some brighter , lensed quasars ; rix et al . 1999 find that their sample of lensed @xmath22 radio - quiet quasars ( de - magnified @xmath25 @xmath2 24 28 ) also had hosts with comparably faint magnitudes . though inconsistent with passive evolution , our finding l@xmath1 hosts at @xmath22 3 agrees fairly well with the bottom - up hierarchical galaxy formation models of kauffman & haehnelt ( 1999 ) ; they predict median host luminosities that are somewhat below present day l@xmath1 for quasars at @xmath26 ( and even fainter at @xmath27 ) , for quasars with the nuclear magnitudes of our sample . these hosts might therefore still be undergoing major mergers which would allow them to evolve into the present day ges associated with low-@xmath6 quasars ; it is unclear in this interpretation , however , why the radio galaxies do not undergo similar mergers and evolution . conversely , these high-@xmath6 quasar hosts could be l@xmath1 galaxies that will not significantly evolve in luminosity . bahcall , j. , kirhakos , s. , saxe , d. , & schneider , d. 1997 , apj , 479 , 642 dickinson et al . 1999 , in preparation . eales , s. , rawlings , s. , law - green , d. , cotter , g. , & lacy , m. 1997 , mnras , 291 , 593 kauffmann , g. & haehnelt , m. 1999 , mnras , submitted ( astro - ph/9906493 ) loveday , j. , peterson , b. , efstathiou , g. , maddox , s. 1992 , apj , 390,338 . magorrian j. , et al . , 1998 , aj , 115 , 2285 mclure r.j . , dunlop j.s . , kukula m.j . , baum s.a . , odea c.p . , hughes d.h . , aj , submitted ( astro - ph/9809030 ) mcleod , k. , & rieke , g. 1995 , apj , 454 , l77 mcleod , k. , rieke , g. , & storrie - lombardi , l. 1999 , apjl , 511 , l67 van der marel r.p . , 1999 , aj , 117 , 744 zitelli , v. , mignoli , m. ; zamorani , g. ; marano , b. ; boyle , b. j. , 1992 , mnras , 256 , 349	we will present the results of a nicmos h - band imaging survey of a small sample of @xmath02 3 radio - quiet quasars . we have resolved extension in at least 4 of 5 objects and find evidence for a wide range in the morphologies and magnitudes of these hosts . the host galaxy luminosities range from sub - l@xmath1 to about 4 l@xmath1 , with most of the hosts having luminosities about l@xmath1 . these host galaxies have magnitudes and sizes consistent with those of the ly break galaxies at similar redshifts and at similar rest wavelengths , but are about a magnitude fainter than the comparable 6c radio galaxies . one residual host component is not centered on the quasar nucleus , and several have close companions ( within @xmath210 kpc ) , indications that these systems are possibly in some phase of a merger process .
sensors are sometimes distributed in a regular fashion to monitor an area . we assume that the sensors use wireless communication . most wireless communication protocols allow the sensors to send at arbitrary times . however , this can cause the following _ collision problems _ : if two distinct sensors @xmath0 and @xmath1 send at the same time and @xmath1 is within the interference range of @xmath0 , then frequently hardware limitations prevent @xmath1 from receiving the message of @xmath0 correctly . in addition , if two distinct sensors @xmath0 and @xmath1 send at the same time and a sensor @xmath2 is within interference range of both @xmath0 and @xmath1 , then @xmath2 will not be able to correctly receive either message . in these cases , the sensors @xmath0 and @xmath1 need to resend their messages , which is evidently a waste of energy . let us assume that the sensors have access to the current time , represented by an integer @xmath3 . one can assign each sensor node an integer @xmath4 and set up a periodic schedule such that a node with integer @xmath4 is allowed to broadcast messages at time @xmath3 if and only if @xmath5 . the goal of this paper is to give a convenient combinatorial formulation using lattice tilings that allows one to assign optimal schedules with minimal number of time slots @xmath6 such that no two sensors that are scheduled to broadcast simultaneously have intersecting interference ranges ; we call such schedules _ collision - free_. [ [ related - work . ] ] related work . + + + + + + + + + + + + + since most communication protocols for wireless sensor networks are probabilistic in nature , there exist few prior works that are directly related to our approach . however , there exist a few notable exceptions that we want to discuss here . suppose for the moment that we are given a finite set of @xmath4 sensors that share the same frequency band for communication . the simplest way to ensure that the communication will be collision - free , is to use a time division multiple access ( tdma ) scheme . here each of the @xmath4 sensors is assigned a different time slot and scheduling is done in a round robin fashion . because of its simplicity , this scheme is used in many systems , see e.g. ( * ? ? ? * chapter 3.4 ) . the obvious disadvantage of tdma is that it does not scale : if the number @xmath4 of sensors is large , then the sensors can not communicate frequently enough . the basic tdma scheme does not take advantage of the fact that each sensor typically affects only small number of neighboring sensors by its radio communication . this prompts the question whether one can modify the tdma scheme and find a schedule with @xmath6 time slots that is collision - free . to answer this question , consider a directed graph that has a node for each sensor and an edge from vertex @xmath7 to vertex @xmath8 if and only if @xmath8 is affected by the radio communication of @xmath7 a valid schedule with @xmath6 time slots corresponds to a distance-2 coloring with @xmath6 colors , that is , all vertices of distance @xmath9 must be assigned a different color (= time slot ) to avoid collision problems . therefore , the number of time slots @xmath6 of an optimal collision - free schedule coincides with the chromatic number of a distance-2 coloring . the distance-2 coloring problem is also known as the broadcast scheduling problem in the networking community . mccormick has shown that the decision problem whether a given graph has a distance-2 coloring with @xmath6 colors is np - complete @xcite . lloyd and ramanathan showed that the broadcast schedule problem even remains np - complete when restricted to planar graphs and @xmath10 time slots @xcite . due to these intractability results , much of the subsequent research focused on heuristics for finding optimal schedules ; for instance , wang and ansari used simulated annealing @xcite , and shi and wang used neural networks @xcite to find optimal schedules . another popular direction of research are approximation algorithms for broadcast scheduling algorithms , see e.g. @xcite . [ [ contributions . ] ] contributions . + + + + + + + + + + + + + + the main contributions of this paper can be briefly summarized as follows ( the terminology is explained in the subsequent sections ) : we develop a method that allows one to derive an optimal collision - free schedule from the tiling of a lattice . our scheme scales to an arbitrary number of sensors ; in fact , we formulate our schedules for an infinite number of sensors . schedules for a finite number of sensors are obtained by restriction , and these schedules remain optimal under very mild conditions ( given in the conclusions ) . our assumption on the set of prototiles ensures that an optimal schedule is obtained regardless of the chosen tiling . in section 4 , we show that if our assumption on the set of prototiles is removed , then in general one will not obtain an optimal schedule . we formulate our results for arbitrary lattices in arbitrary dimensions , since the proofs are not more complicated than in the familiar case of the two - dimensional square lattice . for the square lattice , there are polynomial - time algorithms available to check whether a given prototile can tile the lattice ; thus , despite the fact that finding optimal schedules is np - hard in general , one can use our method to easily construct optimal schedules in the case of a single prototile . this method of creating simple instances of an np - hard problem might be of independent interest . a euclidean lattice @xmath11 is a discrete subgroup of @xmath12 that spans the euclidean space @xmath12 as a real vector space . in other words , there exist @xmath13 vectors @xmath14 in @xmath11 that are linearly independent over the real numbers such that @xmath15 and for each vector @xmath7 in @xmath11 there exists an open set containing @xmath7 but no other element of @xmath11 . in particular , the group @xmath11 is isomorphic to the additive abelian group @xmath16 . two examples of lattices in two dimensions are illustrated in figure [ fig : lattice ] . ( 40,40 ) z1 = ( 1,0 ) ; z2 = ( 0,1 ) ; s = 12 ; drawarrow sz1+sz2 2sz1+sz2 ; drawarrow sz1+sz2 sz1 + 2sz2 ; pickup pencircle scaled 3 ; for k = 1 step s until 5s : draw kz1 + 1z2 scaled 2 ; endfor ; for k = 1 step s until 5s : draw kz1+(s)z2 ; endfor ; for k = 1 step s until 5s : draw kz1+(2s)z2 ; endfor ; for k = 1 step s until 5s : draw kz1+(3s)z2 ; endfor ; for k = 1 step s until 5s : draw kz1+(4s)z2 ; endfor ; label.lrt(btex @xmath17 etex , sz1+sz2 ) ; label.lrt(btex @xmath18 etex,2sz1+sz2 ) ; label.lrt(btex @xmath19 etex , sz1 + 2sz2 ) ; label.lrt(btex @xmath20 etex,2sz1 + 2sz2 ) ; @xmath21 ( 40,40 ) z1 = ( 1,0 ) ; z2 = ( 0.5,0.866 ) ; drawarrow sz1+sz2 2sz1+sz2 ; drawarrow sz1+sz2 sz1 + 2sz2 ; pickup pencircle scaled 3 ; for k = 1 step s until 5s : draw kz1 + 1z2 scaled 2 ; endfor ; for k = 1 step s until 5s : draw kz1+(s)z2 ; endfor ; for k = 1 step s until 5s : draw ( -s+k)z1+(2s)z2 ; endfor ; for k = 1 step s until 5s : draw ( -s+k)z1+(3s)z2 ; endfor ; for k = 1 step s until 5s : draw ( -2s+k)z1+(4s)z2 ; endfor ; label.lrt(btex @xmath17 etex , sz1+sz2 ) ; label.lrt(btex @xmath22 etex,2sz1+sz2 ) ; label.lrt(btex @xmath23 etex , sz1 + 2sz2 ) ; label.lrt(btex @xmath24 etex,2sz1 + 2sz2 ) ; our goal is to find a deterministic collision - free periodic schedule for sensors located at the points of a lattice @xmath11 that is optimal in the number of time slots , _ i.e. _ , no periodic schedule with a shorter period can be found that is collision - free . we call a finite subset @xmath25 of @xmath11 a _ prototile _ or a _ neighborhood _ of the point @xmath17 if and only if it contains @xmath17 itself . the particular nature of @xmath25 will be determined for instance by the type of antenna and by the signal strength used by the sensor . the elements in @xmath25 are the sensors affected by wireless communication of the sensor located at the point @xmath17 ( that is , only the elements in @xmath25 are within interference range of the sensor located at the point @xmath17 ) . we will first assume a homogeneous situation , namely the neighborhood affected by communication of the sensor located at a point @xmath3 in @xmath11 is of the form @xmath26 where the addition denotes the usual addition of vectors in @xmath12 . the set @xmath27 contains @xmath3 , since @xmath17 is contained in @xmath25 . some examples of neighborhoods @xmath25 are given in figure [ fig : neighborhoods ] . ( 50,50 ) t = 12 ; z3 = t(1,0 ) ; z4 = t(0,1 ) ; def cross(expr a , b ) = begingroup pair g ; g = az3+bz4 ; pickup pencircle scaled 2 ; draw ( g-(1,1))(g+(1,1 ) ) ; draw ( g-(1,-1))(g+(1,-1 ) ) ; endgroup ; enddef ; pickup pencircle scaled 3 ; for k = 1 step 1 until 7 : for l = 1 step 1 until 7 : draw kz3+lz4 ; endfor ; endfor ; label.lrt(btex 0 etex , 4z3 + 4z4 ) ; for k = 3 upto 5 : for l = 3 upto 5 : cross(k , l ) ; endfor ; endfor ; pickup pencircle scaled 1 ; pair tg ; tg = 2.5(z3+z4 ) ; draw ( ( 0,0)(0,1)(1,1)(1,0)cycle ) scaled 36 shifted tg withcolor 0.8white ; @xmath28 ( 50,50 ) z3 = t(1,0 ) ; z4 = t(0,1 ) ; def cross(expr a , b ) = begingroup pair g ; g = az3+bz4 ; pickup pencircle scaled 2 ; draw ( g-(1,1))(g+(1,1 ) ) ; draw ( g-(1,-1))(g+(1,-1 ) ) ; endgroup ; enddef ; pickup pencircle scaled 3 ; for k = 1 step 1 until 7 : for l = 1 step 1 until 7 : draw kz3+lz4 ; endfor ; endfor ; label.lrt(btex 0 etex , 4z3 + 4z4 ) ; for k = 3 upto 5 : cross(k,4 ) ; cross(4,k ) ; endfor ; pickup pencircle scaled 1 ; draw ( ( 0,1)(0,2)(1,2)(1,3)(2,3)(2,2)(3,2)(3,1)(2,1)(2,0)(1,0)(1,1)cycle ) scaled ( t ) shifted ( 2.5(z3+z4 ) ) withcolor 0.8white ; @xmath28 ( 50,50 ) z3 = t(1,0 ) ; z4 = t(0,1 ) ; def cross(expr a , b ) = begingroup pair g ; g = az3+bz4 ; pickup pencircle scaled 2 ; draw ( g-(1,1))(g+(1,1 ) ) ; draw ( g-(1,-1))(g+(1,-1 ) ) ; endgroup ; enddef ; pickup pencircle scaled 3 ; for k = 1 step 1 until 7 : for l = 1 step 1 until 7 : draw kz3+lz4 ; endfor ; endfor ; for k = 5 upto 6 : for l = 3 upto 5 : cross(k , l ) ; endfor ; endfor ; cross(3,4 ) ; cross(4,4 ) ; pickup pencircle scaled 1 ; draw ( ( 0,1)(0,2)(2,2)(2,3)(4,3)(4,0)(2,0)(2,1)(0,1 ) ) scaled t shifted ( 2.5(z3+z4 ) ) withcolor 0.8white ; label.lrt(btex 0 etex , 3z3 + 4z4 ) ; our schedule will be a deterministic periodic schedule , that is , each sensor is assigned a certain time slot and it is only allowed to send during that time slot . since our schedule is required to be free of collision problems , it follows that the sensors located at distinct points @xmath29 and @xmath3 in @xmath11 can not broadcast at the same time unless @xmath30 let @xmath31 denote a subset of @xmath11 . we say that @xmath31 provides a _ tiling _ of @xmath11 with neighborhoods ( or tiles ) of the form @xmath25 if and only if the following two conditions hold : @xmath32 , @xmath33 for all distinct @xmath34 in @xmath31 . the set @xmath31 contains all the vectors that translate the prototile @xmath25 . condition * t1 * says that the whole lattice @xmath11 is covered by the translates @xmath27 of the prototile @xmath25 , when @xmath3 ranges over the elements of @xmath31 . condition * t2 * simply says that the translates of the tile @xmath25 do not overlap . the tilings provide us with an elegant means to construct an optimal deterministic schedule . [ th : thm1 ] let @xmath31 be a tiling of a euclidean lattice @xmath11 in @xmath12 with neighborhoods of the form @xmath25 . then there exists a deterministic periodic schedule that avoids collision problems using @xmath35 time slots . the schedule is optimal in the sense that one can not achieve this property with fewer than @xmath6 time slots . suppose that @xmath36 is the neighborhood of @xmath17 . for @xmath4 in the range @xmath37 , we schedule the sensors located at the points @xmath38 at time @xmath5 . we first notice that each sensor located at a point in @xmath11 is scheduled at some point in time , since @xmath39 by property * t1 * of a tiling . seeking a contradiction , we assume that the schedule is not collision - free . this means that at some time @xmath4 in the range @xmath40 there exist sensors located at the positions @xmath41 and @xmath42 with distinct @xmath29 and @xmath3 in @xmath31 such that @xmath43 . however , this would imply that @xmath44 for distinct @xmath29 and @xmath3 in @xmath31 , contradicting property * t2 * of a tiling . it follows that our schedule is collision - free . it remains to prove the optimality of the schedule . seeking a contradiction , we assume that there exists a schedule with @xmath45 time slots that is collision - free . this means that for some time slot @xmath4 in the range @xmath46 two elements @xmath47 and @xmath48 of @xmath25 must be scheduled . however , this would imply that the element @xmath49 is contained in both sets @xmath50 and @xmath51 , contradicting the assumption that the schedule with @xmath52 time slots is collision - free . we illustrate some aspects of the proof of the previous theorem in figure [ fig : thm1a ] . ( 50,50 ) z3 = t(1,0 ) ; z4 = t(0,1 ) ; picture tile ; vardef pnt(expr a , b ) = ( az3+bz4 ) enddef ; pickup pencircle scaled 1 ; path p ; p = ( ( 0,1)(0,2)(2,2)(2,3)(4,3)(4,0)(2,0)(2,1)(0,1)cycle ) ; draw p scaled t shifted ( 0.5z3 + 0.5z4 ) withcolor 0.8white ; label(btex @xmath53 etex , pnt(1,2 ) ) ; label(btex @xmath54 etex , pnt(2,2 ) ) ; label(btex @xmath55 etex , pnt(3,3 ) ) ; label(btex @xmath56 etex , pnt(4,3 ) ) ; label(btex @xmath57 etex , pnt(3,2 ) ) ; label(btex @xmath58 etex , pnt(4,2 ) ) ; label(btex @xmath59 etex , pnt(3,1 ) ) ; label(btex @xmath60 etex , pnt(4,1 ) ) ; tile : = currentpicture ; currentpicture : = nullpicture ; draw tile ; draw tile shifted pnt(2,2 ) ; draw tile shifted pnt(0,4 ) ; draw tile shifted pnt(4,0 ) ; draw tile shifted pnt(4,4 ) ; draw tile shifted pnt(6,2 ) ; draw tile shifted pnt(8,0 ) ; draw tile shifted pnt(8,4 ) ; label(btex @xmath61 etex , pnt(0,4 ) ) ; label(btex @xmath61 etex , pnt(13,4 ) ) ; @xmath21 ( 50,50 ) z3 = t(1,0 ) ; z4 = t(0,1 ) ; picture tile ; vardef pnt(expr a , b ) = ( az3+bz4 ) enddef ; pickup pencircle scaled 1 ; path p ; p = ( ( 0,1)(0,2)(2,2)(2,3)(4,3)(4,0)(2,0)(2,1)(0,1)cycle ) ; draw p scaled t shifted ( 0.5z3 + 0.5z4 ) withcolor 0.8white ; label(btex @xmath53 etex , pnt(1,2 ) ) ; label(btex @xmath54 etex , pnt(2,2 ) ) ; label(btex @xmath55 etex , pnt(3,3 ) ) ; label(btex @xmath56 etex , pnt(4,3 ) ) ; label(btex @xmath57 etex , pnt(3,2 ) ) ; label(btex @xmath58 etex , pnt(4,2 ) ) ; label(btex @xmath59 etex , pnt(3,1 ) ) ; label(btex @xmath60 etex , pnt(4,1 ) ) ; tile : = currentpicture ; currentpicture : = nullpicture ; draw tile ; draw tile shifted pnt(2,2 ) ; draw tile shifted pnt(0,4 ) ; draw tile shifted pnt(4,0 ) ; draw tile shifted pnt(4,4 ) ; draw tile shifted pnt(6,2 ) ; draw tile shifted pnt(8,0 ) ; draw tile shifted pnt(8,4 ) ; label(btex @xmath61 etex , pnt(0,4 ) ) ; label(btex @xmath61 etex , pnt(13,4 ) ) ; draw p scaled t shifted pnt(3.5,2.5 ) dashed evenly ; our concept of tiling a lattice with translates of a prototile @xmath25 turned out to be convenient for our purposes . in this section , we relate the tilings of a lattice to tilings of the euclidean space @xmath12 , so that we can benefit from the large number of results that are available in the literature . any tiling of a lattice @xmath11 can be converted into a tiling of @xmath12 as follows . let @xmath62 denote the union of the closed voronoi regions about the points in @xmath25 . then the translates @xmath63 with @xmath3 in @xmath31 yield a tiling of @xmath12 . conversely , any tiling of @xmath12 with translates of a tile consisting of the union of voronoi regions of points in @xmath11 evidently yields a tiling of the lattice @xmath11 in our sense . figure [ fig : quasipolyominoes ] shows some two - dimensional examples of voronoi regions . ( 40,40 ) z1 = ( 1,0 ) ; z2 = ( 0,1 ) ; pickup pencircle scaled 1 ; fill ( ( 0,0)(1,0)(1,1)(0,1)cycle ) scaled s shifted ( 6.4(z1+z2 ) ) withcolor 0.8white ; pickup pencircle scaled 3 ; for k = 1 step s until 3s : draw kz1 + 1z2 scaled 2 ; endfor ; for k = 1 step s until 3s : draw kz1+(s)z2 ; endfor ; for k = 1 step s until 3s : draw kz1+(2s)z2 ; endfor ; @xmath64 ( 40,40 ) z1 = ( 1,0 ) ; z2 = ( 0.5,0.866 ) ; vardef pnt(expr a , b ) = ( az1+bz2 ) enddef ; pickup pencircle scaled 1 ; fill ( pnt(0,0)pnt(1,0)pnt(1,1)pnt(0,2 ) pnt(-1,2)pnt(-1,1)pnt(0,0)cycle ) scaled ( 0.5s ) shifted ( z1+pnt(1s,0.5s ) ) withcolor 0.8white ; pickup pencircle scaled 3 ; for k = 1 step s until 3s : draw kz1 + 1z2 scaled 2 ; endfor ; for k = 1 step s until 3s : draw kz1+(s)z2 ; endfor ; for k = 1 step s until 3s : draw ( -s+k)z1+(2s)z2 ; endfor ; the union of voronoi regions about points in a lattice are also known as quasi - polyforms . a quasi - polyform that is homeomorphic to the unit ball in @xmath12 is known as a polyform . the books by grnbaum and shepherd @xcite and by stein and szab @xcite contain numerous examples of tilings obtained by translating quasi - polyforms ( and especially polyforms ) . the polyforms in the square grid @xmath65 are called polyominoes , the most well - known type of polyforms ; see golomb s book @xcite . by abuse of language , we will also refer to a prototile @xmath25 in @xmath66 as a polyomino if the union of the voronoi regions of @xmath25 form a polyomino . a prototile @xmath25 in a lattice @xmath11 that admits a tiling is called _ exact_. it is natural to ask the following question : when is a given prototile @xmath25 exact , _ i.e. _ , when does there exist a subset @xmath31 of @xmath11 such that the conditions * t1 * and * t2 * are satisfied ? beauquier and nivat gave a simple criterion that allows one to answer * q1 * for polynominos in the square lattice @xmath66 . roughly speaking , their criterion says that if @xmath25 can be surrounded by translates of itself such that there are no gaps or holes , then @xmath25 is exact ; see @xcite for details . in particular , it immediately follows that each prototile shown in figure [ fig : neighborhoods ] is exact . algorithmic criteria for deciding the question * q1 * are particularly interesting . for polyominoes in the square lattice @xmath66 , one can decide this question in time polynomial in the length of the boundary of the polyomino ( described by a word over the alphabet @xmath67 , which is short for up , down , left , and right ) , as wijshoff and van leeuwen have shown @xcite . the characterization of exactness of a polyomino by beauquier and nivat @xcite mentioned above leads to an @xmath68 algorithm , where @xmath69 is the length of the word describing the boundary . recently , gambini and vuillon @xcite derived an improved @xmath70 algorithm for this problem . less is known for arbitrary ( not necessarily connected ) prototiles in a general lattice . szegedy @xcite derived an algorithm to decide whether a prototile @xmath25 in a lattice @xmath11 is exact assuming that the cardinality of @xmath25 is a prime or is equal to 4 . we have seen that the conditions for tiling a lattice with a single prototile are somewhat restrictive . for example , we might want to allow different rotated versions of the tile if the radiation pattern of the antenna used by a sensor is asymmetrical . we might want to consider different tiles corresponding to various different signal strength settings . furthermore , we might want to allow sensors with various different styles of antenna . we can accommodate all these different situations by allowing translates of several prototiles instead of just a single one . in this section , we show that one can still obtain an optimal periodic schedule which guarantees that the schedule is collision - free , as long as sensors of the same type and setting are deployed within each tile and a constraint on the tiles is satisfied . let @xmath11 be a lattice in @xmath12 . let @xmath71 be prototiles in the lattice @xmath11 , that is , @xmath72 is a subset of @xmath11 that contains @xmath17 for @xmath73 . let @xmath74 be pairwise disjoint nonempty subsets of @xmath11 . we say that @xmath75 provide a tiling of @xmath11 with prototiles @xmath76 if and only if the following two conditions are satisfied : @xmath77 for all @xmath78 , we have @xmath79 for all @xmath80 in @xmath81 and @xmath82 in @xmath83 such that @xmath84 . condition * gt1 * ensures that the lattice @xmath11 is covered by translates of the prototiles @xmath71 . condition * gt2 * ensures that two distinct tiles will not overlap . the set @xmath81 contains all vectors that are used to translate the tile @xmath72 , that is , the set @xmath85 contains all shifted versions of @xmath72 that occur in the tiling of @xmath11 . since the sets @xmath75 are pairwise disjoint , it is clear that @xmath86 whenever @xmath87 . condition * gt2 * requires further that the translates of the prototile @xmath72 with elements in @xmath81 do not overlap . we will call a tiling of @xmath11 _ respectable _ if and only if the prototile @xmath88 contains all other prototiles @xmath72 , that is , @xmath89 for @xmath90 . if this is the case , then we call @xmath88 the _ respectable prototile_. suppose that we are given a tiling @xmath74 of @xmath11 respectively with neighborhoods of the form @xmath76 . we will assume that the sensors are deployed in the following fashion : a sensor at location @xmath80 in the neighborhood @xmath91 of an element @xmath92 in @xmath81 affects precisely the neighbors @xmath93 by interference , where @xmath4 is in the range @xmath73 . loosely speaking , condition * d1 * says that all elements in the neighborhood @xmath91 have neighborhood type @xmath72 . [ th : thm2 ] let @xmath74 be a respectable tiling of a euclidean lattice @xmath11 with neighborhoods of the type @xmath71 . suppose that the sensors are deployed according to the scheme * d1. then there exists a deterministic periodic schedule that avoids collision problems using @xmath94 time slots . the schedule is optimal in the sense that one can not achieve this property with fewer than @xmath6 time slots . the periodic schedule is specified as follows . let @xmath95 . for all @xmath96 in the range @xmath97 , we schedule the elements @xmath98 at time @xmath5 if and only if @xmath99 is contained in the neighborhood @xmath100 . notice that all elements in @xmath11 will be scheduled at some point in time by property gt1. furthermore , condition gt2 * ensures that an element in @xmath11 is not scheduled more than once within @xmath6 consecutive time steps . we claim that this schedule is collision - free . seeking a contradiction , we assume that two distinct elements in @xmath11 are scheduled at the same time , but yield a collision problem . in other words , there must exist integers @xmath4 and @xmath96 in the range @xmath101 , an element @xmath102 such that @xmath69 is contained in both @xmath72 and @xmath100 , and elements @xmath103 and @xmath104 with @xmath84 such that @xmath105 . this implies that @xmath106 for @xmath84 , contradicting property * gt2. therefore , our schedule is collision - free . without loss of generality , we may assume that the point @xmath17 in @xmath11 has a respectable neighborhood @xmath88 ( otherwise , simply shift the tiling such that this condition is satisfied ) . seeking a contradiction , we assume that there exists a deterministic periodic schedule with @xmath45 time slots that is collision - free . it follows that there must exist two distinct elements @xmath47 and @xmath48 in @xmath88 that are scheduled at the same time . however , this would imply that the element @xmath49 is contained in both @xmath107 and @xmath108 ; thus , @xmath109 , contradicting the fact that the schedule with @xmath52 time slots is collision - free . the previous theorem is a natural generalization of theorem [ th : thm1 ] . a salient feature of theorems [ th : thm1 ] and [ th : thm2 ] is that the optimal schedule is independent of the nature of the tiling of @xmath11 . notice that one can obtain a collision - free periodic schedule even when there does not exist a respectable prototile . in fact , the respectable prototile was only used in the last part of the proof of theorem [ th : thm2 ] to establish the optimality of the schedule . therefore , one might wonder what will happen in the non - respectable case . let us agree on some ground rules . we would like to maintain the fact that for each translated version of a prototile the schedule is the same , as this simplifies configuring the sensor network . however , in the non - respectable case we might have different prototiles of the same size , so we allow that the schedules in the different prototiles can be independently chosen , as long as this does not lead to collision problems . figure [ fig : non - respectable ] shows that the number of time steps in an optimal schedule depends on the chosen tiling when the tiling is non - respectable . ( 50,50 ) z3 = t(1,0 ) ; z4 = t(0,1 ) ; picture ltile , rtile ; vardef pnt(expr a , b ) = ( az3+bz4 ) enddef ; pickup pencircle scaled 1 ; path p ; p = ( ( 0,1)(0,3)(1,3)(1,2)(2,2)(2,0)(1,0)(1,1)(0,1)cycle ) ; draw p scaled t shifted ( 0.5z3 + 0.5z4 ) withcolor 0.8white ; label(btex @xmath53 etex , pnt(1,2 ) ) ; label(btex @xmath54 etex , pnt(1,3 ) ) ; label(btex @xmath57 etex , pnt(2,2 ) ) ; label(btex @xmath58 etex , pnt(2,1 ) ) ; ltile : = currentpicture ; currentpicture : = nullpicture ; path q ; q = p reflectedabout ( ( 2,0),(2,1 ) ) ; draw q scaled t shifted ( -1.5z3 + 0.5z4 ) withcolor 0.8white ; label(btex @xmath56 etex , pnt(1,1 ) ) ; label(btex @xmath55 etex , pnt(1,2 ) ) ; label(btex @xmath53 etex , pnt(2,2 ) ) ; label(btex @xmath54 etex , pnt(2,3 ) ) ; rtile : = currentpicture ; currentpicture : = nullpicture ; draw ltile shifted pnt(0,5 ) ; draw ltile shifted pnt(1,2 ) ; draw ltile shifted pnt(3,0 ) ; draw ltile shifted pnt(4,1 ) ; draw ltile shifted pnt(5,2 ) ; draw ltile shifted pnt(4,5 ) ; draw ltile shifted pnt(2,7 ) ; draw ltile shifted pnt(1,6 ) ; draw rtile shifted pnt(3,3 ) ; draw rtile shifted pnt(2,4 ) ; @xmath21 ( 50,50 ) z3 = t(1,0 ) ; z4 = t(0,1 ) ; picture ltile , rtile ; vardef pnt(expr a , b ) = ( az3+bz4 ) enddef ; pickup pencircle scaled 1 ; path p ; p = ( ( 0,1)(0,3)(1,3)(1,2)(2,2)(2,0)(1,0)(1,1)(0,1)cycle ) ; draw p scaled t shifted ( 0.5z3 + 0.5z4 ) withcolor 0.8white ; label(btex @xmath53 etex , pnt(1,2 ) ) ; label(btex @xmath54 etex , pnt(1,3 ) ) ; label(btex @xmath55 etex , pnt(2,2 ) ) ; label(btex @xmath56 etex , pnt(2,1 ) ) ; ltile : = currentpicture ; currentpicture : = nullpicture ; path q ; q = p reflectedabout ( ( 2,0),(2,1 ) ) ; draw q scaled t shifted ( -1.5z3 + 0.5*z4 ) withcolor 0.8white ; label(btex @xmath56 etex , pnt(1,1 ) ) ; label(btex @xmath55 etex , pnt(1,2 ) ) ; label(btex @xmath53 etex , pnt(2,2 ) ) ; label(btex @xmath54 etex , pnt(2,3 ) ) ; rtile : = currentpicture ; currentpicture : = nullpicture ; for j = 0 step 2 until 4 : for i = 0 step 2 until 2 : draw rtile shifted pnt(i , j ) ; endfor ; endfor ; for j = 0 step 2 until 4 : for i = 4 step 2 until 6 : draw ltile shifted pnt(i , j ) ; endfor ; endfor ; draw pnt(4.5,0)pnt(4.5,9 ) dashed evenly ; we have introduced a deterministic periodic schedule for sensors using wireless communication that are placed on the points of a lattice . we have shown that the schedule is optimal assuming that there exists a respectable prototile . a natural question is whether the schedule remains optimal if one restricts the schedule from the lattice @xmath11 to a finite subset @xmath110 of @xmath11 . this question has an affirmative answer if @xmath110 contains a translate of the set @xmath111 , as the latter set consists of the respectable prototile @xmath88 and its neighbors , in which case our optimality proof carries over without change . another natural question is whether one can extend the method to the case of mobile sensors . this question has an affirmative answer . indeed , one straightforward way is to use our schedule to assign time slots to the locations rather than to the sensors . let us assume that the lattice points are spaced fine enough to ensure that only one sensor is within a voronoi region of a lattice point . if the time slot @xmath4 is assigned to a lattice point @xmath112 , then a sensor @xmath29 within the open voronoi region about @xmath112 can send at time @xmath3 if and only if @xmath5 and the interference range of @xmath29 fits within the tile of @xmath112 . clearly , this yields a collision - free schedule for mobile sensors . however , it should be stressed that there are many other solutions possible , but a comparison of such methods is beyond the scope of this paper .	suppose that wirelessly communicating sensors are placed in a regular fashion on the points of a lattice . common communication protocols allow the sensors to broadcast messages at arbitrary times , which can lead to problems should two sensors broadcast at the same time . it is shown that one can exploit a tiling of the lattice to derive a deterministic periodic schedule for the broadcast communication of sensors that is guaranteed to be collision - free . the proposed schedule is shown to be optimal in the number of time slots . * keywords : * distributed computing , scheduling sensors , lattice tiling , wireless communication .
the appearance of initial singularities in the classical cosmological models which better describe the universe we live in constitutes a big puzzle to all cosmologists . until now , singularities are out of the scope of any physical theory . if we assume that a physical theory can describe the whole universe at every instant , even at its possible moment of creation ( which is the best attitude because it is the only way to seek the limits of physical science ) , then these classical singular points must be avoided . indeed , no one expects that classical general relativity continues to be valid under extreme situations of very high energy density and curvature . in particular , it is very plausible that quantum gravitational effects become important under these conditions , eliminating the singularities that appear classically . to see if this is indeed the case , we should construct a theory of quantum cosmology . however , any quantum theory when applied to cosmology presents new profound conceptual problems . how can we apply the standard probabilistic copenhaguen interpretation to a single system as the universe ? where in a quantum universe can we find a classical domain where we could construct our classical measuring apparatus to test and give sense to the quantum theory ? who are the observers of the whole universe ? this is not a problem of quantum gravity alone , because there is no problem with the concept of an ensemble of black holes and a classical domain outside it , but it is specific of quantum cosmology . as we can not apply the copenhaguen interpretation to quantum cosmology , we will adopt an alternative non - probabilistic interpretation , which circumvents the measurement problem because it is an ontological interpretation of quantum mechanics : it is not necessary to have a measuring apparatus or a classical domain in order to recover physical reality . it is the causal interpretation of quantum mechanics @xcite . in this letter , we apply the causal interpretation of quantum mechanics to some specific quantum states already found in the literature @xcite , and show that the bohmian ( quantum ) trajectories are non - singular , tending to the classical trajectories when the scale factor is large . the quantum potential is the responsible for this behavior . when the scale factor is small , the quantum potential becomes large creating an effective repulsive force avoiding the singularity . when the scale factor is large , the quantum potential becomes negligible and the classical potential dominates . the minisuperspace models we study are constituted of a friedman - robertson - walker ( frw ) metric with either dust or radiation perfect fluids @xcite . the adm quantization procedure is performed because in these cases there is a preferable time variable which rends the quantum equations into a very simple shroedinger form . the quantum solutions obtained are gaussians @xcite . the dust case is rather academic but the radiation case is important because it can mimic quite well the matter content at the very early universe . in all cases we obtain non - singular eternal models without any particle horizon . this letter is organized as follows . in the next section we make a summary of the causal interpretation and its application to quantum cosmology . in section 3 we abridge the results of refs . @xcite which concern this letter . in section 4 we apply the causal interpretation to the solutions presented in section 3 , obtaining the bohmian trajectories . we compare our results with those of refs . we end with comments and conclusions . in this section , we will review the causal interpretation of quantum mechanics . let us begin with the schrdinger equation , in the coordinate representation , for a non - relativistic particle with the hamiltonian @xmath0 : @xmath1 \psi ( x , t ) .\ ] ] writing @xmath2 , and substituting it into ( [ bsc ] ) , we obtain the following equations : @xmath3 @xmath4 the usual probabilistic interpretation takes equation ( [ bpr ] ) and understands it as a continuity equation for the probability density @xmath5 for finding the particle at position @xmath6 and time @xmath7 if a measurement is performed . all physical information about the system is contained in @xmath5 , and the total phase @xmath8 of the wave function is completely irrelevant . in this interpretation , nothing is said about @xmath8 and its evolution equation ( [ bqp ] ) . however , examining equation ( [ bpr ] ) , we can see that @xmath9 may be interpreted as a velocity field , suggesting the identification @xmath10 . hence , we can look to equation ( [ bqp ] ) as a hamilton - jacobi equation for the particle with the extra potential term @xmath11 . after this preliminary , let us introduce the causal interpretation of quantum mechanics , which is based on the _ two _ equations ( [ bqp ] ) and ( [ bpr ] ) , and not only in the last one as is the copenhaguen interpretation : \i ) a quantum system is composed of a particle _ and _ a field @xmath12 ( obeying the schrdinger equation ( [ bsc ] ) ) , each one having its own physical reality . \ii ) the quantum particles follow trajectories @xmath13 , _ independent on observations_. hence , in this interpretation , we can talk about trajectories of quantum particles , contrary to the copenhaguen interpretation , where only positions at one instant of time have a physical meaning . \iii ) the momentum of the particle is @xmath10 . \iv ) for a statistical ensemble of particles in the same quantum field @xmath12 the probability density is @xmath14 . equation ( [ bpr ] ) guarantees the conservation of @xmath15 . let us make some comments : \a ) equation ( [ bqp ] ) can now be interpreted as a hamilton - jacobi type equation for a particle submited to an external potential which is the classical potential plus a new quantum potential @xmath16 hence , the particle trajectory @xmath13 satisfies the equation of motion @xmath17 \b ) even in the regions where @xmath12 is very small , the quantum potential can be very high , as we can see from equation ( [ qp ] ) . it depends only on the form of @xmath12 , not on its absolute value . this fact brings home the non - local and contextual character of the quantum potential . this is very important because bell s inequalities together with aspect s experiments show that , in general , a quantum theory must be either non - local or non - ontological . as bohm s interpretation is ontological , it must be non - local , as it is . the quantum potential is responsible for the quantum effects . \c ) this interpretation can be applied to a single particle . in this case , equation ( [ bpr ] ) is just an equation to determine the function @xmath18 , which forms the quantum potential acting on the particle via equation ( [ beqm ] ) . it is not necessary to interpret @xmath5 as a probability density and hence it may not be normalizable . the interpretation of @xmath5 as a probability density is appropriate only in the case mentioned in item ( iv ) above . the causal interpretation is not , in essence , a probabilistic interpretation . \d ) the classical limit is very simple : we only have to find the conditions for having @xmath19 . \e ) there is no need to have a classical domain because this interpretation is ontological . the question on why in a real measurement we do not see superpositions of the pointer apparatus is answered by noting that , in a measurement , the wave function is a superposition of non - overlaping wave functions . the particle will enter in one region , and it will be influenced by the unique quantum potential obtained from the sole non - zero wave function defined on this region . the particle can not jump to other branchs because it can not pass through nodal points of the wave function . in the next section we will perform an adm quantization of friedman - robertson - walker minisuperspace models containing either dust or radiation perfect fluids . in these cases , a prefered time variable can be selected in terms of potentials of the velocity field of the fluids . as a time variable will be chosen before quantization , the quantum equations will be exactly like the schrdinger equation . instead of the particle position , as in the above example , the single degree of freedom of these quantum models will be the scale factor of the universe . the interpretation of the quantum solutions will run analogously to what was described above . the time evolution of the scale factor will be different from the classical one due to the presence of an extra quantum potential term in the modified hamilton - jacobi equation it satisfies . in this section we present the minisuperspace models which we will analyze using the causal interpretation . these models are obtained from the quantization of the dust and radiation fluids through the adm prescription . for the quantization of the radiation filled frw model we follow ref . @xcite . we start with the line element @xmath20 where @xmath21 is the metric of constant curvature three - surfaces . the full action for a perfect fluid is given by @xmath22 where @xmath23 is the pressure , @xmath24 is the scalar curvature , @xmath25 is the 3-metric on @xmath26 @xmath27 is the second fundamental form of the boundary , and we choose @xmath28 . the action ( [ actionnivaldo ] ) can be reduced to @xmath29 , \ ] ] where we are using schutz s fluid variables @xmath30 @xmath31 @xmath32 @xmath33 @xmath8 @xcite , in terms of which the four - velocity of the fluid is written as @xmath34 the quantity @xmath35 is the specific enthalpy . the super - hamiltonian @xmath36 takes the form @xmath37 for details , see refs . @xcite . performing an adm reduction using the conformal - time gauge @xmath38 in the radiation case , the reduced action becomes @xcite @xmath39 , \ ] ] where @xmath40 @xmath41 or @xmath42 , for spherical , flat , or hyperbolic spacelike sections of the three space with metric @xmath21 , respectively . the hamiltonian in the reduced phase space , which has only @xmath43 as a degree of freedom , takes the very simple form @xmath44 the classical solutions are : @xmath45 which are the well known solutions for radiation in conformal time . the quantized hamiltonian , in units where @xmath46 is @xmath47 as @xmath48 , the requirement that @xmath49 must be self - adjoint leads to the restriction @xmath50 on the wave function @xmath51 where @xmath52 is a parameter in the interval @xmath53.$ ] two solutions will be obtained , one for @xmath54 and the other for @xmath55 the propagators for the hamiltonian ( [ ho ] ) are well known . however , we have the extra constraint that @xmath56 for this reason , the hilbert space is restricted to functions in @xmath57 for the case where @xmath58 the propagator is @xmath59 where @xmath60 is the usual harmonic oscillator propagator for a system with mass @xmath61 and angular frequency @xmath62 seting @xmath63 in eq . ( [ boundary ] ) , we take a gaussian wavepacket as the initial normalized state , @xmath64 where @xmath65 @xmath66 and @xmath67 are real , and @xmath68 the wave function at time @xmath7 given by eqs . ( [ propagatora ] ) and ( [ gaussian0 ] ) is @xmath69}\right ] ^{1/2 } \label{psia } \\ & & \times \exp \left\ { \frac{6i\sqrt{k}}{\tan ( \sqrt{k}t)}\left ( 1+\frac{6i% \sqrt{k}}{\cos ^{2}(\sqrt{k}t)[\beta \tan ( \sqrt{k}t)-6i\sqrt{k}]}\right ) a^{2}\right\ } . \nonumber\end{aligned}\ ] ] the expectation value for the scale factor @xmath43 can be computed from ( [ psia ] ) , and is @xmath70 for @xmath71 the propagator is @xmath72 we take as the initial state the wave packet @xmath73 the evolution of ( [ gaussiannot ] ) governed by the propagator ( [ propagatorb ] ) is @xmath74 \left [ \beta -\frac{6i\sqrt{k}}{% \tan ( \sqrt{k}t)}\right ] ^{-3/2 } \label{psib } \\ & & \times r\exp \left\ { \frac{6i\sqrt{k}}{\tan ( \sqrt{k}t)}\left ( 1+\frac{6i% \sqrt{k}}{\cos ^{2}(\sqrt{k}t)[\beta \tan ( \sqrt{k}t)-6i\sqrt{k}]}\right ) r^{2}\right\ } . \nonumber\end{aligned}\ ] ] the expectation value for the scale factor @xmath43 with the wavefunction ( [ psib ] ) is @xmath75 we now turn our attention to the dust filled minisuperspace model presented by gotay and demaret @xcite . once again schutz s variables and a frw metric are used . in this case the field @xmath76 in eq . ( [ veloc ] ) is the only velocity potential which is non null . the time to be chosen will be the proper time of the dust particles defined by @xmath77 , which is equivalent to choose @xmath78 . seting @xmath79 and @xmath80 , the adm reduced super - hamiltonian becomes ( see ref . @xcite ) @xmath81 the reduced hamiltonian , with the above choice of time , is @xmath82 if we perform the canonical transformation @xmath83 the hamiltonian takes the simple form @xmath84 where @xmath85 taking @xmath86 , the classical solution is simple to obtain . it is @xmath87 or equivalently @xmath88 . the quantized hamiltonian for @xmath86 is , in units where @xmath46 @xmath89 once again , the requirement of self - adjointness of @xmath49 yields the boundary conditions @xmath90 . choosing @xmath91 , we can evolve the initial gaussian wave function @xmath92 ^{1/4}\exp ( -\beta x^{2 } ) , \label{gauss2}\ ] ] by applying the corresponding propagator , analogously to what was done in the radiation case . as before , @xmath93 with @xmath94 . the wave function at time @xmath7 is given by : @xmath95 ^{1/4}(1 + 4i\beta t)^{-1/2}\exp \left [ \frac{-\beta x^{2}}{1 + 4i\beta t}\right ] . \label{psic}\ ] ] the expectation value for @xmath6 can easily be computed , and is @xmath96 ^{1/2}. \label{expectaionc}\ ] ] we will now turn our attention to the causal interpretation of the wavefunctions @xmath97 @xmath98 , and @xmath99 the causal interpretation of the above minisuperspace models is straightforward . let us begin by the dust field . in this case , the quantum trajectories are solutions of the following differential equation : @xmath100 where @xmath101 is the phase of the wave function ( [ psic ] ) . the general solution is : @xmath102^{1/2 } , \label{sd0}\ ] ] where @xmath103 is an arbitrary positive integration constant . that the mean value of @xmath6 founded in the previous section is , apart from the integration constant , the same function of time as the solution ( [ sd ] ) is not surprising . mean values in the causal interpretation are the same as in the usual interpretation if we also assume that the amplitude squared of the wave function , @xmath104 , represents a probability density distribution of initial conditions of the quantum trajectories . then , we obtain from eq . ( [ gauss2 ] ) @xmath105^{1/2}dx_0 \nonumber \label{sd } \\ & = & { ( 2\pi b)}^{-1/2}[16b^{2}t^{2}+(1 - 4bt)^{2}]^{1/2 } \ ; , \end{aligned}\ ] ] which is the result ( [ expectaionc ] ) of the previous section . it can be seen from eq . ( [ sd0 ] ) that no quantum trajectory is singular . the scale factor is never zero for any @xmath103 greater than zero . also , the trajectories approach the classical one for large @xmath106 ( remember that @xmath107 ) . this behavior can be seen by examining the quantum potential , @xmath108}{[(1 - 4bt)^{2}+16b^{2}t^{2}]^2}% . \label{qpd}\ ] ] along the quantum trajectory ( [ sd0 ] ) , the quantum potential turns out to be : @xmath109 where @xmath13 is given by ( [ sd0 ] ) . from the above equation we can see that the quantum potential goes to zero when @xmath106 becomes large but it is positive of order @xmath67 when @xmath106 is small . the quantum potential is zero along the trajectory but the quantum force @xmath110 is not . ] . in this situation , it works like a repulsive force around the region @xmath111 . figure 1 shows the curves @xmath13 and @xmath112 ( along the trajectory ) versus @xmath7 . they represent classical universes contracting from infinity to a minimum size , where its behavior is not classical , and then expanding to infinity , getting classical again as the scale factor becomes large . these models have no particle horizon . the integral @xmath113 diverges . let us now turn our attention to the more realistic case of radiation . again , the application of the causal interpretation is straightforward . we have to take the phases of the wave functions ( [ psia],[psib ] ) for the cases @xmath114 , calculate their derivatives with respect to @xmath43 , and equate the result with @xmath115 . the solutions of these first order differential equations are : @xmath116 where @xmath117 is a positive integration constant . like before , they have the same functional behaviour as the mean values encountered in the previous section for the reasons already explained in the dust case . none of these solutions reach the singular point @xmath118 .the quantum potentials are given by @xmath119}{[(6-b\tan ( t))^2\cos ^2(t)+b^2\sin ^2(t)]^2 } & \quad \mbox{for } k=1 \\ \frac{[b(bt^2 - 72a^2)+(6-bt)^2]}{[(6-bt)^2+b^2t^2]^2 } & \quad \mbox{for } k=0 \\ \frac{[b(b\sinh ^2(t)-72a^2)+(6-b\tanh ( t))^2 \cosh ^2(t ) ] } { [ ( 6-b\tanh ( t))^2\cosh ^2(t)+b^2\sinh ^2(t)]^2 } & \quad \mbox{for } k=-1 \end{array } .\right.\ ] ] at the trajectories , the quantum potentials for @xmath120 are the same , and equal to @xmath121 where @xmath122 is given by ( [ abohm ] ) . for @xmath86 and @xmath123 the models are qualitatively similar to the dust case . it can be seen from eq . ( [ abohm ] ) that these universes contract classically from infinity to a minimum size , where their behaviors are not classical , and then expand to infinity , getting classical again as the scale factor becomes large . the case @xmath124 deserves special attention . examining the case with @xmath125 and @xmath126 ( the general case is qualitatively similar ) , we can see that the quantum trajectory oscillates between the minimum value @xmath127 and the maximum value @xmath128 . their ratio is @xmath129 . the ratio between the quantum and classical forces at these points are @xmath130 for the minimum and @xmath131 for the maximum . the universe we live in is large and classical , and it must have gone through a contracted phase where nucleosynthesis took place . this means that the ratio between the maximum and minimum values of the radius of the universe ( which is the inverse ratio of the respective temperatures at these epochs since the fluid is radiation ) must be at least of the order of @xmath132 . hence @xmath67 must be greater than @xmath133 , yielding a very flat initial gaussian wave function . this ensures that the universe had a nucleosynthesis era with a classical behaviour since then . these models do not have particle horizons . in this case the particle horizon is proportional to the integral @xmath134 which in the gauge @xmath135 we are using evidently diverges . figures 2 and 3 show the scale factor and the quantum potential for @xmath86 and @xmath123 . when @xmath43 is large , the quantum potentials go to zero , while when @xmath43 approaches its minimum size , they become important , creating an effective repulsive quantum force around this region . figures 4 and 5 show the scale factor and the quantum potential for @xmath124 and for different values of @xmath67 . note that for larger @xmath67 s the amplitude of oscillation also becomes larger . in this letter we have obtained a class of non - singular cosmological models without particle horizons by applying the causal interpretation to some quantum states of the universe already obtained in the literature @xcite and presented in section 3 . we have shown that the bohmian trajectories are the same functions of time as their corresponding mean - values obtained in refs . @xcite , which makes use of the conventional copenhaguen interpretation . however , it seems to us that only in the causal approach we can arrive at definite conclusions about the existence of singularities . it is not because the mean value of a variable is different from zero at all times that this variable can not be zero sometime . nothing forbids that this be the case for the single universe we live in . the causal interpretation applied to this problem states that each individual trajectory of these quantum states is not singular , a much more stronger and valuable result . this means that if the universe is in one of these quantum states , then it is indeed non - singular because there is no quantum trajectory with singularities . for the more realistic radiation fluid , we have obtained singularity - free models , without particle horizons , which approach the classical behavior as the universe expands . for flat and negative curvature three - surfaces this is the situation in all cases . the universe contracts classically from infinity to a minimum size , where its behavior is not classical , and then expands to infinity , getting classical again as the scale factor becomes large . for the positive curvature case , the classical behavior is achieved only for some values of the parameters . in this case , we would have an eternal periodic universe which is classical when it is large . these examples show that quantum gravitational effects can indeed prevent the formation of cosmological singularities but this result can only be stated with strength along the lines of the causal interpretation . it should be interesting to investigate if the results presented here are stable under small perturbations . then we will have to face new technical and interpretational problems . but this is another story . part of this work was done while npn was a previ ( special program for visiting professor / researcher ) fellow at the physics department of the federal university at juiz de fora . we would like to thank the group of the pequeno seminrio " at cbpf for useful discussions and fapemig for financial support . npn would like to thank cnpq for financial support and the federal university at juiz de fora , ufjf , for hospitality . jab would like to thank the laboratory for cosmology and experimental high energy physics ( lafex ) at the brazilian center for physical research ( cbpf / cnpq ) for hospitality . masl is supported by a scholarship from capes .	we apply the causal interpretation of quantum mechanics to homogeneous and isotropic quantum cosmology where the sources of the gravitational field are either dust or radiation perfect fluids . we find non - singular quantum trajectories which tends to the classical one when the scale factor becomes much larger then the planck length . in this situation , the quantum potential becomes negligible . there are no particle horizons . as radiation is a good approximation for the matter content of the early universe , this result suggests that the universe can be eternal due to quantum effects . pacs number(s ) : 98.80.h , 03.65.bz 6000 6000 40by -1 to epsf # 1 # 1#2#3 = # 1 in to # 1#2#3 # 3
integral field spectroscopy ( ifs ) provides a spectrum simultaneously for each spatial sample of an extended , two - dimensional field . basically , an ifs is located in the focal plane of a telescope and is composed by an integral field unit ( ifu ) and a spectrograph . the ifu acts as a coupler between the telescope and the spectrograph by reformatting optically a rectangular field into a quasi - continuous pseudo - slit located at the entrance focal plane of the spectrograph . therefore , the light from each pseudo - slit is dispersed to form spectra on the detector and a spectrum can be obtained simultaneously for each spatial sample within the ifu field . the ifu contains two main optical sub - systems : the fore - optics and the image slicer . the fore - optics introduces an anamorphic magnification of the field with an aspect ratio of 1@xmath02 onto the set of slicer mirrors optical surfaces . in such way each spatial element of resolution forms a 1@xmath02 pixels image on the detector ( i.e. the width of each slice corresponds to 2 pixels ) , which ensures correct sampling in the dispersion direction ( perpendicular to the slices ) and prevents under - sampling the spectra . this anamorphism can be avoided if under - sampling spectra is acceptable by science ( for example , the snap @xcite project ) or if a spectral dithering mechanism is included in the spectrograph in order to recover for the under - sampling . the image slicer optically divides the anamorphic ( or not ) two - dimensional field into a large number of contiguous narrow sub - images which are re - arranged along a one - dimensional slit at the entrance focal plane of the spectrograph . an image slicer is usually composed of a slicer mirror array located at the image plane of the telescope and associated with a row of pupil mirrors and a row of slit mirrors . the slicer mirror array is constituted of a stack of several thin spherical mirrors ( called `` slices '' ) which `` slice '' the anamorphic field and form real images of the telescope pupil on the pupil mirrors . the pupil mirrors are disposed along a row parallel to the spatial direction . each pupil mirror then re - images its corresponding slice of the anamorphic field on its corresponding slit mirror located at the spectrograph s focal plane ( slit plane ) . the slit mirrors are also diposed along a row parallel to the spatial direction . finally , each slit mirror which acts as a field lenses , re - images the telescope pupil ( pupil mirrors ) onto the entrance pupil of the spectrograph . the principle of an image slicer is presented in fig . [ fig : principle ] . [ cols="^ " , ] as an example in fig . [ fig : wfe ] , we consider the distribution of the theoretical wfe ( t - wfe ) of both fan - shaped and classical image slicers . the t - wfe of the fan - shaped image slicer is about 45.5@xmath1 nm and the t - wfe of the classical image slicer is about 76.1@xmath2 nm . it is interesting to note that the fore - optics ( see section [ sect : design ] ) provides a mean t - wfe of about 45.7 nm over the whole fov ( marked by a dashed line in fig . [ fig : wfe ] ) . thus , the fan - shaped image slicer preserves the image quality of the fore - optics over the whole fov ( all channels ) while the classical image slicer degrades the t - wfe by a factor two in certain channels . furthermore , the distribution of the t - wfe of the fan - shaped image slicer is sharp since the range of values is about 8 nm over all channels allowing to guaranty a negligible level of differential aberration in the field . this article presents an original concept of image slicer called `` fan - shaped '' . its design delivers good and homogeneous image quality over all ifu elements . we successfully apply its design to jwst / nirspec . here we did nt discuss about manufacturing aspects since the performance aspects were preponderant however further investigations are under studying to drastically reduce costs and manufacturing aspects in such a design by preserving performances . furthermore , a prototyping of the ifs ( ifu and spectrograph ) for the snap application is undergoing at lam @xcite .	integral field spectroscopy ( ifs ) provides a spectrum simultaneously for each spatial sample of an extended , two - dimensional field . it consists of an integral field unit ( ifu ) which slices and re - arranges the initial field along the entrance slit of a spectrograph . this article presents an original design of ifu based on the advanced image slicer concept @xcite . to reduce optical aberrations , pupil and slit mirrors are disposed in a fan - shaped configuration that means that angles between incident and reflected beams on each elements are minimized . the fan - shaped image slicer improves image quality in terms of wavefront error by a factor 2 comparing with classical image slicer and , furthermore it guaranties a negligible level of differential aberration in the field . as an exemple , we are presenting the design lam used for its proposal at the nirspec / ifu invitation of tender .
understanding the frequency and mass function of free - floating planets ( ffps ) is integral for a complete comprehension of the formation and evolution of planetary systems . using deep photometric imaging primarily in near - infrared bands and , in some cases , astrometric and/or spectroscopic follow - up data , jupiter - mass ffps and ffp candidates have been detected in the trapezium cluster @xcite , the @xmath10 orionis open cluster @xcite , the ab doradus moving group @xcite , the @xmath11 pictoris moving group @xcite , the upper scorpius association @xcite , and in the field @xcite . yet detections from photometric surveys require and depend sensitively on an independent constraint for the object s age while the dynamical association of objects in moving groups is intrinsically more uncertain . in a broader statistical sense , @xcite examined two years of microlensing survey data from the microlensing observations in astrophysics ( moa ) collaboration . they found an excess of short - timescale events ( @xmath122 days ) above expectations based on an extrapolation of the stellar mass function down to low - mass brown dwarfs . from this they inferred a population of ffp candidates that outnumber main sequence stars by a ratio of @xmath13 . however , mass measurements do not exist for any of the candidate objects upon which the inference rests . the formation mechanisms for ffps remain an open theoretical question . one possible avenue is that these objects were originally formed in protoplanetary disks and were subsequently ejected . simulations by @xcite of the formation and evolution of planetary systems without eccentricity or inclination damping eject planets at a rate that is @xmath350 times lower than is needed to explain the moa result . indeed , @xcite assert that planet - planet scattering itself is insufficient to reproduce the abundance of ffp candidates seen by moa . in the context of open clusters , n - body dynamical simulations find that roughly half of star - planet and planet - planet interactions occur within the first 30 myr @xcite and that 80@xmath14 of the resulting ffps are ejected from the cluster while the remaining @xmath620@xmath14 become concentrated in the central @xmath62 pc @xcite . a second option is that ffps form via direct collapse of molecular clouds . @xcite found that opacity - limited fragmentation of collapsing clouds could , in principle , produce a fragment with a minimum mass as low as @xmath60.01@xmath15 . additionally , turbulent shocks could cause protostellar cores with masses in the brown dwarf regime to become gravitationally unstable and hypothesized that turbulent density fluctuations smaller than a critical mass could induce collapse into objects with the mass of giant planets @xcite . gravitational microlensing does not rely on the flux output from the lensing object and so is well - positioned to explore ffp demographics across a wide range of planet masses . the mass of an ffp lens , @xmath16 , is given by : @xmath17 where @xmath0 is the angular einstein radius , @xmath18 , and @xmath2 is the microlens parallax : @xmath19 here @xmath20 is the lens - source relative parallax , defined as @xmath21 , where @xmath22 and @xmath23 are the distances from the observer to the lens and source , respectively , and @xmath24 is the physical size of @xmath0 projected onto the plane of the observer . the most salient observable for a light curve arising from a single lensing mass is the einstein timescale @xmath25 , which measures the time required for the lens - source angular separation to change by one angular einstein ring radius and is defined as : @xmath26 where @xmath27 is the lens - source relative proper motion . the timescale is routinely measured for single - lens events . however , several physical properties are encoded within @xmath25 , including @xmath22 , @xmath16 , and @xmath27 . thus , in order to determine that the lensing object giving rise to the short timescale is indeed planetary - mass , it is crucial to measure both @xmath0 and @xmath2 . the angular einstein radius is typically determined by combining a measurement of @xmath1 , the angular radius of the source star normalized to @xmath0 , with multiband photometry to determine the color and , ultimately , the angular size of the source @xcite . lastly , a short - timescale event definitively due to a planetary - mass lens must be proven to be free - floating , which requires high - resolution photometry to search for flux from a possible lens host star . @xcite studied bound planets that are widely separated from their host star . they found that if the projected separation between a planet and its parent star is @xmath320 au , the resulting light curve can mimic that of an ffp , as the alignment between the source trajectory and the binary lens axis can cause the primary microlensing event due to the star _ not _ to be observed , allowing a bound planet to masquerade as an ffp . they also investigated the ability to identify the bound nature of a planet in the case of isolated planetary - mass lensing events via the caustic induced by the shear of the planet s host star , which would introduce magnification structure to the light curve of the event . fully characterizing an ffp requires that all of these criteria are met . however , the intersection of all aforementioned constraints is far from guaranteed for a given space mission . the goal of this paper is to map the regions of parameter space for which a given satellite will be sensitive to ffps , particularly _ k2 _ s campaign 9 ( _ k2_c9 ; see [ sec : kepler ] ) and _ wfirst _ ( see [ sec : wfirst ] ) , and to identify the dominant limiting factors in each regime . throughout the paper we will assume that the source is located in the bulge at a distance of @xmath28 kpc @xcite and focus on @xmath22 as an independent variable . it is possible to explore the demographics of ffps statistically without requiring that the parameters @xmath2 , @xmath1 , and @xmath4 each be measured for all systems . that said , the primary focus of this paper is on the subset of ffps that can be fully characterized via measurements of all three . the expected yield of ffps for a given microlensing survey depends , intrinsically and sensitively , on their underlying event rate . understanding the relative rates for different ffp populations can influence the observational strategy . the microlensing rate as a function of the ffp physical parameters , derived from the generic rate formula @xmath29 , is given by @xcite : @xmath30 here @xmath31 is the local space density of ffps , @xmath32 is the lens - source relative proper motion probability distribution , and @xmath33 is the ffp mass function . while @xmath34 is proportion to @xmath35 and @xmath36 can be assumed to be similar to that for the general galactic stellar population , @xmath31 and @xmath33 are precisely what the experiment is trying to measure and are hitherto unconstrained . furthermore , the kinematics for ffps will depend on the fraction that are the result of dynamical ejections versus those that form in situ . thus , in this paper we can not and do not predict the relative rates for different ffp populations but instead study the feasibility of characterizing ffps with various physical parameters , assuming that an event has occurred . our methodology builds upon and extends previous studies of lens characterization for ffps and also bound planets . @xcite examined the ability of a joint ground- and space - based microlensing survey to constrain the mass of ffps . considering a generic satellite at the earth - sun l2 point paired with a ground - based survey , they found that microlensing events arising from planetary - mass lenses can be detected for planet masses down to that of the earth . @xcite subsequently explored how to distinguish between isolated lensing events due to widely separated , bound planets from ffp events using interferometry to search for the astrometric signature of a lens host star . he found that for bright sources , with @xmath37 , a centroid shift can be measured given an astrometric precision that is of - order one microarcsecond . more generally , @xcite investigated how to measure lens masses and distances solely from measurements of the microlens parallax @xmath2 and the lens flux @xmath4 , circumventing the need to measure @xmath0 . she considered spacecraft in earth - trailing solar orbits and at earth - sun l2 , similar to those of _ kepler _ , _ spitzer _ , and _ wfirst_. naturally , lens flux can not be used as a tool for characterizing ffps , necessitating an additional constraint . @xcite advocate for a simultaneous ground - based survey to accompany the _ wfirst _ microlensing observations in order to obtain two - dimensional vector microlens parallax measurements . these would provide complete solutions , and thus masses and distances , for a substantial fraction of ffp events . we present the first combined treatment of constraints from @xmath2 , @xmath1 , and @xmath4 to explore regimes of detectability and characterization for ffps across a range of lens distances , source star properties , and satellite characteristics . in [ sec : satellites ] we briefly describe the parameters and goals of six relevant space missions . measuring @xmath2 with a space telescope requires a delicate balance between @xmath24 and the projected separation @xmath38 of the earth and the satellite as seen from the lens - source line - of - sight . we investigate the interplay between @xmath38 and @xmath24 for these satellite missions in [ sec : dproj_reproj ] . in [ sec : rho ] we discuss the probability of measuring @xmath1 through the detection of finite - source effects . finally , in [ sec : fl ] we detail the constraints from measuring the lens flux . we summarize our findings and discuss their applications in the context of _ k2_c9 and _ wfirst _ in [ sec : discussion ] , wherein we identify the regimes of parameter space to which each will ultimately be able to robustly characterize ffp events . in this section we provide an overview of all space telescopes that are potentially relevant for monitoring microlensing events simultaneously with ground - based resources . _ spitzer _ , a 0.85 m infrared telescope , was the first satellite used to conduct real - time monitoring of a microlensing event simultaneous with ground - based facilities @xcite . a 100-hour campaign in 2014 expanded upon this in an effort to make such measurements systematically for multiple events . the 2014 program led to the first satellite parallax measurement for an isolated star @xcite and a microlensing - discovered exoplanet @xcite . a larger 832-hour program followed in 2015 , for which target events were selected using objective criteria in order to maximize planet sensitivity and number of detections @xcite . the resulting discoveries include mass and distance measurements for a cold neptune in the galactic disk @xcite , a massive remnant in a stellar binary @xcite , and several isolated objects , including a brown dwarf @xcite , demonstrating the wide range of astrophysical populations that can be probed with such a technique . these campaigns are the first steps for measuring the galactic distribution of bound planets @xcite . two additional _ spitzer _ programs will take place in 2016 , one to explore the galactic distribution of exoplanets using high - magnification microlensing events @xcite and the other to conduct a two - satellite microlensing experiment @xcite by observing in conjunction with _ k2_c9 ( see [ sec : kepler ] ) . however , the short - timescale events that are indicative of ffps are generally inaccessible with _ spitzer _ due to the several - day lag between target selection and upload and the first possible observations ( see figure 1 of @xcite ) . nevertheless , we include it here because it is possible to have a combination of physical parameters @xmath22 , @xmath16 , and @xmath27 that produces a longer timescale for an ffp , and also for the sake of completeness . _ kepler _ is 0.95 m telescope in an earth - trailing solar orbit whose primary mission was to explore exoplanet demographics using the transit method . the mechanical failure of the second of its four reaction wheels in 2013 signaled an end to the primary mission but heralded the genesis of its extended _ k2 _ mission , which is in the midst of a series of @xmath680-day campaigns performing high - precision photometry for targets along the ecliptic @xcite . _ k2_c9 will conduct the first microlensing survey from the ground and from space , covering 3.74 deg@xmath39 from 7/april through 1/july of 2016 in concert with a vast array of ground - based resources @xcite . approximately @xmath3120 events will occur in the _ k2_c9 microlensing survey field ( termed superstamp " ) during the campaign dates , of - order 10 of which are estimated to be short - timescale ffp candidates . _ s campaign 11 will also point toward the galactic bulge from 24/september through 8/december of 2016 . while it will not perform an automated microlensing survey , it will yet be possible to measure @xmath2 for events detected by ground - based telescopes that will have timescales long enough that they will be ongoing when _ k2 _ observations take place , though this excludes short - timescale ffp candidates . _ wfirst _ , which was recently approved for phase a development , will be a 2.4 m telescope equipped with a wide - field imager with a field of view of 0.28 deg@xmath39 and six imaging filters that span 0.762.0 microns @xcite . set to launch in the mid-2020s , it will be inserted into a solar orbit at earth - sun l2 . _ wfirst _ will conduct a @xmath6432-day microlensing survey toward the galactic bulge divided equally between six 72-day seasons . these seasons will be split between the beginning and end of the mission to maximize the ability of _ wfirst _ to measure @xmath27 for the detected events . furthermore , its orbital placement and observational parameters will make it the most advanced experiment in terms of ffp detection and characterization . the scientific focus of _ euclid _ will be to explore the nature of dark energy . however , an exoplanetary microlensing survey could be conducted as part of the mission s legacy program and would , in principle , be sensitive to ffps @xcite . _ euclid _ , which is scheduled to launch in 2020 , will be placed in a halo orbit at earth - sun l2 . so , while its 1.2 m aperture is smaller than that of _ wfirst _ s , the orbital configuration and resulting geometric sensitivity will be quite similar to that of _ wfirst _ , which we derive here . as such , we do not explicitly consider _ euclid _ as a separate mission throughout the remainder of the text and instead note that _ wfirst _ is a good proxy for its capabilities . during the 2015 _ spitzer _ campaign , the microlensing event ogle-2015-blg-1395 ( ob151395 ) was observed with _ swift _ , a 0.3 m telescope that can observe from optical to @xmath40-ray , while it was highly magnified , with @xmath41 . ob151395 is thus the first event with observations from two satellites while the event was ongoing ( shvartzvald et al . , in prep . ) . while the rapid response of _ swift _ is ideal for short - timescale ffps , it is uncommon to have events reach a sufficiently bright magnitude to be detectable for the ultra violet / optical telescope . furthermore , _ swift _ is on a low - earth orbit , so its distance from the surface of the earth is within the range @xmath6560576 km , leading to an extremely small projected separation ( see [ sec : dproj ] ) that moreover varies significantly during its @xmath696-minute orbital period . that said , _ swift _ s orbit can be advantageous for events sensitive to terrestrial parallax . _ new horizons _ completed a six - month flyby of pluto in the summer of 2015 and is continuing into the outer reaches of the solar system to study the kuiper belt . its projected separation is , by a factor of several , the largest of the satellites we will discuss here . while this indicates that it will be all but impossible for _ new horizons _ to detect ffps , we include it as an upper limit and note that it is potentially well - suited to measure @xmath2 for lens systems that are much more massive than ffps and generally have longer timescales and larger einstein radii , such as stellar remnants . a microlensing event due to a single lensing mass leads to a light curve defined by three microlensing observables @xcite . the first is @xmath42 , the time of closest approach of the source to the lens . second is the impact parameter @xmath43 , which measures the angular distance of the closest approach of the source to the lens and is normalized to @xmath0 . third is the timescale @xmath25 , defined in equation ( [ eq : te ] ) . measuring the microlensing parallax @xmath2 for a short - timescale microlensing event requires measuring the shift in @xmath42 , @xmath43 , or both , in the light curve as it is seen from two ( or more ) locations ( for a recent review on this method see @xcite ) . together these observables define the ratio of @xmath38 , the projected separation between two observers , to @xmath24 via : @xmath44 in principle , @xmath2 can also be measured by a single , accelerated observer for long - timescale events , for which @xmath25 is of - order the orbital period ( i.e. , @xmath45 d ) , though this is not relevant for ffps . figure [ fig : dproj ] shows the projected separation @xmath38 between earth and each of the six satellite missions discussed in [ sec : satellites ] as seen from the center of the _ k2_c9 superstamp throughout a generic year . in all cases we compute orbits using the jpl ` horizons ` web interface . we use the center of the _ k2_c9 superstamp , ( ra , dec ) = @xmath46 , as the approximate center of bulge observations , coupled with the earth s geocenter when computing @xmath38 for all satellites . over the course of its projected separation will span @xmath47 . the visibility windows for _ spitzer _ are set by the requirement that the viewing angle @xmath48 between the sun and the target falls within the range : @xmath49 . this highlights the synergy between _ kepler _ and _ spitzer _ that will occur during the last two weeks of _ k2_c9 , though it is unlikely that _ spitzer _ will obtain data for any ffp candidates . for _ swift _ , we show the maximum @xmath38 it reaches over the course of its @xmath696-minute low - earth orbit . we note that its projected separation varies significantly over the course of its orbit and also as a displacement measured from any observatory on the surface of the earth . for _ new horizons _ , the annual parallactic oscillation in its projected separation is primarily due to the orbital motion of the earth . _ wfirst _ will not observe the bulge continuously but will feature six 72-day observing seasons dedicated to microlensing . the most recent sun angle pointing requirements for the planned spacecraft indicate that the microlensing seasons will be centered on the equinoxes . this indicates that @xmath38 will be near its maximum for much of these seasons and also that simultaneous ground - based observations will be restricted due to the visibility of the galactic bulge . the _ wfirst _ spacecraft will be inserted into the earth - sun l2 lagrange point . due to the dynamically unstable nature of l2 , _ wfirst _ will trace out a halo orbit around the l2 point . currently no orbital data exist on jpl s ` horizons ` for _ wfirst _ , so we investigate the orbits of two other l2 spacecraft as proxies . _ jwst _ will also be inserted into a halo orbit at earth - sun l2 , but its periapsis and apoapsis will each be larger than that of _ wfirst _ by a factor of @xmath62 . _ gaia _ is on a lissajous orbit around l2 with a periapsis and apoapsis that are each within a few tens of percent of that of _ wfirst_. while halo orbits are periodic and lissajous orbits are quasi - periodic , the l2 orbits of _ jwst _ , _ wfirst _ , and _ gaia _ all feature a periodicity of @xmath6180 days . we thus use the orbit of _ gaia _ as a proxy for that of _ wfirst _ and note that the morphology and extrema for the final true orbit of _ wfirst _ will be slightly altered from what is presented here . the projected physical size of the einstein ring is related to the physical parameters @xmath16 and @xmath20 by : @xmath50 figure [ fig : reproj ] shows @xmath24 for several planet masses as well as the brown dwarf mass limit ( 13 @xmath51 , @xmath64100 @xmath52 ) , as a function of @xmath22 . we consider a microlensing event to be detectable if @xmath53 from both the ground and space . in principle , though , it is possible to detect an event with @xmath54 if the source is very bright . defining @xmath55 as the angle between the lens - source relative proper motion @xmath56 and the earth - satellite projected separation @xmath38 , we establish the following relation : @xmath57 the criterion for detectability , when coupled with equation ( [ eq : delta_uzero ] ) , means that even if @xmath53 from one observatory , there will exist values of @xmath58 for which @xmath54 for the second observatory , leading to an undetectable event . furthermore , the existence of a preferred direction for @xmath56 will necessarily select a locus of @xmath55 and consequentially , via equation ( [ eq : delta_uzero ] ) , a range of @xmath58 . as a result , even for an assumed distribution of @xmath43 for the first observatory that is uniform , the distribution of @xmath43 for the second observatory would _ not _ be uniform and , as mentioned above , would include a population of undetectable events . a lower limit of @xmath58 , which is implied in the case of a non - detection of the event by one observer , places a lower limit on @xmath2 and hence an upper limit on the lens mass , if @xmath0 is measured . given this , we note that even with a sample of events that are detected only from the ground _ or _ from space , and for which there exists only a lower limit on @xmath2 , it is still possible to statistically explore the ffp mass function . in the milky way there is a preferred direction of @xmath56 defined by the direction of galactic rotation , so here we determine the probability that an event will be detectable as a function of @xmath59 . the probability distribution of the proper motion @xmath56 is a two - dimensional ( galactic coordinates : @xmath60 ) gaussian with expectation value : @xmath61 where @xmath62 , @xmath63 , and @xmath64 are the velocity of the lens , source , and sun , in galactic coordinates . we model each as a gaussian and list the mean and standard deviation for each object in table [ tab : v_dist ] . since all satellites considered here are on the ecliptic and the bulge microlensing fields are generally close to the ecliptic ( @xmath55@xmath65 ) , we rotate the coordinates by 60@xmath65 ( the angle between the ecliptic and the galactic plane ) and derive @xmath55 . the correction for fields off of , but close to , the ecliptic under this approximation should thus be @xmath55@xmath65 . moreover , events due to lenses in the bulge have a negligible preferred direction ( see table [ tab : v_dist ] ) , so this sets an upper limit for the error on our assumed rotation . we note that the preferred direction of @xmath56 also suggests a preferred sign of @xmath66 , primarily for disk lenses . on average , events will be detected first by a western observer , e.g. , an earth - trailing satellite . this can have implications when considering events alerted from one observer location and followed up by a second observer . for _ wfirst _ this would indicate a shift in the origin of first - detection alerts for a season centered on the vernal equinox compared to a season centered on the autumnal equinox . ccccc disk lens & 0 & 220 & 20 & 30 + bulge lens & 0 & 0 & 100 & 100 + source & 0 & 0 & 100 & 100 + sun & 7 & 242 & 0 & 0 [ tab : v_dist ] with the velocity distributions in hand , we compute the probability of a detectable event , i.e. , an event that is seen from both the earth and a satellite . we first draw @xmath43 for one observer ( the order of observatories is immaterial in this context ) from a probability distribution of @xmath67 ( i.e. , detected by the first observer ) . next , we draw a direction from the proper motion probability distribution for lenses in the disk with @xmath22 = 1 and 4 kpc and bulge lenses with @xmath22 = 7 kpc . using equation ( [ eq : delta_uzero ] ) we then derive @xmath58 for a range of @xmath38/@xmath24 , specifically , @xmath68 . finally , after adding @xmath58 to the @xmath43 of the first observer , we check if @xmath53 also for the second observer . in principle , @xmath67 should be uniform . however , as @xcite and @xcite showed , there is a selection effect that , for intrinsically faint sources , favors the detection of high - magnification events ( i.e. , small @xmath43 ) due to the limiting magnitudes of the surveys . this bias will persist for current ground and space surveys . in the case of _ , most of the sources will be bright enough to be detected at baseline from space and , potentially , from the ground as well , though the resources contributing to a simultaneous ground - based survey are currently unknown . additionally , since , for a significant fraction of the ffp parameter space , the maximum possible magnification is relatively low ( see [ sec : rho ] ) , this bias will be less prevalent . therefore , we calculate the results both for a uniform distribution of @xmath43 and for the observed distribution found by @xcite and @xcite , which favors small @xmath43 . figure [ fig : prob_dproj_reproj ] shows the probability of a detectable event ( i.e. , @xmath53 from both observer locations ) as a function of @xmath38/@xmath24 for the three lens distances and two @xmath43 distributions we explore . for @xmath69 , the probability is @xmath7060@xmath14 for all possibilities , with the observed @xmath43 distributions yielding a higher probability by up to 8@xmath14 as expected . for @xmath71 the probability falls to 2036@xmath14 , with a higher probability for the uniform distribution by @xmath62@xmath14 . this arises from the fact that on average @xmath72 , so a higher probability of having a smaller @xmath43 for one observer suggests that for the other @xmath73 . beyond @xmath74 the probability declines gradually . an additional requirement for a secure ffp characterization is the determination of @xmath0 , which must be done via the measurement of finite - source effects , given the absence of detectable lens flux and high - precision astrometric data . the geometric probability @xmath75 that the projected separation between the source and the ffp lens will be sufficiently small that the angular size of the source is measurable can be cast in terms of the physical parameters as : @xmath76 where @xmath77 is the angular radius of the source star . in general , it is necessary that @xmath78 , for at least one of the observers , in order for the finite angular size of the source to be measured . we note that for @xmath79 , the probability can be boosted by a factor of @xmath62 for two observers with @xmath69 , assuming a uniform distribution for @xmath43 , and up to a factor of @xmath64 , given the observed @xmath43 distribution . however , as @xmath1 increases , the probability converges to equation ( [ eq : rho_prob ] ) . figure [ fig : rho ] shows @xmath1 as a function of @xmath22 for several ffp masses . each panel represents a typical physical source radius @xmath80 for different source spectral types : @xmath80 = 13@xmath81 for a clump giant , 3@xmath81 for a subgiant , 1@xmath81 for a main sequence solar - type source , and 0.5@xmath81 for a late k / early m dwarf . it is crucial to note that the probability of detecting finite - source effects computed in equation ( [ eq : rho_prob ] ) is counterbalanced by the fact that the maximum possible magnification decreases for larger @xmath1 . @xcite found that the maximum magnification @xmath82 as a function of @xmath1 is : @xmath83 the right - hand y - axis of figure [ fig : rho ] provides the values of @xmath82 . as @xmath1 increases , not only is the maximum possible magnification suppressed , but the light curve broadens . consequentially , the fractional deviation of the wings of the light curve from that expected for a point - like source can be substantial . however , it becomes increasingly possible to reproduce the light curve with a range of impact parameters , a degeneracy that ultimately impacts the ability to robustly measure @xmath2 . for giant sources , the maximum possible magnification across all ffp planet masses is @xmath5100 and never exceeds 2 for earth - mass planets . this will make it difficult to securely detect the event from two observatories and to also robustly measure @xmath2 . the limiting factor shifts when considering main sequence sources , with @xmath84 . a smaller @xmath1 indicates a higher possible magnification but a correspondingly lower probability for measuring finite - source effects . given a solar - type source with @xmath85 and an ffp lens at @xmath22 = 4 kpc , there is only a @xmath630@xmath14 chance of measuring @xmath1 for a planet with @xmath86 , and this drops to @xmath61@xmath14 for @xmath87 . lastly , once a planetary - mass lens has been identified , it is necessary to exclude the possibility of the ffp being bound by searching for and constraining the flux contribution from a possible host star . this requires a measurement of the source flux and at least one epoch of high - resolution data . the assumed procedure is as follows . the flux @xmath88 of a microlensing event at time @xmath89 is given by : @xmath90 where @xmath91 is the flux of the source , @xmath92 is the magnification of the source at @xmath89 , and @xmath93 is the blend flux of all other stars that are not resolved . it is relatively straightforward to fit a microlensing model for a well - sampled light curve of a single - lens microlensing event , giving @xmath92 for all times @xmath89 . this is true regardless of observing bandpass , since microlensing is achromatic . then , for every additional band beyond the primary one with which @xmath92 is computed , in principle only two data points , taken at two different magnifications , are required to measure @xmath91 and @xmath93 in that band . a high - resolution image of the microlensing event taken at any time will resolve out all stars not dynamically associated with the event to a high probability , meaning that @xmath93 will consist only of the lens and companions to the lens and/or source . taking a high - resolution observation at baseline , when the event is over and the source is no longer magnified , optimizes the chances of detecting light from any luminous lenses , since the flux contribution from the source will be minimized and no ambient interloping stars will be contributing to @xmath93 . we will refer to such a high - resolution flux measurement as @xmath94 , since only light from the microlensing target the lens and source systems will be included in the point - spread function ( psf ) . this can be combined with the known source flux to determine @xmath95 , a measure of any flux in excess of the known source flux in the high - resolution image , via : @xmath96 regarding ffps , @xmath95 = 0 provides a limit on the brightness of a possible lens host star . since the distances to the ffps will be derived from the measurements of @xmath2 and @xmath0 , this allows for the strongest constraints on the lens host star flux @xmath4 , which in turn provides yields a limit on the lens host star mass @xmath97 . this same principle has been used to characterize a handful of bound planetary microlensing events @xcite . the high - resolution data are typically taken in @xmath98-band , which balances the seeing - limited angular resolution with the attainable strehl ratio and the sky background . hereafter we will consider only @xmath98-band fluxes of the source , lens , and target , @xmath99 , @xmath100 , and @xmath101 , respectively , but note that this technique works for any observational band . estimating the mass above which it is possible to rule out the presence of a lens host star requires a careful treatment of the relevant uncertainties . here we consider two observational regimes . for our fiducial simulation , we assume the observational resources that will be implemented during _ a proposal was accepted to take target - of - opportunity ( too ) observations of short - timescale ffp candidates using nirc2 on keck 2 . each too will be used to obtain a high - resolution measurement of the @xmath98-band flux of a microlensing target while the source is magnified , which we will refer to as @xmath102 . we also assume the existence of a future high - resolution epoch of the target when the source is at its baseline flux , which we will denote as @xmath103 . as described in [ sec : fl_method ] , these two points are sufficient to measure the source flux ( @xmath104 ) and blend flux ( @xmath105 ) in the high - resolution data , given a light curve model derived using other bandpasses . equation ( [ eq : deltaf ] ) refers to the general case for which only the baseline epoch of high - resolution data exists . in this scenario , however , obtaining a magnified point _ and _ an eventual baseline epoch mean that a measurement of @xmath106 establishes a limit on the brightness of a possible lens host star . under this procedure , the associated uncertainties are thus : 1 . @xmath107 : the statistical uncertainty of the microlensing target taken using the high - resolution facility when the source is magnified ; 2 . @xmath108 : the statistical uncertainty of the microlensing target taken using the high - resolution facility when the event is over and the source has returned to its baseline brightness ; 3 . @xmath109 : the uncertainty of photometrically aligning the two high - resolution images in order to compute @xmath105 ; 4 . @xmath110 : the statistical uncertainty of the blend brightness , which includes its covariance with the other parameters from the light curve model ; and 5 . @xmath111 : the uncertainty required to calibrate to a photometric standard system . for our fiducial computation , we assume @xmath112 and @xmath113 . we assume that @xmath114 @xcite and consider four different @xmath98-band source magnitudes , each of which corresponds to one of the four spectral types explored in [ sec : rho ] . assuming a source distance of @xmath23 = 8.2 kpc @xcite and an @xmath98-band extinction of @xmath115 , we take @xmath99 = 14.0 for a giant source , @xmath99 = 16.5 for a subgiant , @xmath99 = 19.0 for a solar - type main sequence source , and @xmath99 = 21.0 for a late k / early m dwarf . we add all uncertainties in quadrature and determine the @xmath98-band apparent magnitude of a possible lens host star , @xmath100 , that can be excluded at the three - sigma level , from an upper limit of @xmath105 = 0 . finally , using a 5 gyr padova isochrone @xcite , we convert @xmath100 into a limit on the lens host star mass , @xmath97 , as a function of lens distance @xmath22 . the isochrone extends only to 0.09@xmath15 , so we extrapolate down to the deuterium - burning limit at 0.08@xmath15 , which we establish as the threshold for stellar - mass ffp host star exclusion . our results are shown in figure [ fig : fl ] . giant and subgiant sources are generally too bright to be able to robustly rule out the presence of an m - dwarf ffp host star , particularly at distances of several kiloparsecs . however , for fainter main sequence sources , for which the flux contribution from a potential lens host star will be a significantly larger fraction of the total light measured in the baseline high - resolution observation @xmath94 , the situation is much different . assuming @xmath116 , stellar hosts of any mass can be excluded up to @xmath61.8 kpc , and even for ffps in the bulge the most massive host star that could go undetected is @xmath60.25 @xmath15 . finally , as a more conservative limit , we set @xmath117 , the results for which are also shown in figure [ fig : fl ] . this decreases the limiting distance closer than which hosts of any mass can be excluded to @xmath61.3 kpc . correspondingly , the mass limit that can be ruled out for an ffp lens at any distance increases to @xmath60.37 @xmath15 . we also perform a second simulation with an optimal set of assumptions regarding the relevant uncertainties . the primary band in which _ wfirst _ will conduct its microlensing survey , w149 , will have a psf of @xmath60.14@xmath118 @xcite . at this angular resolution , the probability of having an ambient interloping star with @xmath119 blended with the microlensing event is @xmath1215@xmath14 @xcite . thus , for the vast majority of microlensing events that _ wfirst _ will detect , having additional high - resolution observations taken with a different facility , both while the source is magnified and while it is at baseline , is unnecessary . moreover , during a microlensing season images will be taken in w149 every @xmath615 minutes , producing light curves with exquisite photometric precision and sampling . microlensing predicted detection rates peak between @xmath120 = 2223 ( penny et al . , in prep . ) , much fainter than for _ k2_c9 or current ground - based surveys . in summary , the _ wfirst _ data set will largely be self - contained . given these factors , for our optimistic simulation we remove @xmath109 and @xmath107 , set @xmath121 , and assume @xmath114 . figure [ fig : fl ] shows the results for our optimal scenario . for subgiant and giant sources it is still difficult to set meaningful mass limits for a lens host star , particularly as @xmath22 increases . however , for solar - type main sequence sources with @xmath122 , the distance to which stellar hosts can be excluded increases to @xmath62.4 kpc , and for lenses in the bulge we can probe down to @xmath60.2@xmath15 this improves dramatically for late k / early m dwarf sources with @xmath123 . stellar lens host stars can be excluded out to @xmath66.1 kpc , and even in the central bulge the highest mass star that would remain undetected is only @xmath60.1@xmath15 . for our simulations and results regarding lens flux constraints we have operated in the regime of prompt follow - up photometry , i.e. , high - resolution images taken shortly after the event is over . even with facilities such as nirc2 on keck and _ wfirst _ , the lens and source will remain unresolved . however , it is possible to set even stronger @xmath4 constraints by waiting for the lens and source to separate according to their relative proper motion @xmath27 . then the possible lens host star can be treated as an isolated object , allowing for more robust @xmath4 and , ultimately @xmath97 constraints when compared to searching for an increase in flux on top of the ( known ) source flux . @xcite finds that for ffp candidate events it is feasible to exclude the presence of lens host stars with separations comparable to the oort cloud using existing adaptive optics facilities by waiting @xmath650 years after the event is over . this technique still requires that @xmath2 and @xmath0 ( from a determination of @xmath1 ) are measured , such that the mass of the ffp will be known . by coupling these parameters with @xmath25 , the lens - source relative proper motion vector @xmath56 can be measured . it is then straightforward to estimate the timescale on which the lens and source could be spatially resolved using a given facility as well as the direction of separation ( see @xcite for a more thorough discussion of spatially resolving lenses and sources ) . it would still be optimal to obtain a prompt high - resolution epoch shortly after the event is over . given the stellar surface density toward the bulge , it is best to identify and exclude any ambient interloping stars that could possible be confused for the microlensing target , especially if the proper motion is not well constrained . here we present the implications our findings have for the characterization of ffp candidate events . it is eminently possible to explore the frequency , galactic distribution , and even mass function of ffps without robustly measuring @xmath2 and @xmath0 ( via @xmath1 ) , which together provide mass and distance measurements for individual events . nor is fully excluding stellar - mass hosts a requirement for gaining traction in understanding ffp demographics . in this section , however , we will focus on the full set of constraints as they can be measured using data taken during _ k2_c9 or by _ wfirst_. the projected separation @xmath38 of _ k2_c9 reaches its minimum value of 0.07 au one week into the campaign , on 14/april/2016 , after which it increases monotonically to 0.81 au on 1/july . this significantly affects its ability to measure @xmath2 for ffp lenses with different masses and distances throughout the campaign . figure [ fig : mffp_dl ] shows @xmath16 as a function of @xmath22 . by assuming a value of @xmath24 that is equal to @xmath38 it is possible to determine the masses and distances of ffps that will yield detectable events at a given time during _ k2_c9 . as discussed in [ sec : reproj ] , the probability that events with @xmath124 will be detectable by the pair of observer locations is @xmath660@xmath14 , and this probability declines steeply as @xmath59 increases . we thus use this as a benchmark for sensitivity to @xmath2 . immediately we see that toward the end of _ k2_c9 @xmath2 can only be measured for lenses with mass lower than the deuterium - burning limit for @xmath125 kpc . near the beginning of the campaign it will be possible to measure a parallactic shift for nearby jupiter - mass lenses , with its sensitivity extending down to earth - mass ffps as the lens distance increases to @xmath126 kpc . but @xmath2 is only one of two ingredients necessary for measuring @xmath16 and @xmath22 . the other is @xmath1 , which helps determine @xmath0 . by rearranging the right - hand side of equation ( [ eq : rho_prob ] ) we compute @xmath16 as a function of @xmath22 for different values of @xmath1 , given an assumed physical source radius @xmath80 ( and thus angular source radius @xmath77 , assuming @xmath127 kpc ) . to compute @xmath1 for _ k2_c9 ( left panel of figure [ fig : mffp_dl ] ) we assume @xmath128 . we select this as the benchmark value of @xmath80 because @xmath129 , corresponding to @xmath130 , is likely too faint for _ k2_c9 and @xmath116 allows for the otherwise strongest lens flux constraints . from the left - hand side of equation ( [ eq : rho_prob ] ) , the probability of detecting finite - source effects , and thus measuring @xmath1 and , ultimately , determining @xmath0 , scales approximately as @xmath1 itself . thus , the value of @xmath1 at a given ( @xmath22 , @xmath16 ) coordinate gives the probability of measuring @xmath1 for an ffp with those physical properties . rephrased , @xmath131 indicates , roughly , the number of events with that specific ( @xmath22 , @xmath16 ) combination that would need to be detected in order to measure @xmath1 for at least one of them . the constraints for measuring @xmath2 and @xmath1 work in opposing directions . for a fixed @xmath38 , the mass to which the satellite is sensitive decreases as @xmath22 increases . however , as @xmath22 increases , the probability of measuring @xmath1 _ increases _ , for a fixed ffp mass . finally , we incorporate the flux constraint for ruling out the presence of lens host stars . from the left panel of figure [ fig : mffp_dl ] , the largest @xmath22 at which stellar - mass lens host stars can be excluded for sources with @xmath116 is @xmath61.8 kpc from our fiducial simulation . when considering all three factors , @xmath2 , @xmath1 , and @xmath4 , the optimal regime for characterizing ffps occurs for the smallest @xmath38 , when the sensitivity extends to ffps with the lowest masses , for a fixed @xmath22 , and for when the probability of measuring @xmath1 is the highest , i.e. , larger values of @xmath1 . this drives the sensitivity of _ k2_c9 toward the first few weeks of the campaign , when @xmath132 au , as is furthermore supported by the goal of excluding stellar - mass host stars , which requires that @xmath133 kpc . in this regime , however , while stellar hosts can be excluded at the three - sigma level from @xmath4 constraints , the probability of measuring both @xmath2 and @xmath1 is only @xmath61.5@xmath14 . at the end of _ k2_c9 , when @xmath7 au , the lowest mass lens for which it will be possible to measure @xmath2 and also constrain the flux will be @xmath134 , above the brown dwarf mass limit . the current plan is for each of _ wfirst _ s 72-day observing seasons to be centered on an equinox . its projected separation would thus be restricted to @xmath135 . nevertheless , given the uncertain nature of the orbital parameters ( see [ sec : dproj ] ) and the exact timing of the observing seasons , we presume @xmath136 au and @xmath137 au to bracket the full range possible for _ wfirst_. these are shown in the right panel of figure [ fig : mffp_dl ] . from the detectability limit of @xmath124 discussed in [ sec : reproj ] , this indicates that _ wfirst _ will be able to measure @xmath2 for earth - mass planets for @xmath138 kpc and for lower - mass planets at larger distances . were the sun - angle restrictions relaxed such that _ wfirst _ could observe toward the bulge near the june solstice , when its projected separation could reach as small as 10@xmath139 au , it would be sensitive to planets @xmath62 orders - of - magnitude less massive ( i.e. , free - floating moons ) . the typical source stars for _ wfirst _ will have @xmath130 , meaning that the probability to measure @xmath1 is @xmath310@xmath14 for earth - mass ffps with @xmath140 kpc . the corresponding source magnitude of @xmath129 , combined with the optimal lens flux simulation ( which is appropriate for _ wfirst _ ) , suggests that it will be possible to exclude host stars for ffps detected by _ wfirst _ for @xmath141 kpc . this means that _ wfirst _ will be sensitive to ffps at least down to the mass of super - earths throughout the galaxy . one possible limiting factor , however , could be the capabilities of the accompanying ground - based facilities . a source with @xmath129 will have an optical magnitude of @xmath142 2224 . this is below the sensitivity of current microlensing surveys . additionally , the maximum magnification limitation from having a larger @xmath1 means that many of these events may not get sufficiently bright to be detected even at peak . however , if future surveys will be conducted by larger telescopes , such as that proposed for subaru , this limit is no longer relevant . secondly , the bulge is visible only for @xmath617 hours from ctio ( as a representative for the southern hemisphere ) during the currently planned _ wfirst _ campaigns ( as shown on the bottom panel of figure [ fig : dproj ] ) . for telescopes in the northern hemisphere , such as subaru on mauna kea , the bulge is visible for only 60@xmath14 of each _ wfirst _ observing season and even then only up to a maximum of @xmath63 hours per night . work by cbh and ys was supported by an appointment to the nasa postdoctoral program at the jet propulsion laboratory , administered by universities space research association through a contract with nasa . we thank david ciardi , chas beichman , and chris gelino for useful discussions regarding adaptive optics observations . finally , we thank scott gaudi , dan maoz , matt penny , and sebastiano calchi novati for a careful and fruitful reading of the manuscript .	simultaneous space- and ground - based microlensing surveys , such as those planned with _ k2 _ s campaign 9 ( _ k2_c9 ) and , potentially , _ wfirst _ , facilitate measuring the masses and distances of free - floating planet ( ffp ) candidates . ffps are initially identified as events arising from a single lensing mass with a short timescale , ranging from one day for a jupiter - mass planet to a few hours for an earth - mass planet . measuring the mass of the lensing object requires measuring the angular einstein radius @xmath0 , typically by first determining the finite size of the source star @xmath1 , as well as the microlens parallax @xmath2 . a planet that is gravitationally bound to , but widely separated from , a host star ( @xmath320 au ) can produce a light curve that is similar to that of an ffp . this tension can be resolved with high - resolution imaging of the microlensing target to search for the lens flux @xmath4 from a possible host star . here we investigate the accessible parameter space for each of these components @xmath2 , @xmath1 , and @xmath4 considering different satellite missions for a range of ffp masses , galactic distances , and source star properties . we find that at the beginning of _ k2_c9 , when its projected separation from the earth ( as viewed from the center of its survey field ) is @xmath50.2 au , it will be able to measure @xmath2 for jupiter - mass ffp candidates at distances larger than @xmath62 kpc and to earth - mass lenses at @xmath68 kpc . at the end of its campaign , when @xmath7 au , it is sensitive to planetary - mass lenses for distances @xmath33.5 kpc , and even then only to those with mass @xmath3m@xmath8 . from lens flux constraints we find that it will be possible to exclude _ all _ stellar - mass host stars ( down to the deuterium - burning limit ) for events within @xmath62 kpc , and for events at any distance it will be possible to exclude main sequence host stars more massive than @xmath60.25 m@xmath9 . together these indicate that the ability to characterize ffps detected during _ k2_c9 is optimized for events occurring toward the beginning of the campaign . _ wfirst _ , on the other hand , will be able to detect and characterize ffps with masses at least as low as super - earths throughout the galaxy during its entire microlensing survey .
the diffusion process of a particle in a potential with the presence of a sink studied through solution of the smoluchowski equation@xcite is of keen interest to many scientists in chemical dynamics as it serves a reference model for a wide variety of dynamical processes . various attempts has been made to study the diffusion processes with a suitable position of the sink@xcite which can find application in diffusion controlled reactions@xcite . such a model is used by wilemski and fixman@xcite to calculate the rate of diffusion controlled reactions as well as cyclization of polymer chain in solutions . ovchinnikova@xcite , zusman@xcite , marucs and nadler@xcite have used such a model for electron transfer reactions in polar solvents . marcus@xcite recently used a diffusive equation with a sink term to develop a theory of uni - molecular reactions in clusters . pressure influence on isomerization reactions is also explained by sumi@xcite by using such type of model.bagchi , fleming@xcite uses a model of this kind to analyze barrier less electron relaxation in solution . exact analytical results for diffusion problems helps in understanding the different parameters like friction and provide an insight to different approximations . as per the authors knowledge , there are no cases where the smoluchowski equation with a time dependent sink is solved by analytical methods . even there is no analytical solution available for the simplest possible case _ i.e. , _ for free particle . there are large number of works on diffusion under time independent sink . in recent work from our group we have given an analytical model in which the problem of particle undergoing diffusive motion under the influence of two potentials where the coupling is time independent@xcite . most of that has focussed on the time evolution / propagator derivation of the case of one or more dirac delta function sinks with constant strength in time . in contrast , this paper presents an work that deals mainly with a dirac delta sink whose strength varies with time and we are the first one to consider this effect explicitly . there are two common methods that one may think of for solving this problem . one method is using a path integral method of the feynman type and the other is using laplace transforms . we use the latter method though the two methods are closely related . our method closely follows the method used for solving schrodinger wave equation . we would like to solve the smoluchowski equation with a time - dependent sink as shown in figure 1 . @xmath2 in the above equation @xmath3 is the diffusion coefficient . it is related to the friction coefficient @xmath4 by @xmath5 . by integrating eq . ( 1 ) from @xmath6 to @xmath7 , we find that the eq.(1 ) reduces to . @xmath8 this is just the case of zero potential with an additional time - dependent boundary condition . if we take the laplace transform @xcite of eq . ( 2 ) , we get @xmath9.\ ] ] for solving eq . ( 3 ) we first consider the homogeneous solution in order to satisfy the boundary condition at the origin . the solution is given below @xmath10 where @xmath11 is the integration constant . by applying the boundary condition at the origin to eq . ( 4 ) we find . @xmath12\\ \nonumber & \rightarrow a(s ) = \frac{l\left[k(t)p(0,t)\right]}{\sqrt { s / a}}.\end{aligned}\ ] ] we now set @xmath13 in eq . ( 4 ) to get @xmath14}{\sqrt { s / a}}+\frac{1}{2 a \sqrt{s / a}}\int dx'(\sqrt{s / a}\|x'\|)p(x',0),\ ] ] using the convolution theorem for laplace transforms @xcite , we invert eq . ( 6 ) to get @xmath15 the eq.(7 ) may be used to determine the probability distribution at the origin ( _ i.e. , _ @xmath0 ) . once @xmath0 is known , one may find @xmath1 by inverting eq.(4 ) which has the following result : + @xmath16 eq . ( 7 ) and eq.(8 ) together completely determine the problem . as we can see knowing the probability distribution at the origin _ @xmath17 is essentially equivalent to knowing the probability distribution everywhere ( @xmath1 ) . in some cases it may be advantageous to write down the solution in terms of propagator as follows @xmath19 where @xmath20 is propagator . to find the propagator , we use eq.(5 ) and eq.(6 ) to get @xmath21 using eq . ( 4 ) and eq . ( 10 ) , we get @xmath22 so that @xmath23 we now invert eq . ( 11 ) to get @xmath24 \times erfc\left(k_0 \sqrt { t / a}+\frac{\|x\|+\|x'\|}{2\sqrt { t / a } } \right)\nonumber \\ + \frac{1}{2 a^2 \sqrt{\pi t / a}}exp \left ( a\frac{(x - x')^2}{4t}\right).\end{aligned}\ ] ] we can see easily that the same propagator can be found by using feynman approach . let us assume that @xmath25 , so that @xmath26 and hence the detailed expression for @xmath1 can be @xmath27 \times erfc\left(k_0 \sqrt { t / a}+\frac{\|x\|+\|a\|}{2\sqrt { t / a } } \right)\nonumber \\ + \frac{1}{2 a^2 \sqrt{\pi t / a}}exp \left(a\frac{(x+a)^2}{4t}\right).\end{aligned}\ ] ] we now consider the case , where @xmath28 has the following form : @xmath29 here @xmath30 is a real number . we take the initial condition as @xmath31 where @xmath32 and @xmath30 are real positive numbers . using eq.(6)and the identity @xmath33 = - d l[t\times f(t)]/ds$ ] , we get @xmath34 we can solve this equation under the condition that the laplace transform vanishes at s=@xmath35 and that it must also be bounded at @xmath36 .the last condition is required for the inversion integral to exist . so that we get @xmath37 this laplace domain expression can not be converted to time domain in terms of elementary or special functions . however , it can be done using the integral @xmath38 in the above expression @xmath39 is a positive number taken to be larger than the right - most pole in the complex plane . so we are unable to do the last step analytically , although numerically it is not a problem . we choose @xmath40 , where @xmath30 is a real positive constant . the laplace transform of @xmath28 does not exist . interestingly , the laplace transform of @xmath41 does exist if we choose an initial condition that vanishes at the origin ! our derivation in this subsection is valid under this retriction . we first use the fact , @xmath42=\int^\infty_s ds ' l[f(t),s'].\ ] ] now by using eq . ( 21 ) and eq . ( 6 ) we get @xmath43 now we define a new function @xmath44 as given below @xmath45 now eq.(22 ) can be written as @xmath46 on solving eq . ( 24 ) we get , @xmath47 by equation 29 we find , + @xmath48 we may now invert equation 30 to find , + @xmath49 we now seek @xmath50.to find this quantity , we first write , + @xmath51 to find g , we use results of equations ( 30 ) , ( 27 ) , ( 5 ) and ( 4 ) to find , + @xmath52\ ] ] we may now invert this expression to find , + @xmath53 @xmath54= l[f\;(t),s\;-a].\ ] ] by equation ( 31 ) and ( 6 ) we find @xmath55 to obtain a solution , we could iterate this expression repeatedly to obtain the series solution : + @xmath56 but since we already know the form of the serise , we may substitute equation(37)directly into equation(36 ) . after doing this nad solving for the a s by equating terms with like powers of @xmath39 , we find that @xmath57 ( s ) is given by + @xmath58 @xmath59 in the current work the authors had attempted to find the exact solution for the smoluchowski equation with a time dependent delta function sink in many special cases . these special cases include linear , inversely and exponential dependence . the case of sinusoidal time dependent sink can be treated in a similar manner as the exponential time dependent sink by representing the sine function as a sum of complex exponentials . th authors want to thank iit mandi for professional development fund as well as htra scholarship . h. risken , the fokker planck equation ( springern , berlin 1984 ) . a. samanta and s. k. ghosh , j. chem . 97(12 ) ( 1992 ) 9321 . b. bagchi j. chem . 87 , 5393 ( 1987 ) s. a. rice , diffusion - limited reactions , edited by c. h. bamford , c. f. h. tipper and r. o. compton ( elseiver , amsterdam , 1985 ) , vol 25 . g. wilemsky and m. fixman , j. chem . phys . 84 ( 1986)4894 . g. wilemsky and m. fixman , j. chem . phys . 60 ( 1974)878 . m. ya . ovchinnikova , teor . khim 17 ( 1981)651 . l. d. zusman , chem . 80 ( 1983)29 . h. sumi and r. a. marcus , j. chem . 84 ( 1986)4894 r. a. marcus , j. chem . ( 1996)5446 . h. sumi , j. mol . 65/66 ( 1995)65 . b. bagchi , g. r. fleming , j. chem . 78 ( 1983)7375 . b. bagchi , g. r. fleming , j. phys . 94 ( 1984)9 . a. chakraborty . phys . 139 ( 2013)094101 . j. campbell , j. phys . a : math . theor . 42 ( 2009)365212 . m. abramowitz and i. stegun , _ handbook of mathematical functions _ ( dover , new york , 1970 ) .	the smoluchowski equation for a free particle with a time dependent sink is solved exactly for many special cases . in this method by knowing the probability distribution at the origin @xmath0 , one may derive the probability distribution at all positions _ i.e. , _ @xmath1 .
in many physics analyses , one seeks to extract information from a reaction with multi - dimensional kinematics . fits to binned data are often limited by statistical uncertainties ; thus , use of the unbinned maximum likelihood method is generally preferred . this method is well suited to obtaining estimators for unknown parameters in a hypothesis probability density function ( p.d.f . ) . it is also an excellent way of discriminating between hypotheses ; however , it does not provide a means of determining how well the best hypothesis describes the data . in this paper , we describe an _ ad - hoc _ procedure for obtaining a @xmath0 goodness - of - fit measurement of a hypothesis whose parameters were determined using an unbinned maximum likelihood fit without having to bin the data . this quantity is not meant to replace the likelihood . the likelihood should still be used to determine which hypothesis best describes the data and to obtain estimators for any unknown parameters , _ i.e. _ extract any physical observables . the goodness - of - fit quantity obtained using our method would simply be used to judge how well the best hypothesis describes the data . in our approach , event - by - event standardized residuals are used to obtain the global @xmath0 . these residuals can also be used for diagnostic purposes to study the goodness - of - fit as a function of kinematics or detector components . determining where in phase space a fit fails to describe the data is vital to understanding what physics has not been accounted for in the p.d.f . identifying regions or components of the detector where a fit fails to describe the data could help diagnose unaccounted for inefficiencies in the acceptance calculation . other authors have also developed methods for obtaining goodness - of - fit quantities from unbinned data . the methods closest to ours are based on the nearest neighbors to a given event @xcite . however , there are several other recent articles on other procedures as well @xcite . while these and other authors have developed methods for obtaining goodness - of - fit quantities from unbinned maximum likelihood fits , the advantage of our approach lies in the event - by - event residuals . it is also straight - forward to apply our method to most experimental physics analyses . consider a data set composed of @xmath1 total events , each of which is described by @xmath2 coordinates , @xmath3 . the coordinates can be masses , angles , energies , _ etc_. two separate hypotheses have been proposed to describe the data . their probability density functions will be denoted by @xmath4 where @xmath5 are the functional dependencies on the coordinates , @xmath3 , and ( possible ) unknown parameters , @xmath6 , of hypothesis @xmath7 @xmath8 . using these p.d.f.s , unbinned maximum likelihood fits can easily be performed to obtain estimators , @xmath9 , for the unknown parameters , @xmath6 , of each hypothesis . the likelihoods from these fits , @xmath10 , can be used to determine which hypothesis provides the better description of the data ; however , the following simple question about this hypothesis can not yet be answered : `` how well does it describe the data ? '' the aim of this procedure is to obtain a @xmath0 goodness - of - fit measurement of a hypothesis whose parameters were determined using an unbinned maximum likelihood fit . we first need to define a metric for the space spanned by the `` relevant '' coordinates of @xmath3 , _ i.e. _ all coordinates on which the p.d.f has functional dependence . a reasonable choice is to use @xmath11 , where @xmath12 is the maximum possible difference between the coordinates , @xmath13 , of any two events in the sample . using this metric , the distance between any two events , @xmath14 , is given by @xmath15 ^ 2,\ ] ] where the sum is over all relevant coordinates ( discussed above ) . for each event , we then compute the distance , @xmath16 , to the @xmath17th closest event using ( [ eq : dist ] ) . the value of @xmath17 is discussed below . thus , a hypersphere centered at the @xmath18 event with radius @xmath16 contains @xmath17 data events ( excluding the event itself ) . in this way , we can calculate the density of data events at the point @xmath19 . by comparing these density calculations to those predicted by a given hypothesis , we can determine how well the hypothesis describes the data without resorting to binning . for each event , the standardized residual , @xmath20 , for any hypothesis is then calculated according to @xmath21 where @xmath22 is the number of measured ( predicted ) events contained within a hypersphere of radius @xmath16 centered at @xmath19 and @xmath23 is the uncertainty on @xmath22 . for the case where the data set contains no background events ( dealing with backgrounds is discussed in section [ section : example : bkgd ] ) , @xmath24 and @xmath25 , provided @xmath17 is chosen to be large enough for the gaussian approximation to hold . the number of predicted events , @xmath26 , is calculated for hypothesis @xmath7 as @xmath27,\ ] ] where the integral in the numerator is over the hypersphere of radius @xmath16 centered at @xmath19 discussed above . in principle , there are cases where the integrals in ( [ eq : method - num - predicted ] ) can be performed analytically ; however , once detector acceptance is included , for which an analytic expression typically is not known , they must be done using the monte carlo technique . for these cases , @xmath28 can be calculated by generating @xmath29 monte carlo events according to @xmath3 phase space which leads to the following approximation of ( [ eq : method - num - predicted ] ) : @xmath30,\ ] ] where the heaviside function , @xmath31 , is used to restrict the sum in the numerator to monte carlo events within the hypersphere . the statistical uncertainty in the monte carlo calculation of @xmath28 would be @xmath32 , where @xmath33 is the number of monte carlo events within the hypersphere . we note here that if @xmath33 is small , then the uncertainty is better approximated by @xmath34 @xcite . if the hypothesis does in fact provide a good description of the data , then the values of @xmath35 obtained from ( [ eq : chi2-event ] ) will follow a @xmath36 distribution with one degree of freedom . the overall goodness - of - fit can be obtained as follows : @xmath37 where @xmath38 is the number of free parameters present in the hypothesis p.d.f . hypotheses which describe the data well will yield values of @xmath39 , while those which do not describe the data well will yield larger values . as an example , we will consider the reaction @xmath40 in a single @xmath41 bin , _ i.e. _ a single center - of - mass energy and production angle bin ( extending the example to avoid binning in production angle , or @xmath42 , is discussed below ) . the @xmath43 decays to @xmath44 about 90% of the time ; thus , we will assume we have a detector which has reconstructed @xmath45 events . below we will analyze a simulated @xmath43 photoproduction data set . the goal of our model analysis is to extract the @xmath43 polarization observables known as the _ spin density matrix elements _ , denoted by @xmath46 ( discussed below ) . in terms of the mass of the @xmath44 system , @xmath47 , the @xmath43 events were generated according to 3-body phase space weighted by a voigtian ( a convolution of a breit - wigner and a gaussian ) to account for both the natural width of the @xmath43 and detector resolution . for this example , we chose to use @xmath48 mev / c@xmath49 for the detector resolution . the goal of our analysis is to extract the three measurable elements of the spin density matrix ( for the case where neither the beam nor target are polarized ) traditionally chosen to be @xmath50 , @xmath51 and @xmath52 . these can be accessed by examining the distribution of the decay products ( @xmath44 ) of the @xmath43 in its rest frame . for this example , we chose to work in the helicity system which defines the @xmath20 axis as the direction of the @xmath43 in the overall center - of - mass frame , the @xmath53 axis as the normal to the production plane and the @xmath54 axis is simply given by @xmath55 . the decay angles @xmath56 are the polar and azimuthal angles of the normal to the decay plane in the @xmath43 rest frame . the decay angular distribution of the @xmath43 in this frame is then given by @xcite @xmath57 we chose to use the values @xmath58 , @xmath59 and @xmath60 for our simulated data set . we will begin by considering the simplest case where the signal is able to be extracted without any background contamination and our detector efficiency is 100% . for this study , a simulated data set of 10,000 events was generated following the criteria described above ( see figure [ fig : gen - no - acc - no - bkgd ] ) . to facilitate the calculation of ( [ eq : method - num - predicted - mc ] ) , 100,000 monte carlo events were generated using the same @xmath47 distribution as the data ; however , the decay distribution was generated according to flat phase space in @xmath61 and @xmath62 . the likelihood function required to extract the spin density matrix elements is simply @xmath63 where @xmath64 is the decay angular distribution defined in ( [ eq : schil ] ) . to obtain estimators for the @xmath46 elements , we minimize @xmath65 which yields the values [ eq : rho - no - acc - no - bkgd ] @xmath66 @xmath67 @xmath68 these values are in good agreement with the values used to generate the data . at this point in our model analysis , we have extracted the physical observables we are interested in . the spin and parity of the @xmath43 meson are well established ; however , what if we were analyzing a less well - known meson ? in that case , we may not be certain that @xmath69 , _ i.e. _ the use of ( [ eq : schil ] ) might not be justified . prior to publishing our results , it would be reassuring to know that the fit performed using ( [ eq : schil ] ) does in fact provide a good description of our data . for this analysis , the relevant coordinates are the decay angles @xmath70 and @xmath62 , since these are the only kinematic quantities for which the p.d.f . defined in ( [ eq : schil ] ) has dependence . the metric defined in ( [ eq : dist ] ) is then given by @xmath71 for this analysis . for each simulated event , we find the distance to the @xmath17th nearest neighbor event , @xmath16 . for now , we ll proceed using @xmath72 , determining what to use for this value is discussed in detail in section [ section : example : nc ] . the phase space monte carlo events we generated are then used to obtain the predicted number densities according to @xmath73,\ ] ] using the @xmath46 values given in ( [ eq : rho - no - acc - no - bkgd ] ) . for this example , the 100% acceptance makes explicitly computing the sum in the denominator unnecessary ; however , in the examples that follow it will be required . the standardized residuals , @xmath20 , for this hypothesis are obtained from ( [ eq : chi2-event ] ) as @xmath74 where @xmath26 is obtained from ( [ eq : npi - no - bkgd - no - acc ] ) . figure [ fig : chi2-no - acc - no - bkgd ] shows the @xmath36 , pull and confidence level distributions obtained for our simulated data set . the values are in excellent agreement with the theoretical distributions , _ i.e. _ they do follow a @xmath36 distribution . the total @xmath36 is obtained by summing over all events . for this example , the value is 10344 ; thus , @xmath75 ; a clear indicator that this hypothesis provides an excellent description of the data . it is also instructive to examine hypotheses which do not perfectly describe the data . if the likelihood is maximized while requiring @xmath76 , then the spin density matrix elements extracted are @xmath77 and @xmath78 . figure [ fig : chi2-no-1 - 1-no - acc - no - bkgd ] shows the @xmath36 , pull and confidence level distributions obtained for our simulated data set under this hypothesis . since the data were generated with @xmath59 , this fit provides a fair description of the data ; the total @xmath79 . maximizing the likelihood while requiring @xmath80 yields @xmath81 . figure [ fig : chi2-no - off - no - acc - no - bkgd ] shows the @xmath36 , pull and confidence level distributions obtained for our simulated data set under this hypothesis . this fit clearly provides a poor description of the data ; the total @xmath82 . the total @xmath0 values do provide a goodness - of - fit value which accurately indicates how well the fit describes the data . thus , these values can be used to ascertain the quality of the hypotheses . it is also interesting to note that the pull and confidence level distributions are also good indicators . one could plot these quantities _ vs. _ kinematic variables to determine where a given hypothesis fails to describe the data . in this section , we will examine the value of @xmath17 ( the number of nearest neighbor events used to determine the values of the radii @xmath16 ) . figure [ fig : nc_ratio](a ) shows the @xmath83 value obtained for two choices of @xmath17 ( @xmath84 and @xmath85 ) for various event sample sizes . the data sets ( with sample sizes ranging from 50 to 10,000 events ) were all generated from the same parent distributions as in section [ section : example : simple ] . each data set was fit individually to extract the @xmath46 values used to obtain the @xmath36 s . figure [ fig : nc_ratio](a ) clearly shows that the @xmath83 value obtained does depend on the choice of @xmath17 , if @xmath17 is chosen to be greater than about 2% of the total number of events . for values of @xmath17 less than 2% of the event sample size , the @xmath0 is quite stable . this behavior is expected . if @xmath17 is large relative to @xmath1 , then the method is averaging over large fractions of phase space ; thus , finer structure in the physics will not be properly accounted for . we also note here that @xmath86 as @xmath87 regardless of the quality of the fit , due to how @xmath28 is calculated ( see ( [ eq : method - num - predicted ] ) ) . figure [ fig : nc_ratio](b ) shows the value of @xmath83 obtained for the same two choices of @xmath17 for various event sample sizes for fits to the data which constrain @xmath88 . the values of @xmath0 are again stable when @xmath17 is chosen to be less than 2% of the data ; however , the values depend on the choice of @xmath17 . this behavior is again expected ; a similar effect can be seen in fits to binned data for which the value of @xmath83 depends on the number of bins . the larger the value of @xmath17 , the larger the @xmath0 can be . consider a kinematic region where the number density of the data is high , leading to a small value of @xmath16 . for this case , the largest obtainable value for @xmath35 is @xmath89 thus , there are two competing factors which should be considered when choosing the value of @xmath17 . the ratio of @xmath90 must be small enough to permit a true comparison of the finer structure of the physics ; however , the value of @xmath17 must be large enough such that the relative statistical uncertainties do not forbid larger values of @xmath20 . typically , these considerations will dictate @xmath17 be about 1%-2% of the event sample size . if this results in @xmath17 being less than 50 , _ i.e. _ if @xmath91 , then this method will most likely be less effective . of course , this statement is not all encompassing ; in practice , it will depend on the physics . ( color online ) events generated without background and with 100% acceptance . the quantities plotted are @xmath92 vs. @xmath50 ( red triangles ) and @xmath93 vs. @xmath50 ( black squares ) . , scaledwidth=45.0% ] it is also useful to examine how changes in the likelihood map onto changes in the @xmath36 values obtained using this method . here we trace out both goodness - of - fit quantities near the optimal value of @xmath50 . figure [ fig : diff - comp ] shows the changes in @xmath96 and @xmath36 _ vs _ @xmath50 for the data set used in section [ section : example : simple ] ( @xmath97 and @xmath72 ) . clearly , an increase in @xmath96 does corresponds to an increase in @xmath36 we will now extend our example by including detector acceptance . the physics used in this section will be identical to that of section [ section : example : simple ] ; however , we will now include a detector efficiency , @xmath98 , given by @xmath99 we again generated 10,000 data events and 100,000 monte carlo events from the same parent distributions as in section [ section : example : simple ] convolved with the detector acceptance given in ( [ eq : detector - acc ] ) . figure [ fig : gen - acc - no - bkgd ] shows the decay angular distribution of this data set . the effects of the detector acceptance are clearly visible when compared with figure [ fig : gen - no - acc - no - bkgd ] . an unbinned maximum likelihood fit was performed and the following values were obtained for @xmath50 : [ eq : rho - acc - no - bkgd ] @xmath100 @xmath101 @xmath102 to obtain @xmath36 values , we apply the procedure just as in section [ section : example : simple ] . figure [ fig : chi2-acc - no - bkgd ] shows the @xmath36 , pull and confidence level distributions obtained for this simulated data set . the values are again in excellent agreement with the theoretical distributions . the total @xmath103 , indicating that this fit does provide an excellent description of the data ( as it should ) . ( color online ) @xmath62 ( radians ) vs @xmath61 : events generated without background and with detector acceptance given by ( [ eq : detector - acc ] ) . , scaledwidth=45.0% ] in many physics analyses , there is a sample of non - interfering background events which can not be separated from the signal . we will now deal with this situation . the signal sample will simply be that of section [ section : example : acc ] ; thus , we will be including detector acceptance effects as well . for the background , we chose to generate it according to 3-body phase space weighted by a linear function in @xmath47 and @xmath104 in the decay angles . the number of background events generated , @xmath105 , was 10,000 ( the background events were also subjected to the detector acceptance given in ( [ eq : detector - acc ] ) ) . figure [ fig : decay - bkgd - acc](a ) shows the @xmath44 mass spectrum for all generated events and for just the background . the generated decay angular distributions for all events , along with only the signal and background are shown in figures [ fig : decay - bkgd - acc](b ) , ( c ) and ( d ) . there is clearly no way to separate out the signal events through the use of a cut . the method used to extract the signal events is described in detail in @xcite . it accurately preserves all kinematic correlations in the data while separating signal and background events . each event is assigned a signal weight factor , @xmath106 , or equivalently , a background weight factor , @xmath107 . these @xmath106-factors are then used to weight each event s contribution to the `` log likelihood '' during unbinned maximum likelihood fits to extract physical observables . the method works in a very similar way to the one presented in this paper . a metric is defined in the space of all relevant kinematic variables and the @xmath17 nearest neighbor events ( we chose @xmath72 ) are selected . for this example , the metric defined in ( [ eq : example - metric ] ) can again be used . since each subset of events occupies a very small region of phase space , the @xmath47 distributions can be used to determine each event s @xmath106-factor while preserving the correlations present in the remaining kinematic variables ( @xmath61 and @xmath62 ) . to this end , unbinned maximum likelihood fits were carried out for each event , using its nearest neighbors , to determine the parameters @xmath108 in the p.d.f . @xmath109 where , @xmath110 parameterizes the signal as a voigtian ( convolution of a breit - wigner and a gaussian ) with mass @xmath111 gev / c@xmath49 , natural width @xmath112 gev / c@xmath49 and resolution @xmath113 . the parameter @xmath114 sets the overall strength of the signal . the background , in each small phase space region , was parameterized by the linear function @xmath115 the @xmath106-factor for the event was then calculated as @xmath116 where @xmath117 is the event s @xmath118 mass and @xmath119 are the estimators for the parameters obtained from the @xmath18 event s fit . figures [ fig : decay - bkgd - acc](e ) and ( f ) show the extracted signal and background distributions , _ i.e. _ they show the events weighted by @xmath106 and @xmath107 respectively . the agreement with the generated distributions ( see figures [ fig : decay - bkgd - acc](c ) and ( d ) ) is very good . we conclude this section by noting that the full covariance matrix obtained from each event s fit , @xmath120 , can be used to calculate the uncertainty in @xmath106 as @xmath121 the likelihood function used to obtain the spin density matrix elements , defined in ( [ eq : log - likelihood ] ) , can be modified to account for the presence of background events as follows : @xmath122 thus , the @xmath106-factors are used to weight each event s contribution to the likelihood . using the @xmath106-factors obtained in section [ section : example : q - factors ] , minimizing ( [ eq : log - l - q ] ) yields @xmath123 @xmath124 @xmath125 which are in good agreement with the generated values . the metric used to obtain the @xmath17 nearest neighbor events is again ( [ eq : example - metric ] ) ; however , the number of measured events contained within each hypersphere is now given by @xmath126 with uncertainty @xmath127 where the sums ( @xmath128 ) are over the @xmath17 events used to calculate the @xmath18 event s @xmath106-factor and @xmath129 is the correlation factor between events @xmath130 and @xmath131 . this factor is equal to the fraction of shared nearest neighbor events used in calculating the @xmath106-factors for these events . thus , ( [ eq : n_m - error ] ) accounts for both the statistical uncertainty in the signal yield , along with the uncertainty in the signal - background separation , _ i.e. _ the uncertainty in the @xmath106-factors . the number of predicted events , along with its uncertainty , is calculated the same way as in section [ section : example : simple ] and the @xmath20 values are again obtained using ( [ eq : chi2-event ] ) . figure [ fig : chi2-acc - bkgd ] shows the @xmath36 , pull and confidence level distributions obtained for our simulated data set . the values are again in excellent agreement with the theoretical distributions . this indicates that we have properly handled the uncertainties in the signal . the total @xmath132 . calculating the correlation factor , @xmath129 , in ( [ eq : n_m - error ] ) is very cpu intensive due to the multiple loops which need to be performed over the data events . in @xcite , we noted that assuming 100% correlation for all @xmath106-factors provides a reasonable upper limit on the signal yield in any kinematic region . under this assumption , which eliminates much of the bookkeeping and cpu cycles needed to properly calculate @xmath129 , the uncertainty on @xmath133 becomes @xmath134 where the sum is again over the @xmath17 nearest neighbor events used to calculate @xmath135 . figure [ fig : chi2-cor - acc - bkgd ] shows the @xmath36 , pull and confidence level distributions obtained for our simulated data set under this assumption . the deviations from the theoretical distributions are relatively small . the total @xmath136 . this value is smaller than 1 as expected ; however , it is still reasonably close to the value obtained when the errors are handled rigorously . thus , this is a reasonable approximation and could be used effectively in situations where calculating the values of @xmath129 is not possible ( or , perhaps , desirable ) . to extend this example to allow for the case where the data is not binned in production angle , we would simply need to include @xmath137 or @xmath42 in the vector of relevant coordinates , @xmath3 . to perform a full partial wave analysis on the data , we would also need to include any additional kinematic variables which factor into the partial wave amplitudes , _ e.g. _ the distance from the edge of the @xmath44 dalitz plot ( typically included in the @xmath43 decay amplitude ) . we would then construct the likelihood from the partial waves and minimize @xmath96 using the @xmath106-factors obtained by applying our signal - background separation procedure , including the additional coordinates . the estimators from these event - based fits would then serve as input into a higher dimensional version of the method presented in this example . the procedure for obtaining @xmath20-values , however , would remain almost unchanged . we would simply need to account for the additional relevant kinematic variables in our metric . if the fit provides a good description of the data , then the @xmath0 should be approximately one and the squares of the standardized residuals should follow a @xmath36 distribution . in this paper , we have presented an _ ad - hoc _ procedure for obtaining @xmath0 values from unbinned maximum likelihood fits which does not require binning the data . this makes it very applicable to multi - dimensional problems . we have shown that these @xmath0 values accurately reflect how well a given hypothesis describes the data . this work was supported by grants from the united states department of energy no . de - fg02 - 87er40315 and the national science foundation no . 0653316 through the `` physics at the information frontier '' program . 99 j. friedman , phys . rev . * d9 * , 11 ( 1973 ) . mark f. schilling , the annals of statistics , * 11 * , 13 , ( 1983 ) . mark f. schilling , journal of the american statistical association , * 81 * , 799 ( 1986 ) . b. aslan and g. zech , arxiv : hep - ex/0203010 ( 2003 ) . k. kinoshita , arxiv : physics/0312014 ( 2003 ) . r. raja , arxiv : physics/0401133 ( 2003 ) . v. h. regener , phys . rev . * 84 * , 161 ( 1951 ) . k. schilling , p. seyboth and g. wolf , nucl b15 * , 397 ( 1970 ) . m. williams , m. bellis and c. a. meyer , arxiv:0804.3382 ( submitted to nim ) .	a common goal in an experimental physics analysis is to extract information from a reaction with multi - dimensional kinematics . the preferred method for such a task is typically the unbinned maximum likelihood method . in fits using this method , the likelihood is a goodness - of - fit quantity in that it effectively discriminates between available hypotheses ; however , it does not provide any information as to how well the best hypothesis describes the data . in this paper , we present an _ ad - hoc _ procedure for obtaining @xmath0 values from unbinned maximum likelihood fits . this method does not require binning the data , making it very applicable to multi - dimensional problems . and 11.80.et,29.85.fj
extensive neutron - scattering experiments clarified a close relation between the spin correlations and the superconductivity in high-_t_@xmath13 superconductors@xcite . in the prototypical high-_t_@xmath13 superconductor la@xmath5sr@xmath2cuo@xmath4 ( lsco ) , in which the doping - dependence of physical properties can be easily investigated by changing sr concentration , following results were obtained : ( i ) the incommensurate ( ic ) spin correlations were observed in a whole superconducting ( sc ) phase with 0.055@xmath14@[email protected]@xcite . ( ii ) the direction of the ic wavevector in the sc phase differs by 45@xmath15 in angle from that observed one in the spin - glass phase for @[email protected]@xcite . ( iii ) the incommensurability ( @xmath17 ) is proportional to _ t_@xmath13 in underdoped region@xcite . these experimental facts naturally suggest that the origin of ic spin correlations are important to be clarified for the understanding of role of magnetism in the mechanism of high-_t_@xmath13 superconductivity . on the other hand , it is known that the distortion of cuo@xmath0 planes can affect the superconductivity and the spin correlations . systematic neutron - scattering of la@xmath18ba@xmath19sr@xmath2cuo@xmath4 , in which the crystal structure varies from low - temperature - tetragonal ( ltt ) phase to low - temperature - orthorhombic ( lto ) phase upon increasing @xmath11 , clarified the enhancement of low - energy component of spin fluctuations and the stability of charge stripe order in the ltt phase , where the @xmath12 is well suppressed@xcite . furthermore , a resistivity measurement performed under high pressures showed the strong correlation between the corrugation of cuo@xmath0 planes and @xmath12 in the lto phase and suggests a higher @xmath12 in the structure with flat cuo@xmath0 planes @xcite . therefore , it is interesting to study the relationship between the spin correlation and the corrugation of cuo@xmath0 planes in order to extract an intrinsic interplay between the superconductivity and the spin correlation . motivated by the above reason , we have performed neutron - scattering experiment on + la@xmath1sr@xmath2ce@xmath3cuo@xmath4 ( lscco ) , which is co - doped system of sr@xmath20 and ce@xmath21 ions into la@xmath0cuo@xmath22 . the corrugation of cuo@xmath0 planes is relaxed with increasing sr concentration@xcite , and ce doping is considered to reduce the total hole concentration without a remarkable change in the corrugation of cuo@xmath0 planes . thereby , we can prepare less - corrugated cuo@xmath0 planes in lscco compared with lsco system at similar hole concentration ( @xmath6 ) . = 4mev and @xmath10=8mev observed in lscco ( a ) @xmath11=0.14 , ( b ) @xmath11=0.18 , and ( c ) @xmath11=0.24 ( topan ) . the vertical scale is normalized by neutron counts monitor . total counting time corresponds to @xmath75.9 minutes ( @xmath10=4mev ) and @xmath75.4 minutes ( @xmath10=8mev ) . no clear magnetic signal is observed on @xmath11=0.18 and @xmath11=0.24 , in contrast , well - defined incommensurate peak is observed on @xmath11=0.14 at @xmath10=4mev , width=241 ] for the experiment , we prepared the single crystals of lscco with @xmath11=0.14 ( @[email protected] ) , @xmath11=0.18 ( @[email protected] ) , and @xmath11=0.24 ( @[email protected] ) , corresponding to underdoped , optimally - doped , and slightly - overdoped sample , respectively . the crystals grown by using the traveling - solvent floating - zone method with the typical length of @xmath7100 mm and diameter of @xmath78 mm were annealed in o@xmath23 gas flow and cut into @xmath735mm - length pieces . two of cut samples were assembled so that the cuo@xmath0 planes in the horizontal scattering plane , and rest part of samples are further sliced for the magnetic susceptibility measurement . = 0.5mev , and ( c ) @xmath10=1.0mev observed in lscco @xmath11=0.14 ( her ) . all magnetic spectra show well - defined incommensurate signals.,width=241 ] inelastic neutron - scattering measurements for 2 mev@xmath14@xmath10@xmath1412 mev were performed on the thermal neutron triple - axis spectrometer topan installed in the reactor of jrr-3 at japan atomic energy agency ( jaea ) . we fixed the final neutron energy ( @xmath24 ) at 13.5mev and used a typical horizontal collimation sequence of 50@xmath25 - 100@xmath25-sample-60@xmath25 - 180@xmath25 . to reduce the flux of high energy neutrons and cut the higher - order wavelengths neutrons , sapphire crystal and the pyrolitic graphite were placed before and after sample , respectively . elastic and low - energy ( @xmath10@xmath142 mev ) inelastic neutron - scattering measurements with high experimental resolution were done on lscco @xmath11=0.14 by using the cold neutron triple - axis spectrometer her installed in the guide - hall of jrr-3 . in the measurement at her , @xmath24 of 5mev and the horizontal collimation sequence of 32@xmath25 - 40@xmath25-s-80@xmath25(180@xmath25)-80@xmath25(180@xmath25 ) for elastic ( inelastic ) measurement were selected . in the elastic measurement , be - filter was put before the sample . in this paper , crystallographic indexes are denoted as ( @xmath26 @xmath27 0 ) in the tetragonal i4/mmm notation . in order to determine the superconducting transition temperature _ t_@xmath13 , we measured the magnetic susceptibility with using a superconducting quantum interference device magnetometer . evaluated _ t_@xmath13 s were @xmath730k ( @xmath11=0.14 ) , @xmath735k ( @xmath11=0.18 ) , and @xmath725k ( @xmath11=0.24 ) , respectively . we , furthermore , determined @xmath28 s for several crystals with the @xmath11-range from 0.10 to 0.30 to make a phase diagram for the lscco system . from a comparison of phase diagram of the present lscco and lsco systems , we concluded that 6% ce doping reduces @xmath74% hole concentration . in addition , it was confirmed by our neutron - scattering measurement that the ce - doping does not affect both degree of in - plane orthorhombic distortion at low temperature and the structural transition temperature from high - temperature - tetragonal ( htt ) to lto @xmath29 . therefore , the superconductivity occurs on the less - corrugated cuo@xmath0 planes in lscco compared with lsco with similar hole concentration . detailed results will be presented elsewhere@xcite . figure 1 shows the constant-@xmath10 spectra measured at topan . well - defined ic peaks were observed in all samples at @xmath10=8 mev ( closed circles ) . in contrast , no clear magnetic signal was observed in the spectra measured at @xmath10=4 mev ( open circles ) for the @xmath11=0.18 and @xmath11=0.24 , while the @xmath11=0.14 sample shows sharp ic peaks at this energy . the incommensurability ( @xmath17 ) , corresponding to the half distance between the two peaks , is @xmath7 0.12 r.l.u . in all samples . . corresponds to 1.67 @xmath30@xmath31 . ) thus , @xmath17 is independent on @xmath11 in the measured concentration range . absence of well - defined magnetic signal at low - energy in the @xmath11=0.18 and 0.24 samples is consistent with the opening of a spin - gap , which is reported for the optimally - doped lsco . on the other hand , the persistence of magnetic signal with @xmath10 down to 2 mev in the @xmath11=0.14 sample is rather analogous to the results for the underdoped lsco exhibiting a short range magnetic order at low temperature . we next pay our attention to the low - energy spin fluctuations below 2 mev and the possible static magnetic order . elastic magnetic signal was searched for all samples using the thermal neutron spectrometer and the signal was observed only in the @xmath11=0.14 sample . then we investigated the low - energy and the static spin correlations in the @xmath11=0.14 with higher experimental resolution ( @xmath32@[email protected] and @xmath32@[email protected]@xmath34 ) at the cold triple - axis spectrometer . figure 2 shows the elastic and the representative inelastic spectra . existence of ic magnetic peaks at @xmath10=0.5mev suggests a gap - less spin excitation spectrum in the @xmath11=0.14 sample ( fig.2 ( b ) ) . furthermore , as seen in fig . 2(a ) , the elastic magnetic intensity was reconfirmed to appear at low temperatures . from the temperature - dependence of peak - intensity , the onset temperature for the appearance of intensity ( @xmath35 ) was determined to be @xmath720 k. @xmath9(@xmath10 ) on ( a)@xmath11=0.14 , ( b)@xmath11=0.18 , and ( c)@xmath11=0.24 . filled circle and open circle represents using the data on topan and her , respectively . clear gap structure exist in spin excitation on @xmath11=0.18 and @xmath11=0.24 . otherwise , @xmath11=0.14 has gap - like structure in the magnetic excitation.,width=264 ] for the evaluation of local spin susceptibility @xmath8@xmath9(@xmath10 ) , which corresponds to @xmath33-integrated intensity after correcting the thermal factor , we take into account four isotropic peaks at ( 0.5@xmath36@xmath17 0.5 0 ) and ( 0.5 0.5@xmath36@xmath17 0 ) in the analysis . then the following gaussian function was fitted to the observed spectra with convoluting experimental resolution , @xmath37 where @xmath33@xmath38 and @xmath39@xmath40 represent af wavevectors of ( 0.5 0.5 0 ) and ic wavevectors of ( @xmath17 0 0)/(0 @xmath17 0 ) , respectively , and @xmath41 is the peak - width ( half - width at half - maximum ) . @xmath10-dependence of @xmath8@xmath9(@xmath10 ) for three samples is plotted in fig . the results obtained for the @xmath11=0.14 sample at thermal and cold spectrometers are normalized with the values at 2 mev . the magnetic excitation spectra in the @xmath11=0.18 and @xmath11=0.24 samples show a clear gap - structure , while @xmath8@xmath9(@xmath10 ) in the @xmath11=0.14 sample has finite values in the all measured @xmath10-range below 12 mev . in the latter , though the decrease of @xmath8@xmath9(@xmath10 ) below @xmath7 7 mev is resemble to the gap - structure observed in the @xmath11=0.18 and 0.24 samples , the intensity start to develop below 2 mev with decreasing @xmath10 . here , we discuss the structural effects on the spin correlations from the comparison between the results for lscco and lsco with comparable hole concentrations . first , we consider the gap - structure in the lscco with @xmath11=0.18 and 0.24 samples . the energy spectra for both samples are nearly the same with that for the optimally - doped lsco , although the in - plane lattice distortion is smaller in the lscco system@xcite . therefore , the gap - structure near the optimally doped region is not affected by a small lattice distortion . we note that the appearance of the energy - gap by ce doping into the sr- overdoped sample is one of strong evidence of the reduction of hole concentration by the formation of ce@xmath21 state . we hence emphasize that the absence of spin - gap in the overdoped lsco ( @[email protected])@xcite is not an extrinsic chemical disorder effect of dopants but an inherent property in overdoped region . second , the gap - like structure observed in the underdoped lscco with @xmath11=0.14 is contrastive to the gapless structure in the lsco with @xmath11=0.10 , although the hole concentration in these two samples is comparable . difference in the @xmath8@xmath9(@xmath10 ) between the two samples is prominent in the low - energy part below around 7 mev . therefore the absence of spin - gap in the underdoped lsco sample can be ascribed by the slowing down of spin fluctuations . we speculate carrier localization on the corrugated cuo@xmath0 planes is one of the main reasons for the slowing down of spin fluctuations . the stronger carrier localization occurs when the carrier concentration is smaller . when spin and charge stripes are formed , the carrier localization corresponds to the pinning of stripes by the lattice potential on the corrugated cuo@xmath0 planes . pinning of stripes could take place even in the lto phase , although the pinning effect in the lto phase is weaker than that in the ltt phase . in this context , the observed gap - like structure in @xmath8@xmath9(@xmath10 ) for the lscco with @xmath11=0.14 sample suggests that spin - gap states intrinsically exist even in the underdoped la214 system , which easily becomes unclear due to the structural effect . the low - energy spin fluctuations below 2 mev and the static short range order observed in the present lscco with the @xmath11=0.14 sample would be induced by residual corrugation of cuo@xmath23 planes and/or disorder of large amount of doped sr and ce ions@xcite . evidence for the existence of the low - eneryg spin fluctuations below the gap energy in the lsco system is recently reported from the inelastic neutron scattering measurement done under magnetic fields@xcite . to clarify the intrinsic spin correlations in the high-@xmath12 superconductors , further systematic neutron - scattering experiments are needed . kastner , r.j . birgeneau , g. shirane , and y. endoh , rev . * 70 * ( 1998 ) , 897 . birgeneau , d.r . gabbe , h.p . jenssen , m.a . et al_. , phys . b * 38 * ( 1988 ) , 6614 . k. yamada , c.h . lee , k. kurahashi , j. wada _ et al_. , phys . b * 57 * ( 1998 ) , 6165 . m. matsuda , m. fujita , k. yamada , r.j . et al_. , phys . b * 65 * ( 2002 ) , 134515 . m. fujita , h. goka , k. yamada , and m. matsuda , phys . * 88 * ( 2002 ) , 167008 . f. nakamura , t. goko , j. hori , y. uno _ et al_. , phys . b * 61 * ( 2000 ) , 107 . fleming , b. batlogg , r.j . cava , and e.a . rietman , phys . b * 35 * ( 1987 ) , 7191 . m. enoki , m. fujita , and k. yamada , in preparation . lee , k. yamada , h. hiraka , c.r . venkasteswara rao , and y. endoh . phys . rev . b * 67 * ( 2003 ) , 134521 . lee , k. yamada , y. endoh , g. shirane _ et al_. , j. phys . * 69 * ( 2000 ) , 1170 . s. wakimoto , h. zhang , k. yamada , i. swainson _ et al_. , phys . * 92 * ( 2004 ) , 217004 . t. goko , k. shiraishi , m. kawazu , s. takeshita _ et al_. , j. mag . mag . mat . * 310 * ( 2007 ) e99 . j. chang , a.p . schnyder , r. gilardi , h.m . et al_. , phys . * 98 * ( 2007 ) 077004 .	we performed a neutron - scattering experiment to investigate the effect of distortion of cuo@xmath0 planes on the low - energy spin correlation of la@xmath1sr@xmath2ce@xmath3cuo@xmath4 ( lscco ) . due to the carrier - compensation effect by co - doping of sr and ce , lscco has a smaller orthorhombic lattice distortion compared to la@xmath5sr@xmath2cuo@xmath4 ( lsco ) with comparable hole concentration @xmath6 . a clear gap with the edge - energy of 6@xmath77 mev was observed in the energy spectrum of local dynamical susceptibility @xmath8@xmath9(@xmath10 ) for both @xmath11=0.18 ( @[email protected] ) and @xmath11=0.24 ( @[email protected] ) samples as observed for optimally - doped lsco ( @[email protected] ) . for the @xmath11=0.14 ( @[email protected] ) sample , in addition to the gap - like structure in @xmath8@xmath9(@xmath10 ) we observed a low - energy component within the gap which develops below 2@xmath73mev with decreasing the energy . the low - energy component possibly coincides with the static magnetic correlation observed in this sample . these results are discussed from a view point of relationship between the stability of low - energy spin fluctuations and the distortion of cuo@xmath0 planes . high-@xmath12 superconductivity , spin fluctuations , neutron - scattering
in the first paper in this series @xcite , we developed a _ t_-matrix theory of light scattering by spherical but optically anisotropic either radially or uniformly scatterers . whereas for radial anisotropic scatterers it is possible to derive closed form expressions for the elements of the t - matrix , in the uniformly anisotropic case this is no longer true . to further complicate matters , in this latter case the _ t_matrix elements can only be derived numerically by inverting a rather difficult set of matrix equations . because the spherical symmetry of a usual scattering problem is reduced in this case to cylindrical symmetry , the scattering involves angular momentum mixing . the consequence is that the set of equations to be inverted is in principle infinite in number . in this paper we continue this study . we find that notwithstanding the daunting nature of the formal structure of the _ t_matrix theory , in fact this theory provides a viable and efficient strategy for calculating the elements of the _ t_matrix , and thus computing the scattering properties of these objects . throughout this paper we shall refer to the first paper in this series as i , and write eq . ( n ) of i as eq . ( n.i ) . a number of approaches are available to study light scattering by complex objects . a brief summary is as follows . the scattering amplitudes can be described using green s function techniques @xcite , but these involve solving complex integral equations over infinite domains . under some circumstances one can approximate the kernels of these equations either as the incident wave or as a semi - classical perturbed wave , leading to the well - known rayleigh - gans ( r - g ) and anomalous diffraction approximations ( ada ) . these have been used by umer and coworkers to examine the problems we consider in this paper @xcite . however , the approximations are only valid over certain wavelength and optical contrast regimes . the century - old mie strategy and its modern _ t_matrix extensions yield exact solutions , but unfortunately this approach does not work in every case . finally one can of course use real space finite element approaches , but these are notoriously inefficient at reproducing known solutions . for a more comprehensive review we refer the reader to chap . 2 in @xcite and references therein . the _ t_-matrix theory _ is _ known to be a computationally efficient approach to study light scattering by nonspherical optically isotropic particles @xcite . one may thus expect that a _ t_matrix approach to geometrically spherical but optically non - spherical scatterers can at the very least enable scattering properties to be evaluated when the approximate methods can not be applied . in addition , whereas the region of validity of the approximate methods such as r - g and ada in the case of isotropic scatterers is reasonably well - understood , in the case of anisotropic scatterers this problem has not been been studied in any detail . in i we have discussed composite spherical scatterers , consisting of a central isotropic core plus a surrounding annular layer in which the optical tensor is anisotropic : @xmath0 $ ] , where @xmath1 is the anisotropy parameter . for radial anisotropy the optical axis , @xmath2 , is directed along the radius vector , @xmath3 , and for the uniform anisotropy the optical axis is parallel to the @xmath4-axis , @xmath5 . these cases present different mathematical challenges to the theorist . light scattering from the radially anisotropic annular layer was first studied long ago by roth and digman @xcite using the technique normally known as debye potentials . in an earlier paper @xcite we have recovered this solution as a special case of a more general set of anisotropies . a crucial step in the derivation of this result involves writing the so - called modified _ t_-matrix ansatz ( see eqs . ( 26.i)-(27.i ) ) . the spherical symmetry of the scatterer requires the modes ( 27.i ) entering the ansatz ( 26.i ) to be proportional to the corresponding vector spherical harmonics . as a result the scattering does not mix different angular momenta . t_matrix is then diagonal over the angular momentum indices and the azimuthal numbers and the elements of the _ t_-matrix are expressible in closed form ( see sec . iv.a of i ) . the uniform anisotropy case is much more difficult . the light scattering problem for a uniformly anisotropic spherical scatterer is not exactly solvable @xcite . as an alternative , it has been studied by using r - g and ada @xcite . in i we approached this problem by formulating it as a suitably modified _ t_-matrix theory . it is necessary to relate the plane wave packet representation to expansions of electromagnetic fields over vector spherical harmonics . the net result is to define a set of wave functions representing _ exact solutions _ of maxwell s equations in an anisotropic layer that are at the same time deformed spherical harmonics . the coefficient functions that enter the expressions for these wave functions describe angular momentum mixing and are computationally accessible . in this paper we find numerical results for the total scattering cross - section in the limiting case of a droplet , i.e. when the radius of the isotropic core of the scatterer , @xmath6 , is negligible ( @xmath7 ) . the scattering by a droplet presents fewer technical difficulties than scattering by the annular layer which was our principal focus in i. anisotropy effects are our primary concern and for this reason we pay special attention to the case where the scattering can be solely attributed to the presence of the anisotropy . specifically , in our subsequent calculations we consider the special case for which the ordinary wave refractive index and the refractive index of the material are matched , @xmath8 or @xmath9 . the paper is organised as follows . in sec . [ sec : t - matrix ] we use the general _ t_-matrix formalism developed in i to write down some necessary formulae relevant to the special case of droplets . in sec . [ sec : num - res ] we make brief comments on the numerical strategy and present some numerical results . finally in sec . [ sec : concl ] we make some brief concluding remarks . in this section we adapt the key theoretical relations derived in i so that they can be used in the case of droplets . in addition , we rewrite the expressions for the total scattering cross - section in a more convenient form . we first introduce some notation . radially anisotropic droplets present an isotropic face to the world , to the extent that the scattering properties are independent of the the direction of the incident wave . in the case of the uniformly anisotropic droplets , this is no longer true . in this case the scattering geometry is shown in fig . [ fig : droplet ] . the angle of incidence @xmath10 is the angle between the direction of incidence @xmath11 and the direction of the uniform anisotropy . the @xmath12 direction is perpendicular to the plane made from these two directions . we shall show explicitly that the scattering process does not involve the @xmath12-component , @xmath13 , of the incoming plane wave @xmath14 provided the refractive indices @xmath15 and @xmath16 are matched . in this paper we put the magnetic permittivity equal to the unit and slightly change the notations : @xmath17 and @xmath18 instead of @xmath19 and @xmath20 . this corresponds to the physical condition in which there is a matching condition for the refracted ordinary wave inside the scattering droplet . since the electromagnetic field must be regular at the origin , the harmonics inside the droplet are now given by [ eq : core ] @xmath21 where @xmath22 and @xmath23 is the optical contrast . the modes @xmath24 and @xmath25 have been given by eqs . ( 27.i ) for radial anisotropy and by eqs . ( 43.i ) for uniform anisotropy . following eqs . ( 32.i ) , the continuity conditions at the outside of the droplet , @xmath26 , can then be written in matrix notation as follows : @xmath27 @xmath28 ' & 0\\ 0 & j_j(\rho)\\ 0 & -n^{-1 } [ j_j(\rho ) ] ' \end{pmatrix}\ , , \quad \tilde{{\boldsymbol{\gamma}}}^{\,j}(r)= \begin{pmatrix } h_j^{(1)}(\rho ) & 0\\ n [ h_j^{(1)}(\rho ) ] ' & 0\\ 0 & h_j^{(1)}(\rho)\\ 0 & -n^{-1 } [ h_j^{(1)}(\rho ) ] ' \end{pmatrix}\ , , \label{eq : gamma}\ ] ] where the index 1 indicates that matrix elements are calculated at the boundary of droplet , @xmath26 . in the case of a radially anisotropic droplet the matrix on the left hand side of eq . is diagonal over angular momentum numbers @xmath29 ' & 0\\ 0 & j_{\tilde{j}}(\rho_o)\\ 0 & -n_o^{-1 } [ j_{\tilde{j}}(\rho_o ) ] ' \end{pmatrix}\ , , \quad \tilde{j}(\tilde{j}+1)=j(j+1)/(1+u)\ , . \label{eq : r_rad}\ ] ] when the droplet is uniformly anisotropic , the matrix @xmath30 is no longer diagonal over angular momentum quantum numbers . in this case it takes the form : @xmath31 where the coefficient functions are given by eqs . ( c3.i)-(c10.i ) in appendix c of i. the system can be then simplified by multiplying both sides by the matrices [ eq : aux_matr ] @xmath32 ' & -n^{-1}\,j_j(\rho ) & 0 & 0\\ 0 & 0 & n^{-1}\,[j_j(\rho ) ] ' & j_j(\rho ) \end{pmatrix}\ , , \\ & \tilde{{\mathbf{h}}}^{\,j}(r)= -i\,\rho^{2}\ , \begin{pmatrix } [ h_j^{(1)}(\rho ) ] ' & -n^{-1}\,h_j^{(1)}(\rho ) & 0 & 0\\ 0 & 0 & n^{-1}\,[h_j^{(1)}(\rho ) ] ' & h_j^{(1)}(\rho ) \end{pmatrix}\ , .\end{aligned}\ ] ] using the fact that the wronskian for spherical bessel functions is given by @xcite : @xmath33 we derive a system equivalent to eqs . in the following form : [ eq : sys_fin ] @xmath34 where @xmath35 and @xmath36 . these are the equivalent of the equations ( 51.i ) for scattering by a spherical annulus . the equations are considerably simpler , as a result of what might be called mode decoupling . the crucial point is that the normal modes inside the droplet are all regular at the origin , and thus two types of modes which appear in the case of the annulus do not appear here . this reduces the number of variables in the problem by one half , even in the case of the uniform anisotropy which , at least in principle , presented a certain number of problems in the annular case . from eq . we can immediately derive an expression for the _ t_-matrix : @xmath37_{jj';\,m}\ , . \label{eq : t_matr}\ ] ] for radial anisotropy all the matrices on the right hand sides of eqs . are diagonal . so , it is easy to write down the result for the _ t_-matrix : @xmath38 where [ eq : t_matr_el_rad ] @xmath39_1\,[j_j(\rho)]'_1- m_o\,[j_j(\rho_o)]'_1\,[j_j(\rho)]_1 } { [ j_j(\rho_o)]_1\,[h_j^{(1)}(\rho)]'_1- m_o\,[j_j(\rho_o)]'_1\,[h_j^{(1)}(\rho)]_1 } \ , , \label{eq : t11_rad}\\ & t_j^{22}=-\frac { m_o\,[j_{\tilde{j}}(\rho_o)]_1\,[j_j(\rho)]'_1- [ j_{\tilde{j}}(\rho_o)]'_1\,[j_j(\rho)]_1 } { m_o\,[j_{\tilde{j}}(\rho_o)]_1\,[h_j^{(1)}(\rho)]'_1- [ j_{\tilde{j}}(\rho_o)]'_1\,[h_j^{(1)}(\rho)]_1 } \ , . \label{eq : t22_rad}\end{aligned}\ ] ] these formulae bear close resemblance to the well known mie expressions @xcite . we have seen in i that the total scattering cross - section for the uniformly anisotropic scatterer depends on the angle of incidence @xmath40 . the scattering cross - section also depends on the polarisation of the incident wave . in order to emphasise this , let us express the coefficients of the incident wave @xmath41 and @xmath42 ( see eq . ( 9.i ) ) as follows @xmath43 where @xmath44^{1/2 } d_{jm}^{(y)}({\mathbf{\hat k}}_{inc})\ , , \quad \alpha_{jm;\,y}^{(inc)}= i^{j+1}\ , [ \pi(2j+1)]^{1/2 } d_{jm}^{(x)}({\mathbf{\hat k}}_{inc})\ , , \notag\\ & \tilde\alpha_{jm;\,x}^{(inc)}=n\ , i^{j+1}\ , [ \pi(2j+1)]^{1/2 } d_{jm}^{(x)}({\mathbf{\hat k}}_{inc})\ , , \quad \tilde\alpha_{jm;\,y}^{(inc)}=-n\,i^j\ , [ \pi(2j+1)]^{1/2 } d_{jm}^{(y)}({\mathbf{\hat k}}_{inc})\ , .\end{aligned}\ ] ] from eq . ( 19.i ) the scattering efficiency then can be rewritten in the following form : @xmath45\ , , \label{eq : eff_comp}\\ \begin{pmatrix } \beta_{jm;\,\alpha}^{(sca)}\\ n^{-1}\ , \tilde{\beta}_{jm:\,\alpha}^{(sca ) } \end{pmatrix } = \sum_{j ' } { \mathbf{t}}_{jj';\,m } \begin{pmatrix } \alpha_{j'm;\,\alpha}^{(inc)}\\ n^{-1}\ , \tilde{\alpha}_{j'm;\,\alpha}^{(inc ) } \end{pmatrix}\ , . \label{eq : matr_uni}\end{gathered}\ ] ] note that for radial anisotropy @xmath46 . by contrast , in the case of uniformly anisotropic droplet we have rather strong dependence of the scattering efficiency on the polarisation of the incoming wave . when the refractive indices are matched , @xmath8 , it is expected that the scatterer does not change the @xmath12 component of the incident wave , which simply transforms into the ordinary wave inside the droplet without being affected by the scattering process . the algebraic interpretation of this fact is that the amplitudes of the scattered wave @xmath47 and @xmath48 are equal to zero . however , it is not straightforward to see that the system is consistent with this conclusion . we show this in appendix [ sec : append ] ; the proof involves using relations ( 46.i ) in which the electric and magnetic fields inside the droplet are expressed in terms of appropriate normal modes . in this section we present numerical results for the scattering efficiency defined by eqs . and . we are primarily interested in anisotropy - induced scattering . in order to concentrate on this test case , we consider the case when the refractive indices @xmath49 and @xmath50 are equal . we shall present a more comprehensive analysis of all possible cases , including the results for the angular distribution of the scattered waves , elsewhere . we begin with brief comments on numerical procedure and then proceed with the description of the calculated dependencies . it is rather straightforward to perform computations for radially anisotropic droplets . the expressions for the elements of _ t_-matrix are known and given by eqs . - . we can thus evaluate the scattering efficiency by explicitly computing the sum in the expression ( 22.i ) . for uniformly anisotropic droplets _ t_-matrix can only be computed numerically by solving the system of equations . some highlights of the results are presented below . the dependence of the scattering efficiency on the angle of incidence is shown in fig . [ fig : uni_ang_u ] . if the size parameter , @xmath51 , is not very large , the scattering efficiency @xmath52 is a monotonically increasing function of the angle of incidence , @xmath40 , in the region from 0 to @xmath53 . by symmetry @xmath54 , and so the scattering efficiency decreases in the range from @xmath53 to @xmath55 . in fig . [ fig : uni_ang_kr ] we show what happens for shorter wavelength and thus higher values of @xmath51 . now , for for relatively large values of the size parameter , the cross - section dependence on the angle is incidence is no longer monotonic . for example , at @xmath56 , the angle at which the scattering efficiency @xmath52 reaches its maximum value is no longer at @xmath53 . [ fig : eff_vs_kr ] shows the scattering efficiencies @xmath57 ( for uniform anisotropy ) and @xmath58 ( for radial anisotropy ) versus the size parameter . the scattering efficiency of uniformly anisotropic droplet has a pronounced peak located at about @xmath59 and exhibits strongly non - monotonic behaviour . by contrast , the corresponding dependence for the radially anisotropic droplet is monotonically increasing . in this case the first maximum is reached at @xmath60 , outside the range of @xmath51 shown in fig . [ fig : eff_vs_kr ] . the scattering efficiencies as a function of the anisotropy parameter , @xmath61 , at different values of the size parameter are plotted in fig . [ fig : eff_u_uni ] and fig . [ fig : eff_u_rad ] for the cases of radial and uniform anisotropies respectively . in both cases an increase in the size parameter leads to the appearance of peaks in this range of @xmath62 . as compared to radially anisotropic scatterers , the uniformly anisotropic droplets seem to be more sensitive to changes both in the size and in the anisotropy parameters . in this work we have described some of the numerical results calculated using the _ t_-matrix theory developed in paper i. in particular , we have studied the scattering efficiency of radially and uniformly anisotropic droplets in which the ordinary refractive index matches the refractive index of the material surrounding them . the assumption in which the ordinary refractive index of the droplet matches the isotropic dielectric constant in the surrounding medium is not taken in order to simplify the numerical treatment . rather in this paper we wish to study the light scattering properties which can be solely attributed to the _ anisotropic _ part of the dielectric tensor . thus we have the anisotropy effects separated out to concentrate on differences between isotropic and anisotropic optical axis distributions . clearly , the difference in symmetry causes the most crucial differences in the light scattering . for uniformly anisotropic droplets the scattering efficiency depends on the angle of incidence and the polarisation of incoming wave , whereas for radially anisotropic scatterers it does not . in other words , the scattering from radially anisotropic droplet shares some features of isotropic scatterers . the graphs plotted in figs . [ fig : eff_vs_kr]-[fig : eff_u_rad ] indicate that uniformly anisotropic droplets are more sensitive to changes in the wavelength and anisotropy parameters than are radially anisotropic droplets . our results are also consistent with results of previous studies @xcite that the internal spatial distribution of the optical axis is a factor which strongly affects light scattering from anisotropic scatterers . the results of this work can be regarded as the first step towards more comprehensive study of light scattering by anisotropic scatterers . we have demonstrated that the _ t_-matrix approach developed in i can be used in an efficient numerical treatment of the scattering problem . it is thus reasonable to expect that further progress can be made by applying this theory to more complex problems involving light scattering by optical anisotropic liquid crystalline systems and other related problems , as discussed further in the last section of i. in this appendix we show mathematically , suing our formalism , the physically obvious result that if the ordinary refractive index of a droplet matches that of the scattering medium , then there will be no scattering of the polarisation component out of the plane of the incident wave and the uniform anisotropy in the droplet . in order to do this , we first extend algebraic relations that follow from eqs . these equations give the expansion of plane wave propagating in a uniformly anisotropic medium . we can rewrite them for the plane wave inside the droplet : @xmath63=\notag\\ & = \exp(i\rho_e\,{\mathbf{\hat k}}_{inc}\cdot{\mathbf{\hat r } } ) e_x({\mathbf{\hat k}}_{inc})\bigl [ { \mathbf{e}}_x({\mathbf{\hat k}}_{inc})+\frac{u}{1+u}\,\sin\theta_{inc } \,{\mathbf{\hat z}}\bigr]+ \exp(i\rho_o\,{\mathbf{\hat k}}_{inc}\cdot{\mathbf{\hat r } } ) e_y({\mathbf{\hat k}}_{inc})\,{\mathbf{e}}_y({\mathbf{\hat k}}_{inc})\ , , \label{eq : a.1}\\ & \sum_{j'm ' } \left[\ , \tilde\alpha_{j'm'}^{(inc)}\,{\mathbf{q}}^{(m)}_{j'm'}(\rho_o,{\mathbf{\hat r } } ) + n_o\ , \alpha_{j'm'}^{(inc)}\,{\mathbf{q}}^{(e)}_{j'm'}(\rho_o,{\mathbf{\hat r } } ) \,\right]=\notag\\ & = n_o\left(\ , \exp(i\rho_e{\mathbf{\hat k}}_{inc}\cdot{\mathbf{\hat r}})\ , m_e^{-1}\,e_x({\mathbf{\hat k}}_{inc}){\mathbf{e}}_y({\mathbf{\hat k}}_{inc})- \exp(i\rho_o\,{\mathbf{\hat k}}_{inc}\cdot{\mathbf{\hat r } } ) e_y({\mathbf{\hat k}}_{inc}){\mathbf{e}}_x({\mathbf{\hat k}}_{inc } ) \,\right)\ , , \label{eq : a.2}\end{aligned}\ ] ] @xmath68 = \notag\\ & = \bigl\langle\ , { \mathbf{y}}_{jm}^{(\alpha)\,}({\mathbf{\hat r}})\cdot \sum_{j'm ' } \left[\ , \alpha_{j'm'}^{(inc)}\,{\mathbf{p}}^{(m)}_{j'm'}(\rho_o,{\mathbf{\hat r } } ) - n_o^{-1}\ , \tilde\alpha_{j'm'}^{(inc)}\,{\mathbf{p}}^{(e)}_{j'm'}(\rho_o,{\mathbf{\hat r } } ) \,\right ] \,\bigr\rangle_{{\mathbf{\hat r}}}\ , , \label{eq : a.3}\\ & \sum_{j'\ge \|m\| } \left [ n_o\,q_{jj';\,m}^{\,(\alpha , e)}(\rho_o)\ , \alpha_{j'm}^{(inc ) } + q_{jj';\,m}^{\,(\alpha , m)}(\rho_o)\ , \tilde{\alpha}_{j'm}^{(inc ) } \right ] = \notag\\ & = \bigl\langle\ , { \mathbf{y}}_{jm}^{(\alpha)\,}({\mathbf{\hat r}})\cdot \sum_{j'm ' } \left[\ , \tilde\alpha_{j'm'}^{(inc)}\,{\mathbf{q}}^{(m)}_{j'm'}(\rho_o,{\mathbf{\hat r } } ) + n_o\ , \alpha_{j'm'}^{(inc)}\,{\mathbf{q}}^{(e)}_{j'm'}(\rho_o,{\mathbf{\hat r } } ) \,\right ] \,\bigr\rangle_{{\mathbf{\hat r}}}\ , , \ ; \alpha\in\{m , e\}\ , , \label{eq : a.4}\end{aligned}\ ] ] it is seen that the elements of the column on the right hand side of this equation are the sums from the left hand sides of eqs . and . on the other hand , from eqs . - , the square bracketed sums on the right hand sides of eqs . - are the plane waves . so , the elements of the column can be evaluated as scalar products of the vector spherical functions and the vector plane waves by using eqs . ( b6.i ) and ( b8.i ) of appendix b in i.	this is the second paper in a series on light scattering from optically anisotropic scatterers embedded in an isotropic medium . the apparently complex t - matrix theory involving mixing of angular momentum components turns out to be an efficient approach to calculating scattering in these systems . we present preliminary results of numerical calculations of the scattering by spherical droplets in some simple cases . the droplets contain optically anisotropic material with local radial or uniform anisotropy . we concentrate on cases in which the scattering is due only to the local optical anisotropy within the scatterer . for radial anisotropy we find non - monotonic dependence of the scattering cross - section on the degree of anisotropy can occur in a regime for which both the rayleigh and semi - classical theories are inapplicable . for uniform anisotropy the cross - section is strongly dependent on the angle between the incident light and the optical axis , and for larger droplets this dependence is non - monotonic .
in recent literature a lot of attention has been given to systems that are able to sense electromagnetic near fields ( evanescent waves ) and even to `` amplify '' them . the superlens proposed by pendry@xcite is one of such systems . his superlens is based on a veselago medium@xcite slab . the real parts of the permittivity and the permeability of the veselago slab are both negative at a certain wavelength . thus , the eigenwaves in the slab are _ backward _ waves , i.e. the wave phase and group velocities are antiparallel . this provides negative refraction and focusing of waves in a planar slab , as was outlined by veselago.@xcite however , pendry discovered that there was also a possibility to excite surface plasmon - polaritons on the slab surfaces and , due to that , amplify near fields@xcite . the slab thickness can be of the order of the wavelength , so that the plasmon - polaritons excited at both sides of the slab are strongly coupled . under certain conditions the plasmon - polariton excited at the back surface of the slab has a much stronger amplitude than that at the front surface . such amplification of evanescent waves is the key principle in subwavelength imaging . known experimental realizations of volumetric artificial materials with negative parameters are highly anisotropic structures that utilize dense arrays of parallel thin conducting wires and variations of split - ring resonators@xcite . proposed isotropic arrangements use orthogonal sets of split rings@xcite and also three - dimensional arrays of wires@xcite . at the same time , there have been achievements in modeling of veselago materials ( and pendry lens ) with the help of @xmath0-circuits or transmission - line ( tl ) based structures.@xcite these networks do not rely on resonant response from particular inclusions , and the period of the mesh can be made very small as compared with the wavelength . these features allow realization of broadband and low - loss devices , which is extremely difficult if resonant inclusions are used . the transmission - line network approach has been successfully realized in one- and two - dimensional networks , but up to now there have been doubts if it is possible to design a three - dimensional ( 3d ) circuit analogy of the veselago medium . the difficulties arise from the fact that such a 3d network requires a common ground connector . any realization of such a ground will effectively work as a dense mesh of interconnected conductors that blocks propagation of the electromagnetic waves practically the same way as a solid metal does ( the structural period must be much less than the wavelength in order to realize an effectively uniform artificial material ) . in this paper we introduce isotropic three - dimensional transmission - line networks that overcome this difficulty . in the tl - based networks that we study , the electromagnetic energy propagates through tl sections . the inside of every tl section is effectively screened from the inside of the other sections and from the outer space . this can be very naturally imagined with a 3d cubic - cell network of interconnected coaxial cable segments : the inner conductors of the segments are soldered at the network nodes ; the same is done for the outer conductors . the whole system appears as a 3d pipe network where every pipe holds a central conductor and those conductors are crossing at the node points inside the pipes . a loaded tl network can be realized now by placing loading elements inside the `` pipes '' . to couple the waves propagating inside the tl sections with the free - space waves one will have to apply a kind of antenna array with every antenna feeding a particular tl segment . when using transmission lines loaded with bulk elements we speak of waves in the meaning of discrete waves of voltages and currents defined at the loading positions . let us note that in the tl sections _ as such _ the usual , forward waves propagate . only because of the loading the discrete voltage and current waves appear as backward ones when appropriate loading impedances are used . while completing this manuscript , we learned about another possible design of a 3d transmission - line analogy of a backward - wave material described in ref . that design is based on kron s formal representation of maxwell s equations as an equivalent electric circuit.@xcite in ref . 12 only 1d propagation was studied analytically and 3d properties were analyzed numerically . the proposed structure of 3d super - resolution lens consists of two forward - wave ( fw ) regions and one backward - wave ( bw ) region . the 3d forward - wave networks can be realized with simple transmission lines and the 3d backward - wave network with inductively and capacitively loaded transmission lines . one unit cell of the bw network is shown in fig . [ 3d_zy_tl_unit_cell ] ( the unit cell enclosed by the dotted line ) . in the 3d structure there are impedances _ z_/2 and transmission lines also along the @xmath1-axis ( not shown in fig . [ 3d_zy_tl_unit_cell ] ) . in view of potential generalizations , the loads are represented by series impedances _ z_/2 and shunt admittances _ y _ , although for our particular purpose to realize a backward - wave network , the loads are simple capacitances and inductances . the unit cell of the fw network is the same as in fig . [ 3d_zy_tl_unit_cell ] but without the series impedances @xmath2 and shunt admittance @xmath3 . the equations that will be derived for these structures can be used in various implementations , but this paper will concentrate on the case when @xmath4 and @xmath5 . first we will derive the dispersion relation for a simplified 3d bw network , i.e. we will not take into account the transmission lines . such approximation is possible at low frequencies . without the transmission line segments this derivation is quite simple and can be done by summing up all the currents that flow to the node ( @xmath6 ) and equating this sum with the current flowing to the ground ( through admittance @xmath3 ) , see fig . [ 3d_zy_tl_unit_cell ] . the result is ( u_+1,,+u_,+1,+u _ , , + 1+u_-1,,+u_,-1,+u_,,-1 - 6u _ , , ) = u_,,y simple_1 . we look for a solution of the form @xmath7 ( @xmath8 ) , and if we use @xmath9 , @xmath10 and @xmath11 , ( 1 ) can be reduced to ( e^-jq_x+e^-jq_y+e^-jq_z+e^+jq_x+e^+jq_y+e^+jq_z-6)=y , or ( q_x)+(q_y)+(q_z)=+3 . if we now insert @xmath4 and @xmath5 we get the dispersion relation for the _ lc_-loaded network : ( q_x)+(q_y)+(q_z)=-+3 . simple_3 next we want to take the transmission lines into account . the effect of the transmission lines can be derived by first evaluating a part of the three - dimensional network as the one shown in fig . [ 1d_deriv ] and deriving the relation between the current that flows into a node and the voltages of adjacent nodes . if current @xmath12 is flowing towards node ( @xmath6 ) and the current that goes into node ( @xmath6 ) from the left is @xmath13 , then using the abcd - matrix for a transmission line we get : u_1=a_tu_2+b_ti_2 u_1 , i_1=c_tu_2+d_ti_2 i_1 , where ( ccc a_t & b_t + c_t & d_t ) = ( ccc ( k_0d/2 ) & jz_0 ( k_0d/2 ) + jz_0 ^ -1 ( k_0d/2 ) & ( k_0d/2 ) ) . abcd @xmath14 in is the wavenumber of waves in the transmission lines . from and we can solve @xmath12 and @xmath15 as functions of @xmath16 and @xmath17 : i_1= , i_2=. similarly for @xmath18 and @xmath13 we get : i_3= , i_z=. i_z next we can derive two equations for the current flowing through the series impedance @xmath19 and solve @xmath17 from both of them : i_2=z^-1(u_2-u_3)=,u_2= ; i_3=z^-1(u_2-u_3)= i_3 , u_2=. u_2 if we let @xmath17 in both equations be equal , we can solve @xmath20 as a function of @xmath16 and @xmath21 : u_3= . u_3 in order to derive an equation for @xmath13 [ the current that flows into node ( @xmath6 ) from the direction of node ( @xmath22 ) ] as a function of @xmath16 and @xmath21 , we insert into and get i_z =- u_4 . if we use @xmath23 and @xmath24 , then @xmath25 . because of the symmetry we can derive the dispersion relation exactly the same way as in , and for the case @xmath4 , @xmath5 the result is ( q_x)+(q_y)+(q_z)=-3 , disp_rel_bw where s_bw= s_bw , k_bw=- k_bw . to derive the dispersion relation for the forward - wave network , we can use the equations derived for the backward - wave network letting @xmath26 and @xmath27 . this way we get from and the following equations for @xmath28 and @xmath29 : s_fw= , k_fw=-- . from we get the dispersion relation : ( q_x)+(q_y)+(q_z)=-3 . disp_rel_fw dispersion curves for backward - wave and forward - wave networks can be plotted if the values of the transmission line parameters and _ l _ and _ c _ are fixed . let us choose the parameters of the tls and the lumped components as : @xmath30 nh , @xmath31 pf , @xmath32 m ( the period of the network ) , @xmath33 ohm , @xmath34 ohm ( characteristic impedances of the tls ) . see figs . [ dispersion_bw ] and [ dispersion_fw ] for examples of dispersion curves when a @xmath1-directed plane wave is considered ( i.e. @xmath35 ) . @xmath36 , where @xmath37 is the speed of light in vacuum . notice that the bw network supports backward - waves only in the region where 0.32 [email protected] ghz . above that frequency band and the following stopband , the bw network works as a normal fw network until the next stopband appears . by tuning the capacitance @xmath39 ( or inductance @xmath40 ) , the stopband between the bw and fw regions shown in fig . [ dispersion_bw ] can be closed , see fig . [ dispersion_cvar]a , where @xmath41 pf . as can be seen from fig . [ dispersion_cvar]b , by changing the value of @xmath39 from this `` balanced '' case , the stopband is formed either by moving the edge of the fw region up ( @xmath42 ) or by moving the edge of the bw region down ( @xmath43 ) . next the characteristic impedance of the backward - wave network is derived . if we assume that the interface between the two networks is in the center of capacitor _ c _ ( see fig . [ 1d_deriv ] , where @xmath4 ) , then we can define the characteristic impedance as z_0,bw==. first we have to express @xmath17 , @xmath20 and @xmath18 ( or optionally @xmath15 ) as functions of @xmath16 and @xmath21 . we can use equations . if we insert into and , we find @xmath17 , @xmath20 and @xmath18 as functions of @xmath16 and @xmath21 . therefore we can present @xmath17 , @xmath20 and @xmath18 simply as : @xmath44 , @xmath45 and @xmath46 . because @xmath47 for a wave moving along the @xmath48-direction , the characteristic impedance can be expressed as z_0,bw== , z0_bw where a_bw= , b_bw=- , c_bw= , d_bw=- , e_bw= , f_bw=- f_bw . to derive the characteristic impedance of the forward - wave network , we can use the equations derived for the backward - wave network if we insert @xmath26 in them . if this condition applies , we get from : z_0,fw== , z0_fw where a_fw = c_fw= , b_fw = d_fw=- , e_fw= , f_fw=- . @xmath49 and @xmath50 can be plotted from and as functions of the frequency if the transmission line parameters and _ l _ and _ c _ are fixed . let us choose the parameters of the tls and the lumped components as : @xmath31 pf , @xmath30 nh , @xmath32 m , @xmath51 , @xmath33 ohm , @xmath52 ohm ( characteristic impedances of the tls ) . see figs . [ z0_fvar_bw ] and [ z0_fvar_fw ] for examples of the characteristic impedances when a @xmath1-directed plane wave is considered ( i.e. @xmath35 ) . the effect of changing @xmath39 on the characteristic impedance can be seen in fig . [ z0_fvar_bw ] , and the effect of changing @xmath53 on the characteristic impedance is shown in fig . [ z0_fvar_fw ] . notice that for the bw network the characteristic impedance is continuous only in the `` balanced '' case ( @xmath54 pf here ) , because in the stopbands the real part of the impedance is zero . we consider a perfect lens with axis parallel to the @xmath1-axis . to have perfect imaging , the lens should support all spatial harmonics ( i.e. waves with all possible transverse wavenumbers @xmath55 ) of the source field and for those values of @xmath56 , @xmath57 and @xmath58 should be equal in magnitude but opposite in sign . from fig . [ f_qzvar_bw_and_fw ] we can conclude that the matching of @xmath59 and @xmath60 ( which corresponds to a relative refraction index of @xmath61 ) can be achieved only at one frequency ( depending on the parameters of the forward - wave and backward - wave networks ) . in fig . [ f_qzvar_bw_and_fw ] this frequency is @xmath62 ghz . at this frequency the dispersion curves of the forward - wave and backward - wave networks intersect . in the analytical form this means that -3=-3 , disp_ideal as can be seen from and . the dispersion curves in fig . [ f_qzvar_bw_and_fw ] are plotted so that @xmath63 and @xmath64 are zero ( a plane wave moving along the @xmath1-axis ) . from fig . [ f_qzvar_bw_and_fw ] it is seen also that for @xmath62 ghz @xmath65 when @xmath66 . this means that the absolute value of the maximum wavenumber for propagating waves is approximately @xmath67 m@xmath68 . in addition to matching the wavenumbers ( refractive indices ) , to realize an ideal `` perfect lens '' the interfaces between the forward - wave and backward - wave networks should be also impedance - matched . if the two regions would not be matched , reflections from the interface would distort the field patterns both inside and outside the lens . as can be seen from figs . [ z0_fvar_bw ] and [ z0_fvar_fw ] , the characteristic impedances of the backward - wave network and the forward - wave network are about 40 ohms at the frequency where the wavenumbers are matched ( and when @xmath31 pf and @xmath52 ohm ) . notice that the impedances of the transmission lines in the forward - wave network have been lowered from 85 ohm to 70 ohm to achieve impedance matching of the forward - wave and backward - wave networks . next the effect of nonzero @xmath55 on the matching is considered ( at the optimal frequency ) . the minimum and maximum values of @xmath55 can be found from @xmath69 and @xmath70 . from and we can plot @xmath58 and @xmath57 as functions of the transverse wavenumber ( @xmath71 ) if we fix the frequency . now @xmath58 and @xmath57 are surfaces with variables @xmath72 and @xmath73 . by comparing these surfaces , it was seen that they are practically the same for all possible values of @xmath55 . this happens only at the frequency @xmath62 ghz , where the dispersion curves of the forward - wave and backward - wave networks intersect . because the characteristic impedances are functions of the abcd - matrices and @xmath74 ( and @xmath74 is a function of @xmath56 ) , we can plot @xmath50 and @xmath49 as functions of @xmath55 if we fix the frequency . now @xmath49 and @xmath50 are surfaces with variables @xmath72 and @xmath73 . by comparing these surfaces , it was seen that they are almost the same ( less than one percent difference ) for all possible values of @xmath55 at @xmath62 ghz . see fig . [ z0diff_ktvar ] for a 2d cut of relative difference between such surfaces ( now @xmath75 and therefore @xmath76 ) . the ideal `` perfect lens '' situation is achieved with @xmath77 ohm , as can be seen also from fig . [ z0diff_ktvar ] . when the frequency deviates from the optimal value ( for which is true ) , the wavenumbers and characteristic impedances are no longer matched . the effect of this is distortion of the image seen in the image plane of the lens . the transmission coefficient of the lens can be solved by considering the incident and reflected fields in the lens system . let us assume that the incident field outside the lens has the unit amplitude , the reflected field outside the lens has amplitude @xmath78 , the incident field inside the lens has amplitude @xmath79 , and the reflected field inside the lens has amplitude @xmath80 ( @xmath1 is the distance from the front edge of the lens ) . from these values we can form the following equations ( the length of the bw slab is @xmath81 and the transmission coefficient of the bw slab is @xmath82 , see fig . [ perfect_lens ] ) : 1+r = a+b t_1 , = t_2 , t_lens = ae^-jk_z , bw l+be^+jk_z , bw l t_3 , = . the resulting equation for the transmission coefficient @xmath83 is : t_lens(k_t)=. tthe total transmission from the source plane to the image plane ( see fig . [ perfect_lens ] ) is then ( distance from source plane to lens is @xmath84 and distance from lens to image plane is @xmath85 ) t_tot(k_t)=t_lens(k_t)e^-jk_z , fw(s_1+s_2 ) . t_tot the longitudinal wavenumber @xmath74 as a function of @xmath55 can be found from the dispersion relations . let us choose @xmath86 and @xmath75 so we can plot curves instead of surfaces . @xmath87 as a function of @xmath55 can now be plotted if the frequency is fixed . let us choose the lengths of the lens system as the following : @xmath88 m , @xmath89 m , @xmath90 m. now we can choose the frequency at which we want to calculate @xmath87 . let us study the transmission properties at the matching frequency @xmath62 ghz . from and we can plot the magnitude and phase of @xmath87 as a function of @xmath55 , see fig . [ t_tot ] , case 1 . from fig . [ t_tot ] it is seen that the `` lens '' works quite well for the propagating modes ( @xmath91 m@xmath92 m@xmath68 ) , see fig . [ source_image ] for an example of phase correction in the image plane . according to fig . [ t_tot ] , for evanescent modes ( @xmath93 m@xmath94 m@xmath68 ) the `` lens '' works only in a limited range of @xmath55 , where the absolute value of the transmission coefficient @xmath87 is greater than zero . one can notice that for evanescent modes a mismatch in @xmath74 affects mostly the phase of @xmath95 in the area of propagating waves and a mismatch in the characteristic impedances affects primarily the absolute value of @xmath95 . to improve the effect of evanescent waves enhancement , the characteristic impedances should be matched better in the evanescent wave area of @xmath55 ( i.e. @xmath96 m@xmath94 m@xmath68 ) . there are several ways to achieve a better matching of the characteristic impedances . first , there is of course a possibility to change the impedances of the transmission lines , but this is probably not practically realizable because it would require very accurate manufacturing ( even a very small deviation from the ideal impedance values destroys the effect of growing evanescent waves ) . the tuning of @xmath53 was tested and using the exact impedance required ( see fig . [ z0diff_ktvar ] ) , the transmission of evanescent waves was clearly improved . the resonance peaks in fig . [ t_tot ] were moved further away from the center and the absolute value of @xmath87 was larger than or equal to unity approximately for @xmath97 m@xmath98 m@xmath68 ) . second , there is a possibility to change the frequency and study if the impedance matching can be made better that way ( this also means that the matching of wavenumbers @xmath74 is made worse which can also destroy the effect of growing evanescent waves ) . this was tested and the best results were obtained using frequency @xmath99 ghz . the region of transmitted @xmath55 s was again increased , i.e. the resonance peaks in fig . [ t_tot ] were moved further away from the center and the absolute value of @xmath100 was larger than or equal to unity approximately for @xmath97 m@xmath98 m@xmath68 ) . the third way to enhance the growth of evanescent waves is to change the length of the `` lens '' . from it is seen that the growth of evanescent waves is destroyed by the term @xmath101 in the denominator . this term can be made smaller by decreasing the length of the `` lens '' @xmath81 . see fig . [ t_tot ] , case 2 ( @xmath102 is larger than or equal to unity approximately for @xmath103 m@xmath104 m@xmath68 ) and fig . [ source_image_evanescent ] , where the distances equal @xmath105 m , @xmath106 m and @xmath107 m. from fig . [ source_image_evanescent ] one can conclude that there is a significant growth of evanescent waves in the lens . by using the shortened lens _ and _ at the same time tuning the frequency appropriately , it was seen that the transmission coefficient could be made practically ideal ( i.e. @xmath108 and @xmath109 for all possible values of @xmath55 ) . using the shortened lens ( same values as in fig . [ source_image_evanescent ] ) and frequency @xmath110 ghz , the absolute values of evanescent fields were indeed almost the same in the image plane and in the source plane ( less than one percent difference ) . how to manufacture three - dimensional transmission line networks ? the main problem is the ground plane , which should exist in all three dimensions . one solution would be to use coaxial transmission lines ( regular in the forward - wave network and loaded with lumped _ l_- and _ c_-components in the backward - wave network ) as shown in fig . [ 3d_coax_and_mstrip_unit_cell]a . this structure is realizable , but we propose a simpler structure based on microstrip lines , as presented in fig . [ 3d_coax_and_mstrip_unit_cell]b . the problem with microstrip lines is of course the design of intersections where the transmission lines from six directions meet . this problem can be overcome by using ground planes which have holes in them at these intersection points . this way the conducting strip can be taken through the substrate and thus connection of the vertical conducting strips becomes possible . the proposed structure ( see fig . [ 3d_coax_and_mstrip_unit_cell]b ) has been simulated in ansoft hfss ( version 9.2.1 ) . due to complexity of the structure and the limited calculation power available , the three - dimensional structure was simulated only near the first lens interface of fig . [ perfect_lens ] . the simulated model had @xmath111 ( @xmath112 ) unit cells in the forward - wave region and @xmath113 unit cells in the backward - wave region . the properties of the transmission lines and lumped elements were the same as in fig . [ f_qzvar_bw_and_fw ] . the edges of the system were terminated with matched loads to prevent reflections ( @xmath114 ohm in the backward - wave region and @xmath115 ohm in the forward - wave region ) . different types of source fields ( plane waves with different incidence angles and a point source ) were tested and in all cases negative refraction was observed at the interface between the forward - wave and backward - wave networks at the expected frequency ( @xmath62 ghz ) . a two - dimensional cut of the proposed structure was simulated as a complete `` lens '' system . again negative refraction was seen at both interfaces , and therefore also focusing of propagating waves was observed . see fig . [ 2dplot ] for the plot of the phase of the electric field in the two - dimensional simulation . the source field is excited at the left edge of the system in fig . [ 2dplot ] . when the field magnitude @xmath116 is plotted and animated as a function of phase , it is clearly seen that the phase propagates to the right in the forward - wave regions and to the left in the backward - wave region . the energy of course propagates to the right in all regions . in this paper we have introduced and studied a three - dimensional transmission - line network which is a circuit analogy of the superlens proposed by pendry . the structure is a 3d - network of interconnected loaded transmission lines . choosing appropriate loads we realize forward - wave ( fw ) and backward - wave ( bw ) regions in the network . the dispersion equations and analytical expressions for the characteristic impedances for waves in fw and bw regions have been derived . a special attention has been given to the problem of impedance and refraction index matching of fw and bw regions . from the derived dispersion equations it has been seen that there exist such a frequency at which the corresponding isofrequency surfaces for fw and bw regions coincide . theoretically this can provide distortion - less focusing of the propagating modes _ if _ the wave impedances of fw and bw regions are also well matched . impedance matching becomes even more important when the evanescent modes are taken into account . in this paper we have shown that the wave impedances can be matched at least within 1% accuracy or better if the characteristic impedances of the transmission lines are properly tuned . however , from the practical point of view an accuracy better than 1% becomes hardly realizable . it has been shown that decreasing the thickness of the bw region reduces the negative effect of the impedance mismatch , while the amplification of the evanescent modes is preserved . we have also outlined a couple of prospective designs of the perfect lens discussed in this paper and numerically simulated their performance . this work has been done within the frame of the _ metamorphose _ network of excellence and partially funded by the academy of finland and tekes through the center - of - excellence program . the authors would like to thank dr . mikhail lapine for bringing paper @xcite to their attention and for helpful discussions . eleftheriades , a.k . iyer , and p.c . kremer , `` planar negative refractive index media using periodically _ l - c _ loaded transmission lines , '' _ ieee trans . microwave theory and techniques _ , vol . 2702 - 2712 , dec . 2002 . a. grbic and g. v. eleftheriades , `` periodic analysis of a 2-d negative refractive index transmission line structure , '' _ ieee trans . antennas and propagation _ , vol . 51 , no . 10 , pp . 2604 - 2611 , oct . a. grbic and g.v . eleftheriades , `` negative refraction , growing evanescent waves and sub - diffraction imaging in loaded transmission - line metamaterials , '' _ ieee trans . microwave theory and techniques _ , 2297 - 2305 , dec . 2003	an isotropic three - dimentional perfect lens based on cubic meshes of interconnected transmission lines and bulk loads is proposed . the lens is formed by a slab of a loaded mesh placed in between two similar unloaded meshes . the dispersion equations and the characteristic impedances of the eigenwaves in the meshes are derived analytically , with an emphasis on generality . this allows designing of transmission - line meshes with desired dispersion properties . the required backward - wave mode of operation in the lens is realized with simple inductive and capacitive loads . an analytical expression for the transmission through the lens is derived and the amplification of evanescent waves is demonstrated . factors that influence enhancement of evanescent waves in the lens are studied and the corresponding design criteria are established . a possible realization of the structure is outlined .
the evolution of lattice gauge theory techniques has greatly enhanced our understanding of quark - gluon dynamics in qcd . heavy - light mesons provide an ideal laboratory for lattice qcd studies . the static approximation ( @xmath0 ) in which the heavy quark propagator is replaced by a straight time - like wilson line provides a framework which allows a quantitative study of masses , decay constants , mixing amplitudes , and electroweak form factors.[1 ] since heavy - light mesons have only one dynamical light ( valence ) quark , these systems are also well suited to the study of constituent quark ideas [ 2 ] and the chiral quark model [ 3 ] . in view of the success of the nonrelativistic ( nr ) potential model for heavy @xmath1 mesons , one interesting question for heavy - light systems is the nature and extent of the deviation from the nr potential picture as one of the quarks becomes light . here we present results of a numerical lattice study of this question . our findings support a surprisingly simple answer . the coulomb gauge wave functions obtained in lattice qcd agree , within the accuracy of our calculations , with the results of a simple relativistic generalization of the nr quarkonium potential model . it is only necessary to replace the nr kinetic energy term in the hamiltonian by its relativistic form , leaving the nr potential unchanged . the only adjustable parameter is the quark mass parameter @xmath2 . this description holds down to fairly small values of the current quark mass , corresponding to a pion mass of approximately @xmath3 , well into the region where the nr description fails . in lattice qcd , the properties of hadronic states are studied using correlation functions of operators which couple to the state . originally local operators were used . more recently smeared ( non - local ) operators have been found to improve the ability to extract the masses of meson and baryon ground states [ 4 ] . many of the present studies have been done with configurations and propagators fixed in coulomb gauge and operators which smear the position of the quark field uniformly over a spatial cube of variable size . however a constant cube of any size is a very crude approximation to the ground state wave function [ 5 ] . hence , the propagator generally has significant contamination from higher states out to times large compared to the inverse of the energy splitting between the ground state and the lowest excited state . this is a particular problem in the study of heavy - light correlators because they become noisy rather rapidly in time . unfortunately , this is an unavoidable feature of heavy - light systems [ 6,7 ] . recently a multistate smearing technique has been proposed [ 6 ] which allows the extraction of the properties of heavy - light states from relatively short times . the details of the multistate smearing method have been presented elsewhere[6 ] . by choosing an appropriate orthonormal set of smearing functions and diagonalizing the corresponding matrix of correlators , one obtains the wave functions of not only the lowest lying state in a given channel , but also of radially excited states . here we define the wave function to be the vacuum - to - one - particle matrix element , @xmath4 where @xmath5 is the state of interest . the sum is over color , and spin labels are supressed . here we discuss the wave functions obtained for the 1s and 2s levels of the s - wave pseudoscalar meson as well as a preliminary study of the 1p state.[8 ] the main emphasis will be on the remarkable quantitative agreement between the lattice qcd wave functions and those obtained from a simple relativistic quark model hamiltonian . the results for heavy - light decay constants and spectroscopy will be presented at this conference by eichten.[9 ] the investigation used an existing set of 50 configurations ( each separated by 2000 sweeps ) generated by acpmaps on a @xmath6 lattice at @xmath7 . the configurations were fixed to coulomb gauge and light quark propagators with @xmath8 were used . only the four lowest energy smearing functions were included ( @xmath9 ) . the optimized wave functions obtained from our lattice data by the multistate smearing method turn out to be , within errors , the same as the eigenfunctions of a lattice version of the spinless , relativistic quark model hamiltonian , which we will now define . in the absence of gauge fields , the free quark hamiltonian can be exactly diagonalized by introducing momentum space creation and annihilation operators for quarks and antiquarks . in the continuum , @xmath10\label{eq : h0}\end{aligned}\ ] ] where the sum is over spin and color labels . in terms of the covariant quark propagator , the particle and antiparticle operators are associated with propagation forward and backward in time , respectively . since @xmath11 contains no pair creation ( @xmath12 ) terms , it is possible to formulate the eigenvalue problem as that of a one - body operator , @xmath13 if we now turn on the gauge interaction and introduce a heavy - quark , static color source , the description of the bound light quark becomes , in principle , drastically more complicated . we know that , in the limit @xmath14 where the dynamical quark becomes heavy , the primary effect of the color source is to introduce a static , confining potential @xmath15 whose form is well - measured and consistently given by both @xmath16 phenomenology and lattice qcd , @xmath17 at this stage , the hamiltonian can still be regarded as a one - body operator[10 ] . as the mass of the quark becomes light , one expects more complicated effects arising from the gauge interaction which render the hamiltonian eigenvalue problem intractable . these effects include the creation of gluons and light @xmath18 pairs , as well as the exchange of transverse and non - instantaneous gluons with the static source . from the numerical results presented in the next section , we conclude that these effects are relatively small , and that the heavy - light meson system is well - described by the hamiltonian ( 6 ) , which we will refer to as the spinless relativistic quark model ( srqm ) . the construction of explicit eigenfunctions of the srqm hamiltonian is easily accomplished by a numerical procedure . first the operator @xmath19 is discretized on a 3d lattice by replacing the spatial derivatives with finite differences the potential energy v(r ) is just the static energy measured on the same configurations used to study the heavy - light spectrum . then the resolvent operator @xmath20 acting on a source vector @xmath21 is computed by a numerical matrix inversion ( conjugate gradient ) algorithm . finally , the parameter @xmath22 is varied to find the poles in the output vector @xmath23 . the location of the pole is an eigenvalue of @xmath19 , and its residue is the corresponding eigenfunction . in the next section we compare the wave functions obtained in this way from the srqm hamiltonian with the lattice qcd results . [ fig : toosmall ] using the four state smeared correlator described in section 2 an initial study for the s - wave channel was carried out . after some iterative improvement of the smearing functions , it was found that the value @xmath24 for the dimensionless mass parameter in the srqm hamiltonian gave the best agreement with the lattice qcd wave functions with @xmath25 . in fig . 1 the lqcd wave function is plotted with the srqm wave function . for comparison , the nonrelativistic ( nr ) schrodinger wave function ( obtained by replacing the relativistic kinetic term by @xmath26 ) is also plotted . the mass parameter in the nr hamiltonian was adjusted to give the same slope at the origin in the ground state wave function . notice that , for large r , the qcd and srqm wave functions both fall exponentially . on the other hand , the nr wave function falls faster than exponentially ( @xmath27 ) , as expected from the behavior of the analytic solution in a pure linear potential ( airy function ) . remarkably , by including the relativistic kinetic term , the srqm wave functions are brought into excellent agreement with those of lattice qcd , without changing the potential from its nonrelativistic form . in fig . 2 we plot the excited 2s state from lqcd along with the corresponding wave functions from the srqm and the nr model . the qcd wave function is somewhat more peaked at the origin , however , the overall agreement between qcd and the srqm is excellent . here , there are no adjustable parameters , @xmath28 being already fixed from the 1s state fit . finally , in fig . 3 we show some preliminary results of a study of the 1p state . here the solid line is the 1p wave function from the srqm . the data points depict the evolution of the p - wave lqcd radial wavefunction extracted from time slices @xmath29 ( + s ) , @xmath30 ( boxes ) , and @xmath31 ( @xmath32 s ) , starting with an approximate guess for the initial smearing function . the ansatz for the initial smearing function used here was a simple @xmath33 form . as the lqcd wave function evolves in euclidean time , it appears to approach a true eigenstate whose wavefunction again agrees remarkably well with the srqm result , with no adjustable parameters . additional studies are in progress using a variety of lattice sizes , gauge coupling strengths , and light quark masses . preliminary results of these studies are fully consistent with the conclusions presented here . the agreement of lattice qcd with the srqm wave functions suggests that the relativistic propagation of the light valence quark is the most important effect which must be included in a description of heavy - light mesons . other field theoretic effects such as the presence of multibody components of the wavefunction ( containing gluons along with light @xmath18 pairs arising , in quenched approximation , from the propagation of the valence quark backward in time ) are of less quantitative importance in determining the shape of the valence quark wave function . further numerical studies of the connection between lattice qcd and the relativistic quark model are planned . we thank george hockney , aida el khadra , andreas kronfeld , and paul mackenzie for joint lattice efforts without which this analysis would not have been possible . we also thank chris quigg for useful discussions . the numerical calculations were performed on the fermilab acpmaps computer system developed by the cr&d department in collaboration with the theory group . this work was supported in part by the u. s. department of energy under grant no . de - as05 - 89er40518 . 99 for a review of heavy - light systems on the lattice see : e. eichten , in _ lattice 90 _ , nucl . phys . b(proc.suppl . ) * 20 * ( 1991 ) 475 . j. d. bjorken , slac preprint slac - pub-5278 ( 1990 ) . a. manohar and h. georgi , nucl . b234 * ( 1984 ) 189 . p. bacilieri et . al . , nucl . * b317 * ( 1989 ) 509 . s. hashimoto and y. saeki , mod . phys . lett . * a7 * ( 1992 ) 387 . a. duncan , e . eichten , and h. thacker , nucl . b ( proc . suppl ) * 26 * ( 1992 ) 391 ; * 26 * ( 1992 ) 394 . a.kronfeld , in _ lattice 89 _ , nucl.phys . b ( proc.suppl . ) * 17 * ( 1990 ) 313 . for a recent study of light hadron wave functions , see m. w. hecht and t. a. degrand , phys . * d46 * ( 1992 ) 2155 , and references therein . e. eichten , talk presented at this conference . e. e. salpeter , phys . rev . * 87 * , ( 1952 ) 328 .	using a multistate smearing method , coulomb gauge wave functions of heavy - light mesons are studied in lattice qcd . wave functions for the ground state , the first radially excited s - wave state , and the lowest p - wave states of a heavy - light meson are calculated in quenched approximation . the results are found to be in remarkably good agreement with the predictions of a simple relativistic quark model .
the paper deals with the estimation problem in the heteroscedastic nonparametic regression model @xmath0 where the design points @xmath1 , @xmath2 is an unknown function to be estimated , @xmath3 is a sequence of centered independent random variables with unit variance and @xmath4 are unknown scale functionals depending on the design points and the regression function @xmath5 . typically , the notion of asymptotic optimality is associated with the optimal convergence rate of the minimax risk ( see e.g. , ibragimov , hasminskii,1981 ; stone,1982 ) . an important question in optimality results is to study the exact asymptotic behavior of the minimax risk . such results have been obtained only in a limited number of investigations . as to the nonparametric estimation problem for heteroscedastic regression models we should mention the papers by efromovich , 2007 , efromovich , pinsker , 1996 , and galtchouk , pergamenshchikov , 2005 , concerning the exact asymptotic behavior of the @xmath6-risk and the paper by brua , 2007 , devoted to the efficient pointwise estimation for heteroscedastic regressions . heteroscedastic regression models are largely used in financial mathematics , in particular , in problem of calibrating ( see e.g. , belomestny , reiss , 2006 ) . an example of heteroscedastic regression models is given by econometrics ( see , for example , goldfeld , quandt , 1972 , p. 83 ) , where for consumer budget problems one uses some parametric version of model with the scale coefficients defined as @xmath7 where @xmath8 , @xmath9 and @xmath10 are some unknown positive constants . the purpose of the article is to study asymptotic properties of the adaptive estimation procedure proposed in galtchouk , pergamenshchikov , 2007 , for which a non - asymptotic oracle inequality was proved for quadratic risks . we will prove that this oracle inequality is asymptotically sharp , i.e. the asymptotic quadratic risk is minimal . it means the adaptive estimation procedure is efficient under some the conditions on the scales @xmath4 which are satisfied in the case . note that in efromovich , 2007 , efromovich , pinsker , 1996 , an efficient adaptive procedure is constructed for heteroscedastic regression when the scale coefficient is independent of @xmath5 , i.e. @xmath11 . in galtchouk , pergamenshchikov , 2005 , for the model the asymptotic efficiency was proved under strong the conditions on the scales which are not satisfied in the case . moreover in the cited papers the efficiency was proved for the gaussian random variables @xmath3 that is very restrictive for applications of proposed methods to practical problems . in the paper we modify the risk . we take a additional supremum over the family of unknown noise distributions like to galtchouk , pergamenshchikov , 2006 . this modification allows us to eliminate from the risk dependence on the noise distribution . moreover for this risk a efficient procedure is robust with respect to changing the noise distribution . it is well known to prove the asymptotic efficiency one has to show that the asymptotic quadratic risk coincides with the lower bound which is equal to the pinsker constant . in the paper two problems are resolved : in the first one a upper bound for the risk is obtained by making use of the non - asymptotic oracle inequality from galtchouk , pergamenshchikov , 2007 , in the second one we prove that this upper bound coincides with the pinsker constant . let us remember that the adaptive procedure proposed in galtchouk , pergamenshchikov , 2007 , is based on weighted least - squares estimates , where the weights are proper modifications of the pinsker weights for the homogeneous case ( when @xmath12 ) relative to a certain smoothness of the function @xmath5 and this procedure chooses a best estimator for the quadratic risk among these estimators . to obtain the pinsker constant for the model one has to prove a sharp asymptotic lower bound for the quadratic risk in the case when the noise variance depends on the unknown regression function . in this case , as usually , we minorize the minimax risk by a bayesian one for a respective parametric family . then for the bayesian risk we make use of a lower bound ( see theorem 6.1 ) which is a modification of the van trees inequality ( see , gill , levit , 1995 ) . the paper is organized as follows . in section [ sec : ad ] we construct an adaptive estimation procedure . in section [ sec : co ] we formulate principal the conditions . the main results are presented in section [ sec : ma ] . the upper bound for the quadratic risk is given in section [ sec : up ] . in section [ sec : lo ] we give all main steps of proving the lower bound . in subsection [ subsec : tr ] we find the lower bound for the bayesian risk which minorizes the minimax risk . in subsection [ subsec : fa ] we study a special parametric functions family used to define the bayesian risk . in subsection [ subsec : br ] we choose a prior distribution for bayesian risk to maximize the lower bound . section [ sec : np ] is devoted to explain how to use the given procedure in the case when the unknown regression function is non periodic . in section [ sec : cn ] we discuss the main results and their practical importance . the proofs are given in section [ sec : pr ] . the appendix contains some technical results . in this section we describe the adaptive procedure proposed in galtchouk , pergamenshchikov , 2006 . we make use of the standard trigonometric basis @xmath13 in @xmath14 $ ] , i.e. @xmath15x)\,,\ j\ge 2\,,\ ] ] where the function @xmath16 for even @xmath17 and @xmath18 for odd @xmath17 ; @xmath19 $ ] denotes the integer part of @xmath20 . to evaluate the error of estimation in the model we will make use of the empiric norm in the hilbert space @xmath21 $ ] , generated by the design points @xmath22 of model . to this end , for any functions @xmath23 and @xmath24 from @xmath21 $ ] , we define the empiric inner product @xmath25 moreover , we will use this inner product for vectors in @xmath26 as well , i.e. if + @xmath27 and @xmath28 , then @xmath29 the prime denotes the transposition . notice that if @xmath30 is odd , then the functions @xmath31 are orthonormal with respect to this inner product , i.e. for any @xmath32 , @xmath33 where @xmath34 is kronecker s symbol , @xmath35 if @xmath36 and @xmath37 for @xmath38 . [ re.ad.1 ] note that in the case of even @xmath30 , the basis is orthogonal and it is orthonormal except the @xmath30th function for which the normalizing constant should be changed . the corresponding modifications of the formulas for even @xmath30 one can see in galtchouk , pergamenshchikov,2005 . to avoid these complications of formulas related to even @xmath30 , we suppose @xmath30 to be odd . thanks to this basis we pass to the discrete fourier transformation of model : @xmath39 where @xmath40 , @xmath41 , @xmath42 and @xmath43 we estimate the function @xmath5 by the weighted least squares estimator @xmath44 where the weight vector @xmath45 belongs to some finite set @xmath46 from @xmath47^n$ ] with @xmath48 . here we make use of the weight family @xmath46 introduced in galtchouk , pergamenshchikov , 2008 , i.e. @xmath49 where @xmath50 and @xmath51 $ ] . we suppose that the parameters @xmath52 and @xmath53 are functions of @xmath30 , i.e. @xmath54 and @xmath55 , such that , @xmath56 & \lim_{{\mathchoice{n\to\infty}{n\to\infty}{\lower.25ex\hbox{$\scriptstylen\to\infty$ } } { \lower0.25ex\hbox{$\scriptscriptstylen\to\infty$}}}}\,{\varepsilon}_{{\mathchoice{n}{n}{\lower.25ex\hbox{$\scriptstylen$ } } { \lower0.25ex\hbox{$\scriptscriptstylen$}}}}\,=\,0 \quad\mbox{and}\quad \lim_{{\mathchoice{n\to\infty}{n\to\infty}{\lower.25ex\hbox{$\scriptstylen\to\infty$ } } { \lower0.25ex\hbox{$\scriptscriptstylen\to\infty$}}}}\,n^{\nu}\,{\varepsilon}_{{\mathchoice{n}{n}{\lower.25ex\hbox{$\scriptstylen$ } } { \lower0.25ex\hbox{$\scriptscriptstylen$}}}}\,=+\infty\ , , \end{array } \right.\ ] ] for any @xmath57 . for example , one can take for @xmath48 @xmath58 where @xmath59 is any nonnegative constant . for each @xmath60 we define the weight vector @xmath61 as @xmath62 here @xmath63 $ ] with @xmath64 where @xmath65 is any nonnegative constant and @xmath66 [ re.ad.2 ] note that the weighted least squares estimators have been introduced by pinsker , 1981 , for continuous time optimal signal filtering in the gaussian noise . he proved that the mean - square asymptotic risk is minimized by weighted least squares estimators with weights of type . moreover he has found the sharp minimal value of the mean - square asymptotic risk , which was called later as the pinsker constant . nussbaum , 1985 , used the same method with proper modification for efficient estimation of the function @xmath5 of known smoothness in the homogeneous gaussian model , i.e. when @xmath12 and @xmath3 is i.i.d . @xmath67 sequence . to choose weights from the set we minimize the special cost function introduced by galtchouk , pergamenshchikov , 2007 this cost function is as follows @xmath68 where @xmath69 and @xmath70 $ ] . the penalty term we define as @xmath71 where @xmath72 is any slowly increasing sequence , i.e. @xmath73 for any @xmath57 . finally , we set @xmath74 the goal of this paper is to study asymptotic ( as @xmath75 ) properties of this estimation procedure . [ re.ad.3 ] now we explain why does one choose the cost function in the form . developing the empiric quadratic risk for estimate , one obtains @xmath76 it s natural to choose the weight vector @xmath77 for which this function reaches the minimum . since the last term on the right - hand part is independent of @xmath77 , it can be droped and one has to minimize with respect to @xmath77 the function equals the difference of two first terms on the right - hand part . it s clear that the minimization problem cannt be solved directly because the fourier coefficients @xmath78 are unknown.to overcome this difficulty , we replace the product @xmath79 by its asymptotically unbiased estimator @xmath80 ( see , galtchouk , pergamenshchikov , 2007 , 2008 ) . moreover , to pay this substitution , we introduce into the cost function the penalty term @xmath81 with a small coefficient @xmath82 . the form of the penalty term is provided by the principal term of the quadratic risk for weighted least - squares estimator , see galtchouk , pergamenshchikov , 2007 , 2008.the coefficient @xmath82 means , that the penalty is small , because the estimator @xmath80 approximates in mean the quantity @xmath79 asymptotically , as @xmath75 . note that the principal difference between the procedure and the adaptive procedure proposed by golubev , nussbaum , 1993 , for a homogeneous gaussian regression , consists in presence of the penalty term in the cost function . [ re.ad.4 ] as it was noted at remark [ re.ad.2 ] , nussbaum , 1985 , has shown that the weight coefficients of type provide the asymptotic minimum of the mean - squared risk at the regression function estimation problem for the homogeneous gaussian model , when the smoothness of the function @xmath5 is known . in fact , to obtain an efficient estimator one needs to take a weighted least squares estimator with the weight vector @xmath83 , where the index @xmath84 depends on smoothness of function @xmath5 and on coefficients @xmath4 , ( see below ) , which are unknown in our case . for this reason , galtchouk , pergamenshchikov , 2007 , 2008 , have proposed to make use of the family of coefficients , which contains the weight vector providing the minimum of the mean - squared risk . moreover , they proposed the adaptive procedure for which a non - asymptotic oracle inequality ( see , theorem [ th.m.1 ] below ) was proved under some weak conditions on the coefficients @xmath4 . it is important to note that due the properties of the parametric family , the secondary term in the oracle inequality is slowly increasing ( slower than any degree of @xmath30 ) . first we impose some conditions on unknown function @xmath5 in the model . let @xmath85 be the set of @xmath86-periodic @xmath87 times differentiable @xmath88 functions . we assume that @xmath5 belongs to the following set @xmath89 where @xmath90 denotes the norm in @xmath21 $ ] , i.e. @xmath91 moreover , we suppose that @xmath92 and @xmath93 are unknown parameters . note that , we can represent the set @xmath94 as an ellipse in @xmath21 $ ] , i.e. @xmath95\,:\ , \sum_{{\mathchoice{j=1}{j=1}{\lower.25ex\hbox{$\scriptstylej=1 $ } } { \lower0.25ex\hbox{$\scriptscriptstylej=1$}}}}^\infty\,a_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}\theta^2_{{\mathchoice{j}{j}{\lower.25ex\hbox{$\scriptstylej$ } } { \lower0.25ex\hbox{$\scriptscriptstylej$}}}}\le r\}\,,\ ] ] where @xmath96 and @xmath97)^{2i}\,.\ ] ] here @xmath13 is the trigonometric basis defined in . now we describe the conditions on the scale coefficients @xmath98 . * _ @xmath99 for some unknown function @xmath100\times { { \cal l}}_{{\mathchoice{1}{1}{\lower.25ex\hbox{$\scriptstyle1 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle1$}}}}[0,1 ] \to \bbr_+$ ] , which is square integrable with respect to @xmath20 such that @xmath101 where @xmath102 . moreover , @xmath103 and @xmath104 _ * _ for any @xmath105 $ ] , the operator @xmath106\to \bbr$ ] is differentiable in the frchet sense for any fixed function @xmath107 from @xmath108 $ ] , i.e. for any @xmath109 from some vicinity of @xmath107 in @xmath108 $ ] , @xmath110 where the frchet derivative @xmath111\to \bbr$ ] is a bounded linear operator and the residual term @xmath112 , for each @xmath105 $ ] , satisfies the following property : @xmath113 where @xmath114 . _ * _ there exists some positive constant @xmath115 such that for any function @xmath5 from @xmath108 $ ] the operator @xmath116 defined in the condition @xmath117 satisfies the following inequality for any function @xmath109 from @xmath108 $ ] : @xmath118 where @xmath119 . _ * _ the function @xmath120 corresponding to @xmath121 is continuous on the interval @xmath47 $ ] . moreover , @xmath122 _ [ re.co.1 ] let us explain the conditions @xmath123@xmath124 . in fact , this is the regularity conditions of the function @xmath125 generating the scale coefficients @xmath4 . condition @xmath123 means that the function @xmath126 should be uniformly integrable with respect to the first argument in the sens of convergence . moreover , this function should be separated from zero ( see inequality ) and bounded on the class ( see inequality ) . boundedness away from zero provides that the distribution of observations @xmath127 is nt degenerate in @xmath26 , and the boundedness means that the intensity of the noise vector should be finite , otherwise the estimation problem hasnt any sens . conditions @xmath117 and @xmath128 mean that the function @xmath129 is regular , at any fixed @xmath130 , with respect to @xmath5 in the sens , that it is differentiable in the frchet sens ( see e.g. , kolmogorov , fomin , 1989 ) and moreover the frchet derivative satisfies the growth condition given by the inequality which permits to consider the example . last the condition @xmath124 is the usual uniform continuity the condition of the function @xmath131 at the function @xmath132 . now we give some examples of functions satisfying the conditions @xmath123-@xmath124 . we set @xmath133 with some coefficients @xmath134 , @xmath135 , @xmath136 . in this case @xmath137 the frchet derivative is given by @xmath138 it is easy to see that the function satisfies the conditions @xmath123@xmath124 . moreover , the conditions @xmath123@xmath124 are satisfied by any function of type @xmath139 where the functions @xmath140 and @xmath141 satisfy the following the conditions : * @xmath140 is a @xmath47\times\bbr\to [ c_{{\mathchoice{0}{0}{\lower.25ex\hbox{$\scriptstyle0 $ } } { \lower0.25ex\hbox{$\scriptscriptstyle0$}}}}\,,\,+\infty)$ ] function ( with @xmath134 ) such that @xmath142 * @xmath141 is a continuously differentiable @xmath143 function such that @xmath144 where @xmath145 is the derivative of @xmath141 . in this case @xmath146 and @xmath147 where @xmath148 . now to estimate the last term in this inequality note that @xmath149 therefore , from the condition we get @xmath150 and through the bounyakovskii - cauchy - schwarz inequality , for any @xmath151 , @xmath152 now , the condition implies @xmath123 . moreover , the frchet derivative in this case is given by @xmath153 one can check directly that this operator satisfies the inequality with @xmath154 . denote by @xmath155 the family of distributions @xmath156 in @xmath26 of the vectors @xmath157 in the model such that the components @xmath158 are jointly independent , centered with unit variance and @xmath159 where @xmath160 is slowly increasing sequence , that is it satisfies the property . it is easy to see that , for any @xmath161 , the centered gaussian distribution in @xmath26 with unit covariation matrix belongs to the family @xmath155 . we will denote by @xmath162 this gaussian distribution . for any estimator @xmath163 we define the following quadratic risk @xmath164 where @xmath165 is the expectation with respect to the distribution @xmath166 of the observations @xmath167 with the fixed function @xmath5 and the fixed distribution @xmath168 of random variables @xmath3 in the model . moreover , to make the risk independent of the design points , in this paper we will make use of the risk with respect to the usual norm in @xmath21 $ ] also , i.e. @xmath169 if an estimator @xmath163 is defined only at the design points @xmath22 , then we extend it as step function onto the interval @xmath47 $ ] by setting @xmath170 , for all @xmath130 , where @xmath171}{[0,x_{1}]}{\lower.25ex\hbox{$\scriptstyle[0,x_{1}]$ } } { \lower0.25ex\hbox{$\scriptscriptstyle[0,x_{1}]$}}}}(x ) + \sum_{k=2}^n\,f(x_k)\chi_{{\mathchoice{(x_{k-1},x_k]}{(x_{k-1},x_k]}{\lower.25ex\hbox{$\scriptstyle(x_{k-1},x_k]$ } } { \lower0.25ex\hbox{$\scriptscriptstyle(x_{k-1},x_k]$}}}}(x)\,.\ ] ] in galtchouk , pergamenshchikov , 2007 , 2008 the following non - asymptotic oracle inequality has been shown for the procedure . [ th.m.1 ] assume that in the model the function @xmath5 belongs to @xmath172 . then , for any odd @xmath48 and @xmath92 , the estimator @xmath173 satisfies the following oracle inequality @xmath174 where the function @xmath175 is such that , for any @xmath57 , @xmath176 [ re.m.1 ] note that in galtchouk , pergamenshchikov , 2007 , 2008 , the oracle inequality is proved for the model , where the random variables @xmath3 are independent identically distributed . in fact , the result and the proof are true for independent random variables which are not identically distributed , i.e. for any distribution of the random vector @xmath157 from @xmath155 . now we formulate the main asymptotic results . to this end , for any function @xmath151 , we set @xmath177 where @xmath178 it is well known ( see e.g. , nussbaum , 1985 ) that the optimal rate of convergence is @xmath179 when the risk is taken uniformly over @xmath180 . [ th.m.2 ] assume that in the model the sequence @xmath181 fulfills the condition @xmath182 . then the estimator @xmath173 from satisfies the inequalities @xmath183 and @xmath184 the following result gives the sharp lower bound for risk and show that @xmath185 is the . [ th.m.3 ] assume that in the model the sequence @xmath181 satisfies the conditions @xmath186 @xmath187 . then the risks and admit the following asymptotic lower bounds @xmath188 and @xmath189 [ re.m.2 ] note that in galtchouk , pergamenshchikov , 2005 , an asymptotically efficient estimator has been constructed and results similar to theorems [ th.m.2 ] and [ th.m.3 ] were claimed for the model . in fact the upper bound is true there under some additional condition on the smoothness of the function @xmath5 , i.e. on the parameter @xmath87 . in the cited paper this additional condition is not formulated since erroneous inequality @xmath190 . to avoid using this inequality we modify the estimating procedure by introducing the penalty term @xmath191 in the cost function . by this way we remove all additional conditions on the smoothness parameter @xmath87 . [ re.m.3 ] in fact to obtain the non - asymptotic oracle inequality , it is nt necessary to make use of equidistant design points and the trigonometric basis . one may take any design points ( deterministic or random ) and any orthonormal basis satisfying . but to obtain the property one needs to impose some technical conditions ( see galtchouk , pergamenshchikov , 2008 ) . note that the results of theorem [ th.m.2 ] and theorem [ th.m.3 ] are based on equidistant design points and the trigonometric basis . in this section we prove theorem [ th.m.2 ] . to this end we will make use of the oracle inequality . we have to find an estimator from the family - for which we can show the upper bound . we start with the construction of such an estimator . first we put @xmath192 where @xmath193 . then we choose an index from the set @xmath194 as @xmath195 where @xmath87 is the parameter of the set @xmath180 and @xmath196 . finally , we set @xmath197 now we show the upper bound for this estimator . [ th.u.1 ] assume that the condition @xmath182 holds . then @xmath198 [ re.u.1 ] note that the estimator @xmath199 belongs to the family - , but we ca nt use directly this estimator because the parameters @xmath87 , @xmath200 and @xmath201 are unknown . we can use this upper bound only through the oracle inequality proved for procedure . now theorem [ th.m.1 ] and theorem [ th.u.1 ] imply the upper bound . to obtain the upper bound we need the following auxiliary result . [ sec : le.u.1 ] for any @xmath202 and any estimate @xmath203 of @xmath151 , @xmath204 where the function @xmath205 is defined in . proof of this lemma is given in appendix [ subsec : a.1 ] . now inequality and this lemma imply the upper bound . hence theorem [ th.m.2 ] . in this section we give the main steps of proving the lower bounds and . in common , we follow the same scheme as nussbaum , 1985 . we begin with minorizing the minimax risk by a bayesian one constructed on a parametric functional family introduced in section [ subsec : fa ] ( see ) and using the prior distribution . further , a special modification of the van trees inequality ( see , theorem [ th.tr.1 ] ) yields a lower bound for the bayesian risk depending on the chosen prior distribution , of course . finally , in section [ subsec : br ] , we choose parameters of the prior distribution ( see ) providing the maximal value of the lower bound for the bayesian risk . this value coincides with the pinsker constant as it is shown in section [ subsec : th.m.3 ] . let @xmath206 be a statistical model relative to the observations @xmath127 governed by the regression equation @xmath207 where @xmath208 are i.i.d . @xmath67 random variables , @xmath209 is a unknown parameter vector , @xmath210 is a unknown ( or known ) function and @xmath211 , with the function @xmath125 defined in the condition @xmath123 . assume that a prior distribution @xmath212 of the parameter @xmath213 in @xmath214 is defined by the density @xmath215 of the following form @xmath216 where @xmath217 is a continuously differentiable bounded density on @xmath218 with @xmath219 let @xmath220 be a continuously differentiable @xmath221 function such that , for any @xmath222 , @xmath223 where @xmath224 let @xmath225 be an estimator of @xmath226 based on observations @xmath127 . for any @xmath227 - measurable integrable function @xmath228 , we set @xmath229 where @xmath230 is the expectation with respect to the distribution @xmath231 of the vector @xmath232 . note that in this case @xmath233 where @xmath234 we prove the following result . [ th.tr.1 ] assume that the conditions @xmath235 hold . moreover , assume that the function @xmath236 with @xmath237 is uniformly over @xmath130 differentiable with respect to @xmath238 , i.e. for any @xmath222 there exists a function @xmath239 $ ] such that @xmath240 where @xmath241 , all coordinates are @xmath242 , except the i - th equals to @xmath86 . then for any square integrable estimator @xmath225 of @xmath226 and any @xmath222 , @xmath243 where @xmath244 , @xmath245 and @xmath246 @xmath247 , the operator @xmath248 is defined in the condition @xmath186 . * proof * is given in appendix [ subsec : a.2 ] . [ re.tr.1 ] note that the inequality is some modification of the van trees inequality ( see , gill , levit , 1995 ) adapted to the model . in this section we define and study some special parametric family of kernel function which will be used to prove the sharp lower bound . let us begin by kernel functions . we fix @xmath249 and we set @xmath250 where @xmath251 is the indicator of a set @xmath252 , the kernel @xmath253 is such that @xmath254 it is easy to see that the function @xmath255 possesses the properties : @xmath256 moreover , for any @xmath257 and @xmath258 @xmath259 we divide the interval @xmath47 $ ] into @xmath260 equal subintervals of length @xmath261 and on each of them we construct a kernel - type function which equals to zero at the boundary of the subinterval together with all derivatives . it provides that the fourier partial sums with respect to the trigonometric basis in @xmath262 $ ] give a natural parametric approximation to the function on each subinterval . let @xmath263 be the trigonometric basis in @xmath264 $ ] , i.e. @xmath265 x\right)\,,\ j\ge 2\,,\ ] ] where the functions @xmath266 are defined in . now , for any array @xmath267 we define the following function @xmath268 where @xmath269 , @xmath270 - 1\,.\ ] ] we assume that the sequences @xmath271 and @xmath272 , satisfy the following conditions . @xmath273 _ the sequence @xmath274 as @xmath75 and for any @xmath57 @xmath275 moreover , there exist @xmath276 and @xmath277 such that @xmath278 _ to define a prior distribution on the family of arrays , we choose the following random array @xmath279 with @xmath280 where @xmath281 are i.i.d . @xmath67 random variables and @xmath282 are some nonrandom positive coefficients . we make use of gaussian variables since they possess the minimal fisher information and therefore maximize the lower bound . we set @xmath283 we assume that the coefficients @xmath284 satisfy the following conditions . @xmath285 _ there exists a sequence of positive numbers @xmath286 such that @xmath287 moreover , for any @xmath288 , @xmath289 _ @xmath290 _ for some @xmath291 @xmath292 _ @xmath293 _ there exists @xmath294 such that @xmath295 _ [ sec : pr.fa.1 ] let the conditions @xmath273@xmath285 . then , for any @xmath57 and for any @xmath296 , @xmath297 [ sec : pr.fa.2 ] let the conditions @xmath273@xmath293 . then , for any @xmath57 , @xmath298 [ sec : pr.fa.3 ] let the conditions @xmath273@xmath293 . then , for any @xmath57 , @xmath299 [ sec : pr.fa.4 ] let the conditions @xmath273@xmath293 . then for any function @xmath300 satisfying the conditions and @xmath124 @xmath301 proofs of propositions [ sec : pr.fa.1][sec : pr.fa.4 ] are given in appendix . now we will obtain the lower bound for the bayesian risk that yields the lower bound for the minimax risk . we make use of the sequence of random functions @xmath302 defined in - with the coefficients @xmath282 satisfying the conditions @xmath273@xmath293 which will be chosen later . for any estimator @xmath203 we introduce now the corresponding bayes risk @xmath303 where the kernel family @xmath304 is defined in , @xmath212 denotes the distribution of the random array @xmath213 defined by in @xmath214 with @xmath305 . we remember that @xmath162 is a centered gaussian distribution in @xmath26 with unit covariation matrix . first of all , we replace the functions @xmath203 and @xmath5 by their fourier series with respect to the basis @xmath306 by making use of this basis we can estimate the norm @xmath307 from below as @xmath308 where @xmath309 moreover , from the definition one gets @xmath310 it is easy to see that the functions @xmath311 satisfy the condition for gaussian prior densities . in this case ( see the definition in ) we have @xmath312 where @xmath313 now to obtain a lower bound for the bayes risk @xmath314 we make use of theorem [ th.tr.1 ] which implies that @xmath315 where @xmath316 and @xmath317 with @xmath318 . in the appendix we show that @xmath319 and @xmath320 this means that , for any @xmath57 and for sufficiently large @xmath30 , @xmath321 where @xmath322 is defined in . therefore , if we denote in @xmath323 we obtain , for sufficiently large @xmath30 , @xmath324 in the appendix we show that @xmath325 where @xmath326 therefore we can write that , for sufficiently large @xmath30 , @xmath327 obviously , to obtain a `` good '' lower bound for the risk @xmath314 one needs to maximize the right - hand side of the inequality . hence we choose the coefficients @xmath328 by maximization the function @xmath329 , i.e. @xmath330 the parameter @xmath331 will be chosen later to satisfy the condition @xmath290 . by the lagrange multipliers method it is easy to find that the solution of this problem is given by @xmath332 with @xmath333 to obtain a positive solution in we need to impose the following condition @xmath334 moreover , from the condition @xmath290 we obtain that @xmath335 where @xmath336 note that by the condition @xmath124 the function @xmath120 is continuous on the interval @xmath47 $ ] , therefore @xmath337 now we have to choose the sequence @xmath338 . note that if we put in @xmath339 we can rewrite the inequality as @xmath340 where @xmath341 it is clear that @xmath342 therefore to obtain a positive finite asymptotic lower bound in we have to take the parameter @xmath343 as @xmath344 with some positive coefficient @xmath345 . moreover , the conditions - imply that , for sufficiently large @xmath30 , @xmath346 moreover , taking into account that for sufficiently large @xmath30 @xmath347 we obtain the following condition on @xmath345 @xmath348 where @xmath349 to maximize the function @xmath350 on the right - hand side of the inequality we take @xmath351 defined in . therefore we obtain that @xmath352 where @xmath353 furthermore , taking into account that @xmath354 we get @xmath355 where @xmath356 this means that to obtain in the maximal lower bound one has to take in @xmath357 it is important to note that if one defines the prior distribution @xmath212 in the bayesian risk by formulas , , and , then the bayesian risk would depend on a parameter @xmath358 , i.e. @xmath359 . therefore , the inequality implies that , for any @xmath358 , @xmath360 where the function @xmath361 is defined in for @xmath121 . now to end the definition of the sequence of the random functions @xmath362 defined by and one has to define the sequence @xmath363 . let us remember that we make use of the sequence @xmath362 with the coefficients @xmath282 constructed in for @xmath351 given in and for the sequence @xmath343 given by and for some fixed arbitrary @xmath358 . we will choose the sequence @xmath363 to satisfy the conditions @xmath273@xmath293 . one can take , for example , @xmath364 + 1 $ ] . then the condition @xmath273 is trivial . moreover , taking into account that in this case @xmath365 we find thanks to the convergence @xmath366 therefore , the solution , for sufficiently large @xmath30 , satisfies the following inequality @xmath367 now it is easy to see that the condition @xmath285 holds with @xmath368 and the condition @xmath293 holds for arbitrary @xmath369 . as to the condition @xmath290 , note that in view of the definition of @xmath370 in we get @xmath371 hence the condition @xmath290 . now we consider the estimation problem of the non periodic regression function @xmath5 in the model . in this case we will estimate the function @xmath5 on any interior interval @xmath372 $ ] of @xmath47 $ ] , i.e. for @xmath373 . it should be pointed out that at the boundary points @xmath374 and @xmath375 , one must to make use of kernel estimators ( see brua , 2007 ) . let now @xmath376 be a infinitely differentiable @xmath47\to\bbr_{{\mathchoice{+}{+}{\lower.25ex\hbox{$\scriptstyle+$ } } { \lower0.25ex\hbox{$\scriptscriptstyle+$}}}}$ ] function such that @xmath377 for @xmath378 and @xmath379 for all @xmath380 , for example , @xmath381}{[a',b']}{\lower.25ex\hbox{$\scriptstyle[a',b']$ } } { \lower0.25ex\hbox{$\scriptscriptstyle[a',b']$}}}}(z)\ , \d z\,,\ ] ] where @xmath141 is some kernel function introduced in , @xmath382 multiplying the equation by the function @xmath383 and simulating the i.i.d . @xmath67 sequence @xmath384 one comes to the estimation problem of the periodic regression function @xmath385 , i.e. @xmath386 where @xmath387 , @xmath388 and @xmath389 is some sufficiently small parameter . it is easy to see that if the sequence @xmath4 satisfies the conditions @xmath390 , then the sequence @xmath391 satisfies these conditions as well with @xmath392 in conclusion , it should be noted that this paper completes the investigation of the estimation problem of the nonparametric regression function for the heteroscedastic regression model in the case of quadratic risk . it is proved that the adaptive procedure satisfies the non asymptotic oracle inequality and it is asymptotically efficient for estimating a periodic regression function . moreover , in section [ sec : np ] we have explained how to apply the procedure to the case of non periodic function . as far as we know , the procedure is unique for estimating the regression function at the model . let us remember once more the main steps of this investigation . the procedure combines the both principal aspects of nonparametric estimation : non asymptotic and asymptotic . non - asymptotic aspect is based on the selection model procedure with penalization ( see e.g. , barron , birg and massart , 1999 , or fourdrinier , pergamenshchikov , 2007 ) . our selection model procedure differs from the commonly used one by a small coefficient in the penalty term going to zero that provides the sharp non - asymptotic oracle inequality . moreover , the commonly used selection model procedure is based on the least - squares estimators whereas our procedure uses weighted least - squares estimators with the weights minimizing the asymptotic quadratic risk that provides the asymptotic efficiency , as the final result . from practical point of view , the procedure gives an acceptable accuracy even for small samples as it is shown via simulations by galtchouk , pergamenshchikov , 2008 . to prove the theorem we will adapt to the heteroscedastic case the corresponding proof from nussbaum , 1985 . first , from we obtain that , for any @xmath168 , @xmath393 where @xmath394 setting now @xmath395 with the function @xmath396 defined in , the index @xmath397 defined in , @xmath398 $ ] , @xmath399 $ ] and @xmath400 we rewrite as follows @xmath401 with @xmath402 note that we have decomposed the first term on the right - hand of into the sum @xmath403 this decomposition allows us to show that @xmath404 is negligible and further to approximate the first term by a similar term in which the coefficients @xmath405 will be replaced by the fourier coefficients @xmath406 of the function @xmath5 . taking into account the definition of @xmath396 in we can bound @xmath407 as @xmath408 therefore , by lemma [ sec : le.a.1 ] we obtain @xmath409 let us consider now the next term @xmath410 . we have @xmath411 now by lemma [ sec : le.a.2 ] and the definition we obtain directly the same property for @xmath410 , i.e. @xmath412 setting @xmath413 and applying the well - known inequality @xmath414 to the first term on the right - hand side of the inequality we obtain that , for any @xmath296 and for any @xmath168 , @xmath415 \label{sec : up.1 - 3 } & + { \widetilde}{\delta}_{{\mathchoice{1,n}{1,n}{\lower.25ex\hbox{$\scriptstyle1,n$ } } { \lower0.25ex\hbox{$\scriptscriptstyle1,n$}}}}+ { \widetilde}{\delta}_{{\mathchoice{2,n}{2,n}{\lower.25ex\hbox{$\scriptstyle2,n$ } } { \lower0.25ex\hbox{$\scriptscriptstyle2,n$}}}}+(1 + 1/\delta)\ , { \widetilde}{\delta}_{{\mathchoice{3,n}{3,n}{\lower.25ex\hbox{$\scriptstyle3,n$ } } { \lower0.25ex\hbox{$\scriptscriptstyle3,n$}}}}\,,\end{aligned}\ ] ] where @xmath416 taking into account that @xmath93 and that @xmath417 we can show through lemma [ sec : le.a.3 ] that @xmath418 therefore , the inequality yields @xmath419 and to prove it suffices to show that @xmath420 first , it should be noted that the definition and the inequalities - imply directly @xmath421 moreover , by the definition of @xmath422 in , for sufficiently large @xmath30 , for which @xmath423 we find @xmath424 therefore , by the definition of the coefficients @xmath425 in @xmath426 furthermore , in view of the definition we calculate directly @xmath427 now , the definition of @xmath180 in and the condition imply the inequality . hence theorem [ th.u.1 ] . in this section we prove theorem [ th.m.3 ] . lemma [ sec : le.u.1 ] implies that to prove the lower bounds and , it suffices to show @xmath428 where @xmath429 for any estimator @xmath203 , we denote by @xmath430 its projection onto @xmath431 , i.e. + @xmath432 . since @xmath180 is a convex set , we get @xmath433 now we introduce the following set @xmath434 where @xmath281 are i.i.d . @xmath67 random variables from and the sequence @xmath286 is given in the condition @xmath285 . therefore , we can write that @xmath435 here the kernel function family @xmath304 is given in in which @xmath436 + 1 $ ] and the parameter @xmath437 is defined in and ; the measure @xmath212 is defined in . moreover , note that on the set @xmath438 the random function @xmath439 is uniformly bounded , i.e. @xmath440 where the coefficient @xmath441 is defined in . thus , we estimate the risk @xmath442 from below as @xmath443 with @xmath444 by making use of the bayes risk with the prior distribution given by formulae , , and for any fixed parameter @xmath358 we rewrite the lower bound for @xmath442 as @xmath445 with @xmath446 in section [ subsec : br ] we proved that the parameters in chosen prior distribution satisfy the conditions @xmath273@xmath293 . therefore propositions [ sec : pr.fa.2][sec : pr.fa.3 ] and the limit imply that , for any @xmath57 , @xmath447 moreover , by the condition @xmath124 the sequence @xmath448 goes to @xmath361 as @xmath75 . therefore , from this , and we get , for any @xmath358 , @xmath449 where @xmath450 is defined in . limiting here @xmath451 implies inequality . hence theorem [ th.m.3 ] . for any @xmath202 , by making use of the elementary inequality @xmath456 one gets @xmath457 moreover , for any @xmath452 with @xmath93 , by the bounyakovskii - cauchy - schwarz inequality we obtain that @xmath458 hence lemma [ sec : le.u.1 ] . for any @xmath459 we set @xmath460 note that due to the condition , the density is bounded , i.e. @xmath461 so through we obtain that @xmath462 therefore , integrating by parts yields @xmath463 now the bounyakovskii - cauchy - schwarz inequality gives the following lower bound @xmath464 to estimate the denominator in the last ratio , note that @xmath465 > from it follows that @xmath466 moreover , the conditions @xmath117 and imply @xmath467 from which it follows @xmath468 this implies inequality . hence theorem [ th.tr.1 ] . first note that , for @xmath130 , we can represent the @xmath469th derivative as @xmath470 where @xmath471 therefore @xmath472 and by the bounyakovskii - cauchy - schwarz inequality we obtain that @xmath473 with @xmath474 and @xmath475 now we show that , for any @xmath476 and @xmath296 , @xmath477 to this end note that @xmath478 therefore , taking into account the definition of the set @xmath479 in , the functions @xmath480 with @xmath481 can be estimated on this set as @xmath482 and by we get , for any @xmath296 and sufficiently large @xmath30 , @xmath483 moreover , for sufficiently large @xmath30 , @xmath484 therefore , the condition @xmath273 implies @xmath485 for any @xmath57 . hence proposition [ sec : pr.fa.1 ] . first of all we prove that for @xmath486 from the condition @xmath290 @xmath487 indeed , putting in @xmath488 we can represent the @xmath87th derivative of @xmath489 as follows @xmath490 with @xmath491 and @xmath492 first , note that , we can estimate the norm of @xmath493 by the same way as in the inequality , i.e. @xmath494 by making use of we obtain that , for any @xmath495 and for any @xmath296 , @xmath496 let us consider now the last term in . taking into account that @xmath497 we get @xmath498 therefore from the condition @xmath290 we get for sufficiently large @xmath30 @xmath499 with @xmath500 we show that for any @xmath57 and for any @xmath296 @xmath501 indeed , by the chebychev inequality for any @xmath502 @xmath503 note now that according to the burkholder - davis - gundy inequality for any @xmath504 there exists a constant @xmath505 such that @xmath506 moreover , by putting @xmath507 we can estimate the random variable @xmath508 as @xmath509 therefore , by the condition @xmath293 , for sufficiently large @xmath30 , @xmath510 where @xmath511 . now the condition @xmath273 implies , for sufficiently large @xmath30 , @xmath512 thus , choosing in @xmath513 we obtain the limiting equality which together with - implies . now it is easy to deduce that proposition [ sec : pr.fa.1 ] yields proposition [ sec : pr.fa.2 ] . first of all , we recall that , due to the condition @xmath285 , @xmath514 therefore , taking into account that @xmath515 we obtain , for sufficiently large @xmath30 , @xmath516 moreover , for any @xmath517 and @xmath518 , we estimate the last term as @xmath519 & + n\,\p(\xi^c_{{\mathchoice{n}{n}{\lower.25ex\hbox{$\scriptstylen$ } } { \lower0.25ex\hbox{$\scriptscriptstylen$ } } } } ) + 2\e\,\zeta^2\,\chi_{{\mathchoice{\{\zeta^2\ge n\}}{\{\zeta^2\ge n\}}{\lower.25ex\hbox{$\scriptstyle\{\zeta^2\ge n\}$ } } { \lower0.25ex\hbox{$\scriptscriptstyle\{\zeta^2\ge n\}$}}}}\,,\end{aligned}\ ] ] where @xmath520 . by applying now proposition [ sec : pr.fa.2 ] and the limit we come to proposition [ sec : pr.fa.3 ] . [ sec : le.a.2 ] for any @xmath524 , @xmath525}n^{-m}\left\|\sum_{l=2}^{n}\,l^m \ , \left ( \phi^2_{{\mathchoice{l}{l}{\lower.25ex\hbox{$\scriptstylel$ } } { \lower0.25ex\hbox{$\scriptscriptstylel$}}}}(x)-1 \right)\ , \right\|\le 2^m\,.\ ] ] first of all , note that proposition [ sec : pr.fa.4 ] , the condition and the condition @xmath124 imply that @xmath532 let us show now that for any continuously differentiable function @xmath109 on @xmath533 $ ] @xmath534 indeed , setting @xmath535 we deduce @xmath536 where @xmath537 + 1 $ ] , @xmath538 $ ] , @xmath539 + 1-n{\widetilde}{x}_{{\mathchoice{m}{m}{\lower.25ex\hbox{$\scriptstylem$ } } { \lower0.25ex\hbox{$\scriptscriptstylem$}}}})/(nh ) \quad \mbox{and } \quad v^*=([n{\widetilde}{x}_{{\mathchoice{m}{m}{\lower.25ex\hbox{$\scriptstylem$ } } { \lower0.25ex\hbox{$\scriptscriptstylem$}}}}+nh]-n{\widetilde}{x}_{{\mathchoice{m}{m}{\lower.25ex\hbox{$\scriptstylem$ } } { \lower0.25ex\hbox{$\scriptscriptstylem$}}}})/(nh)\,.\ ] ] therefore , taking into account that the derivative of the function @xmath109 is bounded on the interval @xmath533 $ ] we obtain that @xmath540 taking into account the conditions on the sequence @xmath272 given in @xmath273 we obtain limiting equality which together with implies . now we study the behavior of @xmath541 . due to the inequality we estimate the frchet derivative as @xmath542 consider now the fisrt term on the right - hand side of this inequality . we have @xmath543 we recall that the sequence @xmath441 is defined in . therefore , property implies @xmath544 as to the second term on the right - hand side of , we get @xmath545 similarly , @xmath546 and , by @xmath547 therefore , @xmath548 and the condition @xmath273 implies . indeed , by the direct calculation it easy to see that , for any @xmath549 and for any vector @xmath550 , @xmath551 where the operator @xmath552 is defined in . moreover , we remember that @xmath553 . therefore , taking into account the property we obtain . brua , j .- . asymptotic efficient estimators for non - parametric heteroscedastic model . _ statistical metodologie _ , accepted to publication , available at _ http://hal.archives-ouvertes.fr/hal/-00178536/fr/ _ galtchouk , l.,@xmath554 pergamenshchikov , s.(2005 ) . efficient adaptive nonparametric estimation in heteroscedastic regression models . preprint of the strasbourg louis pasteur university , irma , 2005/020 available at _ http://www.univ-rouen.fr/lmrs/persopage/pergamenchtchikov _ galtchouk , l.,@xmath554 pergamenshchikov , s.(2007 ) . adaptive nonparametric estimation in heteroscedastic regression models . sharp non - asymptotic oracle inequalities . preprint of the strasbourg louis pasteur university , irma , 2007/09 , available at _ http://hal.archives-ouvertes.fr/hal/-00179856/fr/ _ galtchouk , l.,@xmath554 pergamenshchikov , s.(2008 ) . sharp non - asymptotic oracle inequalities for nonparametric heteroscedastic regression models . _ journal of nonparametric statistics _ , accepted to publication	the paper deals with asymptotic properties of the adaptive procedure proposed in the author paper , 2007 , for estimating a unknown nonparametric regression . we prove that this procedure is asymptotically efficient for a quadratic risk , i.e. the asymptotic quadratic risk for this procedure coincides with the pinsker constant which gives a sharp lower bound for the quadratic risk over all possible estimates .
let @xmath1 denote the symmetric group of permutations acting on a set @xmath2 of cardinality @xmath3 , and @xmath4 denote the uniform law on @xmath1 , @xmath5 a seminal concentration result on @xmath1 obtained by maurey is the following . @xcite[permmau ] let @xmath6 be the hamming distance on the symmetric group , for all @xmath7 , @xmath8 then for any subset @xmath9 such that @xmath10 , and for all @xmath11 , one has @xmath12 where @xmath13 . milman and schechtman @xcite generalized this result to some groups whose distance is invariant by translation . for example , in the above result we may replace ( up to constants ) the hamming distance by the _ transposition distance _ @xmath14 that corresponds to the minimal number of transpositions @xmath15 such that @xmath16 . the distances @xmath17 and @xmath6 are comparable , @xmath18 ( we refer to @xcite for comments about these comparison inequalities ) . a few years later , a stronger concentration property in terms of dependence in the parameter @xmath3 , has been settled by talagrand using the so - called `` convex - hull '' method @xcite ( see also @xcite ) . this property implies maurey s result with a slitly worse constant . let us recall some notations from @xcite . for each @xmath19 and @xmath20 , let @xmath21 be the set of vectors @xmath22 with @xmath23 for @xmath24 let @xmath25 denote the convex hull of @xmath26 in @xmath27 , @xmath28 where @xmath29 denotes the set of probability measures on @xmath30 . talagrand introduced the quantity @xmath31 with @xmath32 , that measures the distance from @xmath33 to the subset @xmath30 . [ tala95]@xcite for any subset @xmath9 , @xmath34 maurey s concentration result easily follows by observing that @xmath35 and applying tchebychev inequality with usual optimization arguments . talagrand s result has been first extended to product of symmetric groups by mcdiarmid @xcite , and then further by luczak and mcdiarmid to cover more general permutation groups which act suitably `` locally '' @xcite . for any @xmath36 , the _ support _ of @xmath33 , denoted by @xmath37 , is the set @xmath38 and the _ degree _ of @xmath33 , denoted by @xmath39 , is the cardinality of @xmath37 , @xmath40 . by definition , according to @xcite , a subgroup @xmath0 of @xmath1 is _ @xmath41-local _ , @xmath42 , if for any @xmath43 and any @xmath44 with @xmath45 , there exists @xmath46 such that @xmath47 , @xmath48 and @xmath49 . as explained in @xcite , any @xmath50-local group is a direct product of symmetric groups on its orbits , the alternating group ( consisting of even permutations ) is @xmath51-local , and any @xmath51-local group is a direct product of symmetric or alternating groups on its orbits . in the present paper , the concentration result by luczak - mcdiarmid and talagrand is a consequence of a weak transport - entropy inequality satisfied by the uniform law on @xmath0 . we also prove weaker type of transport entropy inequalities for a larger class of probability measures on @xmath0 , denoted by @xmath52 . for a better comprehension of the class of measures @xmath52 , let us first consider the case of the symmetric group @xmath1 on @xmath53:=\{1,\ldots , n\}$ ] . let @xmath54 denote the transposition in @xmath1 that exchanges the elements @xmath55 and @xmath56 in @xmath53 $ ] . one may easily check that the map @xmath57 is one to one . the set of measures @xmath52 consists of probability measures on @xmath1 which are pushed forward by the map @xmath58 of product probability measures on @xmath59 , @xmath60),\;\forall i\in [ n]\big\},\ ] ] where by definition @xmath61 for any subset @xmath62 in @xmath1 . the uniform measure @xmath4 on @xmath1 belongs to @xmath52 since @xmath63 with @xmath64 , where for each @xmath55 , @xmath65 denotes the uniform probability measure on @xmath66 $ ] . let us now construct the class of measures @xmath52 for any @xmath41-local group @xmath0 . to clarify the notations , the elements of @xmath2 are labelled with integers , @xmath67 $ ] . the _ orbit _ of an element @xmath68 , denoted by @xmath69 , is the set of @xmath70 connected to @xmath56 by the elements of @xmath0 , @xmath71 let us observe that for any @xmath72 $ ] , @xmath73 since the identity @xmath74 belongs to @xmath0 . let @xmath0 be a @xmath41-local subgroup of @xmath1 . for any @xmath75 let @xmath76 $ ] . there exists a fixed family of permutions @xmath77 with @xmath75 and @xmath78 with @xmath79 and for every @xmath80 @xmath81 , \quad \mbox{and}\quad \deg(t_{i_jj})\leq \ell,\ ] ] such that the map @xmath82 is one to one . such a family @xmath83 is called `` @xmath41-local base of @xmath0 '' in the present paper . as for the symmetric group , the class @xmath52 is made up of all probability measures on @xmath0 which are pushed forward of product probability measures on @xmath84 by a map @xmath85 defined by , @xmath86 where @xmath87 is a @xmath41-local base of @xmath0 . the class @xmath52 contains the uniform law on @xmath0 , obtained by choosing for each @xmath88 the uniform law on @xmath89 . in this paper , the concentration results are derived from weak transport - entropy inequalities , involving the relative entropy @xmath90 between two probability measures @xmath91 on @xmath0 given by @xmath92 if @xmath93 is absolutely continuous with respect to @xmath4 and @xmath94 otherwise . the terminology `` weak transport - entropy '' introduced in @xcite , encompass many kinds of transport - entropy inequalities from the well - known talagrand s transport inequality satisfied by the standard gaussian measure on @xmath95 @xcite , to the usual csizr - kullback - pinsker inequality @xcite that holds for any ( reference ) probability measure @xmath4 on a polish metric space @xmath96 , namely @xmath97 where @xmath98 denotes the total variation distance between @xmath4 and @xmath93 , @xmath99 where the supremum runs over all measurable subset @xmath30 of @xmath96 . we refer to the survey @xcite for other examples of weak transport - entropy inequalities and their connections with the concentration of measure principle . the next theorem is one of the main result of this paper . it presents new weak transport inequalities for the uniform measure on @xmath0 or measures in the class @xmath52 , that recover the concentration results of theorems [ permmau ] and [ tala95 ] . we denote by @xmath6 the hamming distance on @xmath0 : for any @xmath100 , @xmath101 and the distance @xmath14 is defined as the minimal number of elements of @xmath0 , @xmath15 , with degree less than @xmath41 , such that @xmath16 . for any measures @xmath102 , the set @xmath103 denotes the set of all probability measures on @xmath104 with first marginal @xmath105 and second marginal @xmath106 . we denote by @xmath107 the wasserstein distance between @xmath105 and @xmath106 according to the distance @xmath108 or @xmath109 . @xmath110 we also consider two other optimal weak transport costs , @xmath111 and @xmath112 defined by @xmath113 and @xmath114 where @xmath115 represents a probability measure such that @xmath116 . by jensen s inequality , these weak transport costs are comparable , namely @xmath117 where the last inequality holds for @xmath108 . [ transportsymetrique ] let @xmath0 be a @xmath41-local subgroup of @xmath1 . 1 . let @xmath4 be a probability measure of the set @xmath52 defined by . then , for all probability measures @xmath105 and @xmath106 on @xmath0 , @xmath118 and @xmath119 where @xmath120 when @xmath4 is the uniform law on @xmath0 , the inequalities and hold with @xmath121 2 . let @xmath4 denotes the uniform law on @xmath0 . then , for all probability measures @xmath105 and @xmath106 on @xmath0 , @xmath122 where @xmath123 . the proofs of these results , given in the next section , are inspired by talagrand seminal work on @xmath1 @xcite , and luczak - mcdiarmid extension to @xmath41-local groups @xcite . * comments :* * here is a more popular dual formulation of the transport - entropy inequality : for all @xmath124-lipschitz functions @xmath125 ( with respect to the distance @xmath126 ) , @xmath127 for the uniform measure on @xmath1 , @xmath128 and this property is widely commented in @xcite ; it is also a consequence of hoeffding inequalities for bounded martingales ( see page 18 of @xcite ) . the concentration result derived from item ( a ) are of the same nature as the one obtained by the `` bounded differences approach '' in @xcite . * similarly , by proposition 4.5 and theorem 2.7 of @xcite and using the identity @xmath129 we may easily show that the weak transport - entropy inequality is equivalent to the following dual property : for any real function @xmath130 on @xmath0 and for any @xmath131 , @xmath132 where the infimum - convolution operator @xmath133 , @xmath11 , is defined by @xmath134 + moreover , let us observe that following our proof of in the next section , for each @xmath135 the inequality can be improved by replacing the square cost function by the convex cost @xmath136 given in lemma [ completegraph ] . more precisely , holds replacing @xmath137 by @xmath138 defined by @xmath139 for any @xmath140 . * proposition 4.5 and theorem 9.5 of @xcite also provide a dual formulation of the weak transport - entropy inequality : for any real function @xmath130 on @xmath0 and for any @xmath131 , @xmath141 where the infimum convolution operator @xmath142 is defined by @xmath143 as explained at the end of this section , the property directly provides the following version of the talagrand s concentration result for the uniform law on @xmath0 . + [ talgen ] let @xmath4 denotes the uniform law on a @xmath41-local subgroup @xmath0 of @xmath1 . then , for all @xmath144 and all @xmath145 , one has @xmath146 with @xmath123 . as a consequence , by tchebychev inequality , for any @xmath145 and all @xmath11 , @xmath147 + for @xmath148 , this result is exactly theorem 2.1 by luczak - mcdiarmid @xcite , that generalizes theorem [ tala95 ] on @xmath1 ( since @xmath1 is a 2-local group ) . by projection arguments , theorem [ transportsymetrique ] applied to the uniform law @xmath4 on the symmetric group @xmath1 , also provides transport - entropy inequalities for the uniform law on the slices of the discrete cube @xmath149 . namely , for @xmath150 , let us denote by @xmath151 , @xmath152 , the slice of discrete cube defined by @xmath153 the uniform law on @xmath151 , denoted by @xmath154 , is the pushed forward of @xmath4 by the projection map @xmath155)},\end{array}\ ] ] where @xmath156):=\{\sigma(1),\ldots,\sigma(k)\}$ ] and for any subset @xmath30 of @xmath53 $ ] , @xmath157 is the vector with coordinates @xmath158 $ ] . in other terms , @xmath159 and @xmath160 for all @xmath161 . let @xmath162 denotes the hamming distance on @xmath151 defined by @xmath163 [ transportslice ] let @xmath154 be the uniform law on @xmath151 , a slice of the discrete cube . 1 . for all probability measures @xmath105 and @xmath106 on @xmath151 , @xmath164 and @xmath165 where @xmath166 is the wasserstein distance associated to @xmath162 , @xmath167 is the weak optimal transport cost defined by with @xmath168 , and @xmath169 . 2 . for all probability measures @xmath105 and @xmath106 on @xmath151 , @xmath170 where @xmath171 with @xmath172 for all @xmath173 . up to constants , the weak transport inequality is the stronger one since for all @xmath174 , @xmath175 the proof of theorem [ transportslice ] is given in section [ sectionslice ] . the transport - entropy inequality is derived by projection from the transport - entropy inequality for the uniform measure @xmath4 on @xmath1 . the same projection argument could be used to reach the results of ( a ) from the transport - entropy inequality of item ( a ) of theorem [ transportsymetrique ] , but it provides worse constants . the constant @xmath176 in ( a ) is obtained by working directly on @xmath151 and following similar arguments as in the proof of theorem [ transportsymetrique ] . the results of theorem [ transportslice ] also extend to the multinomial law . let @xmath177 be a set of cardinality @xmath178 and let @xmath179 be a collection of non - zero integers satisfying @xmath180 . the multinomial law @xmath181 is by definition the uniform law on the set @xmath182 , \left\|\big\{i\in[n ] , x_i = e_l\big\}\right\|=k_l \bigg\}.\ ] ] for any @xmath183 , one has @xmath184 . as a result , the weak transport - entropy inequality holds on @xmath185 replacing the measure @xmath154 by the measure @xmath181 . the proof of this result is a simple generalization of the one on @xmath151 , by using the projection map @xmath186 defined by : @xmath187 if and only if @xmath188 , \,\forall i\in j_l,\ ] ] where @xmath189 , k_0+\cdots+k_{l-1 } < i\leq k_0+\cdots+k_{l}\big\}$ ] , with @xmath190 . the details of this proof are let to the reader . a straightforward application of transport - entropy inequalities of theorem is deviation s bounds for different classes of functions . for more comprehension , we present below deviations bounds that can be reached from theorem [ transportsymetrique ] for the uniform law @xmath4 on a @xmath41-local subgroup of @xmath1 . obviously some of these results extend to the class of measures @xmath52 and a similar corollary can be derived from theorem [ transportslice ] on the slices of the discrete cube . for any @xmath191 , the mean of @xmath192 is denoted by @xmath193 . [ cordev ] let @xmath4 be the uniform law on a @xmath41-local subgroup @xmath0 of @xmath1 . let @xmath194 be a real function on @xmath0 . 1 . assume that there exists a function @xmath195 such that for all @xmath196 , @xmath197 where @xmath109 or @xmath108 . then for all @xmath198 , one has @xmath199 and @xmath200 where the constants @xmath201 and @xmath202 are defined as in item @xmath203 of theorem [ transportsymetrique ] . 2 . assume that there exist functions @xmath204 , @xmath205 such that for all @xmath196 , @xmath206 + then , for all @xmath198 , one has @xmath207 and @xmath208 where @xmath209 and @xmath210 . * the above deviation s bounds of @xmath194 around its mean @xmath211 are directly derived from the dual representations , , of the transport - entropy inequalities of theorem [ transportsymetrique ] , when @xmath212 goes to 0 or @xmath212 goes to 1 . by classical arguments ( see @xcite ) , corollary [ cordev ] also implies deviation s bounds around a median @xmath213 of @xmath214 , but we loose in the constants with this procedure . however , starting directly from properties of type theorem [ talgen ] , one simply get the following bound under the condition of @xmath215 : for all @xmath198 , @xmath216 where @xmath217 , @xmath218 , and @xmath210 . + the idea of the proof is to choose the set @xmath219 of measure @xmath10 and to show that under the conditions of @xmath215 , @xmath220 then , the deviation bound above the median directly follows from theorem [ talgen ] by optimizing over all @xmath145 . with similar arguments , the same bound can be reached for @xmath221 . * in ( a ) , the bound above the mean is a simple consequence of . as settled in ( a ) , this bound also holds for the deviations under the mean , and it can be slitly improved by replacing @xmath222 by @xmath223 . this small improvement is a consequence of the weak transport inequality with stronger cost @xmath224 . the same kind of improvement could be reached for the deviations above the mean under additional lipschitz regularity conditions on the function @xmath225 . * let @xmath226^n\to\r$ ] be a 1-lipschitz convex function and let @xmath227^n$ ] be a fixed vector . for any @xmath43 , let @xmath228 . by applying the results of item @xmath215 ( or even ) , to the particular function @xmath229 , @xmath43 , we recover and extend to any @xmath41 subgroup @xmath0 , the deviation inequality by adamczak , chafa and wolff @xcite ( theorem 3.1 ) obtained from theorem [ tala95 ] by talagrand . namely , since for any @xmath100 , @xmath230 with @xmath231 , corollary [ cordev ] implies , for any choice of vector @xmath227^n$ ] , @xmath232 this concentration property on @xmath1 ( @xmath233 ) plays a key role in the approach by adamczak and al . @xcite , to study the convergence of the empirical spectral measure of random matrices with exchangeable entries , when the size of the matrices is increasing . * the aim of this paper is to clarify the links between talagrand s type of concentration results on the symmetric group and functional inequalities derived from the transport - entropy inequalities . for brevity s sake , applications of these functional inequalities will not be examined in the present paper . however , let us briefly mentioned some other applications using concentration results on the symmetric group : the stochastic travelling salesman problem for sampling without replacement ( see appendix @xcite ) , graph coloring problems @xcite . we also refer to the surveys and books @xcite for other numerous examples of application of the concentration of measure principle in randomized algorithms . we start with the proof of ( b ) . under the conditions on the function @xmath194 , for any @xmath234 @xmath235 let @xmath236 . plugging this estimate into the definition of @xmath237 , we get for any @xmath43 @xmath238 as @xmath212 goes to 1 , applied to the function @xmath239 yields @xmath240 and therefore @xmath241 as @xmath212 goes to 0 , yields @xmath242 and therefore @xmath243 the deviation bounds of ( b ) then follows from and by tchebychev inequality by optimizing over all @xmath236 . the deviation bounds of ( a ) are also obtained from by tchebychev inequality . as above , the improvement for the deviation under the mean is a consequence of applied to @xmath239 , as @xmath212 goes to 0 , and using the estimate @xmath244 take a subset @xmath245 and consider the function @xmath246 which takes the values @xmath247 on @xmath30 and @xmath248 on @xmath249 . it holds @xmath250 } \ { \lambda ( 1-\beta ) + \psi(\beta,\sigma)\},\end{aligned}\ ] ] denoting by @xmath251 so it holds @xmath252 } \ { \lambda ( 1-\beta ) + \psi(\beta,\sigma)\ } ; \inf_{\beta \in [ 1-\varepsilon,1 ] } \ { \lambda ( 1-\beta ) + \psi(\beta,\sigma)\},\right ) \\ & \geq \min\left ( \lambda \varepsilon ; \inf_{\beta \geq 1-\varepsilon } \psi(\beta,\sigma ) \right ) \to \inf_{\beta \geq 1-\varepsilon } \psi(\beta,\sigma),\end{aligned}\ ] ] as @xmath253 it is easy to check that for any fixed @xmath33 , the function @xmath254 is continuous on @xmath255 $ ] , so letting @xmath256 go to @xmath247 , we get @xmath257 on the other hand , @xmath258 for all @xmath248 . this proves that @xmath259 . applying to @xmath246 and letting @xmath260 go to infinity yields to @xmath261 it remains to observe that @xmath262 . let @xmath0 be a @xmath41-local subgroup of @xmath1 and @xmath263 be a @xmath41-local base of @xmath0 . let @xmath4 be a probability measure in the set @xmath52 . by definition of @xmath52 , there exists a product probability measure @xmath264 such that @xmath265 where the map @xmath266 is defined by . each transport - entropy inequality of theorem [ transportsymetrique ] is obtained by induction over @xmath3 and using the following partition @xmath267 of the group @xmath0 : for any @xmath268 , @xmath269 if @xmath270 , then @xmath271 and if @xmath272 , then @xmath273 . in any case , from the definition of the @xmath41-local property , we may easily check that @xmath274 is a @xmath41-local group that realizes a natural embedding of a @xmath41-local subgroup of @xmath275 and @xmath276 is a @xmath41-local base of this subgroup . moreover , it is rather easy to get that for any @xmath268 , @xmath277 is the coset defined by @xmath278 . from the definition of @xmath4 , if @xmath279 , then there exist @xmath280 such that @xmath281 and therefore @xmath282 as a consequence , one has @xmath283 . let @xmath284 denote the restriction of @xmath4 to @xmath277 defined by @xmath285 from the construction of @xmath4 , @xmath286 . moreover , for all @xmath287 , @xmath288 , one has @xmath289 this property is needed in the induction step of the proofs . let us note that if @xmath55 and @xmath290 are elements of @xmath291 , then from the @xmath41-local property of @xmath0 , there exists @xmath292 such that @xmath293 and @xmath294 . we also have @xmath295 . if moreover the measure @xmath4 is the uniform law on @xmath0 , then for any @xmath296 , @xmath297 . in that case we will use in the proofs the following property : for any @xmath298 , @xmath299 and @xmath300 since the distance @xmath166 satisfies a triangular inequality , the transport - entropy inequality follows from the following one : for all probability measures @xmath105 on @xmath0 , @xmath301 a dual formulation of this property given by theorem 2.7 in @xcite and proposition 3.1 in @xcite is the following : for all functions @xmath130 on @xmath0 and all @xmath236 , @xmath302 with @xmath303 we will prove the inequality by induction on @xmath3 . when @xmath304 , either @xmath305 and the inequality is obvious , either @xmath306 is the two points space , @xmath233 and one has @xmath307 in that case , exactly corresponds to the following dual form of the csiszar - kullback - pinsker inequality ( see proposition 3.1 in @xcite ): for any probability measure @xmath93 on a polish space @xmath308 , for any measurable function @xmath309 , @xmath310 with @xmath311 the recurrence step will be also a consequence of . let @xmath312 be the partition of @xmath0 defined by . any @xmath234 admits a unique decomposition defined by @xmath313 thus , we define a probability measure @xmath314 on @xmath291 . in particular , according to the definition of the measure @xmath315 , one has @xmath316 it follows that @xmath317 where the last equality is a consequence of property . now , we will bound the right - hand side of this equality by using the induction hypotheses . for any function @xmath318 and any @xmath319 , let @xmath320 denote the function defined by @xmath321 . for any function @xmath322 and any @xmath298 , let us note @xmath323 the next step of the proof relies on the following lemma . [ gateau ] let @xmath324 , for any function @xmath325 and any @xmath298 , one has 1 . @xmath326 + where @xmath327 if @xmath108 and @xmath328 if @xmath109 . 2 . @xmath329 + where @xmath330 if @xmath108 and @xmath331 if @xmath109 , and @xmath332 denotes an element of @xmath0 with @xmath294 and such that @xmath293 . this lemma is obtained using the decomposition of the measures @xmath234 on the @xmath333 s . let @xmath298 . by the triangular inequality and using the invariance by translation of the distance @xmath126 , one has @xmath334 and therefore , since @xmath335 with @xmath336 if @xmath108 and @xmath328 if @xmath109 , @xmath337 it follows that @xmath338 \hat p(l)+c(\ell)\sum_{l\in \orb(n ) } \1_{l\neq i } \hat p(l ) \right\}\\ & = \inf_{\hat p\in \p({\orb(n)})}\inf_{q_l\in \p(h_n),l\in\orb(n)}\\ & \qquad\qquad \left\ { \sum_{l\in \orb(n)}\left [ \int \varphi^{t_{l , n } } \,dq_l+ \int_{h_n } d(\sigma , \tau ) dq_l(\tau)\right ] \hat p(l)+c(\ell)\sum_{l\in \orb(n ) } \1_{l\neq i } \hat p(l ) \right\}\\ & = \inf_{\hat p\in \p(\orb(n))}\left\{\sum_{l\in \orb(n ) } q^{h_n } \varphi^{t_{l , n}}(\sigma ) \hat p(l ) + c(\ell)\sum_{l\in \orb(n ) } \1_{l\neq i } \hat p(l)\right\}.\end{aligned}\ ] ] the proof of the second inequality of lemma [ gateau ] is similar , starting from the following triangular inequality @xmath339 with @xmath340 if @xmath108 and @xmath331 if @xmath109 . the end of the proof of the second inequality of lemma [ gateau ] is let to the reader . the induction step of the proof of continues by applying consecutively lemma [ gateau ] , the hlder inequality , and the induction hypotheses to the measure @xmath341 . if @xmath342 then @xmath343 and @xmath344 if @xmath345 then @xmath346 and for any @xmath268 , @xmath347\\ & = \exp \left[\lambda\inf_{\hat p\in \p(\orb(n))}\left\{\sum_{l\in\orb(n)}\hat \varphi(l ) \hat p(l)+c(\ell ) \sum_{l\in\orb(n ) } \1_{l\neq i } \hat p(l)\right\}+k_{n-1}c(\ell)^2\frac{\lambda^2}8 \right],\end{aligned}\ ] ] where , by using property , @xmath348 . let us consider again the above infimum - convolution @xmath349 defined on the space @xmath350 , with @xmath351 , one has @xmath352 by applying with the probability measure @xmath353 on @xmath291 , the previous inequality gives @xmath354=\exp\left[\lambda \int \varphi \,d\mu + k_{n}c(\ell)^2\frac { \lambda^2}8\right].\end{aligned}\ ] ] the scheme of the induction proof of , with a better constant @xmath201 when @xmath4 is the uniform measure on @xmath0 , is identical , starting from the second result of lemma [ gateau ] and using the property . this is let to the reader . we now turn to the induction proof of the dual formulation of the weak transport - entropy inequality . the sketch of the proof is identical to the one of . for the initial step @xmath304 , in the non - trivial case @xmath306 and @xmath233 , and one may easily check that @xmath355 in that case , the result follows from the following infimum - convolution property . [ completegraph ] for any probability measure @xmath93 on a polish metric space @xmath308 , for all @xmath135 and all measurable functions @xmath309 , bounded from below @xmath356 where for all @xmath357 , @xmath358 and @xmath359 is the convex function defined by @xmath360 observing that @xmath361 for all @xmath362 $ ] , the above inequality also holds replacing @xmath363 by @xmath364 the proof of this lemma can be founded in @xcite ( inequality ( 4 ) ) . for a sake of completeness , we give in appendix a new proof of this result on finite spaces @xmath96 by using a localization argument ( lemma [ chose ] ) . let us now present the key lemma for the induction step in the proof . for any function @xmath322 and any @xmath298 , we define @xmath365 here , writing @xmath366 , we omit the dependence in @xmath201 to simplify the notations . the proof relies on the following lemma . [ lemmeqtilde ] let @xmath268 . for any function @xmath325 and any @xmath298 , one has 1 . @xmath367 with @xmath327 if @xmath108 and @xmath328 if @xmath109 . 2 . @xmath368 where @xmath330 if @xmath108 and @xmath331 if @xmath109 . the proof of this lemma is similar to the one of lemma [ gateau ] . by and the inequality @xmath369 we get for any @xmath370 , @xmath371 it follows that for any @xmath298 , @xmath372 \hat p(l)+\frac{1}{2(1-s)k_n}\bigg(\sum_{l\in\orb(n ) } \1_{l\neq i } \hat p(l)\bigg)^2 \right\}\\ & = \inf_{\hat p\in \p(\orb(n))}\inf_{q_l\in \p(h_n),l\in\orb(n)}\\ & \quad \left\ { \sum_{l\in\orb(n)}\left [ \int \varphi^{t_{l , n } } \,dq_l+\frac{1}{2c(\ell)^2sk_n } \left(\int_{h_n } d(\sigma,\tau ) dq_l(\tau ) \right)^2\right ] \hat p(l)+\frac{1}{2(1-s)k_n}\bigg(\sum_{l\in\orb(n ) } \1_{l\neq i } \hat p(l)\bigg)^2 \right\}\\ & = \inf_{\hat p\in \p(\orb(n))}\left\{\sum_{l\in\orb(n ) } \widetilde q^{h_n}_{k_{n-1 } } \varphi^{t_{l , n}}(\sigma ) \hat p(l ) + \frac 12 \bigg(\sum_{l\in\orb(n ) } \1_{l\neq i } \hat p(l)\bigg)^2\right\},\end{aligned}\ ] ] where the last equality follows by choosing @xmath373 , which ends the proof of the first inequality of lemma [ lemmeqtilde ] . the second inequality of lemma [ lemmeqtilde ] is obtained identically starting from . we now turn to the induction step of the proof . by the decomposition of the measure @xmath4 on the @xmath277 s , we want to bound @xmath374 where the last equality is a consequence of property . if @xmath271 , then the result simply follows for the induction hypotheses apply to the measure @xmath341 . if @xmath375 , then applying lemma [ lemmeqtilde ] , the hlder inequality , and the induction hypotheses , we get @xmath376\right\}\\ & \leq \inf_{\hat p\in \p(\orb(n ) ) } \left\ { \prod_{l\in\orb(n)}\left(\int e^{-(1-\alpha ) \varphi^{t_{l , n } } } d\mu_n \right)^{-\frac{\hat p(l)\alpha}{1-\alpha}}\exp\left[\frac 12 \bigg(\sum_{l\in\orb(n ) } \1_{l\neq i } \hat p(l)\bigg)^2\right]\right\}\\ & = \exp \left[\alpha \inf_{\hat p\in \p(\orb(n))}\left\{\sum_{l\in\orb(n ) } \hat \varphi(l ) \hat p(l)+\frac 12 \bigg(\sum_{l\in\orb(n ) } \1_{l\neq i } \hat p(l)\bigg)^2 \right\ } \right],\end{aligned}\ ] ] where by property , @xmath377 according to the definition of the infimum convolution @xmath378 on the space @xmath379 given in lemma [ completegraph ] , the last inequality is @xmath380 and therefore lemma [ completegraph ] provides @xmath381 the proof of is completed for any measure @xmath315 . to improve the constant when @xmath4 is the uniform probability measure on @xmath0 , the proof is similar using the second inequality of lemma [ lemmeqtilde ] together with property . we prove the dual equivalent property as a consequence of the stronger following result , for any real function @xmath130 on @xmath0 , for any @xmath382 @xmath383 with where the infimum convolution operator @xmath384 is defined by , for @xmath43 @xmath385\setminus \{j\ } } \left(\int \1_{\sigma(k)\neq y(k ) } dp(y)\right)^2 \right\}\label{defqj}. \end{gathered}\ ] ] the proof of relies on lemma [ completegraph ] and the following . for this proof , for any @xmath43 , we define @xmath386 and for @xmath387 $ ] , @xmath388 \setminus \{j\ } } \left(\int \1_{\sigma(k)\neq y(k ) } dp(y)\right)^2 \right\ } . \end{aligned}\ ] ] [ inverse ] let @xmath389 $ ] . for any @xmath43 , one has @xmath390 where @xmath391 . this result follows from the change of variables @xmath392 in the definition of @xmath393 , one has @xmath394 where for the last equality , we use the fact that the map that associates to any measure @xmath234 the image measure @xmath395 under the application @xmath396 , is one to one from @xmath397 to @xmath398 . here is the key lemma for the induction step of the proof of . [ lemclef ] 1 . let @xmath399 . for any @xmath287 , one has @xmath400 2 . if @xmath401 , let @xmath402 with @xmath403 , and let @xmath404 such that @xmath405 and @xmath406 . we note @xmath407 and @xmath408 . for any @xmath287 , for any @xmath409 $ ] one has @xmath410+\frac12(1-\theta)^2.\end{aligned}\ ] ] the first part of this lemma follows from the fact that @xmath411 and the fact that @xmath412 for @xmath298 and @xmath413 . therefore , according to the definition of @xmath414 , one has for @xmath415 , @xmath416\setminus \{j\}}\left(\int \1_{\sigma t_{i , n}(k)\neq y(k ) } dp(y)\right)^2 \right\}\\ & = \inf_{q\in { \mathcal p}(h_n ) } \left\ { \int \varphi^{t_{j , n } } dq + \frac 1{2c(\ell)^2}\sum_{k\in [ n]\setminus \{j\}}\left(\int \1_{\sigma t_{j , n}(k)\neq yt_{i , n}(k ) } dq(y)\right)^2 \right\ } = q^{h_n}\varphi ^{t_{j , n}}(\sigma ) . \end{aligned}\ ] ] for the proof of the second part of lemma [ lemclef ] , let us consider @xmath417 , a collection of measures in @xmath418 , and @xmath419 ( @xmath420 ) . for @xmath421 $ ] , @xmath422,\ ] ] is a probability measure on @xmath0 . according to the definition of @xmath423 , for any @xmath298 , @xmath424 + \frac 1{2c(\ell)^2}(a+b+c ) , \ ] ] with @xmath425\setminus \supp(t_{i , j } ) } \left(\int \1_{\sigma t_{i , n}(k)\neq y(k ) } dp(y)\right)^2 , \quad b=\sum_{k\in d_i } \left(\int \1_{\sigma t_{i , n}(k)\neq y(k ) } dp(y)\right)^2,\ ] ] and @xmath426 . since @xmath298 and @xmath427 , one has @xmath428 and @xmath429 . it follows that @xmath430 for any @xmath431 $ ] and @xmath432 , let us note @xmath433 by the cauchy - schwarz inequality , one has @xmath434\setminus \supp(t_{i , j } ) } u_i^2(k , l ) + ( 1-\theta ) \sum_{k\in [ n]\setminus \supp(t_{i , j } ) } u_j^2(k)\right].\ ] ] we also have @xmath435\\ & \leq \sum_{k\in d_i } \left [ \frac { 2\theta^2 } d u_i^2(k , k)+2(1-\theta)^2 + \frac{\theta^2 } d\sum_{l\in d_i\setminus \{k\}}u_i^2(k , l)\right]\\ & \leq 2d^2(1-\theta)^2 + \frac \theta d \sum_{l\in d_i } \left [ 2u_i^2(l , l ) + \sum_{k\in d_i\setminus \{l\ } } u_i^2(k , l)\right]\end{aligned}\ ] ] all these estimates together provide @xmath436\setminus \{i , l\ } } u_i^2(k , l)\right ) + ( 1-\theta ) \sum_{k\in [ n]\setminus \supp(t_{i , j } ) } u_j^2(k)\right ] . \end{gathered}\ ] ] since @xmath437 , we get by optimizing over all @xmath438 and all @xmath439 , @xmath440 + \frac 12 ( 1-\theta)^2 , \ ] ] with @xmath441\setminus \{i , l\ } } \int\1_{\sigma t_{i , n}(k)\neq y(k ) } dp_i(y)\right\}\\ & = \inf_{q_i\in\p(h_n)}\left\{\int \varphi^{t_{i , n } } dq_i+\frac 1 { c(\ell)^2}\int \1_{\sigma ( t_{i , n}(l))\neq y(t_{i , n}(l ) ) } dq_i(y ) + \frac1 { 2c(\ell)^2}\sum_{k\in [ n-1]\setminus \{t_{i , n}(l)\ } } \int\1_{\sigma t_{i , n}(k)\neq y(k ) } dp_i(y)\right\}\\ & = q^{h_n , t_{i , n}(l ) } \varphi^{t_{i , n}}(\sigma ) \end{aligned}\ ] ] and @xmath442\setminus \supp(t_{i , j } ) } \int \1_{\sigma t_{i , n}(k)\neq y(k ) } dp_i(y)\right\}\\ & = \inf_{p_j\in\p(h_j)}\left\{\int \varphi dp_j+\frac 1{2 c(\ell)^2 } \sum_{k\in [ n]\setminus \supp(t_{i , j } ) } \int\1_{\sigma t_{i , n}(k)\neq y t_{i , j}(k ) } dp_j(y)\right\}\\ & \leq \inf_{p_j\in\p(h_j)}\left\{\int \varphi dp_j+\frac 1{2 c(\ell)^2 } \sum_{k\in [ n]\setminus \{i\ } } \int\1_{\sigma t_{i , n}(k)\neq y t_{i , j}(k ) } dp_j(y)\right\}\\ & \leq \inf_{q_j\in\p(h_n)}\left\{\int \varphi^{t_{i , n}t_{i , j}^{-1 } } dq_j+\frac 1{2 c(\ell)^2 } \sum_{k\in [ n]\setminus \{i\ } } \int\1_{\sigma t_{i , n}(k)\neq y t_{i , n}(k ) } dp_j(y)\right\}\\ & = \inf_{q_j\in\p(h_n)}\left\{\int \varphi^{t_{i , n}t_{i , j}^{-1 } } dq_j+\frac 1{2 c(\ell)^2 } \sum_{k\in [ n-1 ] } \int\1_{\sigma ( k)\neq y ( k ) } dp_j(y)\right\}\\ & = q^{h_n}\varphi^{t_{i , n}t_{i , j}^{-1}}\end{aligned}\ ] ] where we use successively the following arguments : @xmath443 for any @xmath431\setminus \supp(t_{i , j})$ ] , the set @xmath53\setminus \supp(t_{i , j})$ ] is a subset of @xmath53\setminus \{i\}$ ] , for any @xmath444 one has @xmath445 . this ends the proof of lemma [ lemclef ] . we will know prove by induction over @xmath3 . for @xmath304 , in the non - trivial case , the set @xmath0 is the @xmath50 points space @xmath446 which is @xmath50-local . for @xmath447 , and for any @xmath448 , @xmath449 as a consequence , we get the expected result from lemma [ completegraph ] applied with @xmath450 . let us now present the induction step . we assume that holds at the rank @xmath451 for all @xmath452 . by symmetry , we may assume without loss of generality that @xmath453 in since @xmath4 is the uniform probability measure on @xmath0 . we assume that @xmath401 , otherwise the induction step is obvious . we first apply lemma [ inverse ] , @xmath454 let @xmath455 . according to the decomposition of the measure @xmath4 on the sets @xmath456 , @xmath457 where @xmath284 is the uniform probability measure on @xmath277 . for @xmath458 , let us note @xmath459 we choose @xmath460 such that @xmath461 by property and then applying lemma [ lemclef ] , we get @xmath462 by the induction hypotheses applied to the measure @xmath341 , it follows that @xmath463 let us now consider @xmath403 , @xmath324 , property , the second part of lemma [ lemclef ] and jensen s inequality yield : for any @xmath409 $ ] , @xmath464+\frac{\alpha}2(1-\theta)^2\right\}\\\end{aligned}\ ] ] by the induction hypotheses applied with the uniform measure @xmath341 on the @xmath41-local subgroup @xmath274 , and from property , it follows that @xmath465 according to the definition of the infimum - convolution operator @xmath466 defined on the space @xmath467 , we may easily check that for every @xmath268 , @xmath468 } \left\{\theta \hat g ( i ) + ( 1-\theta)\min_{k\in\orb(n ) } \hat g ( k)+\frac12 ( 1-\theta)^2 \right\}.\ ] ] therefore optimizing over all @xmath421 $ ] , we get from and : for all @xmath268 , @xmath469 finally , from lemma [ completegraph ] applied with the uniform probability measure @xmath470 on @xmath291 , the equality gives @xmath471 the proof of is completed . we adapt to the space @xmath151 the proof of item ( a ) of theorem[transportsymetrique ] . in order to avoid redundancy , we only present the main steps of the proof . by duality , it sufises to prove that for all functions @xmath130 on @xmath151 and all @xmath236 , @xmath472 where @xmath473 and for any @xmath131 , @xmath474 where for @xmath475 , @xmath476 the proof is by induction over @xmath3 and @xmath477 . for any @xmath150 , if @xmath478 or @xmath479 , the set @xmath151 is reduced to a singleton and the inequalities or are obvious . for @xmath304 and @xmath480 , @xmath151 is a two points set , and directly follows from property and lemma [ completegraph ] on @xmath481 . for the induction step , we consider the collection of subset @xmath482 , with @xmath483 , defined by @xmath484 since for any @xmath485 , @xmath486 any probability measure @xmath487 on @xmath151 admits a unique decomposition defined by @xmath488 thus , we define probability measures @xmath489 and a probability measure @xmath314 on the set @xmath490 . for the uniform law @xmath4 on @xmath151 , one has @xmath491 where @xmath492 is the uniform law on @xmath493 , @xmath494 , for any @xmath495 . for any @xmath496 , let @xmath497 denote the map that exchanges the coordinates @xmath498 by @xmath499 and @xmath500 by @xmath501 for any point @xmath161 . this map is one to one from @xmath493 to @xmath502 . for any @xmath503 , the set @xmath504 can be identify to @xmath505 and therefore the induction hypotheses apply for the uniform law @xmath492 on @xmath493 with hamming distance @xmath506\setminus\{i , j\ } } \1_{x_k\neq y_k},\qquad x , y\in \omega^{i , j}.\ ] ] for any function @xmath507 and any @xmath495 , we define @xmath508 and @xmath509 the key lemma of the proof that replaces lemma [ gateau ] and [ lemmeqtilde ] is the following . [ lemslice ] for any function @xmath510 and any @xmath495 , one has @xmath511 and @xmath512 the proof of this lemma is obtained by decomposition of the measures @xmath513 on the sets @xmath493 , and using the following inequality@xmath514 for any @xmath495 , @xmath515 . finally , the proof of the induction step by using lemma [ lemslice ] and the identity @xmath516 is let to the reader . we will explain the projection argument on the dual formulations of the transport - entropy inequalities . according to proposition 4.5 and theorem 9.5 of @xcite , the weak transport - entropy inequality is equivalent to the following property that we want to establish : for any real function @xmath517 on @xmath151 and for any @xmath131 , @xmath518 where @xmath519 let us apply property to the function @xmath520 . since @xmath159 , we get @xmath521 the inequality is an easy consequence of the following result . [ tartine ] for any @xmath36 , @xmath522 it remains to prove this lemma . by definition , one has @xmath523\right\}.\end{aligned}\ ] ] let @xmath524 such that @xmath525 . @xmath526 for @xmath527 , let us note @xmath528 , y_i=1\}$ ] . then @xmath529 if and only if @xmath530)=y$ ] . assume that @xmath531 $ ] , if @xmath530)=y$ ] and @xmath532 then @xmath533 . therefore one has @xmath534)=y , \sigma(j)\not\in y \big\}\subset \big\{\tau , u(\tau)=y,\sigma(j)\neq \tau(j)\big\}.\ ] ] assume now that @xmath535 $ ] , if @xmath530)=y$ ] and @xmath536 then we also have @xmath533 . it follows that @xmath534)=y , \sigma(j)\in y \big\}\subset \big\{\tau , u(\tau)=y,\sigma(j)\neq \tau(j)\big\}.\ ] ] from these observations , we get @xmath537 } \left(\int \1_{u(\tau)=y,\sigma(j)\not\in y } dp(\tau)\right)^2 + \sum_{j\in [ n]\setminus[k ] } \left(\int \1_{u(\tau)=y,\sigma(j)\in y } dp(\tau)\right)^2\\ & = \sum_{j\in [ k ] } \left(\int \1_{\sigma(j)\not\in y } dq(y)\right)^2 + \sum_{j\in [ n]\setminus[k ] } \left(\int \1_{\sigma(j)\in y } dq(y)\right)^2\\ & = \sum_{i\in \sigma([k ] ) } \left(\int \1_{i\not\in y } dq(y)\right)^2 + \sum_{i\not\in \sigma([k ] ) } \left(\int \1_{i\in y } dq(y)\right)^2\\ & = \sum_{i\in \sigma([k ] ) } \left(\int \1_{y_i=0 } dq(y)\right)^2 + \sum_{i\not\in \sigma([k ] ) } \left(\int \1_{y_i=1 } dq(y)\right)^2\\\end{aligned}\ ] ] setting @xmath538 , it follows that @xmath539\\ & = \sum_{i=1}^n \left(\int \1_{y_i\neq x_i } dq(y)\right)^2.\end{aligned}\ ] ] this inequality provides @xmath540 the proof of lemma [ tartine ] and item ( b ) of theorem [ transportslice ] is completed . let @xmath135 and @xmath517 be a real function on the finite set @xmath96 . we want to show that for any probability measure @xmath93 on @xmath96 , @xmath541 we will apply the following lemma whose proof is given at the end of this section . [ chose ] let @xmath542 be a real function on @xmath96 and @xmath543 . let us consider the set @xmath544 if @xmath545 is not empty , then the extremal points of this convex set are dirac measures or convex combinations of two dirac measures on @xmath96 . given a real function @xmath517 on @xmath96 , for any @xmath543 , let @xmath546 one has @xmath547 the supremum of the linear function @xmath548 on the non empty convex set @xmath549 is reached at an extremal point of @xmath549 . therefore , by lemma [ chose ] , we get @xmath550 } \left ( ( 1-\lambda)e^ { \alpha \widetilde r^\alpha f(x)}+\lambda e^ { \alpha \widetilde r^\alpha f(y ) } \right)^{1/\alpha}\left((1-\lambda ) e^{-(1-\alpha)f(x ) } + \lambda e^{-(1-\alpha)f(y)}\right)^{1/(1-\alpha)}\end{gathered}\ ] ] now , let @xmath551 and @xmath552 be some fixed points of @xmath96 . it remains to show that for any real function @xmath517 on @xmath553 and for any @xmath554 , @xmath555 the left - hand side of this inequality is invariant by translation of the function @xmath517 by a constant . therefore , by symmetry , we may assume that @xmath556 . it follows that @xmath557 . therefore we want to check that for any non - negative function @xmath517 on @xmath558 , for any @xmath559 $ ] , @xmath560 or equivalently , setting @xmath561 , @xmath562 since @xmath563 is a convex function on @xmath255 $ ] . so , it suffices to check that @xmath564 , where @xmath565 the function @xmath566 is concave and @xmath567 . for all @xmath218 , one has @xmath568 the function @xmath569 is a bijection from @xmath570 to @xmath571 $ ] . it follows that @xmath572 } \left\{\theta h + c_\alpha ( 1-\theta)\right\},\quad h\geq 0,\ ] ] where @xmath359 is the convex function defined by @xmath573.\ ] ] after computations , we get @xmath574 and therefore we exactly have for any @xmath357 , @xmath575 } \left\{\theta f(x ) + c_\alpha ( 1-\theta)\right\}=\widetilde r^\alpha f(x).\ ] ] the proof of lemma [ completegraph ] is completed . we will show that , if @xmath576 is a convex combination of three probability measures @xmath577 , @xmath578 with @xmath579 , @xmath580 , @xmath581 , and @xmath582 , and @xmath583 , @xmath584 , @xmath585 , then there exists two measures @xmath586 in @xmath545 and @xmath587 $ ] such that @xmath588 we will assume that @xmath592 . the case @xmath593 can be treated identically and the proof in that case is let to the reader . since @xmath592 and @xmath594 , there exists @xmath595 $ ] such that @xmath596 so , we may now assume that @xmath604 and @xmath605 and therefore @xmath606 . in that case , we exactly have @xmath607 let us choose @xmath608 the equalities ensure that @xmath609 and @xmath610 . the proof of lemma [ chose ] ends by checking that @xmath611 . one has @xmath612 according to the definitions of @xmath613 , we may easily check that @xmath614 , and @xmath615 since @xmath616 , one has @xmath617 and therefore @xmath618 as a consequence @xmath619 and according to , we get @xmath620 c. mcdiarmid . on the method of bounded differences . in _ surveys in combinatorics , 1989 ( norwich , 1989 ) _ , volume 141 of _ london math . lecture note ser . _ , pages 148188 . cambridge univ . press , cambridge , 1989 .	following talagrand s concentration results for permutations picked uniformly at random from a symmetric group @xcite , luczak and mcdiarmid have generalized it to more general groups @xmath0 of permutations which act suitably ` locally ' . here we extend their results by setting transport - entropy inequalities on these permutations groups . talagrand and luczak - mc - diarmid concentration properties are consequences of these transport - entropy inequalities . we also consider transport - entropy inequalities on @xmath0 for a larger class of measures . by projection , we derive transport - entropy inequalities for the uniform law on the slice of the discrete hypercube and more generally for the multinomial law . the transport - entropy inequalities settled in this paper are new examples , in discrete setting , of weak transport - entropy inequalities introduced in @xcite .
recent satellite observations indicate that the earth s magnetotail is generally in a state of intermittent turbulence constantly intermixed with localized fast bursty bulk flows ( bbf ) [ _ angelopoulos et al . _ , 1996 ; _ lui _ , 1998 ] . a model of sporadic and localized merging of coherent magnetic structures has been proposed by _ chang _ [ 1999 ; and references contained therein ] to describe the dynamics of the magnetotail . when conditions are favorable , the coherent structures may merge , interact , convect and evolve into forced and/or self- organized critical ( fsoc ) states [ _ chang _ , 1992 ] , a condition satisfied by the observational statistics of the burst duration characteristics of the bbf [ _ angelopoulos _ , 1999 ] . when considered with the realistic magnetotail geometry , the coarse - grained dissipation of turbulent fluctuations produced by the interacting multi - scale coherent structures may trigger `` fluctuation - induced nonlinear instabilities '' and possibly initiate substorms . some of the above ideas have been demonstrated by direct numerical simulations [ _ wu and chang _ , 2000 ] . in this paper , we address the dynamics , particularly the preferential acceleration , of the coherent structures in sheared magnetic field geometry . preliminary comparisons of the theoretical ideas and new numerical results with the observed characteristics of bbf will also be presented . in situ observations [ _ angelopoulos et al_. , 1999 ; _ nagai et al . _ , 1998 and references contained therein ] indicate that bbf are an important means of magnetotail transport . removing the fast bbf , the remaining flow state has a small average convection . the bbf are seen during precursor activity , at substorm expansion and recovery phases . near - earth bbf are primarily earthward and consist of multiple , short - lived flow bursts , i.e. , they are intermittent . distribution functions of mid - tail bbf near the source region of a tailward - retreating reconnection site at late substorm expansion phase show that multiple , localized acceleration sites are responsible for the observed flows [ _ angelopoulos et al_. , 1996 ] . distant tail fast flows are tailward , they are associated with plasmoids , they also have multiple flow peaks and multiple current filaments within them , and they are likely localized in the cross - tail dimension ( e.g. _ ieda et al_. [ 2001 ] and references therein ) . thus although preferential acceleration is observed on either side of a mid - tail region , probably associated with the gross formation of an x - point topology , the flows and the currents within them are rather intermittent . when considered in connection with an observed x - point , the primary directions of convection of bbf are earthward ( tailward ) when they are situated earthward ( tailward ) of the x - point [ _ nagai et al . _ , 1998 ] . it is therefore of preponderate interest to decipher the fundamental physical mechanism that are responsible for the preferential acceleration and convective nature of the observed bbf . it had been suggested by _ chang _ [ 1992 ] that the magnetotail could exhibit dynamic forced and/or self - organized critical ( fsoc ) behavior . statistical studies of the ae index [ _ consolini _ , 1997 ] and global auroral imagery [ _ lui et al . _ , 2000 ] during substorm times seem to indicate consistency with such a dynamic description . more recently , _ angelopoulos et al . _ [ 1999 ] demonstrated specifically that the probability distribution of the bbf duration could also be described by a power law , an indication of dynamic criticality . a physically realistic topological model based on the coherent magnetic structures was proposed by _ chang _ [ 1998 , 1999 ] to explain the multiscale nature of the intermittent turbulence that can lead to the fsoc state in the magnetotail and its low dimensional behavior [ _ klimas et al . _ , 1996 ] . it was also suggested that the coarse - grained dissipation of the fluctuations associated with fsoc could lead to nonlinear instabilities and trigger the development of mean - field x - point magnetic structures in sheared magnetic fields . we shall demonstrate below that these ideas lead naturally to the preferential acceleration and observed convective nature of the fast bbf . there is considerable evidence that many turbulent flows are far from being totally disorganized . indeed they can posses `` coherent structures , '' as indicated by experiments and numerical simulations [ e.g. , _ frisch _ , 1995 ] . _ chang _ [ 1998 ] suggested that the coherent structures are basically current filaments in the neutral sheet region of the magnetotail . most field theoretical discussions begin with the concept of propagation of waves . for example , in the mhd formulation , one can combine the basic equations and express them in the following propagation forms:@xmath0 @xmath1 where the ellipsis represents the effects of the anisotropic pressure tensor , the compressible and dissipative effects , and all notations are standard . equations ( 1,2 ) admit the well - known alfvn waves . for such waves to propagate , the propagation vector @xmath2 must contain a field - aligned component , i.e. , @xmath3 . however , near the singularities where the parallel component of the propagation vector vanishes ( the resonance sites ) , the fluctuations are localized . that is , around these resonance sites ( usually in the form of curves in physical space ) , it may be shown that the fluctuations are held back by the background magnetic field , forming coherent structures in the form of flux tubes [ _ chang , _ 1998 ; 1999 ] . for the neutral sheet region of the magnetotail , these coherent structures are essentially force - free and in the form of current filaments in the cross - tail direction [ _ chang _ , 1998 ] . if the cross - tall current is primarily carried by the coherent magnetic structures , these structures will be strongly affected by the magnetic field due to the lorentz force . as they are being preferentially accelerated they will dynamically deform and dissipate fluctuations to produce smaller structures , while occasionally interact and merge with neighboring structures . consider the coherent magnetic structures discussed in the previous section , which in the neutral sheet region of the plasma sheet are essentially filaments of concentrated currents in the cross - tail direction . as we have argued previously , in a sheared magnetic field , these multiscale fluctuations can supply the required coarse - graining dissipation that can produce nonlinear instabilities leading to x - point - like structures of the average magnetic field lines . for a sheared magnetic field @xmath4 , upon the onset of such fluctuation - induced nonlinear instabilities , the average magnetic field will generally acquire a @xmath5-component . let us choose @xmath6 as the earth - magnetotail direction ( positive toward the earth ) and @xmath7 in the cross - tail current direction . then the deformed magnetic field geometry after the development of an x - point structure will generally have a positive ( negative ) @xmath8 component earthward ( tailward ) of the x - point . near the neutral sheet region , the lorentz force will therefore preferentially accelerate the coherent structures earthward if they are situated earthward of the x - point and tailward if they are situated tailward of the x - point , figure 1 . these results would therefore match the general directions of motion of the observed bbf in the magnetotail [ _ nagai et al . _ , 1998 ] . we have performed two - dimensional numerical simulations to verify these conjectures . the simulations are based on a compressible mhd model that has been used by _ wu and chang _ [ 2000 ] in previously studies of coherent magnetic structures . in the first example we started the numerical simulation initially consisting of randomly distributed magnetic fluctuations with zero mean magnetic field . the calculation was carried out with 256 by 256 grid points in a doubly periodic ( @xmath6 , @xmath5 ) domain of length @xmath9 in both directions . eventually , after some elapsed time , multiscale coherent magnetic structures are formed , which are shown in fig . 2(a ) for @xmath10 and @xmath11 . the maximum magnitude of the magnetic field is 0.014 . in this and following examples , both pressure and density are about 1 ; the system has a large plasma @xmath12 to mimic the current sheet region . at this time , a uniform @xmath13 is introduced . in the middle panels of fig . 2 , the effect of the lorentz force is clearly seen through the motions of the coherent structures after an additional elapsed time of 100 . ( for all numerical examples , the unit of time was based on the wave speeds . for instance , the sound speed is about 1.3 ; thus a sound wave would traverse a distance of @xmath14 in a time of 2.4 . ) if the direction of @xmath8 is reversed , the motions of the coherent structures are also reversed as seen in the bottom panels of fig . 2 . in the second example , we have considered the motions of the coherent magnetic structures that developed in a sheared mean magnetic field upon the initial introduction of random magnetic fluctuations . the calculation was carried out with @xmath15 by @xmath16 grid points in a doubly periodic ( @xmath6 , @xmath5 ) domain of length @xmath17 in the @xmath6-direction and @xmath9 in the @xmath5-direction with a sheared magnetic mean field @xmath18 . the system again has a high plasma @xmath12 with both pressure and density about 1 . these structures were generally aligned in the @xmath6-direction near the neutral line , @xmath19 , after some elapsed time . a positive @xmath20 was then applied . it can be seen ( figure 3 ) that , after an additional elapsed time of 50 , the coherent structures ( mostly oriented by currents in the positive @xmath7-direction ) , are generally accelerated in the positive @xmath6-direction ; with one exception where a pair of magnetic structures with oppositely directed currents effectively canceled out the net effect of the lorentz force on these structures . the acceleration continues in time . we have plotted the @xmath6-component of the flow velocities , @xmath21 , due to the cumulative effect of the lorentz ( and pressure ) forces acting on the flow ( and in particular the coherent structures ) after additional duration of 400 has elapsed ( fig . we note that there are a number of peaks and valleys in the 3-dimensional display , with peak velocities of about 0.02 , nearly approaching that of the alfvn speed ( about 0.025 based on @xmath22 ) mimicking the fast bbf that were observed in the magnetotail . it is to be noted that an individual bbf event may be composed of one , two , or several coherent structures . in the final plot ( fig . 5 ) , we demonstrate in a self - consistent picture what might occur near the neutral sheet region of the magnetotail . we injected randomly distributed magnetic fluctuations in a predominantly sheared magnetic field , prescribed by @xmath23 and @xmath24 . the calculation was again carried out with @xmath15 by @xmath16 grid points in a doubly periodic ( @xmath6 , @xmath5 ) domain of length @xmath17 in the @xmath6-direction and @xmath9 in the @xmath5-direction . eventually , a large - scale x - point like mean field magnetic structure is produced . the coherent structures ( aligned in the neutral sheet region ) are subsequently accelerated away from the x - point in both the positive and negative @xmath6-directions . contour plots of @xmath21 after some elapsed time clearly indicated such effects . we demonstrated that dynamical evolutions of coherent magnetic structures in an initially sheared magnetic field might provide a convenient description of the observed fast bbf in the earth s magnetotail , particularly near the neutral sheet region . in the magnetotail , the coherent structures are essentially current filaments in the cross - tail direction . in an x - point mean magnetic field geometry , selective acceleration of these coherent structures can be accomplished by the lorentz force . simulation results demonstrate that the coherent magnetic structures can indeed be preferentially accelerated this way near the neutral sheet and in the appropriate directions . although the x - point mean magnetic field can be formed by nonlinear instabilities due to the stochastic behavior of the coherent magnetic structures , it is not necessary for the x - point to occur this way . an x - point magnetic field geometry formed by forced magnetic reconnection is likely to cause preferential acceleration of the coherent structures in a similar manner . our work is just at the beginning . in the future , we shall consider more general initial two - dimensional mean magnetic field geometries . moreover , the coherent magnetic structures and interactions are usually three - dimensional , characterized by sporadic and localized current disruptions [ _ lui _ , 1996 ] , and the dynamics of the coherent structures , intermittent turbulence , and global interactions are very complicated . we shall look into such complexities . the characteristics of the ion velocity distributions of the observed bbf generally have a crescent - moon shape in the @xmath25 plane [ _ nagai et al . _ , the localized merging of the coherent structures aligned in the neutral sheet region would produce particle jets at the merging point normal to the neutral sheet both in the north and south directions with velocities of the order of the alfvn speed . we suggest that the combined consequence of such localized merging and the preferential acceleration of the coherent structures would produce ion distributions in the bbf of the form as observed . we shall begin a study of such detailed kinetic effects . angelopoulos , v. , f. v. coroniti , c. f. kennel , m. g. kivelson , r.j . walker , c. t. russell , r. l. mcpherron , e. sanchez , c.i . meng , w. baumjohann , g. d. reeves , r. d. belian , n. sato , e. fris - christensen , p. r. sutcliffe , k. yumoto and t. harris , multipoint analysis of a bursty bulk flow event on april 11 , 1985 , _ j. geophys _ , 101 , 4967 , 1996 . angelopoulos , v. , d.g . mitchell , r.w . mcentire , d.j . williams , a.t.y.lui , s.m . krimigis , r.b . decker , s.p . christon , s. kokubun , t. yamamoto , y. saito , t. mukai , f.s . mozer , k. tsuruda , g.d . reeves , w.j . hughes , e. friis - christensen , and o. troshichev , tailward progression of magnetotail acceleration centers : relationship to substorm current wedge . _ j. geophys . _ , 101 , 24599 , 1996 . chang , t. , low - dimensional behavior and symmetry breaking of stochastic systems near criticality can these effects be observed in space and in the laboratory ? _ ieee trans . on plasma science _ , 20 , 691 , 1992 . chang , t. , sporadic localized reconnections and multiscale intermittent turbulence in the magnetotail , in _ geospace mass and energy flow _ , no . 104 in _ geophysical monograph _ , edited by j. l. horwitz , d. l. gallagher , and w. k. peterson ( american geophysical union , washington , d.c . ) , p. 193 , 1998 . consolini , g. , sandpile cellular automata and magnetospheric dynamicsm in _ proc . _ _ vol 58 _ , `` _ cosmic physics in the year 2000 _ '' , s. aiello , n. iucci , g. sironi , a. treves and u. villante ( eds . ) , sif , bologna , italy , 1997 . lui , a. t. y. , multiscale and intermittent nature of current disruption in the magnetotail , in _ physics of space plasmas ( 1998 ) _ , edited by t. chang and j. r. jasperse ( mit geo / cosmo plasma physics , cambridge , massachusetts ) , no . 15 , p. 233	observations indicate that the magnetotail convection is turbulent and bi - modal , consisting of fast bursty bulk flows ( bbf ) and a nearly stagnant background . we demonstrate that this observed phenomenon may be understood in terms of the intermittent interactions , dynamic mergings and preferential accelerations of coherent magnetic structures under the influence of a background magnetic field geometry that is consistent with the development of an x - point mean - field structure .
gaia is the ambitious flagship mission approved by esa for a launch not later than 2012 , and possibly already by q2 of 2010 . the mission is designed to obtain extremely precise astrometry ( in the _ micro-_arcsec regime ) , multi - band photometry and medium / high resolution spectroscopy for a large sample of stars . the goals call for astrometry and photometry ( 5 broad and 11 medium bands covering the optical range ) to be complete over the whole sky to @xmath6 , and spectroscopy complete to @xmath7 . the spectroscopy is primarily meant to provide the radial velocities and thus the 6@xmath8 component in the phase - space to complement the other five provided by astrometry . each star will be visited around a hundred times during the five year mission life - time , in a fashion similar to the highly successful _ hipparcos _ scanning mode . the spectra , photometry and astrometry acquired at each transit _ epoch _ will be individually accessible . gaia has three telescopes / fields of view , roughly equally spaced over a great circle on the sky , so that every @xmath32 hours along the 6 hours spin period , the same star will cross the field of view of astro-1 ( where astrometric position and magnitude in 5 broad bands are measured ) , of astro-2 ( identical to astro-1 ) , and of the spectro - photometric telescope ( where spectra and photometry in @xmath311 medium bands are collected ) . hipparcos performed photometry in three bands . @xmath9 , a sort of white light covering the johnson s @xmath10 and @xmath11 bands , has been the most accurate , with individual measurements accurate to 0.01 mag for solar type stars brighter than @xmath12 . the other two bands , @xmath13 and @xmath14 , were significantly less accurate . gaia will carry instead a fully featured photometric system , composed of a set of 5 broad and @xmath311 medium bands , which are still under fine optimization . figure 1 illustrates the expected errors on the magnitudes derived at each passage in the gaia fields of view for three sample broad bands , three sample medium bands and for the photometric reading accompanying the _ white light _ astrometric measurement . each epoch measurement in any band and for any star brighter than @xmath15 ( irrespective of the spectral type ) will be accurate to better than 0.01 mag . in some bands ( like in 2b56 or 2f59 ) the same limit is reached at @xmath16 . in white light , all individual measurements of stars brighter than @xmath17 will be more accurate than 0.01 mag . field stars have colors corresponding to @xmath3g0 at @xmath18 , @xmath3k0 at @xmath19 and cooler spectral types at fainter magnitudes . therefore , optimization of spectroscopic observations has to be performed for g - k stars . low metallicity stars of the halo that covers a large fraction of the sky should be properly observed as well . wavelength region satisfying these constraints is that of the caii triplet in the far red ( cf . figure 2 ) . this spectral region is interesting even for hotter stars , because it includes the head of the paschen series and strong multiplets # 1 and 8 of ni ( for a detailed discussion see munari 1999 , 2002 ) . figure 3 illustrates the expected errors of gaia radial velocities for single field stars of g - k spectral types and for a single transit . all stars brighter than @xmath16 will have epoch radial velocities accurate to 10 km sec@xmath5 or better , all those brighter than @xmath20 to 1 km sec@xmath5 or better . previous reviews of sb2 eclipsing binaries as seen by gaia that nicely complement the present one have been given by zwitter ( 2002 , 2003 ) , zwitter & munari ( 2004 ) and milone ( 2003 ) . minimal velocity separation for a firm recovery of the sb2 nature for solar type stars is of the order of 25 - 30 km sec@xmath5 . scaling figure 3 to binary star case we see that radial velocity of each star in an sb2 binary with @xmath21 and @xmath220.5 will be derived with an error not exceeding 10 km sec@xmath5 . with @xmath3100 spectra per star randomly distributed in phase it may be expected that 1/3 of them will contribute to precise definition of semi - amplitudes @xmath23 , @xmath24 . if their accuracy is better than 10 km sec@xmath5 , the resulting @xmath23 and @xmath24 will be accurate to 1 km sec@xmath5 or better . figure 1 shows that at @xmath25 the epoch photometry in any of the 5 + 11 bands is accurate to 0.01 mag and in white light to an amazing 0.0006 mag for any of the considered stellar types . therefore _ all sb2 eclipsing binaries with @xmath26 will be accurately observed by gaia_. it is important to check if gaia will collect enough photometric observations to properly constrain the shape of the light - curve , and of the eclipse phases in particular . the measurements in white light and broad bands are essentially simultaneous and altogether @xmath3100 observations scattered over the five year mission lifetime will be collected ( the exact number depends on the ecliptic coordinates and is generally comprised between 60 and 200 ) . gaia will also collect @xmath27 measurements in the medium bands ( again nearly simultaneous in all bands ) . moreover , rvs sky mapper measurements obtained just before spectroscopic observations will provide additional @xmath3100 photometric points in a further photometric band covering the 8480 - 8740 spectroscopic wavelength range . therefore , combining the 2b80 broad band , the 2f89 medium band and the rvs sky mappers , an average of about 370 independent photometric measurements will be obtained at wavelengths close to the cousins @xmath28 band , and 270 measurements for all other bands . such a number of photometric points should suffice for most of the cases . we have embarked in a long term project to assess the accuracy of gaia investigations of sb2 eclipsing binaries ( munari et al . 2001 , zwitter et al . 2003 , marrese et al . 2004 , milone et al . 2004 to be submitted ) . twelve systems have been studied so far , combining ground - based radial velocities acquired in the gaia wavelength range with hipparcos @xmath9 photometry ( to mimic gaia photometric mapping ) . typical light ( and radial velocity ) curves are presented in figure 5 , where orbital solutions are over - plotted . such @xmath9 mapping is generally sufficient to constrain the radii to 2 - 6% , depending of the total number of points and the width of the eclipse phases . with gaia providing 3 to 4times more points , the fraction of sb2 eclipsing binary systems for which the orbital solution will have formal errors in the 1 - 2% range should be remarkable . similar conclusions were reached by niarchos and manimanis ( 2003 ) from their investigation in a gaia - like mode of four over - contact systems . hipparcos derived a frequency of 0.8% of eclipsing binaries among the stars it surveyed . applying this to star counts in the galaxy , we may estimate that gaia will discover 64000 eclipsing binaries with @xmath26 . applying carquillat et al . ( 1982 ) conclusion that 1/4 of ebs are sb2 , we end up with about 16000 sb2 eclipsing binaries brighter than @xmath25 for which gaia should provide orbital solutions formally accurate to 1 - 2% . this is a fantastic number compared to @xmath29100 systems studied at similar accuracy by ground - based observations so far . these 16000 stars will have their parallax measured to 5 micro - arcsec accuracy which corresponds to an error @xmath290.5% in their distance if closer than 1 kpc . such extremely accurate parallaxes will be sed to directly constrain the values of several parameters ( e.g. absolute luminosity ) and physics used in orbital modeling , so that the atmospheric structure and global parameters of eclipsing binaries could be refined to _ absolute _ errors well below 1% . rave ( radial velocity experiment ) is an ambitious program to conduct an all - sky survey ( complete to @xmath11=15.5 mag ) of radial velocities and metallicities of 50 million stars using the 1.2 m uk schmidt telescope of the anglo - australian observatory ( aao ) , together with a northern counterpart , over the period 2006 2010 . a key feature of the rave project is a pilot program aiming to observe @xmath30 stars using the existing 6df facility in unscheduled bright time over the period 20032005 . this pilot program begun on april 11 , 2003 and by the time this conference was held it logged the observations of 20000 stars . the 6df spectrograph brackets the caii triplet region adopted for gaia ( c. figure 2 ) . spectra are taken at a resolution of @xmath31 . the target list of the pilot program includes a large fraction of the @xmath32 hipparcos stars that are accessible from the southern hemisphere as well as some of the @xmath33 stars of the tycho-2 catalog . the survey focuses on stars in the color range @xmath34 . for these stars useful photometric parallaxes can be derived if meaningful trigonometric parallaxes are not available . the rave main survey should utilize a new echidna - style multi fiber spectrograph at the uk - schmidt telescope and its northern counterpart . in the uk - schmidt planned version it consists of a 2250-spine fiber array covering the full field of view ( 40 deg@xmath35 ) . the fiber positions are reconfigurable in @xmath36min . a detailed description of rave and in particular of its science goals , aiming to derive the kinematics and formation history of the milky way , can be found in steinmetz ( 2003 ) . rave observes any given target star only once , even if with remarkable s / n . therefore it is not possible to discover sb1 binaries or trace the motion of components in sb2 systems . however , the respectable resolution and high s / n of rave spectra allow it to detect all sb2 binaries with a suitable luminosity ratio and velocity separation at the time of the observation . given the huge number of stars surveyed by rave , this would put useful statistical constraints ( on velocity separation and luminosity ratio as a function of spectral type and metallicity ) on the population of binary stars in the milky way . figure 7 presents the results of simulations carried to out to infer the sb2 detectability by rave . they show that all f m sb2 binaries with a luminosity ratio @xmath40.5 and observed when velocity separation of their components is @xmath435 km sec@xmath5 will be recognized as such . and new sb2 binaries are already being discovered in the early data secured by rave in 2003 ! carquillat j.m . , nadal r. , ginestet n. , pedoussaut a. 1982 , a&a 115 , 23 marrese , p.m. et al . 2004 , a&a 413 , 635 milone , e.f . 2003 , in ` gaia spectroscopy , science and technology ' , u.munari ed . , asp conf . 298 , pag . 303 munari , u. 1999 , in ` gaia workshop ' , v.straiys ed . , baltic astron . 8 , 73 munari , u. 2001 , in gaia : a european space project " , o.bienaym and c.turon . , edp sciences , pag . 39 munari , u. et al . 2001 , a&a 378 , 477 munari , u. et al . 2003 , in ` gaia spectroscopy , science and technology ' , u.munari ed . , asp conf . 298 , pag . 275 niarchos , p.g . , manimanis , v.n . 2003 , a&a 405 , 263 steinmetz , m. , 2003 , in ` gaia spectroscopy , science and technology ' , u.munari ed . , asp conf . 298 , pag . 381 zwitter , t. 2002 , in iau coll . 187 , asp conf . 279 , pag . 31 zwitter , t. 2003 , in ` gaia spectroscopy , science and technology ' , u.munari ed . , asp conf 298 , pag . 329 zwitter , t. et al . 2003 , a&a 404 , 333 zwitter , t. , munari , u. 2004 , in ` environments and evolution of double and multiple stars ' , iau coll . 191 , revmexaa , c. scarfe ed . , in press ( astro - ph/0306019 )	the expected performance of gaia satellite on eclipsing binaries is reviewed on the basis of ( @xmath0 ) combined hipparcos and ground - based observations mimicking gaia data harvest , and ( @xmath1 ) accurate simulations using the latest instrument model . it is found that for a large majority of the 16000 sb2 eclipsing binaries that gaia will discover at magnitudes @xmath213 , the orbital solutions and physical parameters will be derived with formal accuracies better than 2% . for the same stars the gaia parallax errors will be @xmath35 _ micro_-arcsec , i.e. an error of 0.5% at 1 kpc , which will allow iterative refine of the parameters and physics used in orbital modeling . the detectability of sb2 binaries by the already up and running spectral survey rave is discussed . it is found that all f - to - m sb2 binaries showing a velocity separation @xmath435 km sec@xmath5 and a luminosity ratio @xmath40.5 will be recognized as such . # 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in
because an utterance is best understood in the context in which it is delivered , its interpreters must be able to identify the relevant context and recognize when it is altered , supplanted or revived . the transient nature of speech makes this task difficult . however , the difficulty is alleviated by the abundance of lexical and prosodic cues available to a speaker for communicating the location and type of contextual change . the investigation of the interaction between these cues presupposes a theory of contextual change in discourse . the theory relating attention , intentions and discourse structure@xcite is particularly useful because it provides a computational account of the current context and the mechanisms of contextual change . this account frames the questions i investigate about the correlation between between lexical and prosodic cues . in particular , the theory motivates the selection of the _ cue phrase_@xcite a word or phrase whose relevance is to structural or rhetorical relations , rather than topic and the _ unfilled pause _ ( silent pause ) as significant indicators of discourse structure . to explain the organization of a discourse into topics and subtopics , grosz and sidner postulate three interrelated components of discourse a linguistic structure , an intentional structure and an attentional state@xcite . in the _ linguistic structure _ , the linear sequence of utterances becomes hierarchical utterances aggregate into discourse segments , and the discourse segments are organized hierarchically according to the relations among the purposes or _ discourse intentions _ that each satisfies . the relations among discourse intentions are captured in the _ intentional structure_. it is this organization that is mirrored by the linguistic structure of utterances . however , while the linguistic structure organizes the verbatim content of discourse segments , the intentional structure contains only the intentions that underlie each segment . the supposition of an intentional structure explains how discourse coherence is preserved in the absence of a complete history of the discourse . rather , discourse participants summarize the verbatim contents of a discourse segment by the discourse intention it satisfies . the contents of a discourse segment are collapsed into an intention , and intentions themselves may be collapsed into intentions of larger scope . the discourse intention of greatest scope is the discourse purpose ( dp ) , the reason for initiating a discourse . within this , discourse segments are introduced to fulfill a particular discourse segment purpose ( dsp ) and thereby contribute to the satisfaction of the overall dp . a segment terminates when its dsp is satisfied . similarly , a discourse terminates when the dp that initiated it is satisfied . the _ attentional state _ is the third component of the tripartite theory . it models the foci of attention that exist during the construction of intentional structures . the global focus of attention encompasses those entities relevant to the discourse segment currently under construction , while the local focus ( also called the _ center_@xcite ) is the currently most salient entity in the discourse segment . the local focus may change from utterance to utterance , while the global focus ( i.e. , current context ) changes only from segment to segment . the linguistic , intentional and attentional components are interrelated . in particular , the _ attentional state _ describes the processing of the _ discourse segment _ which has been introduced to satisfy the current _ discourse intention_. the functional interrelation is expressed temporally in spoken discourse the linguistic , intentional and attentional components devoted to one dsp co - occur . therefore , a change in one component reflects or induces changes in the rest . for example , changes ascribed to the attentional state indicate changes in the intentional structure , and moreover , are recognized via qualitative changes in the linguistic structure . it is because of their interdependence and synchrony that i can postulate the hypothesis that co - occurring linguistic and attentional phenomena in spoken discourse cue phrases , pauses and discourse structure and processing are linked . the part of the theory most directly relevant to my investigation are those constructs that model the attentional state . these are the _ focus space _ and the _ focus space stack_. the _ focus space _ is the computational representation of processing in the current context , that is , for the discourse segment currently under construction . within a focus space dwell representations of the entities evoked during the construction of the segment propositions , relations , objects in the world and the dsp of the current discourse segment . a focus space lives on a pushdown stack called the _ focus space stack_. the progression of focus in a discourse is modeled via the basic stack operations pushes and pops applied to the stack elements . for example , _ closure _ of a discourse segment is modeled by _ popping _ its associated focus space from the stack ; _ introduction _ of a segment is modeled by _ pushing _ its associated focus space onto the stack ; _ retention _ of the current discourse segment is modeled by leaving its focus space on the stack in order to add or modify its elements . the contents of a focus space whose dsp is satisfied are accrued in the longer lasting intentional structure . thus , at the end of a discourse the focus space stack is empty while the intentional structure is fully constructed . the focus space model abstracts the processing that all participants must do in order to accurately track and affect the flow of discourse . thus , it treats the emerging discourse structure and the changing attentional foci as publicly accessible properties of the discourse . although the participants themselves may act as if they are manipulating public structures , the informational and attentional properties of a discourse are , in fact , modeled only privately . in explaining certain lexical and prosodic features of discourse , it is often useful to return to these private models . for a speaker s utterance is conditioned both by the state of the her own model and by her beliefs about those of her interlocutors . the time dependent nature of speech emphasizes the importance of synchronizing private models . lexical and prosodic focusing cues hasten synchronization . in particular , they guide the listeners in updating their models ( among them , the focus space stack ) to reflect the attentional changes already in effect for the speaker . for my analysis , the most relevant private model belongs to the _ current speaker _ , whose discourse intentions guide , for the moment , the flow of topic and attention in a discourse and whose spoken contributions provide the richest evidence of attentional state . if cue phrases and unfilled pause durations can be shown to correlate with attentional state ( and by definition , the intentional and linguistic structure ) , the attentional state they reveal belongs to the current speaker , and the attentional changes they denote are the ones the speaker makes in her own private model . the theory of the tripartite nature of discourse frames my hypotheses about the correlation of cue phrases , pause duration and discourse structure . the main hypotheses are these : that particular unfilled pause durations tend to correlate with particular cue phrases and that this correlation is occasioned by changes to the attentional state of the discourse participants , or , equivalently , by the emerging intentional structure of a discourse . changes to the attentional state occur at segment boundaries . cue phrases by definition evince these changes they are utterance and segment - initial words or phrases and they inform on structural or rhetorical relations rather than on topic . thus , for cue phrases , the question is not whether they correlate with attentional state , but how . to answer this question , we ask , for each cue phrase ( e.g. , _ now , to begin with , so _ ) , whether it signals particular and distinct changes to the attentional state . the correlation of unfilled pauses with attentional state is less certain because pauses appear at all levels of discourse structure . they are found within and between the smallest grammatical phrase , the sentence , the utterance , the speaking turn and the discourse segment . their correlation is mainly with the cognitive difficulty of producing a phrase or utterance@xcite . to link this correlation with the task of producing discourse structure , we must posit a variety of attentional operations with corresponding variability in cognitive difficulty . specifically , we construct the chain of assumptions that : @xmath0 @xmath0 @xmath0 to link unfilled pause duration to discourse structure we must first establish that operations on the attentional state can be distinguished sufficiently to explain the different demands that each operation makes on discourse processing and which , therefore , might be reflected in the duration of segment - initial unfilled pauses . the linking of pause duration to the processing of discourse segments motivates some auxiliary hypotheses that refine notions about the kinds of mental operations sanctioned by the focus space model and about the internal structure of a discourse segment . these auxiliary hypotheses are developed in this section . in the theory of discourse structure , changes to the attentional state are modeled as operations on the focus space stack . these operations appear reducible to four distinct sequences of stack operations that correspond to four distinct effects on the attentional state , as follows : @xmath0 @xmath0 @xmath0 @xmath0 the arrangement is asymmetrical in that it is possible to pop more than one focus space per operation , but to push only one , as shown in table [ focusspaceops ] . .*the effect of the four focusing operation on the focus spaces ( fs ) in the pushdown focus space stack . * [ cols="<,^,^ , < " , ] longer pauses were positively correlated with the number of segments opened or closed during one focusing operation ( r = .357 , _ [email protected] ) . this finding might partially explain the long pauses that appear before a _ replace _ , since a _ replace _ is the focusing operation most likely to affect the most focus spaces , by definition , it requires [ almost ] everything to be popped from the focusing before the initiation ( push ) of a new focus space . a correlation of pause duration and the depth of embedding in the linguistic structure ( or equivalently , the number of focus spaces still on the stack ) showed no significant effect on pause duration ( f(1,98 ) = 0.1861 , _ [email protected] ) . the directions dialogue contained too few fragment - initial tokens to calculate meaningful statistics about their relation to focusing operations . therefore , the best course was to select from the raw data ( see table [ mean - values ] ) the patterns that were likely candidates for further testing . for example , _ so _ was never associated with an _ initiate _ operation and also was preceded by the smallest mean pause durations ( 0.13 seconds ) . a filled pause , with a similar mean pause duration ( 0.14 seconds ) was primarily associated with _ initiates _ and _ retains _ but never with _ replace_. and , while _ and _ shared the same focusing operations as a filled pause , its mean value for pause duration was more than twice as large ( 0.33 seconds ) . lccccl initial & & & & & + token & initiate&retain&return&replace&all + & 0.43@xmath2 3 & 0.25@xmath2 2 & 0.25@xmath2 2 & & 0.33@xmath3 7 + * but * & & 0.70@xmath2 1&0.00@xmath2 1&0.10@xmath2 1&0.27@xmath3 3 + * now * & & & & 0.55@xmath2 2 & 0.55@xmath3 2 + * oh * & & 0.00@xmath2 2 & & & 0.00@xmath3 2 + * so * & & 0.15@xmath2 2 & 0.15@xmath2 2 & 0.05@xmath2 1 & 0.13@xmath3 5 + * well * & & & & 0.20@xmath2 2 & 0.20@xmath3 2 + * yknow * & 0.40@xmath2 2 & & & & 0.40@xmath3 2 + * ordinal * & 0.40@xmath2 1 & & & & 0.40@xmath3 1 + * acknow * & 0.10@xmath2 1 & 0.20@xmath2 7 & & 0.90@xmath2 1 & 0.27@xmath3 9 + * ledgment * & & & & & + * filled * & 0.23@xmath2 6 & 0.05@xmath2 4 & 0.00@xmath2 1 & & 0.14@xmath2 11 + * pause * & & & & & + * unmarked * & @xmath4 & @xmath4 & @xmath2 & @xmath2 & @xmath2 + _ all _ & 0.32@xmath4 23&0.21@xmath4 55&0.26@xmath4 11&0.65@xmath4 11&0.29@xmath4 100 + to compensate for the small sample sizes of the cue phrase data , all explicit lexical markers of structure ( cue phrase , acknowledgment , filled pause ) were collapsed into the category , _ marked_. the data in this category were compared to the data for lexically _ unmarked _ fragments . because the longest pauses preceded unmarked _ returns _ and _ replacements _ , i predicted that unmarked operations would in general be preceded by longer pauses than marked . the results are in the direction predicted and are summarized in table [ marked - unmarked ] . the average duration for pauses preceding a marked focusing operation was 0.24 seconds ( standard deviation = 0.24 ) , while the average for pauses preceding unmarked operations was 0.33 seconds ( standard deviation = 0.36 ) . statistically this approaches significance ( t(96 ) = 1.58 , _ p _ = .12 ) . thus far , analysis of the data identifies significantly longer pauses for the _ replace _ operation than for any other and shows that pause duration is positively correlated with the number of segments affected by one focusing operation . these findings begin to distinguish the focusing operations quantitatively , by number of focus spaces affected , and qualitatively , by whether they occur within an established context ( _ initiate , retain , return _ ) or at its beginning ( _ replace _ ) . although , the raw data in table [ rawdata ] appears to show patterns for specific segment - initial tokens , the number of tokens is insufficient for establishing a correlation between cue phrase and focusing operations , let alone a three - way relationship among cue phrase , pause duration and focusing tasks . the categorical classification present particular problems . for , uncertainties arose even with the application of a classification metric . perhaps these uncertainties should have been incorporated into the coding scheme or perhaps the categorical classifications should have been abandoned in favor of additional and quantifiable acoustical and lexical features . only partial conclusions can be drawn from the data . however , the results are useful toward refining the original hypotheses and determining the content of future research . the distinction between the pause data for marked and unmarked fragments is a case in point . for each focusing operation , the difference between mean pause durations at best only approaches significance ( see table [ 24 [ marked - unmarked ] ) . however , because the values for all focusing operations are always greater for unmarked utterances , a hypothesis is suggested : that , given a speech fragment and the focusing operation it evinces , the preceding unfilled pause will be longer if the fragment is lexically unmarked . lccccl speech & & & & & + fragment&initiate&retain&return&replace&all + & 0.30@xmath4 13 & 0.17@xmath4 18&0.13@xmath2 6 & 0.36@xmath2 7&0.24@xmath2 44 + _ unmarked _ & @xmath4 & @xmath4 & @xmath2 & @xmath2 & @xmath2 + _ all _ & 0.32@xmath4 23&0.21@xmath4 55&0.26@xmath4 11&0.65@xmath4 11&0.29@xmath4 100 + if this hypotheses is correct , two accounts can be constructed that would jointly predict the appearance of cue phrases . one account emphasizes the processes involved in choosing and communicating the state of global focus . the other emphasizes the mutually recognized ( by speaker and hearers ) attentional and intentional state of the discourse . together they identify the factors that would impel a speaker to precede an utterance with a cue phrase , an unfilled pause or both . if an unfilled pause preceding a lexically unmarked fragment is significantly longer , we might assume that a particular focusing operation is executed in a characteristic amount of time ( given adequate consideration of other contextual features ) . within this time , we might observe silence , a cue phrase or both . because both pause and cue phrase can appear at the same location in a phrase , we ask if their functions are equivalent , or instead , complementary . my hypothesis selects the second option , that they are complementary in the cognitive processing each reflects and in the discourse functions each fulfills . for , if the duration of an unfilled pause is evidence of the difficulty of a cognitive task , a cue phrase is evidence of its partial resolution . as a communicative device , cue phrases are more cooperative than silence . in silence , a listener can only guess at the current contents of the speaker s models . with the uttering of a cue phrase , the listener is at least notified that the speaker is constructing a response . the minimal cue in this regard is the filled pause . _ bone fide _ cue phrases , however , herald not only an upcoming utterance , but a particular direction of focus and even a propositional relation between prior and upcoming speech . cue phrases serve not only the listener but also the speaker . because they commit to topic structure , but not to specific referents and discourse entities , they buy additional time for the speaker in which to complete a focusing operation and formulate the remainder of the utterance . the account of the influence of the currently observable state of the discourse rests on two patterns in the data : ( 1 ) the difference in pause durations for marked and unmarked _ initiate_s and _ retain_s is minimal ; and ( 2 ) the difference between marked and unmarked _ return_s and _ replace_s is greater . if these patterns can be shown to be significant , they suggest that remaining in the current context is less costly than returning to a former context , or establishing a new one . the corollary is the claim that an expected focusing operation need not be marked , while an unexpected operation is most felicitous when marked . in other words , remaining in the current context or entering a subordinate context is expected behavior , while exiting the current context is not . exiting the current context ( focus space ) carries a greater risk of disrupting a mutual view of discourse structures . the extent of risk is assessed for the listener by the difficulty of tracking the change and for the speaker , by the difficulty of executing it . the risk originates in the nondeterministic definitions of _ return _ and _ replace _ operations both contain in their structure one or more pops . in addition , these operations can be confused because both begin identically , with a series of pops . because closing a focus space is a marked behavior , the clues to changing focus are most cooperative if they guide the listener toward re - invoking a prior context ( i.e. , a _ return _ ) or establishing a new one ( replace ) . thus , certain clues are more likely to mark a return to a former context ( e.g. , _ so , anyway , as i was saying _ ) , while others ( _ now _ , the ordinal phrases ) mark a _ replace_. the goal of future investigations is to establish the bases for predicting the appearance of particular acoustical and lexical features . the speculations presented in this section provide a theoretical framework . if borne out , they can be re - fashioned as characterizations of the circumstances in which cue phrases and unfilled pauses are most likely to be used . the relationships among cue phrases , unfilled pauses and the structuring of discourse are investigated within the paradigm of the tripartite model of discourse . within this model , the postulation of four focusing operations provides an operational framework to which can be tied the discourse functions of cue phrases and the cognitive activity associated with the production of an utterance . especially , the difficulty of utterance production might be explained by the complexity of the co - occurring focusing operation . such a correspondence is , in fact , suggested by the positive correlation of pause duration and the number of focus spaces opened or closed in one operation on the focus space stack . however , because the classification of focusing operations is uncertain , more data and better tests are required to characterize the relationships among the lexical and acoustical correlates of topic and focus . in addition , the aptness of the tripartite model itself is not assured . the idealizations it contains may undergo modification in light of new results , or be augmented by other accounts of discourse processing . on the other hand , the analysis of more quantitative data may confirm the implications of the model , and its appropriateness as the foundation for investigating the lexical and prosodic features of discourse . many thanks to susan brennan who selected and ran the statistical tests on the data and to s. lines for helpful comments . various stages of this work were supervised in turn by chris schmandt and ken haase , both of the m.i.t . media laboratory . pierrehumbert , j. and hirschberg , j. , the meaning of intonation contours in the interpretation of discourse . in _ intentions in communication_. edited by cohen , p. r. , morgan . j. and pollack , m. e. , 1990 , pp . 271 - 311 . sidner , c. l. , focusing in the comprehension of definite anaphora . in _ readings in natural language processing_. ed . by grosz , b. j. , sparck - jones , k. and webber , b. l , morgan kaufman publishers , inc . , 1986 , 363 - 394 . sorensen , j. m. and cooper , w. e. , syntactic coding of fundamental frequency in speech production . in _ perception and production of fluent speech_. ed . by cole , r. a. , published by lawrence erlbaum , 1980 , pp.399 - 440 . walker , m. a. and whittaker , s. , mixed initiative in dialogue : an investigation into discourse segmentation . in _ proceedings of the 28th annual meeting of the association for computational linguistics _ , 1990 , pp.70 - 79 .	expectations about the correlation of cue phrases , the duration of unfilled pauses and the structuring of spoken discourse are framed in light of grosz and sidner s theory of discourse and are tested for a directions - giving dialogue . the results suggest that cue phrase and discourse structuring tasks may align , and show a correlation for pause length and some of the modifications that speakers can make to discourse structure .
how the hadrons are built from quarks and gluons , the fundamental degrees of freedom of qcd , is one of the main open questions in the theory of strong interactions . an efficient way to probe the hadron structure is to study the hadron form factors as these quantities include direct information about the hadron structure . therefore , the form factors have recently received considerable attention both in theory and experiment . like other form factors , the tensor form factors encode the important information about the quark - gluon structure of baryons . the quark distributions in the leading twist are given by the unpolarized distribution @xmath8 , helicty distribution @xmath9 and transversity distribution @xmath10 function of the quark . one of these functions , the transversity distribution -which describes the probability of finding a transversely polarized quark with longitudinal momentum fraction @xmath11 in a transversely polarized baryons @xcite by its chiral - odd nature- is not easy to measure . the strong interactions have approximate chiral symmetry and electroweak interactions conserve the chirality , so the transversity distributions can not be extracted in inclusive deep inelastic scattering ( dis ) . it needs to couple to another chiral - odd quantity in the cross - section . it can be obtained drell - yan processes and semi - inclusive deep inelastic scattering ( sidis ) , because distributions of transversity do come out at leading twist in the cross - section . in ref.@xcite , it was extracted using the experimental data on azimuthal asymmetries in sidis , from belle @xcite as well as data for the nucleon from the hermes @xcite and compass @xcite collaborations . subsequently , in ref.@xcite , isovector nucleon tensor charge obtained @xmath12 at a renormalization scale @xmath13 . in ref.@xcite anselmino et al . updated their nucleon isovector tensor charge result @xmath14 at renormalization scale of @xmath15 . on the theoretical side , isovector tensor charge of the nucleon has been studied in the framework of lattice qcd @xcite , the chiral quark soliton model @xmath16qsm ) @xcite , quark model @xcite , skyrme model @xcite , axial vector dominance model @xcite and dihadron production @xcite qcd sum rules and light cone qcd sum rules @xcite . for octet hyperon isovector tensor transition form factors have been studied chiral quark soliton model @xmath16qsm ) @xcite . the hyperon sector is interesting because it provides for an ideal system in which to study @xmath17 flavor symmetry breaking by replacing up or down quarks in nucleons with strange ones @xcite . tensor form factors play an important role in our understanding of the `` tomography '' of baryons . the hyperon tensor form factors a missing part of the this tomography . in the present work , we calculate the isovector tensor transition form factors of the @xmath1 , @xmath0 and @xmath18 . in order to study the tensor form factors , one needs to use a nonperturbative method . one of the most powerful nonperturbative methods is traditional qcd sum rules ( qcdsr ) , which is more reliable and predictive in calculating the properties of hadrons @xcite . an alternative to the traditional qcd sum rules is the light cone qcd sum rules ( lcsr ) @xcite . in this method , the hadronic properties are expressed with regards to the properties of the vacuum and the light cone distribution amplitudes of the hadrons in the process . since the form factors are expressed with regards to the properties of the qcd vacuum and the distribution amplitudes , any uncertainty in these parameters reflects to the uncertainty of the predictions of the form factors . our aim in this work is to study the tensor form factors in the framework of light - cone qcd sum rules for the octet hyperon . we give the formulation of the hyperon form factors on the light cone and derive our sum rules . in the last section , we present our numerical analysis and discussion . the matrix element of the tensor current between two hyperon states is parametrized by three form factors as presented @xcite @xmath19 u(p),\end{aligned}\ ] ] where @xmath20 = @xmath5 , @xmath6 and @xmath7 baryons and @xmath21 is the tensor current , @xmath22 , @xmath23 and @xmath24 $ ] are the spin operator , and @xmath25 is the spinor of the hyperon with mass @xmath26 and momentum p. in order to calculate three tensor form factors within lcsr , we start our analysis with the following correlation function : @xmath27\|h(p)\rangle,\ ] ] where @xmath28 are the hyperon interpolating fields for the @xmath5 and @xmath6 . in this work , we choose the general form of the interpolating fields for @xmath5 and @xmath6 as @xmath29 where @xmath30 , @xmath31 and @xmath32 , which is an arbitrary parameter . choosing @xmath33 makes the interpolating fields , which are known as ioffe currents . @xmath34 , @xmath35 , @xmath36 are the color indices and @xmath37 is the charge conjugation operator . besides , the u , d , s - quark fields are presented as @xmath38 , @xmath39 and @xmath40 , respectively . in order to calculate sum rules for the tensor form factors , we need to represent the correlation function in two different forms , the first one is calculated in terms of quark and gluon degrees of freedom , and the second one is obtained using hadrons and the related correlation function that is written in terms of a dispersion relation . then , these two forms of correlation functions are equated . in order to suppress the higher states and continuum contributions , we also apply the borel transformation . the hadronic side of the correlation function can be determined as follows , @xmath41 where @xmath42 are the @xmath5 , @xmath6 and @xmath7 baryon mass and dots represents contributions from higher states and continuum . the matrix element of the interpolating current is between the vacuum and hyperon states , defined as @xmath43 where @xmath44 is the hyperon overlap amplitude . inserting the matrix element of the tensor current in eq . ( [ mat ] ) and the matrix element of the interpolating current in eq . ( [ rezi ] ) into the correlation function in eq . ( [ phys ] ) , we get @xmath45 . \end{aligned}\ ] ] also , the correlation function is determined in terms of the quark and gluon on the qcd side . the interpolating fields in eq . ( [ cur ] ) are inserted into the correlation function in eq . ( [ corf ] ) . + we obtain for @xmath0 , @xmath46 and @xmath1 transitions , @xmath47\nonumber\\ & & 4\epsilon^{abc}\langle 0\|q_{1\sigma}^a(0 ) q_{2\theta}^b(x ) q_{3\phi}^c(0)\|h(p)\rangle\end{aligned}\ ] ] where @xmath48 ( @xmath49 ) denotes the quark fields , and @xmath50 represents the quark propagator which is given as @xmath51,\end{aligned}\ ] ] in this expression , @xmath52 is the quark condensate , @xmath53 is defined in terms of the mixed quark gluon condensate as @xmath54 and @xmath55 is the strong coupling constant . @xmath56 is the gluon field strength tensor . the terms proportional to @xmath56 are expected to give negligibly small contributions as they are related to four and five - particle distribution amplitudes @xcite , and hence we will get neglect these terms in further analysis . moreover , the terms proportional to @xmath57 are removed by borel transformations and , finally only the first term , which gives the hard - quark propagator , will be considered for our discussion . then , we need to know matrix elements of the local three - quark operator , @xmath58 where @xmath59 , @xmath60 and @xmath61 are real numbers . this matrix element can be written in terms of distribution amplitudes ( das ) using the lorentz covariance , the spin and the parity of the baryon @xcite @xmath62 where @xmath63 is the spinor of the baryon , m is the mass of the baryon , @xmath37 is the charge conjugation matrix , and @xmath64 $ ] . the `` calligraphic '' expressions can be expressed in terms of functions of the definite twist : @xmath65 where @xmath66 and @xmath67 are scalar , pesudoscalar , vector , axialvector and tensor das , respectively . the expansion of the matrix element is basically an expansion in increasing twists of the das . the twist of a da is defined as the dimension minus the spin of the operators contributing to a given da . the das @xmath68 , @xmath69 and @xmath70 have twist three , @xmath71 , @xmath72 , @xmath73 , @xmath74 , @xmath75 , @xmath76 , @xmath77 , @xmath78 and @xmath79 have twist 4 , @xmath80 , @xmath81 , @xmath82 , @xmath83 , @xmath84 , @xmath85 , @xmath86 , @xmath87 and @xmath88 are of twist 5 , and @xmath89 , @xmath90 and @xmath91 have twist 6 . the das @xmath92 can be written as @xmath93 where @xmath94 with @xmath95 corresponds to longitudinal momentum fractions carried by the quarks inside the baryon . the explicit form of the hyperon das ( @xmath96 ) is studied in detail in refs . the das of the octet hyperons up to twist-6 are investigated the basis of the qcd conformal partial wave expansion approach . for the @xmath5 and @xmath7 baryons calculations are carried out to the next - to - leading order @xcite and for the @xmath6 baryon this calculation is carried out to the leading order of conformal spin accuracy @xcite . the nonperturbative parameters ( e.g. shape parameters of das ) are obtained using the qcd sum rules @xcite . note that the hadronic representation , eq . ( [ res1 ] ) , and the qcd representation are obtained in different kinematical regions . the two expression can be related to each other by using the spectral representation of the correlation functions . quite generally , the coefficients of various structures in the correlation function can be written as : @xmath97 where @xmath98 is called the spectral density . the spectral density can be calculated both using the hadronic representation of the correlation function , @xmath99 , or using the qcd representation , @xmath100 . once @xmath98 is obtained , the spectral representation allows one to evaluate the correlation function in all kinematical regions for @xmath101 and @xmath102 . the lcsr are obtained by matching the expression of the correlation function in terms of qcd parameters to its expression in terms of the hadronic properties , using their spectral representation . in order to do this , we choose the structures proportional to structures @xmath103 , @xmath104 and @xmath105 for the form factors @xmath106 , @xmath107 and @xmath108 , respectively . choosing the coefficients of these structures and applying the borel transformation with respect to the variable @xmath109 we obtain tensor form factors for @xmath110 transition , @xmath111(x_1,x_2,1-x_1-x_2)\nonumber \\ & & -2\int_0 ^ 1 d_\beta \frac { m_\sigma^3}{(q -p\beta)^4}\int_0^{\beta}d\alpha \int_\alpha^{1}dx_2 \int_0^{1-x_2}dx_1\bigg[t_1-t_2-t_5+t_6 - 2t_7 - 2t_8\bigg]\nonumber\\&&(x_1,x_2,1-x_1-x_2)\nonumber\\ \nonumber\\ e_t(q^2 ) \frac{\lambda_\sigma } { { m^2_{\sigma}-p'^2 } } & = & 2 \int_0 ^ 1 dx_2 \frac { m_\sigma}{(q -p x_2)^2 } \int_0^{1-x_2}dx_1 \bigg[-s_1+p_1 + 2t_1-t_3-t_7\bigg](x_1,x_2,1-x_1-x_2)\nonumber \\ & & -2\int_0 ^ 1 d_\beta \frac { m_\sigma^3}{(q -p\beta)^4}\int_0^{\beta}d\alpha \int_\alpha^{1}dx_2 \int_0^{1-x_2}dx_1\bigg[t_1-t_2-t_5+t_6 - 2t_7 - 2t_8\bigg]\nonumber\\&&(x_1,x_2,1-x_1-x_2)\nonumber\\ \nonumber\\ \tilde h_t(q^2 ) \frac{\lambda_\sigma } { { m^2_{\sigma}-p'^2 } } & = & -4m_\sigma^{3 } \int_0 ^ 1 d\alpha \frac{1-\alpha}{(q -p\alpha)^4 } \int_\alpha^{1}dx_2 \int_0^{1-x_2}dx_1\bigg[t_1-t_3-t_7\bigg](x_1,x_2,1-x_1-x_2)\end{aligned}\ ] ] for @xmath112 transition @xmath113(x_1,x_2,1-x_1-x_2)\nonumber \\ & & -2\int_0 ^ 1 d_\beta \frac { m_\lambda^3}{(q -p\beta)^4}\int_0^{\beta}d\alpha \int_\alpha^{1}dx_2 \int_0^{1-x_2}dx_1\bigg[t_1-t_2-t_5+t_6 - 2t_7 - 2t_8\bigg]\nonumber\\&&(x_1,x_2,1-x_1-x_2)\nonumber\\ \nonumber\\ e_t(q^2 ) \frac{\lambda_\sigma } { { m^2_{\lambda}-p'^2 } } & = & 2 \int_0 ^ 1 dx_2 \frac { m_\lambda}{(q -p x_2)^2 } \int_0^{1-x_2}dx_1 \bigg[-s_1 + p_1 + 2t_1-t_3-t_7\bigg](x_1,x_2,1-x_1-x_2)\nonumber \\ & & -2\int_0 ^ 1 d_\beta \frac { m_\lambda^3}{(q -p\beta)^4}\int_0^{\beta}d\alpha \int_\alpha^{1}dx_2 \int_0^{1-x_2}dx_1\bigg[t_1-t_2-t_5+t_6 - 2t_7 - 2t_8\bigg]\nonumber\\&&(x_1,x_2,1-x_1-x_2)\nonumber\\ \nonumber\\ \tilde h_t(q^2 ) \frac{\lambda_\sigma } { { m^2_{\lambda}-p'^2 } } & = & -4m_\lambda^{3 } \int_0 ^ 1 d\alpha \frac{1-\alpha}{(q -p\alpha)^4 } \int_\alpha^{1}dx_2 \int_0^{1-x_2}dx_1\bigg[t_1-t_3-t_7\bigg](x_1,x_2,1-x_1-x_2)\end{aligned}\ ] ] for @xmath114 transition , @xmath115(x_1,x_2,1-x_1-x_2)\nonumber \\ & & + 2\int_0 ^ 1 d_\beta \frac { m_\xi^3}{(q -p\beta)^4}\int_0^{\beta}d\alpha \int_\alpha^{1}dx_2 \int_0^{1-x_2}dx_1 \bigg[v_1-v_2-v_3-v_4-v_5+v_6-a_1+a_2\nonumber\\ & & -a_3-a_4+a_5-a_6\bigg](x_1,x_2,1-x_1-x_2)\nonumber\\ \nonumber \\ e_t(q^2 ) \frac{\lambda_\xi } { { m^2_{\xi}-p'^2 } } & = & 4m_\xi \int_0 ^ 1 dx_2 \frac{1-x_2}{(q -px_2)^2 } \int_0^{1-x_2}dx_1 \bigg[a_1-a_2-v_1+v_2 \bigg](x_1,x_2,1-x_1-x_2)\nonumber\\ \nonumber \\ & & + \int_0 ^ 1 d_\beta \frac { m_\xi^3}{(q -p\beta)^4}\int_0^{\beta}d\alpha \int_\alpha^{1}dx_2 \int_0^{1-x_2}dx_1 \bigg[v_1-v_2-v_3-v_4-v_5+v_6-a_1\nonumber \\ & & + a_2-a_3-a_4+a_5-a_6\bigg](x_1,x_2,1-x_1-x_2)\nonumber\\ \tilde h_t(q^2 ) \frac{\lambda_\xi } { { m^2_{\xi}-p'^2 } } & = & 4m_\xi^{3 } \int_0 ^ 1 d\alpha \frac{1-\alpha}{(q -p\alpha)^4 } \int_\alpha^{1}dx_2 \int_0^{1-x_2}dx_1 \bigg[v_1-v_2-v_3-a_1+a_2-a_3\bigg]\nonumber\\&&(x_1,x_2,1-x_1-x_2)\end{aligned}\ ] ] in order to eliminate the subtraction terms in the spectral representation of the correlation function , we perform to the borel transformation . after the transformation , contributions coming from excited and continuum states are also exponentially suppressed . clearly , the borel transformation and the subtraction of higher states are achieved by using the following substitution rules ( see e.g. @xcite ) : @xmath116 where , @xmath117 @xmath118 is the borel mass and @xmath119 is the solution of the quadratic equation for @xmath120 : @xmath121 where @xmath122 is the continuum threshold . in this section , we present the numerical results of the octet - octet hyperon isovector tensor transition form factors . in this work , we use the hyperon das which depend on various nonperturbative parameters such as , @xmath123 , @xmath124 and @xmath125 @xcite . the das of the @xmath5 , @xmath6 and @xmath7 baryons have been calculated by employing the qcd sum rules without higher - order terms in ref . then the study including higher - order corrections have been done for @xmath5 and @xmath7 baryon in ref . @xcite . in table [ parameter_table ] , we present the values of the input parameters using the das of @xmath5 , @xmath6 and @xmath7 baryons . for the numerical analysis , we use the values of the hyperon masses as follow ; @xmath126 , @xmath127 , @xmath128 @xcite . besides , we need also specify the values of the residues of , @xmath5 and @xmath6 baryon . the residues can be determined from the mass sum rules as @xmath129 and @xmath130 for @xmath5 and @xmath6 , respectively @xcite . \|l\|l\|l\| + @xmath131 & @xmath132 & @xmath133 + & & + @xmath134 gev@xmath135 & @xmath136 & @xmath137 + @xmath138 gev@xmath135 & @xmath139 & @xmath140 + @xmath141 gev@xmath135 & @xmath142 & @xmath143 + @xmath144 gev@xmath135 & @xmath145 & @xmath146 + @xmath147 & @xmath148 & + @xmath149 & @xmath150 & + @xmath151 & @xmath152 & + @xmath153 & @xmath154 & + @xmath155 & @xmath156 & + @xmath157 & @xmath158 & + @xmath159 & & + @xmath160 & & + in the traditional analysis of sum rules , the spectral density of the higher states and the continuum are parameterized using quark hadron duality . in this approach , the spectral density corresponding to the contributions of the higher states and continuum is parameterized as @xmath161 the predictions for the form factors depend on two auxiliary parameters : the squared of borel mass @xmath162 , and the continuum threshold @xmath122 . the continuum threshold signals the scale at which , the excited states and continuum start to contribute to the correlation function . hence it is expected that @xmath163 , @xmath164 and @xmath165 . one approach to determine the continuum threshold and the working region of the borel parameter @xmath162 is to plot the dependence of the predictions on @xmath162 for a range of values of the continuum threshold and determine the values of @xmath122 for which there is a stable region with respect to variations of the borel parameter @xmath162 . for this reason , in figs.(1)-(3 ) , we plot the dependence of the form factors on @xmath166 for fixed values of @xmath167 and various values of @xmath122 in the region @xmath168 . as can be seen from figures ( in the case of old das and new das ) , for @xmath169 , the results are practically independent of the value of @xmath162 for the shown range . the uncertainty due to variations of @xmath122 in this range is much larger than the uncertainty due to variations with respect to @xmath162 . note that the determined range of @xmath122 is in the range that one would expect from the physical interpretation of @xmath122 . in figs . ( 4)-(6 ) , we present the @xmath167 dependence of the form factors obtained using two das . our observations can be summarized as follows ; 1 . @xmath0 case : + we show the behaviour of the form factors that agree well with our expectations . the values of the tensor form factors decrease quickly as we increase the momentum transfer . in fig . ( 4-a , c and e ) are represent the results of the new das and fig . ( 4-b , d and f ) are represent the results of old das . in both case , the @xmath167 dependence of the form factors are similar but the values of the new das results are larger than old das results . we see that higher - order terms gives dominant contribution . @xmath18 case : + in fig . ( 5-a , c and e ) are represent the results of the new das and fig . ( 5-b , d and f ) are represent the results of old das . in the case of old das , the results of the form factors @xmath170 in negatif region but new das result change the behaviour this form factor . in the case of @xmath171 , the @xmath167 dependence of the form factor is similar behaviour and stable but the values of the new das results are larger than old one . in the case of old das , the results of the form factor @xmath172 in positive region but new das result change the behaviour this form factor . @xmath1 case + the higher - order terms of das of @xmath6 baryon have not yet been calculated . so there is only one das result showing the behavior of the form factors that agree well with our expectations . the values of the tensor form factors decrease quickly as we increase the momentum transfers . unlike other form factors the tensor form factors are renormalization - scale dependent @xcite . the numerical values of das are used at the scale @xmath173 @xmath174 in ref . @xcite , therefore , in present work our predictions correspond to this scale . in order to compare our results , we use the following expressions @xcite : @xmath175\right]f(\mu_i^2),\ ] ] where @xmath176 is the number of flavors , @xmath177 is the initial renormalization scale and @xmath178.\end{aligned}\ ] ] the values of the form factors at zero momentum transfer , @xmath179 , defines the corresponding charges . however , in our case , the working region of the lcsr can not extrapolate to the @xmath179 directly . lcsr results more reliable at @xmath180 . the tensor form factor is parameterized in terms of an exponential form @xmath181\ ] ] which makes a reasonable description of data with a two - parameter fit . our predictions are presented in table 2 . as seen in table 2 , predictions of old das are more reliable and reasonable than the new da result . the results of new das are very large and quite suspicious . when one takes the nonrelativistic limit , the isovector tensor charge becomes identical with the isovector axial vector charge @xcite , which is of order similar to the hyperon isovector axial vector charges [ @xmath182 , @xmath183 @xcite ] . therefore obtained by using new das results unreliable . some of the results obtained using the old das are consistent but some of them are not consistent with this prediction . we give the results for the @xmath184 and @xmath185 obtained from chiral quark soliton model @xcite , which have been calculated at a renormalization scale @xmath186 . in order to compare our results , we use the eq . ( 16 ) to the relate this results of form factors to those at @xmath187 , and we obtained these result ; @xmath188 and @xmath189 . as seen from table 2 , our results are different from the results obtained in the chiral quark soliton model . \|l\|l\|l\|l\|l\|l\|l\|l & & + .the values of exponential fit parameters , @xmath190 and @xmath191 for tensor form factor obtained from the old and new das analysis of sum rules . [ cols="^ " , ] & @xmath192 & @xmath193 & @xmath192 + & @xmath194&@xmath195&@xmath196 & @xmath197 + @xmath0&@xmath198 & @xmath199 & @xmath200 & 0.96 + & @xmath201 & @xmath202 & @xmath203 & 1.48 + & @xmath204 & @xmath205&@xmath206&@xmath207 + @xmath18&@xmath208 & @xmath209 & @xmath2104 & @xmath207 + & @xmath211 & @xmath212 & @xmath213 & @xmath197 + & @xmath214 & @xmath215&@xmath216&@xmath216 + @xmath1&@xmath217 & @xmath218 & @xmath219&@xmath220 + & @xmath221 & @xmath222 & @xmath219&@xmath216 + in conclusion , we have evaluated the isovector tensor form factors of octet - octet hyperons by applying the lcsr . these form factors are related to the transverse polarization which gives an important piece of information on the internal structure of baryons ( e.g. the transverse spin structure of the baryons ) . the @xmath167 dependency of form factors are obtained using the old and new da results . the new da results shows that our predictions of the form factors are very large . our predictions on the isovector tensor charges can be summarized in table 2 . the old da results seems to be more reliable . our results on these form factors are compared with the chiral quark soliton model predictions . the chiral quark soliton model results only exist for the @xmath106 form factor so we can not compare results of other form factors . this work has been supported , by the scientific and technological research council of turkey ( tubitak ) under project no .	light - cone qcd sum rules to leading order in qcd are used to investigate the tensor form factors of the @xmath0 , @xmath1 and @xmath2 transitions in the range @xmath3 @xmath4 . the das of @xmath5 , @xmath6 and @xmath7 baryon have been calculated without higher - order terms . then , studies including higher - order corrections have been done for the @xmath5 and @xmath7 baryon . the resulting form factors are obtained using these two das . we make a comparison with the predictions of the chiral quark soliton model . _ keywords _ tensor form factors , octet - octet hyperon transitions , light cone qcd sum rules
in recent years the area of numerical analysis of stochastic differential equations ( sdes ) has expanded at a fast pace . this interest has been driven by different application areas , such as computational finance , neuroscience or electrical circuit engineering . a large part of research in stochastic numerics has been aimed towards the development and strong and weak convergence analysis of several classes of numerical methods . a further important issue for the investigation of numerical methods consists of examining methods for their ability to preserve qualitative features of the continuous system they are developed to approximate . a linear stability analysis is usually the first step of an analysis in this direction . for this the method of interest is applied to a scalar linear test equation and stability conditions on method parameters and step - size are derived and compared with the stability condition for the test equation . in deterministic numerical analysis the underlying idea for a linear stability analysis is based on the following line of reasoning : one linearises and centres a nonlinear ordinary differential equation @xmath2 around an equilibrium , the resulting linear system @xmath3 ( @xmath4 the jacobian of @xmath5 evaluated at equilibrium ) is then diagonalised and the system thus decoupled , justifying the use of the scalar test equation @xmath6 , for the analysis . we refer to , for example , ( * ? ? ? * chapter iv.2 ) for more detail on this procedure . in the stochastic case the same fundamental problem exists : we wish to preserve the qualitative behaviour of solutions of nonlinear stochastic differential equations following discretisation by a numerical method . once again the starting point is a linear stability analysis . we can linearise and centre a nonlinear sde around an equilibrium solution ( see @xcite for the corresponding theory ) , and this procedure yields a system of sdes with an @xmath7-dimensional driving wiener process of the form @xmath8 research on stability analysis for sdes has focused on the scalar linear test equation @xmath9 , where its solution is called geometric brownian motion . this corresponds to considering the linearised sde system when it can be completely decoupled and with one driving wiener process . relevant references are given by , e.g. , @xcite . some first explorations of a linear stability analysis for systems of sdes have been made in @xcite , and in @xcite suitable linear test systems have been derived . the most common and well - known methods to treat sdes numerically are the euler - maruyama approximation and the milstein scheme developed in the last century . these methods and their drift - implicit counterparts have been studied for their stability behaviour , see for example @xcite , however always only for a one - dimensional noise term in the test equation . our aim in this article is to investigate the influence of multi - dimensional noise on the mean - square stability behaviour of these numerical methods and to compare their stability behaviour . assuming that the matrices @xmath10 and @xmath11 can be diagonalised and the system above decoupled , we propose a scalar linear equation with an @xmath7-dimensional wiener process , that is @xmath12 , as a suitable test equation . we consider the @xmath0-maruyama and the @xmath0-milstein method and study the asymptotic mean - square stability properties of these methods . we find that the stability conditions for the @xmath0-milstein method are not only stronger than those for the @xmath0-maruyama method , but they also become more restrictive when increasing the number of noise terms in the sde . in particular we find it impossible to conclude @xmath4-stability for the @xmath0-milstein method , in the sense that we can not define a value or bound @xmath13 for the parameter @xmath0 , such that the method applied to the test equation would be @xmath4-stable for all @xmath14 _ for all @xmath15 and all parameters @xmath16 , @xmath17_. this indicates that analysing the stability behaviour of numerical methods using test equations with only a one - dimensional wiener process may not provide sufficient insight into the properties of the methods for practical high - dimensional application problems . we then present a modification of the @xmath0-milstein method , where we introduce partial implicitness involving a further parameter @xmath1 in the hope to have a better control on the stability behaviour of the method . the stability analysis indicates that this is the case , although the stability condition still depends on the number of noise terms . numerical experiments illustrate the influence that the choice of method and method parameters have for practical simulation tasks . in section [ s.cc ] we introduce the linear test equation , the @xmath0-maruyama method and the @xmath0-milstein scheme , we give a precise definition of asymptotic mean - square stability and state stability conditions for the test equation in the continuous case . section [ s.dc ] is devoted to the analysis of the stability properties of the methods in terms of the arising stochastic difference equations and we compare the stability regions of both types of methods in section [ s.comp ] . in section [ s.sigma ] we introduce a new class of milstein type methods and analyse their stability behaviour . we illustrate the theoretical findings by some numerical experiments in section [ s.num ] . in this section we introduce the stochastic differential and difference equations , as well as the notions of stability , that we consider in this article . we will be concerned with asymptotic _ mean - square stability _ of the zero solution of a test equation with respect to perturbations in the initial data . as a test equation we employ the following scalar linear stochastic differential equation with multiplicative noise @xmath18 driven by an @xmath7-dimensional standard wiener process @xmath19 given on the probability space @xmath20 with a filtration @xmath21 . by @xmath22 we denote expectation with respect to @xmath23 . we assume that the coefficients @xmath24 and @xmath16 , @xmath25 are complex - valued and without loss of generality suppose that the initial value @xmath26 is non - random ( ( * ? ? ? * chapter 4 ) ) . the solution @xmath27 of is a complex - valued stochastic process , where @xmath28 is a vector - valued wiener process in @xmath29 . alternatively this complex - valued sde can be written as 2-dimensional vector sde in the real and imaginary parts . we denote the real and imaginary part of a complex number @xmath30 by @xmath31 and @xmath32 , respectively , @xmath33 denotes the complex conjugate and @xmath34 stands for the absolute value of some @xmath35 . for any non - zero initial value equation has a non - trivial path - wise unique strong solution which we denote by @xmath36 when we wish to emphasise its dependence on the initial data . for @xmath37 the equation obviously admits the _ zero solution _ , which is a steady state solution of the equation . @xcite ) [ d.stabil ] the zero solution of equation , @xmath38 , is 1 . _ mean - square stable _ , if for each @xmath39 , there exists a @xmath40 such that @xmath41 whenever @xmath42 and @xmath43 ; 2 . _ asymptotically mean - square stable _ , if it is mean - square stable and if there exists a @xmath40 such that whenever @xmath44 @xmath45 the zero solution is called unstable if it is not stable in the mean - square sense . definition [ d.stabil ] is slightly more general than necessary in the present context , as for the simple linear equation given by we can take @xmath46 arbitrarily large and thus it does not play a significant role . [ l.contprob ] @xcite the zero solution of equation is asymptotically mean - square stable if and only if @xmath47 we now discuss the stochastic difference equations that arise by applying the @xmath0-maruyama method and the @xmath0-milstein scheme to the scalar test equation . we consider numerical methods for computing approximations @xmath48 of the solution of the test equation at discrete time points @xmath49 ( @xmath50 ) with constant step - size @xmath51 . the @xmath0-maruyama method applied to the test equation reads @xmath52 where we have replaced the wiener increments @xmath53 by the scaled random variables @xmath54 . here each @xmath55 is one of @xmath7 independent sequences of mutually independent standard gaussian random variables , i.e. , each @xmath56 is @xmath57-distributed . the @xmath0-milstein method applied to the test equation is given by @xmath58 here we have additionally replaced the multiple wiener integrals @xmath59 as follows : ( i ) if @xmath60 the multiple wiener integral @xmath61 can be replaced by @xmath62 for @xmath17 , and ( ii ) if @xmath63 we use the identity @xmath64 to obtain @xmath65 for @xmath66 the method reduces to the _ ( forward ) euler - maruyama scheme _ , which is explicit in the drift as well as in the diffusion part . for @xmath67 the methods and are drift - implicit . for @xmath68 the scheme is known as _ stochastic trapezoidal rule _ and for @xmath69 we obtain the _ backward euler - maruyama method_. the two last terms in the @xmath0-milstein method represent a higher order approximation of the diffusion part in equation in the sense of mean - square convergence . we call a method _ mean - square convergent with order @xmath70 _ ( @xmath71 ) if the global error , @xmath72 , satisfies @xmath73 with a positive error constant @xmath74 , which is independent of the step - size @xmath51 . it is well - known that in the case of multiplicative noise the @xmath0-milstein method is mean - square convergent of order @xmath75 , whereas the @xmath0-maruyama method is mean - square convergent of order @xmath76 . obviously the stochastic difference equations and admit the zero solution for the initial value @xmath37 , which is a steady state solution as well . for any non - zero initial value @xmath26 , the equations have a unique solution provided @xmath77 . we write @xmath78 again when we want to emphasise that the solution of the difference equations depends on the initial data . [ d.stabildiff ] the zero solution of the difference equations and is 1 . _ mean - square stable _ , if for each @xmath39 , there exists a @xmath40 such that @xmath79 whenever @xmath80 and @xmath44 ; 2 . _ asymptotically mean - square stable _ , if it is mean - square stable and if there exists a @xmath40 such that whenever @xmath44 @xmath81 stability conditions will now involve the coefficients @xmath24 , @xmath16 @xmath82 ) of the test equation , as well as the method parameter @xmath0 and the step - size @xmath51 . it will be useful to describe the stability region of a stochastic differential or difference equation . we follow the presentation in @xcite and consider sets of parameters @xmath83 , @xmath84 , @xmath85 for which the zero solutions of the continuous and the discrete equations are asymptotically stable , that is @xmath86 further , we consider the extension of the deterministic notion of a - stability @xcite to the mean - square analysis setting and say that a numerical method is _ a - stable _ in mean - square @xcite , if whenever the zero solution of is asymptotically mean - square stable , then the same is true for the zero solution of the method for any step - size @xmath87 . using the above definitions of the stability regions , we call the @xmath0-maruyama or @xmath0-milstein method a - stable in mean - square if for all @xmath88 we have @xmath89 or @xmath90 for the stochastic difference equations and we now derive stability conditions in dependence on the method parameter @xmath0 and the applied step - size @xmath51 , and compare these conditions with those for the continuous problem given in lemma [ l.contprob ] . we start by rearranging the stochastic difference equation into the following one - step recurrence equation . then by squaring and taking expectations we obtain a recurrence equation for the second moments @xmath91 . as mentioned before , we have to assume that @xmath92 to guarantee the existence of a unique solution of the recurrence equations . rearranging equation yields the recurrence @xmath93 where @xmath94 then the recurrence equation for the second moment @xmath91 , using @xmath95 , @xmath96 , and the complex conjugate equation of , reads @xmath97 from we can immediately read off necessary and sufficient stability conditions in terms of the above parameters . [ lemmarstababc ] the zero solution of the recurrence equation , representing a one - step maruyama - type method applied to the test equation , is asymptotically mean - square stable if and only if @xmath98 now rewriting the stability conditions in terms of the parameters @xmath24 and @xmath16 @xmath99 in the test equation , the method parameter @xmath0 and the step - size @xmath51 , using , we obtain the following result . the zero solution of the stochastic difference equation given by the @xmath0-maruyama method applied to the scalar linear test equation is asymptotically mean - square stable if and only if @xmath100 we note that the first two terms in the left - hand sides of are equal to the left - hand side of , that is they correspond to the stability condition for the continuous problem . now comparing the stability condition for the continuous problem to that of the discrete problem , we immediately obtain from an extension of the result ( * ? ? ? 4.1 ) to the case of driven by a multi - dimensional wiener process . [ c.maruyama ] for all @xmath87 it holds that @xmath101 in particular , for @xmath102 the @xmath0-maruyama method is a - stable in mean - square . for @xmath103 and @xmath104 , the stability condition for the @xmath0-maruyama method is satisfied if and only if @xmath105 we now turn to the @xmath0-milstein method and first follow the same steps as in the previous section to deduce a recurrence equation for the second moments of the solution of and read off the corresponding stability conditions . rearranging the difference equation , we obtain @xmath106 @xmath107 and @xmath108 is given in . note that rewriting the parameter @xmath109 in terms of @xmath108 provides a convenient way to compare the stability conditions for the @xmath0-maruyama and @xmath0-milstein method . with analogous calculations as in the previous section and by additionally using @xmath110 and @xmath111 , we find the recurrence for the second moments of the @xmath0-milstein method as @xmath112 again we can read off from necessary and sufficient stability conditions in terms of the parameters @xmath108 , @xmath113 and @xmath114 . [ lemmilstababc ] the zero solution of the recurrence equation given by the @xmath0-milstein method applied to the test equation , is asymptotically mean - square stable if and only if @xmath115 where @xmath116 . when applied to the linear scalar stochastic differential equation , other variants of one - step maruyama - type methods and one - step milstein - type methods can quite often be rearranged into the recurrence equations and , respectively , with appropriate definitions of the parameters @xmath108 , @xmath113 and @xmath114 . then lemmata [ lemmarstababc ] and [ lemmilstababc ] can be applied to obtain stability conditions interpreted with these definitions of @xmath108 , @xmath113 and @xmath114 . the next corollary follows by rewriting the stability condition in terms of the original parameters using . [ cormilstab ] the zero solution of the stochastic difference equation given by the @xmath0-milstein method applied to the scalar linear test equation is asymptotically mean - square stable if and only if @xmath117 again the first two terms in the left - hand side of are equal to the left - hand side of . however , when comparing the above condition with the stability condition for the @xmath0-maruyama method , we see that the left hand side of contains two additional terms , which are always non - negative and depend on the noise intensities @xmath118 ( @xmath119 ) but are independent of the parameter @xmath0 . thus the precise stability region @xmath120 of the @xmath0-milstein method depends on the noise intensities @xmath16 ( @xmath119 ) . suppose our aim is to use the method parameter @xmath0 to determine the optimal stability region of the @xmath0-milstein method for a given set of parameters in , that is we aim to find a value of @xmath0 such that the sum of the third to the fifth term in vanish for any step - size @xmath51 . then , assuming that holds , we obtain for every set of parameters @xmath121 a different value for that optimal @xmath122 . we define @xmath123 where @xmath124 for this optimal @xmath122 we immediately obtain from the stability condition for the @xmath0-milstein method the following corollary . [ c.milstein ] for all @xmath87 it holds that @xmath125 assuming that @xmath104 , for @xmath126 , the stability condition for the @xmath0-milstein method is satisfied if and only if @xmath127 in particular , condition implies that for @xmath128 and @xmath129 , the stability condition for the @xmath0-milstein method is satisfied for any step - size @xmath88 . + in the case that @xmath130 , and @xmath128 and @xmath131 , then the zero solution of the @xmath0-milstein scheme is also unstable only if condition is imposed on the step - size @xmath51 . intrinsic in the concept of @xmath4-stability is the idea that the property of @xmath4-stability of a method holds for the whole class of differential equations considered as test equations . in the setting of this article this implies that we would need to find a @xmath132 , a value of or a bound on @xmath122 , which is independent of @xmath133 , such that @xmath134 for @xmath135 . in the case of the @xmath0-maruyama method the corresponding value of @xmath132 is @xmath136 , see corollary [ c.maruyama ] . however , considering the simple case of multi - dimensional noise terms having equal noise intensities , we find that , at best , we can find an upper bound @xmath132 independent of the given parameters @xmath133 , but depending on the number of noise sources . to see this , let @xmath137 . then , using the squared stability inequality in the form @xmath138 and @xmath139 in yields @xmath140 thus for example , for one , two and three noise sources ( @xmath141 ) , we can find a @xmath132 as @xmath142 and @xmath143 , respectively . in the one - dimensional noise case this corresponds to the result in ( * ? ? ? but in general there exists no upper bound @xmath132 for @xmath122 such that the @xmath0-milstein method for any @xmath144 is @xmath4-stable for the whole class of equations with arbitrary many noise terms . in the general case of the sde it is possible , essentially using the cauchy - schwarz inequality several times , to derive an upper bound @xmath132 for @xmath122 which also only depends on @xmath7 , but as it involves @xmath145 it becomes pointless for large @xmath7 . thus , it appears that the stability condition for the @xmath0-milstein method becomes very restrictive for an increasing number of noise terms . in this section we aim to illustrate the results of the previous sections by visually comparing the stability regions @xmath146 and @xmath120 . as there are too many parameters in the system to do so in a two - dimensional plot , we only consider the case of the test sde with multi - dimensional noise where all the terms have the same noise intensity , i.e. , @xmath147 , and real - valued coefficients , i.e. , we have @xmath148 . we essentially follow @xcite regarding the scaling of the parameters in the plots . thus we set @xmath149 the stability conditions , and become @xmath150 note that for @xmath151 and @xmath152 condition yields the same stability region . to interpret the figures below , observe that given a test equation with parameter values @xmath24 and @xmath153 and @xmath151 , the point @xmath154 corresponds to the choice of step - size @xmath155 . then varying the step - size @xmath51 corresponds to moving along the ray that connects @xmath156 with the origin , where going on this ray in the direction of the origin corresponds to decreasing the step - size @xmath51 . for @xmath157 the scaling is appropriately adapted . figure [ fig.stabregm ] shows the mean - square stability regions of the zero solutions of the sde ( white area with a dashed border ) , the @xmath0-maruyama approximation ( light - grey area ) and the @xmath0-milstein approximation ( light dark area to dark - grey area ) for different values of the method parameter @xmath0 and illustrates how the stability region of the @xmath0-milstein method decreases with @xmath7 and compares to that of the @xmath0-maruyama method . in particular , one can observe that the mean - square stability regions for the @xmath0-milstein method are always smaller than the stability region of the @xmath0-maruyama method . for the @xmath0-maruyama method the figure illustrates the property of @xmath4-stability of the method for @xmath158 for all @xmath159 , where as for the @xmath0-milstein method we can not conclude @xmath4-stability for a particular @xmath0 and all @xmath15 . in ( * ? ? ? * section 4 ) and ( * ? ? ? * section 3 ) the stability regions of the @xmath0-maruyama method and the @xmath0-milstein method have been plotted separately for the scalar sde and real - valued @xmath160 . to emphasise the different stability properties of both methods even in the case of @xmath151 , we provide in figure [ fig.stabrega ] a plot of the stability regions of both methods together for the same setting as in @xcite and several values of the method parameter @xmath0 . in these cases the stability regions of the @xmath0-maruyama method and the @xmath0-milstein method coincide with that of the sde if @xmath161 and @xmath162 , respectively . when using the euler - maruyama method and the standard milstein method , that is taking @xmath163 , the methods that are most often used , then it appears from figure [ fig.stabrega ] that both stability regions are quite small but of similar size . this is already mentioned in @xcite . however , when taking mean - square accuracy into account and aiming to reduce numerical costs by using the milstein method with a larger step - size , one might easily be prevented from doing so by the more restrictive stability condition of the milstein method . a brief look at the deterministic case , that is @xmath164 ( @xmath119 ) in , , , etc . , reminds us that it is by making the approximation implicit and the introduction of the method parameter @xmath0 that one can control the stability properties of the method by suitably choosing @xmath0 . in this sense it would be useful to develop milstein - type methods that incorporate implicit higher order approximations of the diffusion term to have a method parameter for this purpose . in general , a straightforward introduction of implicitness into approximations of the diffusion term results in the numerical solution becoming unbounded with positive probability , see ( * ? ? ? * chap 1.3.4 ) and @xcite for methods avoiding this problem . in this article our aim is to highlight the effect that an implicit discretisation of the diffusion term can have on the stability properties of the @xmath0-milstein method and thus we take advantage of the simple structure of the test equation . the latter results in the fact that in each term in the second last sum in the @xmath0-milstein method the double wiener integral @xmath165 can be replaced by @xmath62 , for @xmath17 . then the second last term in the method can be written as @xmath166 for each @xmath167 and we introduce an implicit approximation with an additional positive method parameter @xmath1 only in the latter term , which does not contain the random variable @xmath56 . thus we propose , for each @xmath167 , to use @xmath168 we emphasise that this represents a ( partial ) implicit approximation of the diffusion term in contrast to well - known approaches to implicit approximations of the drift term . applied to the scalar linear test - equation the @xmath0-@xmath1-milstein method then takes the form @xmath169 @xmath170 , with the random variables @xmath56 defined as in section [ s.cc ] . the choice of @xmath171 yields the @xmath0-milstein approximation of the stratonovich - sde @xmath172 . for the stability analysis we follow the same procedure as in section [ s.dc ] and first rewrite as a one - step recurrence @xmath173 where the parameters @xmath174 and @xmath175 are given by @xmath176 we define @xmath177 and assume that @xmath178 to guarantee the existence of a solution to equation . now squaring and taking the expectation yields @xmath179 where @xmath180 . hence the zero - solution of the stochastic difference equation is asymptotically mean - square stable if and only if the factor on the right hand side of is less than 1 . rewriting this condition in terms of @xmath181 and rearranging , we obtain the following result : [ c.stabcondsigma ] the zero solution of the stochastic difference equation given by the @xmath0-@xmath1-milstein method applied to the scalar linear test equation is asymptotically mean - square stable if and only if @xmath182 where again @xmath183 . we note that the first terms in the left - hand side of are equal to the left - hand side of . the additional term @xmath184 is negative for @xmath185 , i.e. , when the stability condition for the test equation is satisfied . thus , the stability condition in lemma [ c.stabcondsigma ] is less restrictive than the condition for the @xmath0-milstein method . denoting by @xmath186 the stability region of the @xmath0-@xmath1-milstein method , analogous considerations as for corollary [ c.milstein ] yield the next result . [ cormilstein ] define @xmath187 for all @xmath87 it holds that @xmath188 then , assuming that @xmath104 , for @xmath189 the zero solution of the @xmath0-@xmath1-milstein method is asymptotically mean - square stable if and only if @xmath190 in particular , condition then implies that for @xmath128 and @xmath129 , the stability condition for the @xmath0-milstein method is satisfied for any step - size @xmath88 . as in section [ s.comp ] we illustrate the stability regions of the @xmath0-@xmath1-milstein method applied to the test sde with a single noise ( @xmath151 ) in terms of real - valued coefficients @xmath148 . in this case we can find a bound on @xmath191 as @xmath192 . again , we set @xmath193 and @xmath194 , and the stability inequality in terms of @xmath30 and @xmath195 reads @xmath196 figures [ fig.stabregc ] - [ fig.stabrege ] compare the stability regions for the sde with @xmath151 , the @xmath0-maruyama method and the @xmath0-@xmath1-milstein method . in each figure we have fixed the parameter @xmath0 and only the parameter @xmath1 varies . for @xmath197 the plots would correspond to those in figure [ fig.stabrega ] . we can see that the stability region of the @xmath0-@xmath1-milstein method ( dark - grey area ) is becoming larger when increasing the parameter @xmath1 . moreover , the stability region coincides with the stability region of the test equation ( white area ) if @xmath198 . further , we consider again the sde with identical noise intensities @xmath199 and plot in figure [ fig.stabregf ] the stability regions for @xmath200 ( again the condition below is the same for @xmath151 and @xmath152 ) and @xmath201 corresponding to the scaling @xmath202 and @xmath203 , where condition now reads @xmath204 again the stability region of the @xmath0-@xmath1-milstein method decreases with growing @xmath7 , as in the case of the @xmath0-milstein method , which prevents to conclude @xmath4-stability for a particular value of @xmath205 and all @xmath159 . however , the introduction of the ( partial ) implicitness in the diffusion term clearly provides an improvement of the stability behaviour for the @xmath0-@xmath1-milstein method . in this section we aim to illustrate the effect that the choice of method and the choice of the method parameters @xmath0 and @xmath1 have on practical simulation runs . we consider the test equation with the following parameters : @xmath206 , @xmath207 , @xmath208 , @xmath209 , and integrate for @xmath210 $ ] . the equation has the explicit solution @xmath211 . the following four figures show numerical simulation studies performed with the @xmath0-maruyama method ( figure [ fig.stabresa ] ) , the @xmath0-milstein method ( figure [ fig.stabresc ] ) , the @xmath0-@xmath1-milstein method with @xmath212 ( figure [ fig.stabresd ] ) and the @xmath0-@xmath1-milstein method with @xmath213 ( figure [ fig.stabrese ] ) , all simulations done with a fixed step - size @xmath155 over the interval @xmath214 $ ] with gridpoints @xmath215 . the parameter @xmath0 varies as @xmath216 . the analytical results in the previous sections also suggest choices of step - sizes , such that the zero solution of any one of the numerical methods for a fixed set of parameters is asymptotically mean - square stable and obviously it is possible to illustrate this by performing numerical experiments with fixed parameter sets and varying the step - size @xmath51 . however , the focus of this article is rather on comparing qualitative properties of _ methods _ than on how individual methods behave for various step - sizes . each of the four figures below consists of three pictures . the underlying idea of a stability analysis of numerical methods is to provide guidance for choosing method parameters and a step - size such that the numerical solution represents a good approximation of the true solution . ( note that convergence of a method only guarantees this in the limit for @xmath217 . ) thus , we plot in the left picture of each figure the mean - square error @xmath218 between the exact solution @xmath219 and numerical solution @xmath220 for the corresponding method and choice of method parameters . this allows to compare the impact of the choice of method on actually computing solutions of sdes . however , the quantity of interest in the stability definitions [ d.stabil ] and [ d.stabildiff ] are the second moments of the analytical and numerical solutions . therefore the other two pictures present plots of the second moments of the solutions of , , and , in the middle picture estimated from the numerical simulations of , and , in the right picture computed form the analytical solutions . this allows to compare the effect of the choice of method and parameters on the behaviour of the second moments analytically and also , how well this is approximated by the numerical methods . in detail , the figures present the following quantities , where @xmath221 and @xmath222 denote the values on grid points of a trajectory of the explicit solution of with the above parameters and of the numerical trajectory produced by one of the methods , respectively : @xmath223 { \mathsf{e}}(x_i^2 ) & = & s^i \cdot x_0 ^ 2 \quad \text { for all } i \ , . \label{numsecmom}\end{aligned}\ ] ] the expression for @xmath224 represents an estimator for the second moment of the numerical approximation , the expression for @xmath225 is the solution of the deterministic ode @xmath226 ( see proof of lemma [ l.contprob ] ) at discrete time - points and the last expression for @xmath227 is the result of applying the recurrences , and where @xmath228 denotes the factor in front of @xmath227 in the right - hand side of the corresponding recurrence equation . the number of trajectories computed for the above quantities is @xmath229 for each simulation . figures [ fig.stabresa ] to [ fig.stabrese ] illustrate the behaviour of the methods for the above set of parameters , in particular the same fixed step - size @xmath155 and different choices of @xmath0 and @xmath1 . the @xmath0-maruyama method provides reliable approximations for @xmath230 , but not for @xmath163 , the @xmath0-milstein method in same setting _ does not _ provide reliable approximations for @xmath163 and @xmath231 . in fact , the numerical solutions for @xmath231 in this setting diverge . we can observe the improvement in the stability behaviour when using the @xmath0-@xmath1-milstein method introduced in the previous section . in contrast to the @xmath0-milstein method this method produces reliable approximations for the choice of @xmath232 when also setting @xmath201 . further , for the choice of @xmath233 the @xmath0-@xmath1-milstein approximation behaves satisfactorily for all values of @xmath0 , thus the implicit term in the diffusion approximation is effectively stabilising the method . a linear stability analysis has been performed for the @xmath0-maruyama method and the @xmath0-milstein method , using a linear test equation with several multiplicative noise terms . we have obtained stability conditions guaranteeing asymptotic mean - square stability of the zero solution of the stochastic difference equations resulting from both types of methods applied to the test sde . comparing the stability conditions and for the @xmath0-maruyama and the @xmath0-milstein method , respectively , it is quite obvious that the latter is more restrictive than the former , due to the fourth and fifth term in , i.e. @xmath234 , which is always non - negative . in particular , this term is the more restrictive , the larger the parameters in the diffusion term in are . now , in the case that these parameters are small , that is the so - called small noise case , it is well known from the corresponding mean - square convergence analysis that maruyama - type methods provide sufficiently accurate methods for practicable choices of step - sizes , see @xcite . milstein - type methods include higher order approximations of the diffusion term with the aim that they are accurate methods for more practicable choices of step - sizes just in the case that the diffusion term is large . in other words , when dealing with sdes with larger diffusion terms , numerical efficiency considerations would suggest using a milstein - type method with a larger step - size rather than a maruyama - type method with a small step - size to obtain numerical approximations of the solution of an sde with a similar accuracy . however , corollary [ cormilstab ] indicates that in this case there may be a trade - off and one is faced with restrictions on the step - size for the milstein - type method due to stability reasons . further , we have shown that the precise stability region of the @xmath0-milstein method depends on the number and magnitude of the noise terms , whereas the stability region of the @xmath0-maruyama method is independent of them . in particular , it is not possible to define a value @xmath13 such that the @xmath0-milstein method can be called @xmath4-stable for all @xmath235 for the class of sdes for an arbitrary number of driving wiener processes . for the @xmath0-maruyama method @xmath236 , as in the deterministic case and the case of the test equation with @xmath151 . we provide a modified milstein - type method with a partially implicit diffusion approximation and demonstrate that the resulting stability behaviour can be controlled more favourably . the results highlight that it is necessary to include multi - dimensional noise into test equations and to study their effects on the practical behaviour of the methods . we thank martin riedler ( heriot - watt university , edinburgh ) and lukasz szpruch ( university of strathclyde , glasgow ) for fruitful discussions .	in this article we compare the mean - square stability properties of the @xmath0-maruyama and @xmath0-milstein method that are used to solve stochastic differential equations . for the linear stability analysis , we propose an extension of the standard geometric brownian motion as a test equation and consider a scalar linear test equation with several multiplicative noise terms . this test equation allows to begin investigating the influence of multi - dimensional noise on the stability behaviour of the methods while the analysis is still tractable . our findings include : ( i ) the stability condition for the @xmath0-milstein method and thus , for some choices of @xmath0 , the conditions on the step - size , are much more restrictive than those for the @xmath0-maruyama method ; ( ii ) the precise stability region of the @xmath0-milstein method explicitly depends on the noise terms . further , we investigate the effect of introducing partially implicitness in the diffusion approximation terms of milstein - type methods , thus obtaining the possibility to control the stability properties of these methods with a further method parameter @xmath1 . numerical examples illustrate the results and provide a comparison of the stability behaviour of the different methods . _ keywords : _ stochastic differential equations , asymptotic mean - square stability , @xmath0-maruyama method , @xmath0-milstein method , linear stability analysis . _ ams subject classification : _ 60h10 , 65c20 , 65u05 , 65l20
in the last few years most cosmological parameters have been determined up to a few percent . the values of @xmath1 , @xmath2 , @xmath3 and @xmath4 can now be constrained with an unprecedent degree of accuracy ( see , for example , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * and references therein ) . the next challenge to cosmologists is to test the predictions of cosmological models at a few hundred kpc scales . it turns out that these are just the relevant scales involved in galaxy formation and evolution . galaxy formation and evolution are intriguing open questions whose resolution in connection with the global cosmological model will very likely advance considerably in this decade . even though the field is at its beginnings , the use of numerical methods to study how galaxies are assembled within a cosmological scenario from the field of primordial fluctuations , seems a convenient approach . the main advantage of these approaches ( i.e. , self - consistent hydrodynamical simulations ) , is that the physics is introduced at a very general level , and the system evolves as a consequence . we can follow the evolution of the dynamical and hydrodynamical properties of matter in the universe ; galaxy - like objects ( glos ) appear as a consequence of this evolution . and , so , the building - up of objects ( cosmic network structure formation at high @xmath5 , collapse , interactions , mergers , accretions ) , as well as their hydrodynamical consequences ( instabilities , gas infall from halos to discs at hundred of kpc scales , gas inflow along discs at tens of kpc scales , shocks , cooling , piling - up of gas necessary for star formation ) , can be accurately followed . we get not only the _ properties _ of objects at any @xmath5 , but also an _ insight _ into the _ physical processes _ responsible for their formation and evolution . moreover , numerical hydrodynamical simulations using particles permit very convenient comparisons of glos that form in simulations with observational data . simulations _ directly _ provide us , at each @xmath5 , with the structural and dynamical properties of each individual glo ( position and velocity of each of its particles , gas density and temperature of each of its baryonic constituents ) and with their individual star formation rates histories ( sfrhs ) . chemical abundance and spectrophotometrical data are the current standard to compare models of galaxy formation . it is expected that the next generation of astronomical facilities will make possible a new science : mass measurements for distant galaxies ( see , for example * ? ? ? glos formed in numerical simulations are particularly suited to be compared to this new kind of data . pre - prepared numerical experiments are adequate to describe in detail a particular phase of the formation or evolution of galaxies ( for example , merger events or orbital motions of satellites within halos ) , from initial conditions set by the experimenter . these initial conditions try to model conditions that would have arisen along the evolution of the systems under consideration . they are useful to study basic aspects of the physical processes relevant to evolution . for example , the works by barnes , hernquist and mihos @xcite , which have fundamental importance to understand the role played by mergers in galaxy evolution , have been carried out with this technique . however , in pre - prepared simulations , contrary to the self - consistent approach , the process under consideration is studied _ in isolation _ , and not in connection with the other relevant processes involved in galaxy formation and evolution ( already mentioned ) that , moreover , could interact among themselves in a non - trivial way . these considerations stress the ability of self - consistent hydrodynamical simulations as a tool to learn how galaxies form from the field of primordial fluctuations and evolve into the objects we observe today . to properly handle this problem , a numerical code has to allow for enough mass , time and space resolution as well as a convenient dynamical range . they should be as fast as possible and with memory requirements within the current computer capabilities . a very important issue when designing a numerical code to study galaxy formation and evolution , is to make sure that conservation laws are accurately verified , and , particularly , i ) , that angular momentum is conserved at the scales relevant to disc formation ; otherwise , galaxy disc formation could meet with some difficulties ; and , ii ) , that entropy is conserved in reversible adiabatic processes , because the violation of this physical principle could produce spurious effects at galaxy scales . by the moment , star formation ( sf ) processes have to be modelled , either inspired in kpc or pc scale hydrodynamical simulations @xcite or other considerations @xcite . the first choice to be made when designing this kind of codes is the gravity solver . among current numerical methods , those that employ adaptive techniques in regions of high density , either from a lagrangian ( as ap3 m , * ? ? ? * ; * ? ? ? * ) or eulerian ( as the art and mlapm codes , * ? ? ? * ; * ? ? ? * ) approach , are the most suitable to meet the requirements of resolution and large dynamical range , accuracy and rapidity . a detailed comparison between ap3 m and art codes has been carried out by @xcite . they have found out that these codes produce results that are consistent within a 10% , provided that @xmath6 ( @xmath7 is the number of integration steps ; @xmath8 stands for the dynamical range ) . the choice of the gravity solver in a cosmological simulation depends on its purpose . to study galaxy formation and evolution , lagrangian codes have the advantage over eulerian codes that they permit to go backwards and forwards in time in a very easy way . for example , the constituent particles of a given object can be identified at a given redshift , @xmath9 , and one can then analyze their positions in phase space and the properties of the objects or structures they form at a different redshift , @xmath10 . this is a very convenient method to study evolutionary processes and it motivates our choice of an ap3m - based method as gravity solver for our simulations . to solve the hydrodynamical equations ( and , in general , any hyperbolic system of equations in partial derivatives ) , there are also two basic different techniques : i ) , eulerian methods , and , ii ) , lagrangian methods . eulerian methods are based on the so - called godunov algorithm @xcite . their new formulations , using adaptive mesh refinements @xcite , are particularly well suited to combine with art - like codes when both gravitational and hydrodynamical forces are considered . for a comparison of the performances of a number of hydrodynamical codes of both kinds see @xcite . most of the lagrangian methods used in astrophysics are based on the sph ( smooth particle hydrodynamics ) technique @xcite . given the convenience of this technique to be applied to cosmological simulations , a number of authors developed sph codes to be used in a cosmological context . some of them follow . @xcite first used sph techniques in cosmological simulations ( a p3m - sph code ) . @xcite , as well as @xcite , coupled a sph code to the @xcite tree algorithm . @xcite modelled their treesph code after @xcite and @xcite introduced a parallel version of this code , while @xcite makes use of a special purpose hardware to compute the gravitational forces by direct summation ( grape ) . @xcite coupled sph with a pp algorithm in a code designed to be run on a connection machine , and @xcite incorporated in a tree - sph code the so - called @xmath0 terms ( see below ) . gadget @xcite uses either a tree scheme or grape , with individual integration timesteps , and both , serial and parallel versions . another parallel tree - sph code is gasoline @xcite . as already mentioned , ap3m - based codes are particularly well suited to study galaxy formation through self - consistent cosmological simulations . the first ap3m - sph code was introduced by @xcite ( hydra code , see also * ? ? ? @xcite carried out a second implementation . in these implementations , the integration timestep is _ global _ ( i.e. , at a given time , the same for all particles ) . in cosmological simulations , however , multiple time scales appear , due to their very large dynamical ranges from very dense volumes to very rarefied zones . to get an accurate enough integration scheme , and , at the same time , to avoid that particles in denser volumes slow down the simulations , it is advantageous to introduce _ individual _ integration timesteps , i.e. , at each time , different timesteps for each particle , depending on the density of the region it samples . this is the optimal design of the code to increase the mass resolution . another shortcoming of conventional sph formulations concerns the entropy violation of the dynamical equations , related to the space dependence of the smoothing length of sph particles , @xmath11 , as noted by some authors @xcite . a rigorous formulation of sph requires that additional terms must be included in the particle equations of motion which account for the variability of @xmath12 , usually termed as `` the @xmath0 terms '' . until very recently , they were considered as having a negligible effect on the global dynamics of systems @xcite and , therefore , sph codes ignored such additional terms and focused on energy conservation . alternatively , treatments of hydrodynamics based on the lagrange equations can be formulated that are well behaved in their conservation properties of both , energy and entropy , as that introduced recently by @xcite . the effects of entropy violation in sph codes are not completely clear and they need to be analyzed in much more detail , specially in simulations where galaxies are formed in a cosmological framework . previous works have analyzed this question in the case of the collapse of isolated objects and have found that , if such correction terms are neglected , the density peaks associated to central cores or shock fronts are overestimated at a @xmath1330@xmath14 level @xcite . to make up for these shortcomings when dealing with problems related to galaxy formation and evolution , we introduce a new code , deva , where gravity is solved by means of an ap3m - like technique , and hydrodynamics with a sph technique , with individual integration timesteps . the space dependence of the sph resolution scale , @xmath11 , has been taken into account , in order to ensure the conservative character of the equations of motion , as long as entropy and energy is considered . another important particularity of deva is the attention paid to angular momentum conservation , a key point to enable disc formation in simulations ( * ? ? ? * ; * ? ? ? * and references therein ) . our choice has been to put the stress into conservation laws rather than into saving cpu time . but saving cpu time has also been one of our concerns , so that the code is fast enough that cosmological self - consistent simulations can be run on a modest computer machine . the paper is organized as follows : @xmath151 is the introduction . in @xmath152 , the sph method is briefly reviewed and we present the sph equations when the @xmath0 terms are considered . @xmath15 3 is devoted to the particularities of the deva code , and @xmath15 4 to test whether the code integrates correctly the hydrodynamical and n - body equations ( standard tests ) . in @xmath15 5 we introduce some self - consistent simulations run with deva , compare to one standard of reference for hydrodynamical simulations in a cosmological framework ( the santa barbara cluster comparison project * ? ? ? * ) , and analyze the effects of the @xmath0 terms in these simulations . finally , in @xmath15 6 , we give a summary of the work and discuss deva performances . the basic idea of the sph method @xcite lies in representing the fluid elements by @xmath16 particles which act as interpolation centers to determine the local value of any macroscopic variable @xmath17 . in order to smooth out local statistical fluctuations , this interpolation is performed by convolving the field @xmath17 with a smoothing ( or kernel ) function @xmath18 . for example , the smoothed estimate of the local density is @xmath19 where @xmath20 , @xmath21 is the mass of particle @xmath22 , and @xmath23 is the smoothing length for particle @xmath24 , which specifies the size of the averaging volume . ideally , the individual particle smoothing lengths @xmath23 must be updated so that each particle has a constant number of neighbors @xmath25 . by neighbors we mean those particles @xmath22 with distances @xmath26 . such a condition can be exactly implemented by constructing , for each particle @xmath24 , a list of its @xmath25 nearest neighbors . the smoothing length of @xmath24 is then defined to be @xmath27 where @xmath28 is the position vector of particle @xmath24 s most distant neighbor . since each particle has its own @xmath12 value , it is possible to find couples of particles @xmath29 such that @xmath22 is a neighbor of @xmath24 , but @xmath24 is not a neighbor of @xmath22 . in these cases , it is obvious that the reciprocity principleth particle belongs to the neighbor list of the @xmath30th particle , then it is mandatory that , at this same time , the @xmath30th particle belongs to the neighbor list of the @xmath22th particle ] is not satisfied and , therefore , simulations will not conserve momentum . in order to solve this problem , it is necessary to symmetrize the sph equations by using , for example , averaged kernels @xcite : @xmath31\;.\ ] ] a first consequence of the adopted symmetrization procedure is the specific form for the kernel derivatives . as a matter of fact , ( [ wij ] ) implies that @xmath32 is a function of three variables : @xmath33 , @xmath34 and @xmath35 . consequently , its gradient @xmath36 is given by : @xmath37 \nonumber \\ + \frac{1}{2}\left[\frac{\partial w(r_{jk},h_j)}{\partial h_{j}}{{\mathbf{\nabla}}}_i h_{j } + \frac{\partial w(r_{jk},h_k)}{\partial h_{k } } { { \mathbf{\nabla}}}_i h_{k}\right]\end{aligned}\ ] ] the first part of eq . ( [ nablawij ] ) , which does not involve derivatives of the smoothing lengths , is the usual symmetrized form of @xmath36 . the second part , which involves derivatives of the smoothing lengths , arises because of the spatial and temporal variability of @xmath12 . we shall refer to terms of this type as @xmath0 terms. most implementations of the sph algorithm consider only the first one and neglect the @xmath0 terms . the motion of particle @xmath30 is determined by the momentum and energy equations : @xmath38 where @xmath39 is the local gravitational potential , @xmath40 is the acceleration due to pressure forces , @xmath41 is the acceleration due to viscosity forces , @xmath42 is the specific internal energy , @xmath43 is the pressure ( with @xmath44 being the constant heat ratio ) , and @xmath45 is the power due to non - adiabatic heating or cooling processes . a fully consistent sph expression for pressure forces , satisfying all conservations laws ( including entropy conservation in reversible adiabatic problems ) , was obtained by @xcite : @xmath46 using eqs . ( [ rho ] ) and ( [ nablawij ] ) to compute @xmath47 , one obtains @xmath48 \frac{{\mbox{\it{\bf r}}}_{ij}}{r_{ij } } \nonumber\\ & -&\frac{1}{4}\tilde{r}_i \sum_j m_j\left(\frac{p_i}{\rho_{i}^{2}}+ \frac{p_j}{\rho_{j}^{2}}\right ) \frac{\partial w(r_{ij},h_i)}{\partial h_{i}}\\ & + & \frac{1}{4}\sum_j \tilde{r_j}\delta_{ij_f } \sum_k \frac{m_j m_k}{m_i } \left(\frac{p_j}{\rho_{j}^{2}}+ \frac{p_k}{\rho_{k}^{2}}\right ) \frac{\partial w(r_{jk},h_j)}{\partial h_{j}}\ ; , \nonumber\end{aligned}\ ] ] where @xmath49 . on the other hand , using eqs . ( [ rho ] ) and ( [ nablawij ] ) to compute the @xmath50 derivative appearing in eq . ( [ dedt ] ) , the energy equation becomes : @xmath51 \frac{\textbf{r}_{ij}\cdot\textbf{v}_{ij}}{r_{ij}}\nonumber\\ & + & \frac{1}{4}\frac{p_i}{\rho_{i}^{2}}\sum_j m_j \left[\frac{\partial w(r_{ij},h_i)}{\partial h_i } \breve{r}_i+ \frac{\partial w(r_{ij},h_j)}{\partial h_j } \breve{r}_j\right ] + { \cal{h}}_i\;,\end{aligned}\ ] ] where @xmath52 note that eqs . ( [ accpress ] ) and ( [ energy ] ) have been deduced by using both spatial and time derivatives of the sph density as defined by eq . ( [ rho ] ) _ with the symmetrization specified in eq . ( 3 ) _ , because compatibility with the conservation laws requires that the sph force and energy equations are evaluated in consistency with the density definition . in the case of deva , this requirement increases the cpu time per integration step . in fact , since the density @xmath53 associated to a particle @xmath30 depends on both @xmath34 and @xmath35 , for @xmath54 , ( i.e. , for its @xmath25 nearest neighbors , see eq . [ rho ] ) , the computation of @xmath53 at a given integration step requires the knowledge of @xmath35 for these @xmath25 nearest neighbors at its beginning . value must be kept fixed all along the integration step in order to avoid violating the reciprocity principle ] this can be achieved either by using the @xmath35 values predicted in the previous integration step or by performing , at each step , a first loop over the particles to compute their @xmath35 values and , once it is over , a second loop to compute their hydrodynamical properties . since we look for a high accuracy rather than a high computational speed , we have adopted this latter possibility . as usual in sph , to account for dissipation at shocks , the above equations must be completed by adding an artificial viscous pressure term , @xmath55 . when the @xmath0 terms are considered , @xmath55 is added only to the leading term of equations ( [ accpress ] ) and ( [ energy ] ) , that is , those not involving @xmath0 terms @xcite : @xmath56 where we have adopted the standard viscous pressure proposed by @xcite : @xmath57 where @xmath58 and @xmath59 are constant parameters of order unity , @xmath60 is a softening parameter to prevent numerical divergences , @xmath61 is the local sound speed , and @xmath62 the simulation of a system constituted by @xmath63 particles usually requires a computational effort which considerably varies from some regions ( or particles ) to other . for example , regions of high density and submitted to strong shocks need to be simulated with timesteps much shorter than the rest of the system . in the ap3m+sph codes described in the literature , all the particles in the system are simultaneously advanced at each timestep . the particle needing the highest time resolution determines the timestep length of all the others . consequently , some few particles can slow down the simulation of a system . to make a code more efficient in handling with problems with multiple time scales , the computational effort must be centered on those particles that require it , avoiding useless computations for the remaining particles . in other words , it is necessary to allow for different timesteps for each particle . a pec ( predict - evaluate - correct ) scheme with individual timesteps has been developed and implemented on our code in the following way : 1 . we enter the step @xmath64 ( which corresponds to the time @xmath65 ) with known positions @xmath66 , velocities @xmath67 , and accelerations @xmath68 , for all the @xmath63 particles . furthermore , any integration scheme with individual timesteps needs some information to identify , at each step , those particles needing a recomputation of their acceleration . this information is stored in two vectors @xmath69 and @xmath70 , where @xmath69 is the time at which the last update of @xmath71 was performed , while @xmath72 is the time at which a recomputation of @xmath71 will be necessary in the future . @xmath73 2 . a list is constructed with those particles @xmath22 which will be advanced at the current step . such particles are labelled as _ active_. obviously , the particle @xmath74 with the smallest prediction time , @xmath75 , must be included in this list , and fixes the timestep of the remaining active particles : @xmath76 since each step requires the update of many auxiliary arrays , it is impractical to advance only a single particle . for this reason , we label as _ active _ all particles within a cubic box around @xmath74 . the size of the activation box is chosen , at each position , so that it contains a small fraction of the total number of particles . 3 . for all particles , active or not , we predict the value of @xmath77 and @xmath78 at @xmath79 @xmath80 4 . only for active particles , we compute their accelerations and correct @xmath77 and @xmath78 using : @xmath81 where the choice @xmath82 and @xmath83 maintains accuracy to second order both in positions and velocities . in these expressions , @xmath84 represents the time interval elapsed from the last evaluation of @xmath85 to that performed in the current timestep @xmath86 note that , unlike @xmath87 , the @xmath84 value is different for each active particle . we update the global time @xmath88 , as well as the @xmath89 and @xmath90 values of each active particle . here , in order to maintain the numerical stability of the ap3 m algorithm , the individual timestep @xmath91 must be smaller than the time scale for significant displacements or changes in velocity due to accelerations : @xmath92 where @xmath93 is the gravitational softening . the above integration scheme may easily be extended to include hydrodynamics . the sph processes involve three new independent variables in addition to those listed in eq . ( [ variables ] ) : @xmath94 where @xmath95 is the smoothing length , @xmath96 the specific internal energy , and @xmath97 its derivative . for all particles , we must then predict the value of @xmath98 at @xmath79 @xmath99 and compute , for active particles , both their total acceleration ( @xmath100 and @xmath101 ) as well as their hydrodynamical variables ( @xmath102 , @xmath103 , @xmath104 and @xmath105 ) . these quantities are then used to correct the internal energy of active particles : @xmath106 where the choice @xmath107 maintains accuracy to second order in internal energies . now , the numerical stability requires additional limits on the timestep of each gas particle . a first timestep control is that concerning the time scale for significant displacements or changes in velocity due to accelerations : @xmath108 a second limit on @xmath109 is usually given by a timestep control which combines the courant and the viscous conditions : @xmath110\;.\label{dtcv}\ ] ] when required , radiative cooling is implemented in an integral form @xcite using the fact that , due to the courant condition , the density field is nearly constant over a time - step : @xmath111 where @xmath112 is the power radiated per unit volume and @xmath113 is the change in @xmath114 due to cooling processes . this integral procedure circumvents the need of a control time for cooling , and , hence , it never limits the timestep . the numerical stability of our code requires that cooling effects must be updated at each step for all particles , active or not . otherwise , the simultaneous presence of already cooled and not yet cooled particles in a given object would break the local pressure equilibrium and , as a result , cold particles would fall to the object center causing a non - physical core of very high density ( see @xmath155.2 for an example ) . fig [ cpu ] shows , for a typical cosmological simulation , the ratio of the cpu time consumed by an algorithm with individual timesteps to that consumed when all particles are simultaneously advanced . we see that the use of individual timesteps typically reduces the cpu time per step in a factor of five . in a pentium iv 1.7ghz personal computer , the cpu time typically consumed by our code in a @xmath115 cosmological simulation without the @xmath0 terms is : a ) 25 seconds per step in simulations without radiative cooling ( such as the santa barbara cluster test of 5.1 ) , b ) from 25 ( at high redshifts ) to 70 ( at low redshifts ) seconds per step in cosmological simulations with radiative cooling ( such as those of 5.2 ) . these cpu times are increased by about 150% when the @xmath0 terms are taken into account . deva has been applied to different problems with known analytical or numerical solutions . the aim of such simulations was not only to test our code , but also to analyze the effects of the @xmath0 terms included in it . the one - dimensional shock tube problem proposed by @xcite has become a standard test of all transport and source terms ( including artificial viscosities ) of hydrodynamic algorithms . it considers a perfect gas distributed on the @xmath116-axis . a diaphragm at @xmath117 initially separates two regions which have different densities and pressures . all particles are initially at rest . at time @xmath118 the diaphragm is broken and both regions start to interact . nonlinear waves are then generated at the discontinuity and propagate into each region : a shock wave which moves from the high to the small pressure region , while the associated rarefaction wave moves in the inverse sense . the analytical solution to this problem has been given by @xcite and @xcite . in our simulation , we have considered @xmath119 gas particles initially distributed in the interval @xmath120 according to : @xmath121 dissipational effects , other than those associated with the artificial viscosity ( with @xmath122 and @xmath123 ) , were ignored , as well as gravitational interactions . [ shocktube ] shows our results at @xmath124 . we see from this figure that our results are in excellent agreement with the analytical solutions . the resulting profiles both in the shock wave ( located at @xmath125 ) and in the contact discontinuity ( located at @xmath126 ) are much less rounded than in previous sph computations ( see , e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) as a result of having used a larger number of particles and , hence , a better resolution . we also note the almost complete suppression of post - shock oscillations in our results . these oscillations can be seen in the previous sph simulations of this problem , especially in the velocity field , while no high - frequency vibrations are perceptible in our results . the weak blip observed in the pressure profile at the contact discontinuity ( @xmath126 ) is normal in sph codes . such non - physical blip has been explained by @xcite as due to the fact that the smoothed estimate of pressure is computed by using discontinuous quantities . it is then inevitable that @xmath127 has some slight perturbation at the contact discontinuity , but it has a negligible effect on the motion . in this test , simulations including the @xmath0 terms gave exactly the same results as those neglecting such terms . a 3d - problem usually considered to test hydrodynamical codes is that concerning the adiabatic collapse of a non - rotating gas sphere . this problem has been studied from a finite - difference method by @xcite , and from sph simulations by @xcite and @xcite in order to facilitate the comparison of our results to those obtained by these authors , we have taken their same initial conditions : a gas sphere of radius @xmath128 and total mass @xmath129 , with density profile @xmath130 all the @xmath119 gas particles are initially at rest and have the same specific internal energy @xmath131 . units were taken so that @xmath132 . initially far from equilibrium , the system collapses converting most of its kinetic energy into heat . a slow expansion follows and , at late times , a core - halo structure develops with nearly isothermal inner regions and the outer regions cooling adiabatically . we show in fig . [ hkfig ] different system profiles at end of the simulation . the solid line represents the numerical solution obtained when the @xmath0 terms have been included , the dashed line represents the numerical solution obtained when these terms have been neglected , and the points represent the numerical solution obtained by @xcite . we see that , although the solid and dashed lines are not exactly superposed , both solutions are coincident within the error bars . we can understand why the @xmath0 terms have a negligible effect in the two standard tests reported in this section . the effect of the @xmath0 terms on the thermal energy can be expressed by a time scale defined by @xmath133 where @xmath134 is the ratio of the specific thermal energy @xmath114 of particle @xmath30 to the change in @xmath114 due to the @xmath0 terms @xmath135 when the time elapsed , @xmath136 , is longer than this time scale , the @xmath0 terms will produce a non - negligible effect on the thermal energy . this condition can be expressed as @xmath137 where @xmath138 represents the area contained by the @xmath139 curve between @xmath140 and @xmath141 . when these timescales are computed for the non - adiabatic tests reported in this paper , one obtains @xmath142 ( in the shock tube problem ) , and @xmath143 ( in the collapse of a non - rotating gas sphere ) . the effect of the @xmath0 terms are then expected to be small . testing the effects of the @xmath144 terms on hydrodynamical evolution can be better worked out in isentropic processes . it is not easy , however , to get such kind of processes in simulations of gas evolution because gas easily develops shocks where dissipation must occur . a possibility is considering the adiabatic expansion of a gas sphere from a situation of equilibrium . using expansion rather than collapse has the advantage that shell crossing decreases substantially , so that viscous force terms can be removed and the evolution is isentropic . such simulations as that shown in [ collapse ] lead at late times , @xmath145 , to equilibrium spheres with density profiles as that displayed in fig . [ hkfig ] . we use the adiabatic expansion of such spheres to test the effects of the @xmath0 terms in deva . initial conditions were generated by performing a simulation as that described in [ collapse ] , using different numbers of particles . at @xmath146 , we switched - off its viscous pressures ( by setting @xmath147 ) to ensure that the subsequent evolution conserves the total entropy . the self - gravity was also switched - off at @xmath146 . in absence of gravitational interactions , this system expands fast and , at @xmath148 , its central density has decreased by a factor of @xmath149 . the evolution from @xmath146 to @xmath148 must conserve both the total energy and entropy . table 1 shows the results for this series of simulations . we see that , when the @xmath0 correction terms are neglected , energy is conserved very accurately but there exists a considerable violation of the total entropy ( about @xmath150 in the time interval we have considered ) . in the opposite case , when such correction terms are taken into account , both total energy and total entropy are conserved very accurately ( about @xmath151 ) . these results appear to be independent on the number of particles . we then conclude that the @xmath0 correction terms cure entropy violation , allowing , at the same time , a very good energy conservation . the deva code is particularly well suited to numerically follow , in a cosmological context , the assembly from the field of primordial fluctuations of collapsed objects , such as galaxy clusters or galaxies . to illustrate deva performances in this situation , we briefly analyze some results of self - consistent simulations . self - consistency means that initial conditions are set at high @xmath5 as a montecarlo realization of the field of primordial fluctuations ( i.e. , perturbations , characterized by a spectrum , to a given cosmological model ) , and then they are left to evolve according with newton s laws and the hydrodynamical equations . the santa barbara cluster problem was proposed by @xcite to compare the results obtained from different codes . the formation of a x - ray cluster in a cold dark matter ( cdm ) universe has been simulated using most of the hydrodynamic codes available at that time , setting a standard of reference to test newly proposed hydrodynamic codes . the initial conditions of this test correspond to a @xmath152 peak of the density field smoothed with a gaussian filter of radius @xmath153 mpc according to the algorithm of @xcite . the perturbation was centered on a periodic cubic region of side @xmath154 mpc . the cosmological scenario is a flat cdm universe with @xmath155 km s@xmath156 mpc@xmath156 for the hubble constant ; @xmath157 for the present - day linear rms mass fluctuation in spherical top hat spheres of radius 16 mpc ; and @xmath158 for the baryon density ( in units of the critical density ) . 64@xmath159 dark matter and 64@xmath159 baryon particles have been used with a softening length of 20 kpc . to test the influence of the @xmath0 terms , two different simulations were run with deva . in one of these simulations , the @xmath0 terms have been considered ( sbgh ) while in the other they have been neglected ( sbnogh ) . in figure [ stbfig ] , the density , temperature and entropy profiles of the cluster are plot . the stars represent the results obtained @xcite by jenkins from a high - resolution sph simulation using a parallel version of the hydra code @xcite , while the circles represent the results obtained by bryan & norman from a high - resolution adaptive mesh refinement shock - capturing code , samr , @xcite . as previously remarked by @xcite , we see that the sph and mesh results differ at the central region . this figure also shows the results obtained from our code both when the @xmath0 terms are included ( solid line ) and neglected ( dashed line ) . error bars correspond to the standard deviation of the individual sph data . we see that , now , our results differ slightly depending on whether the @xmath0 terms have been included or not ( the slope of the density profile flattens more rapidly in sbgh than in sbnogh ; the temperature profile is flat in sbgh and decreases within 100 kpc in sbnogh ; the entropy profile is almost flat within 100 kpc in sbgh and decreases continuously in sbnogh ) . moreover , when the @xmath0 terms are neglected , we obtain results that are similar to those of previous sph simulations , and very close to jenkins results , obtained with a much higher resolution . when these terms are taken into account , the results are intermediate between previous sph and grid results . this suggests that , at least in part , the difference between the sph and grid results could be due to the non - physical entropy introduced by sph codes . this non - physical entropy is negative @xcite and , therefore , it produces objects with a smaller central temperature and a higher central density . particular attention deserves the comparison of deva results with those obtained from the entropy conserving sph - tree formulation by springel & hernquist ( 2002a ) , hereafter s - gadget . in figure [ comparison ] we give a comparison of the santa barbara cluster entropy profiles obtained in the sbgh run and s - gadget , kindly provided by y. ascasibar and g. yepes @xcite . both have been run with the same number of particles and gravitational resolution . we see that the agreement is very good within the error bars . so , both techniques compare very well in terms of entropy conservation . the differences found between sbgh and sbnogh simulations can be understood on the basis of the timescale for the @xmath0 terms ( section 4.3 ) . the time integral of @xmath160 ( see eq . [ ts ] ) for this test is now larger than unity ( @xmath161 ) . we now report on several self - consistent simulations run in the context of flat @xmath162cdm cosmological models with the aim of studying galaxy assembly ( see table 2 for a summary ) . we have considered two different @xmath162cdm models with parameters whose values are consistent with their recent determinations from observations : simulations @xmath162cdm1 ( @xmath162cdm2 ) have @xmath163 ( 0.7 ) , @xmath164 ( 0.04 ) , @xmath165 ( 1.00 ) and @xmath166 ( 0.70 ) ( * ? ? ? * ; * ? ? ? * ; * ? ? ? * and references therein ) . both simulations share the same seed for the montecarlo realization of the initial fluctuation field , so that each object formed in one simulation has its counterpart in the other simulation . in each case , we have used @xmath167 dm particles and @xmath167 gas particles , in a periodic box of 10 mpc comoving side . the gravitational softening is @xmath168 kpc , and the minimum allowed smoothing length , @xmath169 , as usual . for each cosmological model , two simulations have been run that are identical ( they have exactly the same initial conditions and the same values of the cosmological parameters ) except that in one case the @xmath0 terms have been included ( simulations @xmath162cdm1-h and @xmath162cdm2-h hereafter ) , while in the second case these terms have not been taken into account ( simulations @xmath162cdm1-noh and @xmath162cdm2-noh hereafter , see table 2 ) . note that in these simulations the cosmological volume is homogeneously sampled , in the sense that no resampling multimass technique has been used to study glo assembly . this work can be considered as an extension of previous works in standard cdm models @xcite , where cosmological simulations had been run with a different code based on a different numerical approach with fixed integration timesteps and particle masses @xcite . relative to these previous works , deva opens the possibility of considering the effects of the @xmath0 terms at scales of galaxies in self - consistent simulations . moreover , the simulations we report here represent an improvement of the baryonic mass resolution by factors of @xmath170 in the number of baryonic particles sampling a glo of a given total mass . also , the time resolution allowed by the multistep technique has been improved by a factor of @xmath171 in the denser areas of the box . cooling has been implemented as described in @xmath15 3.4 , where the cooling curve is that from @xcite and @xcite for an optically thin primordial mixture of h and he ( @xmath172 , @xmath173 ) in collisional equilibrium and in absence of any significant background radiation field . concerning star formation , in this work we report on _ direct _ results ( i.e. , no spectrophotometric ) obtained with the simplest implementation of star formation in the code : through a parameterization , similar to those used by @xcite and @xcite , see @xcite for details , based on the jeans criterion for a collapsing region . gas particles are turned into stars according with an inefficient schmidt - law - like transformation rule ( see * ? ? ? * ; * ? ? ? * ) , @xmath174 where @xmath175 is a dimensionless star - formation efficiency parameter , and @xmath176 is a characteristic time - scale chosen to be equal to the maximum of the local gas - dynamical time @xmath177 , and the local cooling time , @xmath178 . equation ( [ star - rate ] ) implies that the probability @xmath179 that a gas particle forms stars in a time @xmath87 is @xmath180 as usual , we compute @xmath179 at each time step for all eligible gas particles and draw random numbers to decide which particles actually form stars . stellar feedback processes have not been explicitly considered , but a tuning of the efficiency parameters can mimic these feedback effects . galaxy - like objects of different morphologies appear in the simulation : disk - like objects ( dlos ) , early - type - like objects ( etlos ) and irregular objects . dlos contain gas in an extended disk , and most stars in a massive compact central concentration . in simulations with lower @xmath181 values ( not reported here ) , stars form also in the disks , along arms . etlos are very poor in gas and their stellar component have relaxed regular ellipsoidal shapes . irregulars have not defined shapes , and , in most cases , they are the product of a recent merger or interaction event . we note that glos formed in @xmath162cdm2-noh and @xmath162cdm2-h tend to be of later type than their @xmath162cdm1-noh or @xmath162cdm1-h counterparts , because of the lower values the parameters @xmath3 and @xmath182 take in @xmath162cdm2 simulations @xcite . first analyses of glo formation and evolution in simulations run with the deva code are reported in @xcite ; more details will be given elsewhere ( siz el al . 2003 , in preparation ) . here we focus on different aspects related with deva performances . one important issue related to disc formation in hydrodynamical simulations is specific angular momentum conservation at kpc scales @xcite . in fig . [ jespmass ] , the specific total angular momentum at @xmath183 is represented versus mass for dlos identified in @xmath162cdm2-noh and @xmath162cdm2-h with @xmath184 150 km s@xmath156 ( @xmath185 is the disc scalelength , see * ? ? ? the specific total angular momentum is plot for dark haloes , @xmath186 ( open symbols ) , for the inner 83 per cent of the disc gas mass ( i.e. , the mass fraction enclosed by @xmath187 in a purely exponential disc ) , @xmath188 ( filled symbols ) , and for the stellar component of the dlos in our simulations , @xmath189 ( starred symbols ) . we see that , except for the three more massive objects , @xmath190 is of the same order as @xmath186 , so that these gas particles have collapsed conserving , on average , their angular momentum . moreover , dlos formed in our simulations are inside the box defined by observed spiral discs in this plot @xcite . in contrast , @xmath189 is much smaller than either @xmath186 or @xmath190 , meaning that the stellar component in the central parts has formed out of gas that had lost an important amount of its angular momentum in catastrophic events or that had never acquired it . we note that this result is similar to those obtained in simulations run with the code by @xcite , even if the codes are different , as explained above , and that the global cosmological models are also different . the smoothed estimate of the local gas density given by eq . ( 2 ) and the ensuing formulation of sph equations , symmetrized to ensure the reciprocity principle , has made possible @xmath190 conservation in an axisymmetric potential . this is a delicate crucial point in sph codes . as stated in @xmath152.2 , its accurate implementation requires that the individual smoothing lengths must be completely updated at the beginning of each integration step , what increases the cpu time requirements . this complication can not be avoided , however , when an accurate @xmath190 conservation is a key point in the physical processes under study . a different approach to saving cpu time is then necessary . using different timesteps for each particle , useless computations for particles that do not require high time resolution are avoided . this saves considerable amounts of cpu time . multistepping has allowed us to run these simulations , that have a considerable dynamical range , in a modest computing machine ( a pentium iv 1.7ghz personal computer , see above ) . let us now turn to the effects of the @xmath0 terms at kpc scales . as stated in @xmath154.3 , their general effect is to correct the spurious negative entropy introduced in their absence by sph codes , mainly at the central regions of collapsed objects . as a consequence , when the @xmath0 terms are taken into account , dissipation by the gaseous component of a given glo increases , so that they are more disordered or dynamically hotter at their central regions . equilibrium is then attained with lower central baryon concentrations or densities , decreasing the amount of gas infall . but this is not the unique effect of @xmath0 terms on mass distribution . it is a well known effect that dark matter is pulled in by baryons as they lose their energy and fall onto these central volumes of collapsed configurations @xcite . decreasing the amount of gas infall translates , as a consequence , into a decreasing of the amount of dark matter at the glo centers . as an illustration of this effect on both the gaseous and dark components , in figure [ velcir ] ( upper panel ) we show the circular velocity curves for the most massive etlo formed in @xmath162cdm1-h ( thick lines ) and its counterpart formed in @xmath162cdm1-noh ( thin lines ) . in the lower panel , the corresponding circular velocity curves are represented for a dlo formed in @xmath162cdm2-h ( thick lines ) and its counterpart in @xmath162cdm2-noh ( thin lines ) . in these figures , @xmath191 is the radial distance to the glo center of mass , solid lines are the circular velocities , @xmath192 , short - dashed lines and long - dashed lines stand for the dark matter ( dm ) and the baryonic ( bar , both stellar and luminous ) contributions , respectively , given by : @xmath193 with @xmath30 = bar and dm . as can be clearly seen in these figures , the central distributions of both , dark matter and baryons , are different and in any case the concentrations are lower when the @xmath0 terms are included . these central concentrations are often quantitatively estimated in literature through the @xmath194 parameter ( the maximum or peak circular velocity , see * ? ? ? glos in @xmath162cdm1-h or @xmath162cdm2-h have lower @xmath195 values than their @xmath162cdm1-noh and @xmath162cdm2-noh counterparts , respectively . this can be seen in figure [ velcir ] for an etlo and a dlo , but the behavior is general for any glo . other useful parameter to quantitatively characterize circular velocity curves is the logarithmic slope ( ls ) , observationally defined for disc rotation curves as the slope of the straight line that fits @xmath196 in log - log scale from @xmath197 up to the last measured point in the rotation curve ( * ? ? ? * and references therein ; @xmath185 is the disk scalelength ) . lss are a measure of the glo halo compactness at the scales of 10 - 30 kpcs . as illustrated in figure [ velcir ] for an etlo and a glo , a tendency for glos to be less compact also at these scales when the @xmath0 terms are taken into account has been found in the simulations . a second effect of the @xmath0 terms is related to the amounts of stars formed in a given glo at its formation and all along its evolution until @xmath183 . in these simulations , many stars form at the shock fronts , where gas is compressed to very high densities . softer shocks mean less star formation with the same efficiency parameters . to illustrate this effect , in figure [ massstar ] the ratios of the total stellar masses for the 8 ( 7 ) more massive etlos produced in @xmath162cdm1-h ( @xmath162cdm2-h ) over the total stellar masses of their @xmath162cdm1-noh ( @xmath162cdm2-noh ) counterparts , are plot versus their total virial masses , @xmath198 . we see that , except in one case , these ratios are smaller than the unity , as expected . concerning sizes , for etlos we define the intrinsic or 3d cold baryon effective radius , @xmath199 , as the radius of the sphere enclosing half the total etlo mass in cold baryons ( i.e. , cold gas or stars ) , @xmath200 . this is a measure of the etlo size at scales of the baryonic objects . in figure [ restar ] we plot @xmath199 for the @xmath162cdm1-h ( @xmath162cdm2-h ) etlos in units of those of their @xmath162cdm1-noh ( @xmath162cdm2-noh ) counterparts . as expected , @xmath162cdm1-h ( @xmath162cdm2-h ) etlos have larger baryonic sizes than their @xmath162cdm1-noh ( @xmath162cdm2-noh ) counterparts , except for one case . finally , shocks heat the gas , producing an extended diffuse hot component . the effect of the @xmath0 terms in this case is to lower the temperature of this diffuse gaseous halo component of glos relative to the case when these terms are not considered . histograms for the temperature distribution of the gaseous component of two of the most massive glos in @xmath162cdm1-h ( thick lines ) and in @xmath162cdm1-noh ( thin lines ) are shown in figure [ temphist ] . to draw these histograms , all the gas particles within a sphere of radius @xmath201 , centered at the glo center of mass , have been considered ( @xmath201 is the radius where the curve of integrated plasma emission luminosity reaches its asymptotic value ) . as illustrated in the histograms for these two glos , the temperature distributions of the gaseous component for @xmath162cdm1-h and @xmath162cdm1-noh objects do not differ substantially . in both cases , the gas is close to biphasic , but @xmath162cdm1-h objects are slightly colder within @xmath202 than @xmath162cdm1-noh objects . they are also slightly more extended at @xmath201 scales , with the small excess of colder particles placed at the outskirts of the gaseous configurations . the differences found between @xmath162cdm1-h and @xmath162cdm1-noh or between @xmath162cdm2-h and @xmath162cdm2-noh simulations can be understood on the basis of the timescale for the @xmath0 terms ( section 4.3 ) . the time integral of @xmath160 ( see eq . [ ts ] ) is now much larger than unity ( @xmath203 ) , indicating that the @xmath0 terms can not be neglected in this kind of simulations . as remarked in @xmath15[incsph ] , cooling implementation in multistep codes has to be handled with caution : cooling processes need to be updated at each timestep for all particles . otherwise , gas particles involved in shocks suffer from a spurious cooling and non - physical cores of high density appear , giving rise to extremely compact objects . as an example , in figure [ restar ] we plot the 3d cold baryon half - mass radii for a variant of @xmath162cdm1-noh , termed @xmath162cdm1-mcool ( see table 2 for details ) , where only active particles at each timestep have been allowed to cool . in figure [ restar ] ( starred symbols ) we see that now the effective radii for the most massive objects are significantly smaller than in @xmath162cdm1-noh ; the difference becomes less important as the mass of the objects decreases and , as a consequence , the fraction of their constituent particles involved in shocks also decreases . note that , as the comparison of filled points and starred symbols in figure [ restar ] shows , the combined effects of entropy violation and spurious cooling can produce unphysically very small objects , a factor of ten smaller , in some cases , than the values found when these effects are circumvented ( i.e. , the @xmath162cdm1-h simulation ) . we present deva , a multistep ap3m - like - sph code designed to study galaxy formation and evolution in connection with the global cosmological model , that uses a formulation of sph equations ensuring both energy and entropy conservation . multistepping is introduced to save cpu time . in self - consistent cosmological simulations multiple time scales appear , due to their large dynamical ranges . to avoid that particles in denser zones slow down the simulation , and , at the same time , to get a properly accurate integration algorithm , it is then advantageous to use individual time steps . a comparison of the cpu time used in a self - consistent cosmological simulation when it is run with a _ global _ timestep or with a _ multistep _ scheme indicates that in the second case results at an equivalent level of accuracy are produced @xmath2045 times faster . when a multistep scheme is adopted , a delicate issue in the study of galaxy assembly in a cosmological context is the _ cooling _ implementation . in deva , as no cooling timescale is taken into consideration to fix the individual timestep for each particle , cooling must be calculated in a non - multistep way . otherwise , particles involved in strong shocks would spuriously cool and form dense objects characterized by unphysically small baryon half - mass radii . on writting deva , particular attention was paid that conservation laws of physics ( energy , entropy , momentum ) are correctly implemented in the code , so that they hold at scales and under physical conditions relevant for galaxy assembly in a cosmological context . the usual formulations of sph equations focus on energy conservation and they violate , by construction , entropy conservation . as a consequence , a negative entropy is numerically produced in shocks that might ( or might not ) spuriously pollute the results , depending on the problem one studies . different authors have addressed the issue of entropy violation in sph codes @xcite and have given different alternative formulations of its equations . we have implemented a formulation that considers explicitly the effects of the @xmath0 terms in sph equations @xcite . by taking advantage of the structure of the ap3 m algorithm in the neighbor search , the implementation of @xmath0 terms in the code is simple and noiseless . to test the relevance of entropy violation at shock locations under different physical conditions , we have studied problems and run simulations ( namely , the adiabatic santa barbara cluster formation test , @xcite , and fully self - consistent cosmological simulations to study galaxy formation including cooling an star formation ) using identical initial conditions and two different versions of deva , one that takes into account the @xmath0 terms in sph equations and one that does not take them into account . we show that entropy violation has consequences on the thermodynamical properties of the very central regions of the santa barbara cluster and on the structure at kpc scales of galaxy - like objects ( glos ) formed in simulations , but it does not have any appreciable consequence on the standard non - cosmological tests of hydrodynamical codes . to understand the origins of these different behaviors , a criterion is introduced that allows to elucidate when entropy violation is expected to have appreciable consequences on the results . in standard tests , only moderate shocks and time scales are involved . as a consequence of the non - physical negative entropy numerically produced when the @xmath0 terms are neglected , both glos and the cluster present more concentrated baryon density profiles ( either star or gas ) . concerning the santa barbara cluster test , when the @xmath0 terms are not included , we get results that are similar to those of previous sph simulations that focus on energy conservation . however , when these terms are considered , the results are intermediate between those sph and grid results . for example , the temperature profile is decreasing within about 100 kpc of the cluster center when the simulation is run with standard sph codes ( including deva without @xmath0 terms ) , it increases when it is run with grid codes and it is flat when deva + @xmath0 is used . we would like to note that the accuracy figure of entropy conservation obtained with deva + @xmath0 and the entropy - conserving sph - tree code by springel & hernquist ( s - gadget , 2002a ) compare quite satisfactorily , as indicated by the good agreement between the santa barbara cluster entropy profiles obtained with both codes . in cosmological simulations , negative entropy causes galaxy - like objects ( glos ) to be dynamically less hot and gas infall onto their central regions is artificially increased , causing , also , an increase of the amount of dark matter at the glo centers . these results qualitatively agree with those obtained with s - gadget by @xcite in their simulation of a @xmath162cdm cosmological model . no quantitative comparisons are possible by the moment , because a standard of comparison for self - consistent cosmological simulations is unfortunately not available . an important result of this work is that the combined effects of entropy violation and multi - step cooling implementation in cosmological simulations can be particularly dramatic concerning the concentration of mass distribution in the galaxy - like objects they produce . for example , their baryon half - mass radii can be up to a factor of ten smaller than half - mass radii of glos produced in entropy - conserving non - multistep runs with deva . concerning momentum conservation , we have used a formulation of sph equations that is consistent with the smoothed estimate of the local gas density ( eqs ( 1 ) and ( 3 ) ) . equations are symmetrized to ensure that the reciprocity principle holds ( that is , if at a given time the @xmath22th particle belongs to the neighbor list of the @xmath30th particle , then it is mandatory that , at this same time , the @xmath30th particle belongs to the neighbor list of the @xmath22th particle ) , so that momentum and angular momentum are conserved . the implementation of this principle in a sph code increases considerably the cpu time per integration step , because a double loop on gas particles is necessary to evaluate smoothing lengths . to test angular momentum conservation , we have measured the specific angular momentum of discs formed in self - consistent simulations . it has been found that conservation is good enough to obtain simulated discs with observational counterparts , without any need of previous heating , as already @xcite have shown ( see also * ? ? ? the use of a very high number of particles could ensure angular momentum conservation in sph - tree codes @xcite . in this paper it has been shown that codes paying a particular attention to the implementation of conservation laws of physics at the scales of interest can attain a good level of accuracy in conservation laws with more limited resources . it is a pleasure to thank a. knebe , m. norman , v. quilis , j. silk , j. sommers - larsen , p. tissera and g. yepes for useful information and discussions on the topics addressed in this paper . particular thanks are due to h.m.p . couchman for making public his ap3 m code , on which the gravitational part of deva is based , and to y. ascasibar and g. yepes for allowing us to use their results on the santa barbara cluster test . this project was partially supported by the mcyt ( spain ) through grant aya-0973 from the programa nacional de astronoma y astrofsica . we also thank the iberdrola foundation for financial support , and the centro de computacin cientfica ( uam , spain ) for computing facilities . 50 alimi j .- m . , serna a. , pastor c. & bernabeu g. 2002 , j. comput . ( in press ) ascasibar , y. 2003 , phd thesis , universidad autnoma de madrid avila - reese , v. , & vzquez - semadeni , e. 2001 , , 553 , 645 barnes , j.e . 1989 , , 338 , 123 barnes , j.e . , 1992 , , 393 , 484 barnes , j. e. & hernquist , l. e. 1991 , , 370 , l65 barnes , j. e. & hernquist , l. 1992 , , 30 , 705 barnes j. , & hut p. 1986 , 324 , 446 benz , w. & hills , j. g. 1987 , , 323 , 614 bond , j. r. , centrella , j. , szalay , a. s. , & wilson , j. r. 1984 , , 210 , 515 borgani , s. , governato , f. , wadsley , j. , menci , n. , tozzi , p. , quinn , t. , stadel , j. , & lake , g. 2002 , , 336 , 409 bryan , g.l . , norman , m.l . , stone , j.m . , cen , r. , & ostriker , j.p . 1995 , comput . comm . , 89 , 149 bryan , g.l . , & norman , m.l . 1995 , baas , 187 , 9504 couchman , h. m. p. 1991 , , 368 , l23 couchman , h.m.p . , thomas , & pierce , f.r . , 1995 , apj , 452 , 797 courteau , s. 1997 , , 114 , 2402 dalcanton , j. j. , spergel , d. n. , & summers , f. j. 1997 , , 482 , 659 dave , r. , dubinski , j. , & hernquist , l. 1997 , new astronomy , 2 , 277 domnguez - tenreiro , r. , serna , a. , siz , a. , & sierra - gonzlez de buitrago , m. m. 2003 , , in press dom ' inguez - tenreiro , r. , tissera , p. b. , & s ' aiz , a.1998 , , 508 , l123 elmegreen , b. 2003 , , in press evrard , a. e. 1988 , , 235 , 911 fall , s.m . 1983 , in athanassoula e. , ed , iua symp . 100 , internal kinematics and dynamics of galaxies . reidel , dordrecht , p. 391 frenk , c. s. et al . 1999 , , 525 , 554 gingold r. a. & monaghan j. j. 1977 , , 181 , 375 gingold r. a. & monaghan j. j. 1982 , j. comput . 46 , 429 godunov , s. k. 1959 , matematicheskii sbornik , 47 , 271 governato , f. , mayer , l. , wadsley , j. , gardner , j.p . , willman , b. , hayashi , e. , quinn , t. , stadel , j. , & lake , g. 2002 , astro - ph/0207044 preprint hawley , j. f. , smarr , l. l. , & wilson , j. r. 1984 , , 277 , 296 hernquist , l. 1993 , , 404 , 717 hernquist , l. & katz , n. 1989 , , 70 , 419 hoffman , y. & ribak , e. 1991 , , 380 , l5 kang , h. , ostriker , j. p. , cen , r. , ryu , d. , hernquist , l. , evrard , a. e. , bryan , g. l. , & norman , m. l. 1994 , , 430 , 83 katz n. 1992 , 391 , 502 katz , n. , weinberg , d. h. , & hernquist , l. 1996 , , 105 , 19 kennicutt , r. 1998 , , 498 , 541 klein , r.i . , fisher , r.t . , mckee , c.f . & truelove , j.k . 1998 , in proceedings of the international conference of numerical astrophysics , ed . s .. m. miyama et al . ( boston : kluwer academic ) , p. 131 knebe , a. , green , a. , & binney , j. 2001 , , 325 , 845 knebe , a. , kravtsov , a. v. , gottl " ober , s. , & klypin , a. a.2000 , , 317 , 630 kravtsov , a. v. , klypin , a. a. , & khokhlov , a. m. 1997 , , 111 , 73 kritsuk , a. g. & norman , m. l. 2002 , , 569 , l127 lahav , o. 2002 , in proceedings of the xxxviith rencontres de moriond on the cosmological model , astro - ph/0208297 preprint lahav , o. , bridle , l. , percival , w.j . , et al . 2002 , astro - ph/0112162 preprint lucy , l. b. 1977 , , 82 , 1013 mihos , j. c. & hernquist , l. 1994 , , 425 , l13 mihos , j. c. & hernquist , l. 1996 , , 464 , 641 monaghan , j. j. 1992 , , 30 , 543 monaghan j. j. & gingold r. a. 1983 , j. comput . phys . , 52 , 374 monaghan , j. j. & lattanzio , j. c. 1985 , , 149 , 135 navarro , j. f. & white , s. d. m. 1993 , , 265 , 271 nelson , r. p. & papaloizou , j. c. b. 1993 , , 265 , 905 nelson , r. p. & papaloizou , j. c. b. 1994 , , 270 , 1 netterfield , c. b. et al . 2002 , , 571 , 604 norman , m. l. & bryan , g.l . 1998 , in proceedings of the international conference of numerical astrophysics , ed . miyama et al . ( boston : kluwer academic ) , p. 19 padoan , p. , jimenez , r. , nordlund , a. , & boldyrev , s. 2003 , astro - ph/0301026 preprint padoan , p. , juvela , m. , goodman , a. & nordlund , a. 2001 , , 553 , 227 pearce , f. r. & couchman , h. m. p. 1997 , new astronomy , 2 , 411 rasio , f. a. & shapiro , s. l. 1991 , , 377 , 559 siz a. , domnguez - tenreiro r. , & serna , a. 2002 , , 281 , 309 siz a. , domnguez - tenreiro r. , & serna , a. 2003 , , in press siz , a. , domnguez - tenreiro , r. , tissera , p. b. , & courteau , s. 2001 , , 325 , 119 serna a. , alimi j .- m . , & chize j .- p . 1996 , , 461 , 884 silk , j. 2001 , , 324 , 313 sod , g. a. 1978 , j. comput . phys . , 27 , 1 sommer - larsen j. , & dolgov a. 2001 , , 551 , 608 spergel d. n. et al . 2003 , astro - ph/0302209 preprint springel , v. & hernquist , l. 2002 , , 333 , 649 springel , v. & hernquist , l. 2002b , astro - ph/0206393 preprint springel , v. , yoshida , n. & white , s. 2001 , new astronomy , 6 , 79 steinmetz , m. 1996 , , 278 , 1005 steinmetz , m. & navarro , j. 1999 , , 513 , 555 teyssier , r. 2002 , a & a , in press thacker , r. j. & couchman , h. m. p. 2000 , , 545 , 728 thomas p. a. 1987 , ph.d . thesis , cambridge univ . thomas p. a. & couchman h. m. p. 1992 , , 257 , 11 tissera p. b. , lambas d. g. , & abadi m. g. , 1997 , , 286 , 384 tissera p. b. , & domnguez - tenreiro , r. 1998 , , 297 , 177 tissera p. b. 2001 , , 540 , 384 tissera , p.b . , domnguez - tenreiro , r. , scannapieco , c. , & siz , a. 2002 , mnras , 333 , 327 tucker w. h. 1975 , radiation precesses in astrophysics ( new york , wiley ) vzquez - semadeni , e. , ostriker , e. c. , passot , t. , gammie , c. f. , & stone , j. m. 2000 , protostars & planets iv , ed . v. mannings et al . ( tucson : univ . of arizona press ) , in press vedel , h. , hellsten , u. , & sommer - larsen , j. 1999 , , 271 , 743 verheijen , m. , bershady , m. , & andersen d. 2002 , mass galaxies at low and high redshift ( eso workshop , venice ) pp . 24 - 26 wada , k. & norman , c. a. 2001 , , 547 , 172 yepes , g. , kates , r. , khokhlov , a. , & klypin , a. 1997 , , 284 , 235	we describe deva , a multistep ap3m - like - sph code particularly designed to study galaxy formation and evolution in connection with the global cosmological model . this code uses a formulation of sph equations which ensures both energy and entropy conservation by including the so - called @xmath0 terms . particular attention has also been paid to angular momentum conservation and to the accuracy of our code . we find that , in order to avoid unphysical solutions , our code requires that cooling processes must be implemented in a non - multistep way . we detail various cosmological simulations which have been performed to test our code and also to study the influence of the @xmath0 terms . our results indicate that such correction terms have a non - negligible effect on some cosmological simulations , especially on high density regions associated either to shock fronts or central cores of collapsed objects . moreover , they suggest that codes paying a particular attention to the implementation of conservation laws of physics at the scales of interest , can attain good accuracy levels in conservation laws with limited computational resources .
the current decade has witnessed a significant amount of effort towards discovering the networks of protein - protein interactions ( interactomes ) in a number of model organisms . these efforts resulted in hundreds of thousands of individual interactions between pairs of proteins being reported @xcite . repositories such as the biogrid @xcite , intact @xcite , mint @xcite , dip @xcite , bind @xcite and hprd @xcite have been established to store and distribute sets of interactions collected from high - throughput scans as well as from curation of individual publications . depending on its goals , each interaction database , maintained by a different team of curators located around the world includes and annotates interactions differently . consequently , while many interactions of specific interactomes are shared among databases @xcite , no one contains the complete known interactome for any model organism . constructing a full - coverage protein - protein interaction network therefore requires retrieving and combining entries from many databases . this task is facilitated by several initiatives developed by the proteomics community over the years . the imex consortium @xcite was formed to facilitate interchange of information between different primary databases by using a standardized format . the proteomics standards initiative molecular interaction ( psi - mi ) format @xcite allows a standard way to represent protein interaction information . one of its salient features is the controlled vocabulary of terms that can be used to describe various facets of a protein - protein interaction including source database , interaction detection method , cellular and experimental roles of interacting proteins and others . the psi - mi vocabulary is organized as an ontology , a directed acyclic graph ( dag ) , where nodes correspond to terms and links to relations between terms . this enables the terms to be related in an efficient and algorithm - friendly manner . consistently annotated datasets are useful for development and assessment of interaction prediction tools @xcite . furthermore , such datasets also form the basis of interaction networks , for which numerous analysis tools have been developed @xcite . depending on biological aims of a tool , different entities ( nodes ) and potentially weighted interactions ( edges ) may be preferred . the chance of conflicting predictions from different tools can be reduced by starting from a consistently annotated dataset that faithfully represents all available evidences . to maintain a coherent development of biological understanding , it is indispensable to keep the reference datasets up - to - date . we examined several primary interaction databases with the aim of constructing non - redundant , consistently annotated and up - to - date reference datasets of physical interactions for several model organisms . unfortunately , the common standard format used by most primary databases still does not allow direct compilation of full non - redundant interactomes . this mainly results from the fact that different primary databases may use different identifiers for interacting proteins and different conventions for representing and annotating each interaction . combining interaction data from bind @xcite , biogrid @xcite , corum @xcite , dip @xcite , hprd @xcite , intact @xcite , mint @xcite , mpact @xcite , mppi @xcite and ophid @xcite , the irefindex @xcite database represents a significant advance towards a complete and consistent set of all publicly available protein interactions . apart from being comprehensive and relatively up - to - date , the main contribution of irefindex is in addressing the problem of protein identifiers by mapping the sequence of every interactant into a unique identifier that can be used to compare interactants from different source databases . in a further ` canonicalization ' procedure @xcite , different isoforms of the same protein are mapped to the same canonical identifier . by adhering to the psi - mi vocabulary and file format , irefindex provides largely standardized annotations for interactants and interactions . construction of irefindex led to the development of irefweb , a web interface for interactive access to irefindex data @xcite . irefweb allows an easy visualization of evidence for interactions associated with user - selected proteins or publications . recently , the authors of irefindex and irefweb published a detailed analysis of agreement between curated interactions within irefindex that are shared between major databases @xcite . however , aiming to maintain all information from original sources , irefindex consequently , there will be features one desires to have that may not fit well within the scope of irefindex . for example , one may wish to treat interactions arising from enzymatic reactions as directed and to be able to selectively include / exclude certain types of reactions such as acetylation . in many cases , the information about post - translational modifications is available directly from source databases , but is not integrated into irefindex . another issue that propagates into irefindex from source databases has to do with protein complexes . some databases represent experimentally observed complexes as interactions with more than two participants , while others expand them into binary interactions using spoke or matrix model @xcite . @xcite recently observed that this different representation of complexes is responsible for a significant number of disagreements between major databases curating the same publication . from our earlier work @xcite , we found that such expanded complexes may lead to nodes with very high degree and often introduce undesirable shortcuts in networks . to fairly treat the information provided by protein complexes without exaggeration , it is preferable to replace the expanded interactions , either from spoke or matrix models , with a flat list of complex members . additionally , we discovered that the mapping of each protein to a canonical group by irefindex would sometimes place protein sequences clearly originating from the same gene ( for example differing in one or two amino acids ) into different canonical groups . to achieve the goal of constructing non - redundant , consistently annotated and up - to - date reference datasets , we developed a script , called ppitrim , that processes irefindex and produces a consolidated dataset of physical protein - protein interactions within a single organism . our script , called ppitrim , is written in the python programming language . it takes as input a dataset in irefindex psi - mi tab 2.6 format , with 54 tab - delimited columns ( 36 standard and 18 added by irefindex ) . after three major processing steps , it outputs a consolidated dataset , in psi - mi tab 2.6 format , containing only the 36 standard columns . the three processing steps are : ( i ) mapping all interactants to ncbi gene ids and removing all undesired raw interactions ; ( ii ) deflating potentially expanded complexes ; and ( iii ) at each step , ppitrim downloads the files it requires from the public repositories and writes its intermediate results as temporary files . the former is removed because it contains either computationally predicted interactions or interactions verified from the literature using text mining ( i.e. without human curation ) . as a first step , the script seeks to map each interactant to an ncbi entrez gene @xcite identifier . for most interactants , it uses the mapping already provided by irefindex . in the cases where irefindex provides only a uniprot @xcite knowledge base accession , the script attempts to obtain a gene i d is used to query the ncbi entrez gene database for a matching gene record using an eutils interface . if a single unambiguous match is found , the record s gene i d is used for the interactant . every mapped gene i d is checked against the list of obsolete gene ids , which are no longer considered to have a protein product existing _ in vivo_. the interactants that can not be mapped to valid ( non - obsolete ) gene ids are removed along with all raw interactions they participate in . after assigning gene ids , the script considers the psi - mi ontology terms associated with interaction detection method , interaction type and interactants biological roles . using the full psi - mi ontology file in open biomedical ontology ( obo ) format @xcite , it replaces any non - standard terms in these fields ( labeled mi:0000 ) with the corresponding valid psi - mi ontology terms . the terms marked as obsolete in the psi - mi obo file are exchanged for their recommended replacements . the single exception are the interaction detection method terms for hprd ` in vitro ' ( mi:0492 , translated from mi:0045 label in irefindex ) and ` in vivo ' ( mi:0493 ) interactions , which are kept throughout the entire processing . source interactions annotated with a descendant of the term mi:0415 ( enzymatic study ) as their detection method or with a descendant of the term mi:0414 ( enzymatic reaction ) as their interaction type are classified as candidate biochemical reactions . this category also includes any interactions ( including those with more than two interactants ) where one of interactants has a biological role of mi:0501 ( enzyme ) or mi:0502 ( enzyme target ) . in the recent months , the biogrid database has started to provide additional information about the post - translational modifications associated with the ` biochemical activity ' interactions , such as phosphorylation , ubiquitination etc . this information is available from the biogrid datasets in the new tab2 format but is not yet reflected in the psi - mi terms for interaction type provided in the psi - mi 2.5 format or in irefindex . since the post - translational modifications annotated by the biogrid can be directly matched to standard psi - mi terms , the script downloads the most recent biogrid dataset in tab2 format , extracts this information and assigns appropriate psi - mi terms for interaction type to the candidate biochemical reactions from irefindex that originate from the biogrid . any source interaction not classified as candidate biochemical reaction is considered for assignment to the candidate complex categories . this category includes all true complexes ( having edge type ` c ' in irefindex ) , interactions having a descendant of mi:0004 ( affinity chromatography ) as the detection method term or mi:0403 ( colocalization ) as the interaction type , as well as the interactions corresponding to the biogrid s ` co - purification ' category . interactions with interaction type mi:0407 ( direct interaction ) are never considered candidates for complexes . all source interactions not falling into candidate biochemical reaction or candidate complex categories are considered ordinary binary physical interactions . the phase ii script attempts to detect spoke - expanded complexes from ` candidate complex ' interactions and deflate them into interactions with multiple interactants . first , all candidate interactions are grouped according to their publication ( pubmed i d ) , source database , detection method and interaction type . each group of source interactions is turned into a graph and considered separately for consolidation into one or more complexes . when a portion of a group of interactions is deflated , we replace these source interactions by a complex containing all their participants . two procedures are used for consolidation : pattern detection and template matching ( fig . [ fig : complexes ] ) . pattern detection procedure is used only for the interactions from the biogrid . unlike the interactions from the dip , those interactions are inherently directed since one protein is always labeled as bait and other as prey ( in many cases this labeling is unrelated to the actual experimental roles of the proteins ) . the pattern indicating a possible spoke - expanded complex consists of a single bait being linked to many preys . since all interactions in the biogrid s co - purification and co - fractionation categories arise from complexes that are spoke - expanded using an arbitrary protein as a bait ( biogrid administration team , private communication ) , a bait linked to two or more preys can in that case always be considered an expanded complex and deflated . such deflated complexes are assigned the edge type code ` g ' . the remainder of the complex candidate interactions from the biogrid were obtained by affinity chromatography and are , in most cases , also derived from complexes . here we adopted a heuristic that a bait linked to at least three preys can be considered a complex . clearly , some experiments involve a single bait being used with many independent preys , in which case this procedure would generate a false complex . therefore , complexes generated in this way are assigned a different edge type code ( ` a ' ) and the user is able to specify specific publications to be excluded from consideration as well as the maximal size of the complex . the second procedure is based on matching each group of candidate interactions to the complexes indicated by other databases ( templates ) , mostly from intact , mint , dip and bind . in this case , the script checks for each protein in the group whether it , together with all its neighbors , is a superset of a template complex . if so , all the candidate interactions between the proteins within the complex are deflated . the neighborhood graph is undirected for all source databases except the biogrid . the new complexes generated in this way are given the code ` r ' . the scripts also attempts to use complexes generated from the biogrid s interactions through a pattern detection procedure as templates , in which case the newly generated complexes have the code ` n ' . any source interactions that can not be deflated into complexes are retained for phase iii . the dag structure of an ontology naturally induces a partial order between the terms : for two terms @xmath0 and @xmath1 , we say that @xmath0 refines @xmath1 ( @xmath0 is smaller @xmath1 , @xmath0 precedes @xmath1 ) if there exists a directed path in the dag from @xmath0 to @xmath1 . two psi - mi terms can be considered compatible if they are comparable , that is , one refines the other . every nonempty collection of terms @xmath2 can be uniquely split into disjoint sets @xmath3 , such that every @xmath3 has a single maximal element ( an element comparable to and not smaller than any other member ) and contains all members of @xmath2 comparable to its maximal element . every subcollection @xmath3 is then consistent because there exists at least one term within it that can describe all its members , while any two members from different subcollections are incomparable . finest consistent term _ of a subcollection @xmath3 is the smallest member of @xmath3 that is comparable to all its members ( it can also be defined as the smallest member of the intersection of the transitive closures of all the members of @xmath3 . ) . if @xmath3 is a total order , where all members are pairwise comparable , the finest consistent term is the minimal term . on the other hand , the minimal term need not exist ( fig . [ fig : example1 ] ) , so that the finest consistent term is higher in the hierarchy and represents the most specific annotation that can be assigned to @xmath3 as a whole . to produce consolidated interactions from a single cluster , each of its members ( interactions ) is identified with its psi - mi term for information detection method . for every cluster member , the set of all other members with compatible annotations ( ` compatible set ' ) is computed . as a special case , the following detection method tags are treated as smaller than any other : ` unspecified method ' ( mi:0686 ) , ` in vivo ' and ` in vitro ' ( the latter two are from hprd only ) . in this way , non - specific annotations are considered as compatible with all other , more specific evidences . compatible sets are further grouped according to their maximal elements . within each group , the union of the compatible sets produces a subcluster . the finest consistent term for each subcluster is found by considering all psi - mi terms on the paths from the subcluster members to its maximum the search is not restricted to those terms that are within the subcluster ( fig . [ fig : example1 ] ) . this definition takes into account that a source database may report an interaction several times for the same publication , using the same or different interaction detection method . if two databases annotate the same interaction using incompatible terms , this is most likely due to an error or specific disagreement about the appropriate label , rather than that each database is reporting a different experiment from the same publication . to test ppitrim , we applied it to the yeast ( _ s . cerevisiae _ ) , human ( _ h . sapiens _ ) and fruitfly ( _ d . melanogaster _ ) datasets from irefindex release 8.0-beta , dated jan 19th 2011 . when processing the yeast dataset , we accounted for two special cases . first , we specifically removed the genetic interactions reported by @xcite because they were not labeled as genetic for all source databases . second , we excluded the dataset by @xcite from phase ii and retained all its interactions as binary undirected . the results of applying ppitrim to process irefindex 8.0 are shown in tables [ tbl : mapping2 ] [ tbl : consolidate ] . we chose to standardize proteins using ncbi gene identifiers rather than the irefindex - provided canonical ids ( crogids ) for several reasons . ncbi gene records not only associate each gene with a set of reference sequences , but also include a wealth of additional data ( e.g. list of synonyms ) and links to other databases such as gene ontology @xcite that are important when using the interaction dataset in practice . in addition , gene records are regularly updated and their status evaluated based on new evidence . thus , a gene record may be split into several new records or marked as obsolete if it corresponds to an orf that is known not to produce a protein . for network analysis applications , it is desirable that only the proteins actually expressed in the cell are represented in the network and hence the gene status provided by ncbi gene is a valuable filtering criterion . however , crogids do have one advantage over ncbi gene ids in that they are protein - based and hence identical protein products of several genes ( like histones ) are clustered together . around 10% of crogids could not be mapped to gene ids even after processing with ppitrim algorithms . a few interactors ( supplementary table 5 ) have only pdb accessions as their primary ids since their interactions were derived from crystal structures . in such cases , often only partial sequences of participating proteins are available . these partial sequences can not be fully matched to any uniprot or refseq record and hence are assigned a separate i d . hence , an improvement for our procedure , that would account for this case as well as for those unmapped proteins that differ from canonical sequences only by few amino acids , would be to use direct sequence comparison to find the closest valid reference sequence . this task may not be technically difficult ( a similar procedure was applied by @xcite to construct protein databases for mass spectrometry data analysis ) but is beyond the scope of ppitrim , which is intended as a relatively short standalone script . in our opinion , such additional mappings would best be performed at the level of reference sequence databases such as uniprot or refseq , which contain curator expertise to resolve ambiguous cases . protein complexes obtained through chromatography techniques provide information complementary to direct binary interactions . while it is often difficult to determine the exact layout of within - complex pairwise interactions , an identification of an association of several proteins using mass spectroscopy is an evidence for _ in vivo _ existence of that association . unfortunately , in spite of its great importance , the currently available information within irefindex is deficient because of different treatments of complexes by different source databases . our results ( table [ tbl : collapsing ] ) show that the apparently inflated complexity of interaction datasets can be substantially reduced by attempting to collapse spoke - expanded complexes . we feel that the benefits from reduction of interactome complexity outweigh the disadvantages from potentially over deflating interactions . the best way to solve would be at the level of source databases ( biogrid in particular ) , by reexamining the original publications . our complexes from the ` r ' category , where deflated complexes fully agree with an annotated complex from a different database , could serve as a guide in this case . overall , our processing significantly reduced the number of interactions within each of the three datasets considered ( table [ tbl : consolidate ] ) . this indicates a significant redundancy , particularly for protein complexes , original and deflated ( compare table [ tbl : collapsing ] with table [ tbl : consolidate ] ) , and for binary interactions . the directed interactions ( biochemical reactions ) are relatively rarer and largely non - redundant at this stage . given their importance in elucidating biological function , the directed interactions are expected to be discovered more fully with time . upon closer examination ( table [ tbl : conflicts ] ) , it can be seen that most common conflicts arise as instances of few specific labeling disagreements between databases . in many cases , such disagreements arise from using different sub - terms of affinity chromatography ( see fig . [ fig : example1 ] ) and can be resolved by assigning a more general term consistent with both conflicting terms . in many other cases , the conflicts are due to biogrid internally using a more restricted detection method vocabulary than the imex databases ( dip , intact and mint ) . the ppitrim algorithms work best if accurate and fully populated fields for interaction detection method , publication and interaction type are available in its input dataset . this requirement is mostly fulfilled . nevertheless , we have noticed two minor inconsistencies . the first , which will be fixed in a subsequent release of irefindex ( ian donaldson , private communication ) , involves the psi - mi labels for interaction detection method for corum interactions and complexes . these are missing from irefindex although they are present in the original corum source files . the second issue concerns missing or invalid pubmed ids for certain interactions . we found that a number of interactions with missing pubmed ids come from mint . upon inspection of the original mint files , we discovered that in many cases mint supplies a digital object identifier ( doi ) for a publication as its identifier instead of a pubmed i d ( although the corresponding pubmed i d can be obtained from the mint web interface ) . to ensure consistency with other source databases within irefindex , it would be desirable to have the pubmed ids available for these interactions as well . in this paper , we have identified the tasks needed for using combined interaction datasets provided by irefindex as a basis for construction of reference networks and developed a script to process them into consistent consolidated datasets . we see ppitrim as answering a need for a consolidated database and hope that most of the issues that required processing will be eventually fixed in upstream databases and distributed through imex consortium . at this stage we have not addressed the issue of quality of interactions although such information is available in some databases for some publications @xcite . utilizing the quality information in consolidating datasets demands a universal data - quality measure that is not yet existent . 37 [ 1]#1 [ 1]`#1 ` urlstyle [ 1]doi : # 1 de las rivas , j. and fontanillo , c. protein - protein interactions essentials : key concepts to building and analyzing interactome networks . _ plos comput biol _ , 60 ( 6):0 e1000807 , 2010 . stark , c. , breitkreutz , b .- j . , chatr - aryamontri , a. _ et al . _ the biogrid interaction database : 2011 update . _ nucleic acids res _ , 390 ( database issue):0 d698704 , 2011 . aranda , b. , achuthan , p. , alam - faruque , y. _ et al . _ the intact molecular interaction database in 2010 . _ nucleic acids res _ , 380 ( database issue):0 d52531 , 2010 . ceol , a. , chatr - aryamontri , a. , licata , l. _ et al . _ , the molecular interaction database : 2009 update . _ nucleic acids res _ , 380 ( database issue):0 d5329 , 2010 . salwinski , l. , miller , c. s. , smith , a. j. _ et al . _ the database of interacting proteins : 2004 update . _ nucleic acids res _ , 320 ( database issue):0 d44951 , 2004 . alfarano , c. , andrade , c. e. , anthony , k. _ et al . _ the biomolecular interaction network database and related tools 2005 update . _ nucleic acids res _ , 330 ( database issue):0 d41824 , 2005 . isserlin , r. , el - badrawi , r. a. , and bader , g. d. the biomolecular interaction network database in psi - mi 2.5 . _ database ( oxford ) _ , 2011:0 baq037 , 2011 . keshava prasad , t. s. , goel , r. , kandasamy , k. _ et al . _ 2009 update . _ nucleic acids res _ , 370 ( database issue):0 d76772 , 2009 . cusick , m. e. , yu , h. , smolyar , a. _ et al . _ literature - curated protein interaction datasets . _ nat methods _ , 60 ( 1):0 3946 , 2009 . orchard , s. , kerrien , s. , jones , p. _ et al . _ submit your interaction data the imex way : a step by step guide to trouble - free deposition . _ proteomics _ , 7 suppl 1:0 2834 , 2007 . kerrien , s. , orchard , s. , montecchi - palazzi , l. _ et al . _ broadening the horizon level 2.5 of the hupo - psi format for molecular interactions . _ bmc biol _ , 5:0 44 , 2007 . markowetz , f. and spang , r. inferring cellular networks a review . _ bmc bioinformatics _ , 8 suppl 6:0 s5 , 2007 . gomez , s. m. , choi , k. , and wu , y. prediction of protein - protein interaction networks . _ curr protoc bioinformatics _ , chapter 8:0 unit 8.2 , 2008 . kanaan , s. p. , huang , c. , wuchty , s. _ et al . _ inferring protein - protein interactions from multiple protein domain combinations . _ methods mol biol _ , 541:0 4359 , 2009 . lewis , a. c. f. , saeed , r. , and deane , c. m. predicting protein - protein interactions in the context of protein evolution . _ mol biosyst _ , 60 ( 1):0 5564 , 2010 . chautard , e. , thierry - mieg , n. , and ricard - blum , s. interaction networks : from protein functions to drug discovery . a review . _ pathol biol ( paris ) _ , 570 ( 4):0 32433 , 2009 . przytycka , t. m. , singh , m. , and slonim , d. k. toward the dynamic interactome : it s about time . _ brief bioinform _ , 110 ( 1):0 1529 , 2010 . ruepp , a. , waegele , b. , lechner , m. _ et al . _ : the comprehensive resource of mammalian protein complexes2009 . _ nucleic acids res _ , 380 ( database issue):0 d497501 , 2010 . gldener , u. , mnsterktter , m. , oesterheld , m. _ et al . _ : the mips protein interaction resource on yeast . _ nucleic acids res _ , 340 ( database issue):0 d43641 , 2006 . pagel , p. , kovac , s. , oesterheld , m. _ et al . _ the mips mammalian protein - protein interaction database . _ bioinformatics _ , 210 ( 6):0 8324 , 2005 . brown , k. r. and jurisica , i. online predicted human interaction database . _ bioinformatics _ , 210 ( 9):0 207682 , 2005 . razick , s. , magklaras , g. , and donaldson , i. m. : a consolidated protein interaction database with provenance . _ bmc bioinformatics _ , 9:0 405 , 2008 . turner , b. , razick , s. , turinsky , a. l. _ et al . _ : interactive analysis of consolidated protein interaction data and their supporting evidence . _ database ( oxford ) _ , 2010:0 baq023 , 2010 . turinsky , a. l. , razick , s. , turner , b. _ et al . _ literature curation of protein interactions : measuring agreement across major public databases . _ database ( oxford ) _ , 2010:0 baq026 , 2010 . stojmirovi , a. and yu , y .- k . : analyzing information flow in protein networks . _ bioinformatics _ , 250 ( 18):0 24479 , 2009 . maglott , d. , ostell , j. , pruitt , k. d. _ et al . _ : gene - centered information at ncbi . nucleic acids res _ , 390 ( database issue):0 d527 , 2011 . . the universal protein resource ( uniprot ) in 2010 . _ nucleic acids res _ , 380 ( database issue):0 d1428 , 2010 . smith , b. , ashburner , m. , rosse , c. _ et al . _ the obo foundry : coordinated evolution of ontologies to support biomedical data integration . _ nat biotechnol _ , 250 ( 11):0 12515 , 2007 . tong , a. h. y. , lesage , g. , bader , g. d. _ et al . _ global mapping of the yeast genetic interaction network . _ science _ , 3030 ( 5659):0 80813 , 2004 . collins , s. r. , kemmeren , p. , zhao , x .- c . _ et al . _ toward a comprehensive atlas of the physical interactome of saccharomyces cerevisiae . _ mol cell proteomics _ , 60 ( 3):0 43950 , 2007 . gavin , a .- c . , aloy , p. , grandi , p. _ et al . _ proteome survey reveals modularity of the yeast cell machinery . _ nature _ , 4400 ( 7084):0 6316 , 2006 . krogan , n. j. , cagney , g. , yu , h. _ et al . _ global landscape of protein complexes in the yeast saccharomyces cerevisiae . _ nature _ , 4400 ( 7084):0 63743 , 2006 . ashburner , m. , ball , c. a. , blake , j. a. _ et al . _ gene ontology : tool for the unification of biology . the gene ontology consortium . _ nat genet _ , 25:0 2529 , 2000 . alves , g. , ogurtsov , a. y. , and yu , y .- k . : mass - spectrometry based peptide identification web server with knowledge integration . _ bmc genomics _ , 9:0 505 , 2008 . hannich , j. t. , lewis , a. , kroetz , m. b. _ et al . _ defining the sumo - modified proteome by multiple approaches in saccharomyces cerevisiae . _ j biol chem _ , 2800 ( 6):0 410210 , 2005 . croft , d. , okelly , g. , wu , g. _ et al . _ reactome : a database of reactions , pathways and biological processes . _ nucleic acids res _ , 390 ( database issue):0 d6917 , 2011 . blaiseau , p. l. and thomas , d. multiple transcriptional activation complexes tether the yeast activator met4 to dna . _ embo j _ , 170 ( 21):0 632736 , 1998 . this work was supported by the intramural research program of the national library of medicine at the national institutes of health . we thank dr . donaldson for his critical reading of this manuscript and for providing us with the proprietary version of irefindex 7.0 dataset , which was used for initial development of ppitrim . rlp5.0cm > p5.4 cm column & short name & description & example + 1 & uida & smallest gene i d of the interactor a@xmath4 & entrezgene / locuslink:854647 + 2 & uidb & smallest gene i d of the interactor b@xmath5 & entrezgene / locuslink:855136 + 3 & alta & all gene ids of the interactor a@xmath5 & entrezgene / locuslink:854647 + 4 & altb & all gene ids of the interactor b@xmath5 & entrezgene / locuslink:855136 + 5 & aliasa & all canonical gene symbols and integer crogids of interactor a & entrezgene / locuslink : bnr1\| icrogid:2105284 + 6 & aliasb & all canonical gene symbols and integer crogids of interactor b & entrezgene / locuslink : myo5\| icrogid:3144798 + 7 & method & psi - mi term for interaction detection method & mi:0018(two hybrid ) + 8 & author & first author name(s ) of the publication in which this interaction has been shown@xmath6 & tong ah [ 2002]\|tong-2002a-3 + 9 & pmids & pubmed id(s ) of the publication in which this interaction has been shown & pubmed:11743162 + 10 & taxa & ncbi taxonomy identifier for interactor a & taxid:4932(saccharomyces cerevisiae ) + 11 & taxb & ncbi taxonomy identifier for interactor b & taxid:4932(saccharomyces cerevisiae ) + 12 & interactiontype & psi - mi term for interaction type & mi:0407(direct interaction ) + 13 & sourcedb & psi - mi terms for source databases@xmath6 & mi:0000(mpact)\|mi:0463(grid)\| mi:0465(dip)\|mi:0469(intact ) + 14 & interactionidentifier & a list of interaction identifiers@xmath7 & ppitrim : tyugksok231dh3ynsi6gbczjcfe=\| mpact:8233\|dip : dip-11198e\|grid:147506\| intact : ebi-601565\|intact : ebi-601728\| irigid:288990\|edgetype : x + 15 & confidence & a list of ppitrim confidence scores@xmath8 & maxsources:2\|dmconsistency : full\| conflicts : s3oaixt5ta4vvruso1rc1ta9krk= + 16 & expansion & either ` none ' for binary interactions or ` bipartite ' for subunits of complexes & none + 17 & biologicalrolea & psi - mi term(s ) for the biological role of interactor a@xmath6 & mi:0499(unspecified role ) + 18 & biologicalroleb & psi - mi term(s ) for the biological role of interactor b @xmath6 & mi:0499(unspecified role ) + 19 & experimentalrolea & psi - mi term(s ) for the experimental role of interactor a@xmath6 & mi:0496(bait)\|mi:0498(prey)\| mi:0499(unspecified role ) + 20 & experimentalroleb & psi - mi term(s ) for the experimental role of interactor b@xmath6 & mi:0496(bait)\|mi:0498(prey)\| mi:0499(unspecified role ) + 21 & interactortypea & psi - mi term for the type of interactor a ( either ` protein ' or ` protein complex ' ) & mi:0326(protein ) + 22 & interactortypeb & psi - mi term for the type of interactor b ( always ` protein ' ) & mi:0326(protein ) + 29 & hostorganismtaxid & ncbi taxonomy identifier for the host organism & taxid:4932(saccharomyces cerevisiae ) + 31 & creationdate & date when ppitrim was run & 2011/05/11 + 32 & updatedate & date when ppitrim was run & 2011/05/11 + 35 & checksuminteraction & ppitrim i d for an interaction & ppitrim : tyugksok231dh3ynsi6gbczjcfe= + 36 & negative & always ` false ' & false + .@xmath9 interaction type is also adjusted to mi:0403 as recommended in ` psi-mi.obo ` ; @xmath10 hprd terms are treated as a special case , see main text ; @xmath11 mppi interactions in the human dataset . [ cols="<,<,<,<,<",options="header " , ] 8avruhg76vkifn2czgicnzzr00y= & grid & 14759368 & cft2 , ysh1 , pta1 , mpe1 & part of mrna cleavage / polyadenylation complex ( 4/10 proteins ) . + 9ys57j / gbrbolnmmimsveonoraa= & grid & 14759368 & nut1 , med7 , med4 , sin4 , srb4 & part of mediator complex . + ju+eokq6iplh9djkrtgrluvt7vm= & grid , mint & 14759368 & ubp6 , rpt3 , rpn9 , rpt1 , rpn8 , rpn2 , rpn7 , rpn1 & part of proteasome . mint does not contain complexes from the original paper . + httmhgipyfit2vftrz94uww0rsy= & grid & 16429126 & ioc3 , htb1 , hta2 , hhf2 , isw1 , kap114 , itc1 , rps4a , vps1 , nap1 , rpo31 , isw2 , tbf1 , bro1 , mot1 & part of complex # 99 . + lnnzfypgshcg7zkkynu6+fsk2eu= & grid & 16429126 & psk1 , nth1 , bmh2 , rtg2 , bmh1 & part of complex # 147 ( two core proteins plus three attachments ) . + s2i6vrjfmwc6rkkm+oyxwkcg9yq= & grid & 16429126 & rpl4b , mnn10 , mnn11 , hoc1 , mnn9 , anp1 & core complex ( # 111 mannan polymerase ii ) + one attachment protein ( rpl4b ) . + 1frmaapl2ruoqq202yujg55mafo= & grid , mint & 16554755 & rsm24 , rsm28 , mrps5 , mrp13 , mrps35 , rsm27 , rsm7 , rsm25 , mrps17 , mrps12 , rsm19 , mrp4 & part of complex # 1 . + 5tbkyomk / g1h3vaqmionuobhhmq= & grid , mint & 16554755 & cft2 , ysh1 , mpe1 , pap1 & part of complex # 18 . + 9f2dvj2rdgecp53lhonwrmwq14a= & grid , mint & 16554755 & kap95 , rtt103 , vma2 , rai1 , rat1 , rpb2 , srp1 & true experimental association but not part of any derived complex . + avawv51 + 6fqe3dquygd / xfyrxxe= & grid , mint & 16554755 & rrp42 , rrp45 , rrp6 , csl4 , mpp6 , rrp4 , lrp1 , ddi1 & part of complex # 19 . + nolewovavmsfrqedksut / mldemc= & grid , mint & 16554755 & cdc3 , shs1 , cdc11 , cdc12 & part of complex # 121 . + wa51i87lj1wgp / eef1ov / yvbw1y= & grid , mint & 16554755 & gtt2 , trx1 , crn1 , ssa3 , ipp1 , cmd1 , trx2 , tdh1 , rpl40b , cdc21 , oye2 & true experimental association but not part of any derived complex + yn / hqxqvzob5hqrgpzvth28mgsy= & grid , mint & 16554755 & rrp43 , rrp42 , rrp45 , rrp40 , dis3 , rrp6 , rrp4 , lrp1 & part of complex # 19 . + 1lrk+agi8hpgosagkhdznjwsvti= & grid & 20489023 & rtg3 , rtg2 , tor1 , tor2 , cka2 , myo2 , mks1 , kog1 & true experimental association . + xwzvxejfgqjkcihjmqvf5gzhjjq= & dip , grid , mint & 20489023 & puf3 , sam1 , gcd6 , spt16 , mtc1 , ygk3 , lsm12 & true experimental association . + 15vfqtoe5gxgnwpsy3ag0sq6a2u= & grid & 9891041 & ccr4 , hpr1 , paf1 , srb5 , gal11 & not a true complex . this is because of bad annotation of paf1srb5 interaction by the biogrid . completely opposite interpretation was given in the paper . + d79idtwftaenrh8cq+c8cps389y= & grid & 10329679 & ypt1 , vps21 , ypt7 , gdi1 & true complex . this is the only experiment in the paper . + ets4cgpheptqjb / fs5qxyzf0ke8= & grid & 11733989 & cdc39 , ccr4 , cdc36 , caf130 , caf40 , caf120 , pop2 , not5 , mot2 & true complex . caf120 is an unusual member that could almost be left out . + 2koygdwzwywspn5mhk26gccc6lq= & grid & 14769921 & gbp2 , imd3 , tef1 , kem1 , ctk2 , ctk1 , ctk3 & true complex , except that tef1 should be tef2 . this is an error in the irefindex source file ; the biogrid website has the correct assignment . + kd07bbuf07sqy9np3d0lixss / ty= & grid & 15303280 & bud31 , rpl2b , prp19 , cdc13 , atp1 , rps4a , snu114 , mdh1 , mam33 , mrpl3 , mrpl17 , prp8 , prp22 , pab1 , brr2 & true association + zagz / izqker3/ntdlzpedad9cko= & grid & 16179952 & cdc40 , ufd1 , ssm4 , ubx2 & not a true complex , probably due to a typo in annotation . cdc40 can not be found anywhere in the paper and should most likely be cdc48 . + rdu0dspan0qeadfsu5sv05ifihw= & grid & 16286007 & sin3 , rco1 , rpd3 , ume1 , eaf3 & true complex . + vqbn3ddwtpgye9dzbatfnqzdfe0= & grid & 16615894 & vps36 , vps25 , vps28 , snf8 & vps28 binds the other three , which form a complex . + lmdypan9kahbdaslws19x8k7kke= & grid & 20159987 & ubi4 , ufd2 , pex29 , ssm4 & biological association but indicated as ` not a stable complex ' in the paper . + aakrh6qvahgxgvqhe399+faxpva= & grid & 20655618 & pex13 , pex10 , pex8 , pex12 & association is correct , although mutant strain was used to obtain this particular complex . + > p12cmr consolidated terms & count mi:0018 ( two hybrid ) , mi:0045 ( experimental interaction detection ) , mi:0398 ( two hybrid pooling approach ) , mi:0399 ( two hybrid fragment pooling approach ) & 3959 mi:0090 ( protein complementation assay ) , mi:0111 ( dihydrofolate reductase reconstruction ) & 2612 mi:0090 ( protein complementation assay ) , mi:0112 ( ubiquitin reconstruction ) & 2077 mi:0004 ( affinity chromatography technology ) , mi:0676 ( tandem affinity purification ) & 1840 mi:0004 ( affinity chromatography technology ) , mi:0007 ( anti tag coimmunoprecipitation ) & 1408 mi:0018 ( two hybrid ) , mi:0045 ( experimental interaction detection ) , mi:0397 ( two hybrid array ) & 1231 mi:0018 ( two hybrid ) , mi:0045 ( experimental interaction detection ) & 954 mi:0018 ( two hybrid ) , mi:0397 ( two hybrid array ) & 914 mi:0045 ( experimental interaction detection ) , mi:0686 ( unspecified method ) & 628 mi:0004 ( affinity chromatography technology ) , mi:0019 ( coimmunoprecipitation ) & 598 mi:0018 ( two hybrid ) , mi:0398 ( two hybrid pooling approach ) & 506 mi:0004 ( affinity chromatography technology ) , mi:0007 ( anti tag coimmunoprecipitation ) , mi:0676 ( tandem affinity purification ) & 444 mi:0018 ( two hybrid ) , mi:0045 ( experimental interaction detection ) , mi:0686 ( unspecified method ) & 320 mi:0004 ( affinity chromatography technology ) , mi:0096 ( pull down ) & 217 mi:0415 ( enzymatic study ) , mi:0424 ( protein kinase assay ) & 192 mi:0045 ( experimental interaction detection ) , mi:0081 ( peptide array ) & 150 mi:0045 ( experimental interaction detection ) , mi:0676 ( tandem affinity purification ) & 120 mi:0492 ( in vitro ) , mi:0493 ( in vivo ) & 5739 mi:0018 ( two hybrid ) , mi:0398 ( two hybrid pooling approach ) & 5394 mi:0018 ( two hybrid ) , mi:0492 ( in vitro ) , mi:0493 ( in vivo ) & 2796 mi:0096 ( pull down ) , mi:0492 ( in vitro ) , mi:0493 ( in vivo ) & 2760 mi:0096 ( pull down ) , mi:0492 ( in vitro ) & 2134 mi:0018 ( two hybrid ) , mi:0492 ( in vitro ) & 1658 mi:0018 ( two hybrid ) , mi:0493 ( in vivo ) & 1193 mi:0018 ( two hybrid ) , mi:0397 ( two hybrid array ) & 1045 mi:0096 ( pull down ) , mi:0493 ( in vivo ) & 513 mi:0004 ( affinity chromatography technology ) , mi:0006 ( anti bait coimmunoprecipitation ) & 384 mi:0004 ( affinity chromatography technology ) , mi:0019 ( coimmunoprecipitation ) & 309 mi:0004 ( affinity chromatography technology ) , mi:0007 ( anti tag coimmunoprecipitation ) & 195 mi:0114 ( x - ray crystallography ) , mi:0492 ( in vitro ) & 166 mi:0004 ( affinity chromatography technology ) , mi:0096 ( pull down ) & 161 mi:0047 ( far western blotting ) , mi:0492 ( in vitro ) , mi:0493 ( in vivo ) & 106 mi:0018 ( two hybrid ) , mi:0398 ( two hybrid pooling approach ) & 17738 mi:0018 ( two hybrid ) , mi:0399 ( two hybrid fragment pooling approach ) & 1426	+ * * www.ncbi.nlm.nih.gov/cbbresearch/yu/downloads/ppitrim.html
the stability analysis of stochastic switched linear systems has attracted a significant amount of attention in the last two decades . in particular , their mean stability , which requires that the power of the norm of the state variable converges to zero in expectation . some early results on the mean stability of switched linear systems with an independent and identically distributed ( i.i.d . ) switching signal can be found in @xcite . the stability characterizations of switched linear systems driven by an extension of homogeneous markov processes called semi - markov processes @xcite are available in @xcite . it is known that homogeneous markov processes having certain irreducibility and recurrence properties and discrete - time i.i.d . stochastic processes are special cases of a more general class of stochastic processes called _ regenerative processes _ @xcite . firstly introduced by smith @xcite , regenerative processes have found applications especially in queuing systems @xcite and network reliability analysis @xcite . as we will see later in example [ eg : process ] , regenerative processes are also suitable to describe a controlled system under periodic maintenance @xcite . despite the above facts , as far as we are aware of , no effort has been made to investigate switched linear systems with a regenerative switching signal . the aim of this paper is to give the characterization of the mean stability of a switched linear system with a regenerative switching signal , which we call a _ regenerative switched linear system_. we show that , if the exponent of the mean stability is even or the system is positive @xcite , then the mean stability of the system is characterized by the spectral radius of a matrix . the matrix is obtained as the expected value of the lift @xcite of the transition matrix of the system . the proof makes use of a stability - preserving discretization of the system at the embedded renewal process of the underlying regenerative process . the characterization in particular generalizes well - known floquet s theorem for the stability analysis of linear time - periodic systems @xcite . this paper is organized as follows . after preparing necessary notations and conventions , in section [ sec : regsystems ] we recall the definition of regenerative processes and then introduce regenerative switched linear systems . then section [ sec : mainresult ] presents the main result of this paper , which is followed by an example . the proof of the main result is given in section [ sec : proof ] . then section [ sec : disc ] discusses the discrete - time case . let @xmath0 be a probability space . for an integrable random variable @xmath1 on @xmath2 its expected value is denoted by @xmath3 $ ] . the random variables that appear in this paper will be assumed to be integrable . when @xmath4 is nonnegative entrywise we write @xmath5 . the standard euclidean norm on @xmath6 is denoted by @xmath7 . for @xmath8 , the @xmath9-norm on @xmath6 is defined by @xmath10 . the symbol @xmath11 denotes the column vector of length @xmath12 whose entries are all @xmath13 . it is easy to see @xmath14 if @xmath5 . let @xmath15 and @xmath16 denote the identity and the zero matrices , respectively . we say that @xmath17 is schur stable if its spectral radius @xmath18 is less than one . the @xmath9-lift of @xmath19 , denoted by @xmath20}$ ] , is defined @xcite as the real vector of length @xmath21 with its elements being the lexicographically ordered monomials @xmath22 indexed by all the possible exponents @xmath23 such that , where @xmath24 . it holds @xcite that @xmath25 } { \lvert x^{[m ] } \rvert } = { \lvert x \rvert}^m.\ ] ] we then define @xmath26 } \in \mathbb{r}^{n_m\times n_m}$ ] as the unique matrix @xcite satisfying @xmath27 } = a^{[m ] } x^{[m]}$ ] for every @xmath28 . for any matrix @xmath29 it holds that @xmath30 } ( ab)^{[m ] } = a^{[m ] } b^{[m]}\ ] ] provided the product @xmath31 is well defined . we also define @xmath32 } \in \mathbb{r}^{n_m \times n_m}$ ] as the unique real matrix @xcite such that , for every @xmath6-valued differentiable function @xmath33 on @xmath34 satisfying @xmath35 , it holds that @xmath36}}/{dt } = a_{[m ] } x^{[m]}$ ] . it is easy to check that @xmath37 } \bigl(e^{at}\bigr)^{[m ] } = e^{a_{[m]}t}\ ] ] for every @xmath38 . let us first recall the definition of regenerative stochastic processes @xcite . throughout this paper we fix an underlying probability space @xmath0 . a stochastic process @xmath39 is called a _ regenerative process _ if there exists a random variable @xmath40 , called a regeneration epoch , such that the following statements hold . * @xmath41 is independent of @xmath42 ; * @xmath41 is stochastically equivalent to @xmath43 . in the following we quote some consequences of the above definition from @xcite . by repeatedly applying the definition , one can obtain a sequence of independent and identically distributed random variables @xmath44 called _ cycle lengths _ , which can be used to break @xmath45 into independent and identically distributed _ cycles _ @xmath46 , @xmath47 , @xmath48 . then the stochastic process @xmath49 defined by @xmath50 is called the _ embedded renewal process _ of @xmath45 . throughout this paper , for the sake of convenience , we set @xmath51 and call @xmath52 as the embedded renewal process of @xmath45 . the next example presents a regenerative process that is not a homogeneous markov process is of a systems and control theoretical interest . [ eg : process ] consider a dynamical system with a failure - prone controller @xcite . let us model the controlled system as a switched system with the two modes @xmath53 . instead of assuming that the transition of the mode can be described by a homogeneous markov process ( see , e.g. , @xcite ) , let us consider the scenario when the controlled system is under periodic maintenance , which is commonly employed in the literature from reliability theory @xcite . let the stochastic process @xmath52 represent the times at which a maintenance is performed . for simplicity we assume that @xmath54 with probability one for every @xmath55 , i.e. , that every maintenance repairs a failure with probability one within a negligible time period . we set @xmath51 . furthermore we assume that @xmath56 equals @xmath57 , where @xmath58 is a constant and @xmath59 are independent and identically distributed random variables . @xmath60 represents the designed period of the maintenance and @xmath61 models its random perturbation . we assume that the length of the time for which the process @xmath45 stays at mode @xmath13 after the reset at @xmath62 follows an exponential distribution with parameter @xmath63 . in other words we are assuming that , on any interval of a sufficiently small length @xmath64 , the probability of the occurrence of a failure is approximately equal to @xmath65 . then @xmath45 is clearly a regenerative process with a regeneration epoch @xmath66 and the embedded renewal process @xmath52 . notice that @xmath45 is not a markov process because the length of the time while mode @xmath67 is active depends on . then we introduce the class of switched linear systems studied in this paper . let @xmath43 be a regenerative process that can take values in a set @xmath68 . let @xmath69 be a family of real @xmath70 matrices indexed by @xmath68 . then we call the stochastic differential equation @xmath71 as a _ regenerative switched linear system_. we assume that @xmath72 is a constant vector . the stability of @xmath73 is defined in the following standard manner . let @xmath9 be a positive integer . * @xmath73 is said to be _ exponentially @xmath9th mean stable _ if there exist @xmath74 and @xmath75 such that @xmath76 \leq \alpha e^{-\beta \mathchangescript{k}{t}}{\lvert x_0 \rvert}^m$ ] for all @xmath77 . * @xmath73 is said to be _ stochastically @xmath9th mean stable _ if @xmath78\,dt < \infty$ ] for any @xmath77 . we also introduce the notion of positivity for @xmath73 following @xcite . we say that @xmath73 is _ positive _ if @xmath79 implies @xmath80 with probability one for every @xmath81 . for @xmath73 to be positive it is clearly sufficient that all the matrices @xmath82 are metzler , i.e. , the off - diagonal entries of each @xmath82 are all nonnegative @xcite . however it is not necessary , as illustrated in the following non - trivial example . consider a switched linear system with @xmath83 and @xmath84 since @xmath85 , a simple calculation shows the existence of @xmath58 such that if @xmath86 then , for every @xmath87 , the vector @xmath88 is in the sector @xmath89 . then we construct a regenerative process @xmath45 as follows . set @xmath90 and let @xmath64 follow the uniform distribution on @xmath91 $ ] . define @xmath45 on the first cycle @xmath92 by @xmath93 and extend this definition to the whole interval @xmath94 regeneratively . then , since the second mode decreases the argument of the state vector at most @xmath95 , we can see that @xmath96 stays in the positive orthant for every @xmath38 whenever @xmath97 . therefore @xmath73 is positive although @xmath98 is not metzler . this section states the characterization of the mean stability of regenerative switched linear systems and also presents an example to illustrate the result . we state the next assumption . 1 . [ item : evenornon - neg ] either @xmath9 is even or @xmath73 is positive . item : hk < t ] . [ item : mbdd ] the set @xmath69 is bounded . [ item : evenornon - neg ] covers mean square stability ( @xmath99 ) , which has been the central stability notion of stochastic switched linear systems in the literature @xcite . the second condition [ item : hk < t ] on the boundedness of cycle lengths is crucial . similar assumptions were employed for the stability analysis of semi - markov jump linear systems @xcite and stochastic hybrid systems with renewal transitions @xcite . [ item : mbdd ] is only to ensure that the state variable does not diverge in a finite time and thus is not restrictive . in order to state the main result we need fundamental matrices @xcite of the system @xmath73 . for all @xmath100 , @xmath101 , and @xmath102 let us define @xmath103 by the differential equation @xmath104 then define the @xmath105-valued random variables @xmath106 by @xmath107 which expresses the transition of @xmath33 from @xmath62 to @xmath108 . the next theorem is the main result of this paper . [ theorem : main ] the following statements are equivalent . 1 . [ item : expsta ] @xmath73 is exponentially @xmath9th mean stable . [ item : stosta ] @xmath73 is stochastically @xmath9th mean stable . [ item : schsta ] @xmath109}]$ ] is schur stable . based on the theorem and continuing from example [ eg : process ] , the next example presents the stability analysis of a linear time - invariant system with a failure - prone controller under periodic maintenance . consider the internally unstable linear time - invariant system @xmath110 with the failure - prone controller @xmath111 where @xmath112 the stabilizing feedback gain @xmath113 is obtained by solving a linear quadratic regulator problem . we assume that the transition between the modes @xmath114 follows the regenerative process @xmath45 described in example [ eg : process ] . with this labeling we have @xmath115 and @xmath116 . for simplicity and also suppose that each @xmath117 follows uniform distribution on @xmath118 $ ] independently . let the random variable @xmath64 denote the first time in when the transition to mode @xmath67 occurs . then one can see @xmath119 . if we let @xmath120}$ ] ( @xmath121 ) then equations and show @xmath122 } = e^{\bar a_2 \max(0 , r_1-h ) } e^{\bar a_1 \min(r_1,h)}.\ ] ] here we recall that for square matrices @xmath123 with the same dimensions and @xmath38 it holds that @xcite @xmath124 using this identity and the independence of @xmath64 and @xmath66 , since @xmath64 follows the exponential distribution with mean @xmath13 , we can show that @xmath125 } ] & = \int_{0.9 t}^{1.1 t } \int_0^\infty e^{\bar a_2\max(0 , t - s ) } e^{\bar a_1\min(t , s ) } e^{-s}ds\,\frac{dt}{0.2 t } \\ & = \frac{5}{t}\int_{0.9 t}^{1.1 t } \int_0^t e^{\bar a_2(t - s ) } e^{(\bar a_1 - i)s } \,ds\,{dt } + \frac{5}{t}\int_{0.9 t}^{1.1 t } \int_t^\infty e^{\bar a_1 t } e^{-s } \,ds\,{dt } \\ & = \frac{5}{t } \begin{bmatrix } i & o \end{bmatrix } \int_{0.9 t}^{1.1 t } \exp\left(\begin{bmatrix } \bar a_2&i\\o&\bar a_1 - i \end{bmatrix}t\right ) \,dt \begin{bmatrix } o \\ i \end{bmatrix } \\ & \hspace{2cm}+ \frac{5}{t}\int_{0.9 t}^{1.1 t } e^{(\bar a_1 - i)t}\,{dt}. \end{aligned}\ ] ] figure [ figure : stability ] shows the graph of the spectral radius of @xmath126}]$ ] as @xmath60 varies . } ] $ ] as @xmath60 varies , width=306 ] as is expected , instability is caused by making the period of the maintenance longer . we can see that , by theorem [ theorem : main ] , @xmath73 is mean square stable if and only if @xmath127 . the computation of the matrix @xmath126}]$ ] is performed with matlab . figure [ figure : samplepaths ] shows 20 sample paths of @xmath128 when @xmath129 . , width=306 ] theorem [ theorem : main ] extends celebrated floquet s theory @xcite for the stability analysis of linear time - periodic systems . in fact , if the process @xmath45 is a periodic function with the period @xmath60 then we have @xmath130 } ] ) = \rho(m_0^{[m ] } ) = \rho(m_0)^m$ ] . therefore , by theorem [ theorem : main ] , the linear time - periodic system @xmath73 is stable in the standard sense if and only if @xmath131 , which is the main consequence of floquet s theory . the proof of theorem [ theorem : main ] is based on the discretization of @xmath73 at the embedded renewal process of the underlying regenerative process . in order to analyze the stability of the discretization , in the next section we first present the stability analysis of discrete - time switched linear systems with i.i.d . parameters . then in section [ subsec : proof ] we give the proof of theorem [ theorem : main ] . let @xmath132 be independent and identically distributed random variables following a distribution @xmath133 on @xmath105 . consider the discrete - time switched linear system @xmath134 where @xmath135 is a constant vector . the mean stability of @xmath136 is introduced as follows . [ defn : sta : sigmamu ] let @xmath9 be a positive integer . * @xmath136 is said to be _ exponentially @xmath9th mean stable _ if there exist @xmath74 and @xmath75 such that @xmath137 \leq \alpha e^{-\beta k } { \lvert x_0 \rvert}^m\ ] ] for all @xmath77 . * @xmath136 is said to be _ stochastically @xmath9th mean stable _ if @xmath138 < \infty$ ] for any @xmath77 . also we define the positivity of @xmath136 in the same way as for @xmath73 . we say that @xmath136 is positive if @xmath97 implies @xmath139 with probability one for every @xmath140 . when we check the exponential mean stability of a positive @xmath136 , we can without loss of generality assume that its initial state is nonnegative . [ lem : lemma36result ] assume that @xmath136 is positive . then @xmath136 is exponentially @xmath9th mean stable if and only if there exist @xmath74 and @xmath75 such that holds for all @xmath87 . see the proof of ( * ? ? ? * lemma 3.6 ) . the next proposition characterizes the stability of @xmath136 in terms of the spectral radius of a matrix . [ prop : stab : mu : char ] assume that either 1 . [ item : mu : even ] @xmath9 is even or 2 . [ item : mu : invariant ] @xmath136 is positive . then the following conditions are equivalent . 1 . [ item : mu : expsta ] @xmath136 is exponentially @xmath9th mean stable . [ item : mu : stosta ] @xmath136 is stochastically @xmath9th mean stable . [ item : mu : shcsta ] @xmath141}]$ ] is schur stable . we shall show the cycle [ [ item : mu : expsta ] @xmath142 [ item : mu : stosta ] @xmath142 [ item : mu : shcsta ] @xmath142 [ item : mu : expsta ] ] . one can easily see [ [ item : mu : expsta ] @xmath142 [ item : mu : stosta ] ] . [ [ item : mu : stosta ] @xmath142 [ item : mu : shcsta ] ] : let @xmath143 denote the trajectory of @xmath136 with the initial state @xmath77 . since the identity shows @xmath144 } ] = e[f_0^{[m]}]e[x(k;x_0)^{[m]}]$ ] , an induction with respect to @xmath145 yields @xmath146 } ] = e[f_0^{[m]}]^k x_0^{[m]}.\ ] ] now assume that @xmath136 is stochastically @xmath9th mean stable . let @xmath147 be an eigenvalue of @xmath141}]$ ] with a corresponding eigenvector @xmath148 . since the set @xmath149 } : x\in\mathbb{r}^n \}$ ] spans @xmath150 ( ( * ? ? ? * lemma 1.5 ) ) , there exist @xmath151 and @xmath152 such that @xmath153}$ ] . multiplying @xmath141}]^k$ ] to this equation we obtain @xmath154}]$ ] by . therefore , by the triangle inequality and , we can see @xmath155 $ ] . since the right hand side of this inequality is summable with respect to @xmath145 by the stochastic @xmath9th mean stability of @xmath136 and also @xmath156 , we conclude @xmath157 . [ [ item : mu : shcsta ] @xmath142 [ item : mu : expsta ] ] : assume that @xmath141}]$ ] is schur stable . by , there exist and @xmath158 such that @xmath159 } ] \rvert } \leq \alpha e^{-\beta k } { \lvert x_0^{[m ] } \rvert } = \alpha e^{-\beta k } { \lvert x_0 \rvert}^m\ ] ] for every @xmath55 . we shall show that @xmath136 is exponentially @xmath9th mean stable . let @xmath77 and @xmath55 be arbitrary and write @xmath160 . we consider the two cases [ item : mu : even ] and [ item : mu : invariant ] separately . first assume that @xmath9 is even . take positive constants @xmath161 and @xmath162 such that @xmath163 . then @xmath164 \leq c_2^m e[{\lvert y \rvert}^m_m ] = c_2^m \sum_{i=1}^{n } { \left\lvert e[y_i^m ] \right\rvert}.\ ] ] since the random vector @xmath165}$ ] has all the monomials @xmath166 ( @xmath167 , @xmath168 , @xmath9 ) , we can see that @xmath169 \rvert } \leq { \lvert e[y]^{[m ] } \rvert}_1\leq c_1^{-1}{\lvert e[y]^{[m ] } \rvert}$ ] . therefore this inequality together with and shows that @xmath136 is exponentially @xmath9th mean stable . next assume that @xmath136 is positive . notice that , by lemma [ lem : lemma36result ] , without loss of generality we can assume @xmath79 , which implies @xmath170 and hence @xmath165 } \geq 0 $ ] with probability one . let us take a positive constant @xmath171 such that @xmath172 . then we have @xmath173 } \rvert } \leq c_3 { \lvert y^{[m ] } \rvert}_1 = c_3 1_{n_m}^\top y^{[m]}$ ] with probability one . therefore , the schwartz inequality shows @xmath174 \leq c_3 1_{n_m}^\top e[y^{[m ] } ] \leq c_3 { \lvert 1_{n_m } \rvert}\ , { \lvert e[y^{[m ] } ] \rvert}$ ] . this inequality and prove the exponential @xmath9th mean stability of @xmath136 . proposition [ prop : stab : mu : char ] improves the stability condition in @xcite by reducing the computational cost for checking mean stability . the size @xmath175 of the matrix @xmath141}]$ ] is far less than the size @xmath176 of the matrix used in ( * ? ? ? * theorem 5.1 ) . also the proof presented above is simpler than the proof of ( * ? ? ? * theorem 5.1 ) , which needs the approximation of @xmath133 by a sequence of finitely supported probability measures . let @xmath73 be a regenerative switched linear system satisfying the conditions [ item : evenornon - neg ] to [ item : mbdd ] and let @xmath33 be the trajectory of @xmath73 . then the discretized process @xmath177 given by @xmath178 is clearly the solution of the discrete - time system @xmath179 where @xmath180 is defined by . proposition [ prop : stab : mu : char ] immediately gives the next corollary on the stability of @xmath181 . [ cor : char ] the following conditions are equivalent . 1 . @xmath181 is exponentially @xmath9th mean stable . @xmath181 is stochastically @xmath9th mean stable . @xmath109}]$ ] is schur stable . the random variables @xmath182 are independent and identically distributed by the definition of regenerating processes . also [ item : evenornon - neg ] automatically ensures that one of the conditions [ item : mu : even ] and [ item : mu : invariant ] in proposition [ prop : stab : mu : char ] is satisfied . we will also need the next lemma to prove the main result . there exists @xmath183 such that @xmath184 for all @xmath55 and @xmath185 $ ] . let @xmath186 $ ] . by [ item : mbdd ] there exists a constant , say , @xmath187 , such that @xmath188 for every @xmath189 . since @xmath190 we have @xmath191 . then gronwall s inequality and [ item : hk < t ] shows @xmath192 . similarly we can show @xmath193 . this completes the proof . now we prove the main result of this paper . we shall show the cycle [ [ item : expsta ] @xmath142 [ item : stosta ] @xmath142 [ item : schsta ] @xmath142 [ item : expsta ] ] . it is obvious to prove [ [ item : expsta ] @xmath142 [ item : stosta ] ] . [ [ item : stosta ] @xmath142 [ item : schsta ] ] : by inequality and the definition of regenerative processes , for every @xmath145 we can show @xmath194 & \geq c^{- m } e\left[{\lvert x(z_k ) \rvert}^m\int_{z_k}^{z_{k+1 } } dt\right ] \\ & = c^{- m } e[r_{k+1 } ] e[{\lvert x_d(k ) \rvert}^m ] \\ & = c^{- m } e[r_{1 } ] e[{\lvert x_d(k ) \rvert}^m ] , \end{aligned}\ ] ] where we used the definition of regenerative processes . since fubini s theorem shows @xmath78\ , dt = \sum_{k=0}^\infty e\bigl [ \int_{z_k}^{z_{k+1 } } { \lvert x(t ) \rvert}^m\,dt \bigr]$ ] , taking the summation about @xmath145 in yields @xmath195 \sum_{k=0}^\infty e[{\lvert x_d(k ) \rvert}^m ] \leq \int_0^\infty e[{\lvert x(t ) \rvert}^m]\,dt < \infty.\ ] ] therefore @xmath196 is stochastically @xmath9th mean stable because both @xmath197 and @xmath198 $ ] are positive . hence corollary [ cor : char ] implies that @xmath109}]$ ] is schur stable . [ [ item : schsta ] @xmath142 [ item : expsta ] ] : here we employ the idea used in the proof of the sufficiency part for ( * ? ? ? * theorem 2.5 ) . assume that @xmath109}]$ ] is schur stable . then @xmath181 is exponentially @xmath9th mean stable by corollary [ cor : char ] . let @xmath77 and @xmath38 be arbitrary . let us define @xmath199 . since @xmath200 , the inequality gives @xmath201 . therefore @xmath202 \leq c^m e[{\lvert x_d(k_t ) \rvert}^m].\ ] ] on the other hand , since [ item : hk < t ] shows @xmath203 we have @xmath204 . this implies @xmath205 . taking the expectation in this inequality and then using the @xmath9th mean stability of @xmath181 we obtain @xmath206 \leq \alpha ' e^{-\beta ' t } { \lvert x_0 \rvert}^m$ ] , where @xmath207 and @xmath208 . this inequality and show the @xmath9th mean exponential stability of @xmath73 . this section briefly discusses the stability characterization of regenerative switched linear systems in discrete - time . let @xmath209 be a regenerative process taking values in a set @xmath68 and defined on the set of nonnegative integers @xmath210 . let @xmath211 be a family of @xmath212 real matrices . consider the discrete - time regenerative switched linear system @xmath213 the exponential and stochastic mean stability of @xmath214 are defined as in definition [ defn : sta : sigmamu ] . in addition to the assumptions [ item : evenornon - neg ] to [ item : mbdd ] we place the next assumption on @xmath214 : 1 . [ item : d : invmbdd ] @xmath82 is invertible for each @xmath215 and the set @xmath216 is bounded . for each @xmath55 we define the transition matrix @xmath217 for @xmath214 representing the transition of @xmath33 from @xmath62 to @xmath108 in the same way as we defined @xmath180 for continuous - time regenerative switched linear systems in . the next theorem is a discrete - time counterpart of theorem [ theorem : main ] . the following statements are equivalent . 1 . @xmath214 is exponentially @xmath9th mean stable . @xmath214 is stochastically @xmath9th mean stable . @xmath218}\bigr]$ ] is schur stable . let @xmath219 be the embedded renewal process of @xmath45 . using [ item : mbdd ] and [ item : d : invmbdd ] we can show the existence of a constant @xmath220 such that , for every @xmath55 , if @xmath221 then @xmath222 . then we can prove the desired equivalence in the same way as in the proof of theorem [ theorem : main ] . the details are omitted . in this paper we investigated the mean stability of regenerative switched linear systems . a necessary and sufficient condition for the @xmath9th mean stability of regenerative switched linear systems established under the assumption that either @xmath9 is even or the system is positive and that the length of each cycle of the underlying regenerative process is essentially bounded . the proof used a discretization of the system at the embedded renewal process of the underlying regenerative process . a numerical example was presented to illustrate the result . j. bertram , p. sarachik , stability of circuits with randomly time - varying parameters , ire trans . circuit theory 6 ( 1959 ) 260270 . x. feng , k. a. loparo , y. ji , h. j. chizeck , stochastic stability properties of jump linear systems , ieee trans . control 31 ( 1992 ) 3853 . o. costa , m. fragoso , stability results for discrete - time linear systems with markovian jumping parameters , j. math . ( 1993 ) 154178 . o. l. costa , m. d. fragoso , m. g. todorov , continuous - time markov , springer , 2013 . m. ogura , c. f. martin , http://arxiv.org/abs/1309.2720[stability analysis of positive semi - markovian jump linear systems with state resets ] , siam j. control optim . ( accepted ) . arxiv:1309.2720 [ math.oc ] d. antunes , j. p. hespanha , c. silvestre , stochastic hybrid systems with renewal transitions : moment analysis with application to networked control systems with delays , siam j. control optim . 51 ( 2013 ) 14811499 . k. sigman , r. wolff , a review of regenerative processes , siam rev . 35 ( 1993 ) 269288 . w. l. smith , regenerative , proceedings of the royal soc . a : math . 232 ( 1955 ) 631 . j. zhang , b. zwart , steady state approximations of limited processor sharing queues in heavy traffic , queueing syst . 60 ( 2008 ) 227246 . d. logothetis , k. trivedi , the effect of detection and restoration times for error recovery in communication networks , j. netw . manag . 5 ( 1997 ) 173195 . r. v. canfield , cost optimization of periodic preventive maintenance , ieee trans . 35 ( 1986 ) 7881 . t. nakagawa , periodic and sequential preventive maintenance policies , j. appl . probab . 23 ( 1986 ) 536542 . r. w. brockett , lie theory and control systems defined on spheres , siam j. appl . math . 25 ( 1973 ) 213225 . a. barkin , a. zelentsovsky , method of power transformations for analysis of stability of nonlinear control systems , systems control lett . 3 ( 1983 ) 303310 . y. zhang , j. jiang , bibliographical review on reconfigurable fault - tolerant control systems , annual rev . control 32 ( 2008 ) 229252 . p. bolzern , p. colaneri , g. de nicolao , markov jump linear systems with switching transition rates : mean square stability with dwell - time , automatica 46 ( 2010 ) 10811088 . m. mariton , detection delays , false alarm rates and the reconfiguration of control systems , int . j. control 49 ( 1989 ) 981992 . m. ogura , c. f. martin , http://www.sciencedirect.com/science/article/pii/s0024379513004278[generalized joint spectral radius and stability of switching systems ] , linear algebra appl . 439 ( 2013 ) 22222239 .	this paper investigates the stability of switched linear systems whose switching signal is modeled as a stochastic process called a regenerative process . we show that the mean stability of such a switched system is characterized by the spectral radius of a matrix . the matrix is obtained by taking the expectation of the transition matrix of the system on one cycle of the underlying regenerative process . the characterization generalizes floquet s theorem for the stability analysis of linear time - periodic systems . we illustrate the result with the stability analysis of a linear system with a failure - prone controller under periodic maintenance . switched linear system , regenerative process , mean stability , periodic maintenance
the reissner - nordstrm ( rn ) spacetime describes a vacuum where matter and electric charge exist in a singularity at the center @xmath0 . when the mass is large enough compared to the electric charge , @xmath1 , which is the case we explore here , a black hole engulfs the singularity . this black hole is far more complex than the simple schwarzschild black hole . the rn spacetime has been extensively investigated in the past ( see e.g. @xcite ) . our aim here is to describe the embedding diagrams of typical hypersurfaces of this spacetime . the embedding diagrams of spatial hypersurfaces of the schwarzschild black hole are well known ( see @xcite ) . surprisingly , little is known about the corresponding diagrams of the richer rn ( and the rather similar kerr ) spacetime . we review briefly the concept of embedding diagrams in sec . [ embsec ] and discuss as an example the embedding of the schwarzschild spacetime . we present in sec . [ coosec ] rn coordinate systems that enable us to explore the rn spacetime , and we derive the formulas to calculate the embeddings of rn hypersurfaces . in sec . [ rnembsec ] we present the embedding diagrams of these slices . we summarize our results in sec . [ sumsec ] . embedding diagrams of hypersurfaces of a spacetime are useful tools to visualize the overall geometry . consider a 3-dimensional hypersurface with an induced metric of @xmath2 . when the hypersurface is spherically symmetric ( and @xmath3 on the hypersurface ) , we can select a 2-dimensional plane within this hypersurface and express the metric as : @xmath4 this 2-dimensional plane can now be embedded in a flat 3-dimensional space : @xmath5 this embedding space is euclidean or minkowski depending on whether @xmath6 ( see ) . the embedding diagram is described by the embedding function @xmath7 , such that the surface @xmath7 has the same intrinsic geometry as the original spacetime slice : @xmath8 where the @xmath9 signs are for euclidean and minkowski embeddings respectively . timelike hypersurfaces @xmath10 , which imply @xmath11 everywhere , are always embedded in minkowski space , whereas certain spacelike hypersurfaces @xmath12 are embedded in euclidean space and others in minkowski space . in 1916 flamm @xcite had already obtained the well known schwarzschild wormhole " with an analytic embedding function of @xmath13^{\frac{1}{2}}$ ] . this describes a hypersurface of @xmath14 outside the event horizon , and similar embeddings correspond to horizontal lines , @xmath15 , in the kruskal - szekeres coordinates . an embedding diagram of such a hypersurface is displayed in figure [ fig : sch - hor ] . as a function of the @xmath16 coordinates ( the @xmath17 plane).,height=377 ] the less familiar embedding of the @xmath14 hypersurface inside the schwarzschild event horizon @xmath18 is presented in figure [ fig : sch - ver ] . this is a timelike hypersurface ( described by a vertical line , @xmath19 , in the kruskal - szekeres coordinates ) , and its embedding is in minkowski space . coordinate and display the embedding lift as a function of @xmath20 alone . the left half of the diagram does not refer of course to negative @xmath20 s - it is merely a rotation of 180@xmath21 that is shown in order to ease the visualization of the entire surface . this embedding is denoted by a dashed line to indicate the background minkowski space.,width=491 ] since our spacetime slices always have an axial symmetry , we will use in the following only 2-dimensional diagrams such as figure [ fig : sch - ver ] to describe our surfaces . we adopt here a convention that when the embedding space is minkowski the diagram is described by a dashed line , while it is a solid line in euclidean space . the familiar rn metric is described in schwarzschild - like coordinates as : @xmath22 where the parameters @xmath23 and @xmath24 describe the mass and electric charge . we consider here @xmath25 when a black hole exists . in this case we have two horizons : @xmath26 we get rid of the coordinate singularities on @xmath27 by moving to a kruskal - szekeres - like coordinate system : @xmath28 where @xmath29 . in fact we use two separate coordinate systems - coordinate system a that is analytic at @xmath30 but singular at @xmath31 and coordinate system b that is analytic at @xmath31 but singular at @xmath30 . the coordinate transformations from @xmath32 to @xmath33 will take a different form in the three regions of spacetime : @xmath34 we will use system a to describe regions @xmath35 and @xmath36 and system b to describe regions @xmath36 and @xmath37 . the transformations to coordinate systems a , b in regions @xmath38 are : [ transformations ] @xmath39 \\ \sinh[f(t ) ] \\ \sinh[f(t)]\\ \cosh[f(t)]\\ \end{array}\right\ } \\ \left\{\begin{array}{c } \psi_{a , i } \\ \psi_{a , ii } \\ \psi_{b , ii } \\ \psi_{b , iii } \\ \end{array } \right\}=\left\{\begin{array}{c } c(r ) \\ c(r ) \\ -c(r)^{-1 } \\ c(r)^{-1 } \\ \end{array}\right\ } \left\{\begin{array}{c } \sinh[f(t ) ] \\ \cosh[f(t ) ] \\ \cosh[f(t)]\\ \sinh[f(t)]\\ \end{array}\right\}\end{aligned}\ ] ] where @xmath40 and @xmath41.\ ] ] @xmath42 is an analytic function in each region of spacetime , and @xmath20 is expressed implicitly in terms of the new coordinates using : @xmath43 where the appropriate @xmath9 signs for each region and coordinate system is determined by . we use in the following lower indexes to indicate the coordinate system and region to which we refer . figure [ fig : coosysa ] depicts constant @xmath20 lines in the @xmath33 plane of coordinate system a. horizon . the relation between the original radial coordinate and the new coordinates is displayed for an example of a black hole characterized by @xmath44 . @xmath45 hypersurfaces appear here as hyperbolae with the asymptotes @xmath46 . these asymptotes describe the surfaces @xmath47 , while the surfaces @xmath48 , which are singular in the current coordinates , are located at the upper and lower infinity of this diagram ( where @xmath49).,width=340,height=340 ] the transformations presented in are valid for one set of system a regions - the upper right half of figure [ fig : coosysa ] corresponding to @xmath50 or @xmath51 . a parallel set of regions with identical geometries is obtained by similar transformations @xmath52 . figure [ fig : coosysb ] shows constant @xmath20 lines in the @xmath53 plane . the relation between @xmath20 and the new coordinates is displayed for an example of @xmath44 . @xmath45 hypersurfaces appear here as hyperbolae with the asymptotes @xmath46 . these asymptotes describe the surfaces @xmath48 , while the surfaces @xmath47 , which are singular in the current coordinates , are located at the upper and lower infinity of this diagram.,width=340,height=340 ] the system b transformations in describe the lower left half of figure [ fig : coosysb ] , and again the parallel regions are similarly obtained . the two separate coordinate systems connect of course as @xmath54 goes to @xmath55 , demonstrating the known structure of the rn spacetime , which is in essence an infinite chain of parallel asymptotically flat universes connected by wormholes . we consider embedding diagrams of spacelike and timelike hypersurfaces that are described by horizontal and vertical lines in the @xmath33 coordinate systems . horizontal lines mean @xmath56 , which implies in coordinate system a : @xmath57 \\ \coth[f(t ) ] \\ \end{array } \right\}\left(1-\frac{2m}{r}+\frac{q^2}{r^2}\right)^{-1}dr\ ] ] using , the embedding function of such a hypersurface in euclidean @xmath58 or minkowski @xmath59 space is given by : @xmath60\\ \coth^2[f(t ) ] \\ \end{array } \right\}\right)\left(1-\frac{2m}{r}+\frac{q^2}{r^2}\right)^{-1}}dr\ ] ] repeating this procedure for horizontal lines in region @xmath36 of coordinate system b we get the same embedding formula as the one for coordinate system a in . the embedding function of horizontal lines in region @xmath37 of coordinate system b is identical to the one for region @xmath35 in . the difference is that the parameter @xmath61 is assigned different values than before , as it is determined by solving the equation @xmath62 rather than @xmath63 . we can similarly arrive at the expressions for embedding surfaces of vertical lines @xmath64 . such a vertical line is partly described by coordinate system a and partly described by coordinate system b. in order to get the entire embedded surface ( which passes through regions @xmath35 , @xmath36 and @xmath37 ) we need to connect continuously and smoothly the embedding curves of the two coordinate systems . note that due to the different coordinate transformations , straight lines of @xmath65 in coordinate system a do not necessarily connect with straight lines of the same @xmath65 in coordinate system b. once we choose the connection point between @xmath31 and @xmath30 , we are required to match the same spacetime event in both coordinate systems , determining specific sets of vertical lines . regardless of this arbitrariness , the embedding diagrams show the qualitative features of timelike and spacelike hypersurfaces , and the arbitrary details of the specific hypersurfaces are not important . the embedding functions we get in minkowski space ( timelike surfaces can never be embedded in euclidean space ) are : @xmath66\\ \tanh^2[f(t ) ] \\ \coth^2[f(t ) ] \\ \end{array } \right\}\right)\left(1-\frac{2m}{r}+\frac{q^2}{r^2}\right)^{-1}}dr\ ] ] where @xmath61 is determined by coordinate system a s transformation for region @xmath35 and by coordinate system b s transformation for region @xmath37 . for region @xmath36 @xmath61 can be determined by either systems transformations , depending on our choice of the vertical lines connection point . using the tools developed in the previous section we turn now to the embedding diagrams of different hypersurfaces in the rn spacetime . figure [ fig : penrose ] depicts a penrose diagram of the rn spacetime ( as shown in @xcite ) . we sketch on this diagram lines corresponding to the hypersurfaces for which we present embedding diagrams in this section . we use as an example throughout the discussion here the case @xmath44 . we begin with spacelike hypersurfaces presented as horizontal lines in coordinate system a. the line element of the @xmath67 surface is the same as any @xmath14 surface @xmath68 in the original @xmath32 coordinates . we expect this surface to be similar to that of @xmath69 in the schwarzschild spacetime , as both cases describe a spacelike wormhole connecting at the event horizon two asymptotically flat regions . the spacetime geometry is independent of the @xmath61 coordinate , but @xmath61 becomes a spacelike coordinate when crossing the event horizon . this means that for an observer looking at the black hole from outside the spacetime appears static . to explore the geometry seen by an observer who crosses the event horizon and journeys into the black hole we advance with the @xmath54 coordinate , which is a genuine timelike coordinate in all regions , and inspect the evolution of the spacelike hypersurfaces with @xmath70 . figure [ fig : emb123 ] depicts the embedding diagrams of the @xmath71 , @xmath72 and @xmath73 hypersurfaces . surface where the wormhole is widest . b _ describes the @xmath74 surface , and curve _ c _ belongs to the @xmath75 surface , which nearly describes the wormhole of smallest circumference.,width=491 ] the geometry of the different surfaces varies . this demonstrates the fact that the rn spacetime is actually dynamic . as we advance in @xmath76 the wormhole shrinks this phenomena also occurs when advancing along the timelike kruskal - szekeres coordinate in the schwarzschild spacetime . while @xmath31 is reached only as @xmath77 , the surface described by @xmath73 can be treated with our working accuracy as if it is tangent to the @xmath48 hyperbola ( see figure [ fig : coosysa ] ) . the wormhole with a throat radius of @xmath31 is the narrowest wormhole that the time evolution in coordinate system a allows . this is contrary to the schwarzschild case , where the wormhole continues to shrink until it closes when reaching the singularity . to explore further the rn spacetime we must transform to coordinate system b. horizontal lines in this coordinate system describe spacelike hypersurfaces that begin at the @xmath78 singularity and increase in radial distance until connecting with the parallel region after crossing the @xmath31 horizon . therefore the embedding diagrams of these hypersurfaces will be of a different nature . in the proximity of the @xmath78 singularity we can never embed any surface in euclidean space . the rn line element has @xmath79 and @xmath80 as @xmath81 . every path that approaches @xmath78 will have at a sufficiently small radius @xmath82 . therefore around the singularity we must embed the following surfaces in minkowski space . at a certain distance the embedding becomes euclidean . as we examine the evolution of spacelike surfaces seen by a traveler that enters the black hole , we enter coordinate system b from the bottom and draw in figure [ fig : emb4 ] the embedding diagram of the surface described by @xmath83 . hypersurface . for brevity the figure depicts two different surfaces that describe the embeddings of parts of this hypersurface in different background spaces . the portion described by dashed lines is embedded in minkowski space and the portion described by solid lines is embedded in euclidean space . we can see an inflection point at @xmath48 . note that the minkowski portion starts at @xmath84.,width=491 ] the connection between the parallel regions @xmath85 occurs now when increasing the radial distance ( up to @xmath86 in the example above ) . figure [ fig : coosysb ] shows that the two inner parallel regions are `` closest '' in the surface of @xmath87 , where they touch at @xmath48 . the embedding of this surface is displayed in figure [ fig : emb5 ] . hypersurface . for brevity , here again , the figure depicts two different surfaces that describe the embeddings of parts of this hypersurface in different background spaces.,width=491 ] it is interesting to inspect the resemblance between the geometry of the horizontal lines in coordinate system b and the geometry of the vertical lines in the kruskal - szekeres coordinates of the schwarzschild spacetime . we can imagine a resemblance because the schwarzschild spacetime s vertical lines also emerge from the @xmath78 singularity and connect with a parallel region after crossing the horizon . however , these vertical lines describe timelike hypersurfaces and so their embeddings will all be in minkowski spaces . comparing figures [ fig : emb4],[fig : emb5 ] with figure [ fig : sch - ver ] , we indeed see some resemblance . the region near the singularity in the rn spacetime s horizontal lines has a similar general shape to the schwarzschild spacetime s vertical lines on their way from the singularity to the parallel region . the region in rn with @xmath88 continues towards the parallel region like the schwarzschild surfaces , but this region is embedded in euclidean space while the schwarzschild surfaces continue in minkowski space . we turn now to inspect timelike hypersurfaces of the rn spacetime described by vertical lines in our coordinate systems . the @xmath89 line passes in the two coordinate systems only in regions @xmath36 between @xmath48 and @xmath47 . the embedding diagram of this surface ( in minkowski space of course ) is displayed in figure [ fig : emb6 ] . the diagram shows just one of the periodic segments of the infinite surface . the enlargement demonstrates the smoothness of the embedding at the @xmath31 wormhole throat.,width=491 ] we can see in the figure a connection ( the timelike wormhole presented in coordinate system b ) between parallel regions at @xmath48 . additionally , we can see in coordinate system a a timelike wormhole connecting parallel regions at @xmath47 , and this should be visible in the embedding by connecting the upper and lower ends of figure [ fig : emb6 ] with identical diagrams . the complete surface is achieved by connecting an infinite chain of such diagrams . when passing through the @xmath31 horizon to a parallel region there appears to be a cusp in the embedding . this is not a consequence of performing a wrong connection between the coordinate systems , since the entire calculation here was done in a single coordinate system . however , a close inspection of the embedding function reveals that its slope is in fact continuous at @xmath48 . when approaching the wormhole throat at @xmath31 , @xmath90 diverges , as is required for a smooth connection ( see the inset in figure [ fig : emb6 ] ) . we now move the vertical line to the side ( either positive or negative @xmath91 ) and examine a hypersurface that touches the singularity at @xmath78 ( see figure [ fig : emb7 ] ) . surface.,width=491 ] this timelike surface passes regions @xmath35 , @xmath36 and @xmath37 . it begins by touching the @xmath78 surface , where the wormhole throat is closed . its radius increases and it crosses the @xmath31 horizon ( where the embedding seems again to break ) . it continues past the @xmath30 horizon and reaches a maximal radius where it connects to a new parallel region . the complete surface is , of course , described by an infinite chain of identical diagrams . it should be noted that at @xmath48 we were careful not to switch between coordinate systems ( we have a freedom to connect the vertical lines of the two coordinate systems somewhere between @xmath48 and @xmath47 ) . the surface is smooth at the point of the coordinate systems connection ( @xmath92 in this example ) . again , the surface is in fact also smooth at @xmath48 , as a detailed inspection of this area shows . the timelike wormhole around the singularity is widest ( the throat has a radius of @xmath31 ) for the hypersurface described by @xmath89 . this is similar to the surface described by the horizontal line @xmath93 in the kruskal - szekeres schwarzschild coordinates ( described in figure [ fig : sch - hor ] ) , where the spacelike wormhole is widest ( with a throat radius of 2 m ) . when we move in the kruskal - szekeres coordinates towards horizontal lines of larger @xmath94 , the wormhole shrinks until it closes with zero circumference at the @xmath78 singularity . this resembles our case , where moving the vertical lines towards larger @xmath91 caused the wormhole to shrink until it eventually closed at the hypersurface tangent to @xmath78 . a resemblance between the known schwarzschild embedding diagrams and the rn embedding diagrams is indeed apparent . notice again the significant difference that the relevant schwarzschild surfaces are spacelike and embedded in euclidean space , whereas our current surfaces are timelike and embedded in minkowski space . the character of the @xmath31 horizon is different than that of the schwarzschild event horizon , and this is apparent in the diagrams . in addition , the schwarzschild spacetime horizontal lines continue towards radial infinity , whereas our vertical lines cross another horizon at @xmath30 and then reach another wormhole and contract towards @xmath31 . thus , the resemblance between the schwarzschild horizontal lines and the rn vertical lines is limited to the region of coordinate system b , describing the area around the singularity and the @xmath31 horizon in the rn spacetime . the portions of the vertical lines that reside in coordinate system a ( around the @xmath30 horizon ) are in fact similar to vertical lines in the schwarzschild spacetime , as is evident by a comparison to figure [ fig : sch - ver ] . this is not surprising due to the similarity of coordinate system a to the kruskal - szekeres coordinates . we developed tools and displayed embedding diagrams of spacelike and timelike hypersurfaces of the rn spacetime . the hypersurfaces described provide a typical example of spacelike and timelike slices of the rn spacetime and demonstrate the evolution of rn surfaces . the visualization presented for these slices provides an alternative way to understand intuitively the geometry of the rn spacetime and its properties . the rn geometry is similar to the schwarzschild geometry in the region external to the event horizon @xmath68 . in accordance with this the embedding diagrams of hypersurfaces described in the coordinate system that crosses @xmath47 were similar to the embedding diagrams of the analogous schwarzschild hypersurfaces . when we advance in time to different rn spacelike slices described by horizontal lines in this coordinate system , the wormhole shrinks , as is the case when the schwarzschild spacetime evolves in the time coordinate of kruskal - szekeres . however , as expected , when we continue to advance in time and reach the hypersurfaces described in the coordinate system that crosses @xmath48 , the similarity to schwarzschild stops . the @xmath31 horizon has no analogy in the schwarzschild spacetime , and the rn horizontal lines that intersect @xmath48 exhibit a different behavior . these hypersurfaces hit on both sides the @xmath78 singularity , and their geometry near the singularity resembles the geometry of timelike schwarzschild hypersurfaces , presented as vertical kruskal - szekeres lines . finally , vertical lines in the rn coordinates describe a path through a wormhole passing beside the singularity . the geometry of the respective surfaces has some features resembling the geometry of the surfaces describing the schwarzschild wormhole , which are presented there as horizontal lines . the fundamental difference is that the schwarzschild wormhole is a spacelike path that can not be embarked upon , whereas the rn wormhole is an actual timelike path . this is reflected in the evolution of the spacelike hypersurfaces . when we advance in time the schwarzschild wormhole shrinks and closes . however , progressing with the rn spacelike surfaces does not cause the wormhole to close , as apparent from our embedding diagrams , and we are not forced to be drained into the singularity . instead one may choose to take a path like the ones described by our timelike embedding diagrams , and pass through the timelike wormhole to a parallel region and to infinite new universes . we thank nathalie deruelle , joseph katz and an anonymous referee for helpful remarks . this research was supported by the isf center of excellence for high energy astrophysics and by the schwartzmann university chair ( tp ) . 4 misner c w , thorne k s and wheeler j a 1973 _ gravitation _ ( san francisco : freeman ) chandrasekhar s 1983 _ the mathematical theory of black holes _ ( oxford : oxford university press ) graves j c and brill d r 1960 _ phys . rev . _ * 120 * 1507 flamm l 1916 _ physik . z. _ * 17 * 448 carter b 1966 _ phys .	we examine embedding diagrams of hypersurfaces in the reissner - nordstrm black hole spacetime . these embedding diagrams serve as useful tools to visualize the geometry of the hypersurfaces and of the whole spacetime in general .
the `` off '' state of the x - ray pulsar her x-1 was observed with the telescope art - p on board granat three times in 1990 - 1991 . the source persistent flux in this state was detected at the level of few mcrabs ( several percents of its high - state flux ) and was characterized by the absence of x - ray pulsations the @xmath2 upper limit on the pulse fraction in the 3 - 20 kev band varied between 2.9% and 18% in different observations . the source s `` off '' state spectra are well described by a simple power law without evidence of a cutoff in the hard energy band or of emission lines . the spectral slopes differ by a factor of @xmath3 , with the photon flux in the energy band 3 - 20 kev being virtually the same . her x-1 is observed in its `` off '' state if the compact source is hidden behind the outer edge of a warped ( or tilted to the orbital plane ) accretion disk and , as it is believed , the source emission can greatly scattered in a hot corona ( it s height is @xmath4 cm ) above the outer parts of the disk . such scattering of emission from the compact source by coronal electrons can account for the existence of an x - ray flux in the `` off '' state and it allows to explain the absence of pulsations in this state , as any information about variability on a time scale shorter than @xmath5 s , where @xmath6 is the speed of light , must be lost . a significant declination ( by a factor of 6 ) in the 4 - 20 kev source flux was detected during the observation on aug 19 , 1990 . this declination occurred in two steps , latter of which corresponded to an orbital phase @xmath7 at which the x - ray eclipse ingress was observed in other experiments . it is important to note that the intensity decline during the eclipse ingress mainly occurred in the hard energy band and the corresponding hardness ( ratio of the 10 - 20 kev and 4 - 6 kev fluxes ) decrease was equal to @xmath8% . after the eclipse ingress a non - zero and non - pulsing x - ray emission was detected by art - p from the source at the signal - to - noise level of @xmath9 . the pulsar s spectrum during the eclipse was measured with large statistical errors ; nevertheless , a line at energy of @xmath10 kev was very reliable . by assuming that this is the 6.7-kev line of helium - like iron , we estimated its intensity , @xmath11 , and equivalent width , @xmath12 kev . although the source itself was already occulted by the disk of the optical star at this time , part of the surface of the scattering cloud remained in the visibility zone , and we observed precisely this emission scattered in the cloud . note also that the measured spectrum at energies below 5 kev shows an excess of soft x - ray emission , which was previously revealed from the light curve analysis . the observation of this x - ray pulsar on june 15 , 1992 ( orbital phases 0.9180.953 ) was carried out mainly during an eclipse . we divided it into two parts , in accordance with the beginning of the source s eclipse egress . a non - zero flux of approximately 1/10 of the bright - state flux at the confidence levels 4.3 and 4.6@xmath0 was observed from the source during the eclipse and eclipse egress , respectively . during the second part of the observation x - ray pulsations with the period @xmath13 s were detected from the source at a confidence level of @xmath14 . the pulse fraction was equal to @xmath15% and it was a factor of 2 lower than one measured in the source s bright state several days earlier . during the eclipse we failed to detect any pulsations the 3@xmath0 limit on the pulse fraction was 9.6% . the persistent flux , detected during an x - ray eclipse , may be the emission scattered in a fairly dense stellar wind or in the extended atmosphere of the optical star . given that the binary characteristic size , @xmath16 light seconds , is half the pulsation period , photons must undergo several scatterings for the pulse profile to be smeared enough . for the scattering optical depth of the wind @xmath17 , the scattered - to - incident flux ratio allows us to estimate the solid angle @xmath18 at which the scattering - envelope ring is seen from the compact source . accordingly , the envelope height above the optical - star surface is @xmath19	results of granat / art - p observations of three x - ray pulsars in non - pulsing states are presented : 1 ) a statistically significant non - pulsing flux with a simple power - law spectrum was detected during the `` off''- state of her x-1 ; 2 ) a significant ( 13 @xmath0 ) non - pulsing flux with a strong iron emission line at energies 6.7 - 6.9 kev was detected during the eclipse ingress of cen x-3 ; 3 ) a weak non - pulsing flux was detected during the x - ray eclipse of vela x-1 , which probably resulted from scattering of the pulsar emission in the stellar wind of an optical star . # 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in @xmath1
several studies have shown that measures of trabecular bone micro - architecture and bone strength are correlated . @xcite together with loss of bone mass , changes in the trabecular bone micro - architecture occur during ageing , during development of osteopenia and osteoporosis as well as in connection with immobilisation or space flight , and can lead to an increased risk of bone fracture . the vertebral bodies and the epiphyses and metaphyses of the long bones consist mainly of trabecular bone surrounded by a thin cortical shell . @xcite a dramatic change in the state of the trabecular bone leads to an increased fracture risk . @xcite bone mineral density ( bmd ) is the most commonly used predictor of bone strength and fracture risk , and also the most commonly used general descriptor of the state of the bone . non - linear relationships have been established between volumetric bmd and compressive bone strength and elastic modulus . @xcite however , for trabecular bone , it has been established that a part of the variation in the strength of the bone can not be explained by bmd alone , but is instead due to the micro - architecture of the trabecular network . for example , a relationship has been established between the mechanical properties of the bone and the shape , orientation , bone trabecular volume fraction , and thickness of the trabeculae . @xcite a series of new methodologies based on techniques from nonlinear data analysis has also been introduced in order to study the relationship between the complexity of the trabecular bone network and bone strength . @xcite saparin et al . established such a relationship by use of structural measures of complexity based on symbolic encoding . @xcite furthermore , several studies using numerical modelling of the trabecular bone and finite element analysis have also confirmed the importance of the trabecular bone micro - architecture to bone strength . @xcite in the present study we propose a method of analysis of trabecular bone micro - architecture that is new in two ways . firstly , the analysis is based on peripheral quantitative computed tomography ( pqct ) images , so the method is nondestructive and noninvasive . secondly , our image analysis uses a new approach called _ long range node - strut analysis _ , which quantify the apparent nodes and struts . in the present study we propose a new method for analysis of trabecular bone micro - architecture from high - resolution quantitative computed tomography ( qct ) images , as at present it is still not possible to perform micro ct ( @xmath3ct ) on humans in vivo and in situ . in contrast to classical histomorphometry , ct images can be obtained in a nondestructive and noninvasive manner , which is preferable in a clinical setting . our method uses a new approach called long range node - strut analysis that quantifies the apparent nodes and struts in the trabecular network . in the present study the analysis is based on peripheral quantitative computer tomography ( pqct ) images . in contrast to classical histomorphometry , pqct images can be obtained in an nondestructive and noninvasive manner which is preferable for implementation of the method in a clinical setting . furthermore , our analysis is based on a new approach called long range node - strut analysis that is able to quantify the apparent nodes and struts . in contrast to the node - strut analysis of garrahan et al . , @xcite our method emphasises long range connectivity of the trabecular network , over a distance controlled by a parameter in the algorithm . the analysis has the dual aims of describing the shape of the trabecular network and predicting bone strength . the trabecular bone compartment consists basically of two components : bone and marrow . due to the limited resolution of present day ct and peripheral quantitative ct ( pqct ) scanners it is not possible to completely resolve the trabecular bone micro - architecture . this results in variations of the ct values of the trabecular voxels , even if the intrinsic bone density is constant throughout the trabecular network . consequently , our method takes this apparent variation in bone density into consideration rather than segment or binarise the image , which would imply a loss of this additional information . we assume that higher ct values of a given trabecula will , all else being equal , result in a higher compressive bone strength . we note that our method does not require skeletonisation of the image . we apply our method to 2-dimensional pqct images of human proximal tibiae and quantify the trabecular bone micro - architecture at different levels of bone integrity ranging from normal healthy bone to osteoporotic bone ( as assessed by their bmd ) . we compare our results with the node / terminus ratio ( nd / tm ) , @xcite computed using traditional histomorphometry performed on bone biopsies obtained at the same location as the pqct scans . the study population comprised 18 women aged 7598 years and 8 men aged 5788 years . at autopsy , the tibial bone specimens were placed in formalin for fixation . for each specimen a pqct image and a bone biopsy were obtained from the same location . for each proximal tibia , an axial qct slice was acquired 17 mm below the tibial plateau with a stratec xct-2000 pqct scanner ( stratec gmbh , pforzheim , germany ) , with an in - plane pixel size of 200@xmath3 m @xmath4 200@xmath3 m and a slice thickness of 1 mm . in some cases , the scans were performed after the biopsies were taken . therefore , the holes left from the bone biopsy appear in some of the pqct images . a standardised image pre - processing procedure was applied to exclude the cortical shell from the analysis . @xcite one of the resulting images is shown in fig . [ fig : tibia ] . . the hole visible on the left of the image is the result of a cylindrical biopsy . ] cylindrical bone samples with a diameter of 7 mm were obtained 17 mm distal from the centre of the medial facet of the superior articular surface by drilling with a compressed - air - driven drill with a diamond - tipped trephine at either the right or the left proximal tibia . these bone biopsies were embedded undecalcified in methyl methacrylate , cut in 10-@xmath3m - thick sections on a jung model k microtome ( r. jung gmbh , heidelberg , germany ) , and stained with aniline blue ( modified masson trichrome ) . the mounted sections were placed in an flat - bed image scanner and 2540 dpi digital 1 bit images of the sections were obtained as previously described in detail . @xcite the resulting pixel size is 10 @xmath3 m @xmath4 10 @xmath3 m . the trabecular bmd of each pqct slice was calculated using a linear relationship derived on the basis of experimental calibration with the european forearm phantom , as described by saparin et al . @xcite the trabecular bmds of the slices in this study range from 30 to 150mg/@xmath5 , with a median of 97 mg/@xmath5 . also in the above paper by saparin et al . is a recommended pqct `` cut - off value '' of 275 , corresponding to bmd 25 , to be used for a bone vs. marrow classification . this is the pqct value of the densest non - bone tissue tested , plus the standard deviation of the noise . if the mean bmd is calculated over only those pixels classified as bone , then the resulting bmd values for the 26 images range from 111 to 235mg/@xmath5 , with a median of 172mg/@xmath5 . most of this section describes the new image analysis method , _ long range node - strut analysis _ , which includes the new measure , _ mean node strength _ ( ndstr ) . this method was applied to the pqct sections described in section [ sectmaterials ] , after removal of the cortical shell . we also computed a standard measures from the same regions of interest of the same pqct images : trabecular volumetric bone mineral density ( bmd ) , calculated as described by saparin et al . @xcite for further comparison , topological 2-dimensional node - strut analysis @xcite was performed on the histological sections described in section [ sectmaterials ] , using a custom - made computer program . @xcite the trabecular bone profile was iteratively eroded until it was only one pixel thick using a hilditch skeletonisation procedure , @xcite and nodes and termini were automatically detected by inspecting the local @xmath6 neighbourhood . if the centre pixel of the @xmath6 neighbourhood is a skeleton pixel and one and only one of the 8 other pixels is a skeleton pixel the centre pixel is classified as a terminus ( tm ) indicating that the strut ends in this pixel . if the centre pixel is a skeleton pixel and three or more of the 8 adjacent pixels are skeleton pixels the centre pixel is classified as a node ( nd ) indicating that two or more struts join in this pixel . compston et al . have argued that the ratio between nodes and termini ( nd / tm ) is an expression of the connectivity of the trabecular network.@xcite consequently , only the node - to - terminus ratio ( nd / tm ) from this analysis of the histological sections is used in the present study . we start with the algorithm to find _ strands _ in each image . a strand is a connected trabecular path , i.e. , a chain of one or more struts , with the whole path going in approximately the same direction . in our algorithm , we use eight directions , labelled by the points of the compass : north ( n ) , northeast ( ne ) , east ( e ) , southeast ( se ) , south ( s ) , southwest ( sw ) , west ( w ) , or northwest ( nw ) . `` north '' is the anterior direction , which corresponds to `` up '' in the images shown in this article . node _ is a pixel that is joined to strands in at least three of these eight directions , any two of which are at least 90 degrees apart ( e.g. n , e , and sw ; but not n , ne , and e ) . the _ node strength _ of a pixel is 0 if it is not a node at all , but otherwise depends on the lengths of the strands that meet at the node and the pqct values of the pixels in these strands . these basic definitions could be implemented in a variety of ways , depending for example on the definitions of `` connected '' and `` same direction '' , and on how the pqct values of the pixels are used . in our algorithm , the first step is to remove marrow from further consideration , by applying a bone threshold filter . the thresholded pqct values will be referred to in this section as _ ct values of bone _ , denoted @xmath7 . thus if @xmath8 is the pqct value of a pixel , then @xmath9 we choose @xmath10 , corresponding to a bmd value of 24 mg/@xmath5 . this is the soft tissue threshold used in symbol encoding for complexity measures by saparin et al . @xcite the threshold was determined as the pqct value of the densest non - bone tissue tested , plus the standard deviation of the noise ( a calibration protocol for finding the threshold based on a phantom will be developed in the future ) . in order to explain the central algorithm , we begin by considering a fictitious image consisting of just one row . the ct values of the pixels in the row form a sequence @xmath11 we will define a corresponding sequence of _ strand strengths _ , @xmath12 where each @xmath13 describes the pattern of densities to the left of pixel @xmath14 in the row . we begin by setting @xmath15 , and then we recursively define @xmath16 where @xmath17 is a _ transmission constant _ between @xmath18 and @xmath19 . note that if all of the ct values have a common value @xmath7 then @xmath20 since @xmath15 . as @xmath14 increases , @xmath13 approaches an upper bound of @xmath21 . thus , for the simple case of a uniformly dense row , we have defined the strand strength to be @xmath22 times the ct values of bone . this constant @xmath22 can be interpreted as a characteristic length of the method , which we can vary depending on the length of strand that we believe to be most important to the strength of the bone . for the present study , we have chosen @xmath23 , corresponding to a characteristic length of @xmath24 pixels , i.e. 4 mm . for the general case of a variable ct values row , the presence of the transmission constant @xmath17 in the recursive formula ensures that the ct value of pixels closest to the @xmath25 pixel have the greatest effect on @xmath13 . finally , the factor @xmath26 ensures that @xmath13 depends strongly on @xmath27 , so that any weak link in a chain of high - density pixels lowers the strand strength dramatically . the difficulty in generalising this definition to a 2-dimensional image is of course that there are infinitely many directions but only four that are parallel with the pixel grid . we have resolved this in a practical but ad hoc fashion . as already stated , we consider eight directions . consider first the definition of strand strength in the leftwards ( w ) direction . > from pixel @xmath28 ( at row @xmath29 and column @xmath14 ) we consider that there are five length-3 paths leading approximately leftwards : @xmath30 as illustrated in fig . [ fig : path3 ] . . the start pixel @xmath28 is marked in black . ] for each of these paths , we define @xmath31 and @xmath32 as the ct values of the pixels in the path in columns @xmath33 , and @xmath34 . we then apply the recursive formula @xmath35 which is the formula in eq . ( [ e : strandstr ] ) , applied twice . ( we must now set @xmath36 . ) we now have five @xmath13 values , one for each of the five paths , i.e. , @xmath37 . we multiply all but the third ( fig . [ fig : path3](iii ) ) of these five values by a _ bending coefficient _ @xmath38 between @xmath18 and @xmath19 , to penalise strands that bend away from the horizontal direction . then the maximum of these five numbers is our leftwards ( w ) strand strength @xmath39 . strand strengths in the other seven directions @xmath40 are determined in a similar manner . for diagonal directions , the geometry of the five length-3 paths used in the computation is slightly different , so a different bending coefficient @xmath38 is used . we chose @xmath41 for the horizontal and vertical directions and @xmath42 for the diagonal directions . these constants were chosen so as to give mean strand strengths that are approximately invariant to image rotations by an arbitrary angle , which were verified by numerical experiments on tibial images . + finally , we calculate the node strength @xmath43 at each pixel . there are @xmath44 possible ways of choosing 3 out of the 8 directions in such a way that each pair of chosen directions makes an angle of at least 90 degrees . for example , e , n , and sw comprise an allowable choice , but e , n , and ne do not . at each pixel , for each allowable choice of three directions , we calculate the minimum of the three strand strengths in these directions , e.g. , @xmath45 . the _ node strength _ @xmath43 is the maximum of the 16 minima @xmath46 subtracted by a minimum strength constant @xmath47 , @xmath48 pixels with a positive node strength are called _ nodes_. the mean node strength over the region of interest ( roi ) is called the node strength of the region , @xmath49 the purpose of subtracting the minimum strength constant @xmath47 is to allow us to ignore short `` strands '' , which are often just transverse sections of trabeculae with an apparent width of more than one pixel . the trabeculae in our images have an apparent width of approximately two pixels , or 0.4 mm , which is higher than the true average trabecular width , due to the 1 mm thickness of the ct image and the 0.2 mm pixel size . thus , we wish to ignore strands of bone that are only two pixels long . the strength of such a strand depends on the pqct values of the two pixels . following eq . ( [ e : strandstr ] ) , the strength @xmath50 of such a strand with ct value @xmath51 is @xmath52 . based on examination of the resulting node strength images ( such as fig . [ fig : patch ] ( f ) ) , we find ad hoc a minimum strength constant of 225 ( corresponding to a ct value of 500 , eq . ( [ e : threshold ] ) , corresponding to a bmd value of 331mg/@xmath5 , which is higher than the mean ct value for trabecular bone ) . so strands of lower strength than this will be ignored by the algorithm . we note that our complete algorithm involves two thresholding steps : one at the beginning , when pqct values are converted to ct values of bone ; and one at the end , when the minimum strength constant is subtracted . these are not equivalent to one thresholding step with a higher threshold . the purpose of the first threshold filtration is to ignore marrow . the purpose of the second threshold filtration is to ignore short strands , and it can be seen as an alternative to skeletonisation . numerical experiments have shown that relative ( cross - subject ) values of ndstr are stable with respect to the choice of the bone threshold value ( @xmath53 ) and minimum strength constant . specifically , if either of these constants is varied by @xmath54 from the values used in this article , and ndstr is recomputed for all subjects , the resulting ndstr values are very strongly linearly correlated with the original ndstr values ( pearson correlation coefficient @xmath55 in all cases ) . the absolute values of ndstr depend on the choice of these constants : increasing the bone threshold by @xmath56 leads to a @xmath57 mean decrease in ndstr ( percent decrease averaged over all subjects ) , while increasing the minimum strength constant by @xmath56 leads to an average decrease in ndstr of @xmath58 . nevertheless , and as already mentioned , the relative ( cross - subject ) variation of ndstr remains constant . a key property of ndstr is that it depends on both the geometry of the trabecular network and the ct values of the trabeculae . ndstr depends linearly on thresholded ct values , _ if _ the ct values of all pixels are varied by the same factor . it follows that images with the same geometry but different bmd will have different ndstr values . on the other hand , two images with the same bmd but different geometry can have different ndstr values . this is apparent from the description of the method , and confirmed by figs . [ fig4 ] and [ fig5 ] , as discussed below , and also by the results discussed in section [ sectresults ] . to illustrate the analytical method , we now present the results in visual form for an enlarged region near the bottom ( posterior ) of the slice in fig . [ fig : tibia ] . the enlarged region of the original image is shown in fig . [ fig : patch](a ) . parts ( b ) to ( e ) of fig . [ fig : patch ] show directional strand strengths , and part ( f ) shows the final node strength plot . each directional strand strength plot shows the sum of two strand strengths at every pixel , in opposite directions : east / west ; north / south ; northwest / southeast ; and northeast / southwest . in each of the directional strand strength plots , the strands in the given direction are shown with the highest intensity , but most of the trabeculae are still visible , even if faintly . in contrast , in the node strength plot ( part ( f ) ) , most of the trabeculae are invisible . this is because of the subtraction of the minimum strength constant . in this example , there are almost no nodes in the right half of the image . this correctly describes the micro - architecture of the original image in that region , which contains many trabeculae but few that cross each other to make a lattice - like micro - architecture . the left half of the image contains many nodes . notice that , in the node strength plot , the nodes seem to be thicker than in the original image . this is because the trabeculae in the original image are actually slightly thicker than they appear , the outer pixels being dimmer ( i.e. lower ct values ) and thus not easily registered by the eye . since the outer pixels near the apparent nodes in the original image are almost as well - connected as pixels in the centres of the nodes , they have large node strengths , and are very visible in the node strength plot . the two specimens depicted in figs . [ fig4 ] and [ fig5 ] have comparable bone mineral densities , but their trabecular micro - architecture are visibly different . in fig . [ fig4 ] , the specimen has a trabecular bmd of 107mg/@xmath5 , which is near the median value ( 97mg/@xmath5 ) of the specimens in this study . on the left is the original image , and on the right is the node strength plot . notice that there are a lot of nodes in most of the outer areas , with the notable exception of a region near the bottom left . the mean node strength is 71.2 . the specimen in fig . [ fig5 ] has a trabecular bmd of 94 mg/@xmath5 , which is only 12% lower than that of the specimen shown in fig . [ fig4 ] , but it has substantially fewer nodes . the mean node strength is only 42.2 , which is 40% lower than that for the specimen shown in fig . this reflects the lack of a strong lattice - like micro - architecture in the original image . the long range node strength of each pixel of each image was computed as described in the previous section , and a mean node strength , ndstr , was calculated for each image . we also computed the bone mineral density ( bmd ) from the pqct slices , as explained in section [ sectanalysis ] . a scatter plot showing ndstr versus bmd for all 26 specimens is shown in fig . [ fig : nodesbmd ] . there is a strong positive correlation , which we quantified in three ways . firstly , pearson s correlation coefficient is @xmath59 , indicating a very strong linear correlation . secondly , since the scatter plot clearly suggests a nonlinear relationship , we fitted an exponential curve to the data and then found pearson s correlation coefficient to be @xmath60 . thirdly , the spearman rank correlation is @xmath61 , indicating a very strong correlation . the spearman rank correlation coefficient is a robust nonparametric correlation measure that is appropriate when little is known about the distributions and nature of the correlation between the variables . we also compared ndstr with the node - terminus ratio , nd / tm , in the topological node - strut analysis introduced by garrahan et al . @xcite the two measures are similar in philosophy , because they both quantify the nodes in the trabecular network . however , the definition of nd / tm is highly localised : after the skeletonisation process has eroded the trabecular network to a thickness of one pixel , each pixel is classified as node , strut , or terminus depending on its @xmath0 grid of nearest neighbours ( including the original pixel ) . for the present study , the images used to compute nd / tm have a pixel size of @xmath62 m , so the classification is made on the basis of a 30@xmath3 m @xmath4 30@xmath3 m region . in contrast , node strength is semi - global , taking into account longer strands to a degree controlled by the transmission strength constant . in the present study , with this constant set to @xmath63 , the method has a characteristic length of 4 mm , in the sense described in section [ sectanalysis ] . the node - terminus ratio , nd / tm , was calculated using histomorphometry performed on bone biopsies as described in section [ sectanalysis ] . recall that these biopsies were from the same regions of the same donors as the pqct slices from which ndstr has been computed . pearson s correlation coefficient for the relationship between ndstr and nd / tm is @xmath64 , and the spearman rank correlation coefficient is @xmath65 . we also measured the correlation between nd / tm and trabecular bmd : pearson s correlation coefficient is @xmath66 , and the spearman rank correlation coefficient is @xmath67 . using either measure of correlation , mean node strength is more strongly correlated with trabecular bmd than the node - terminus ratio is correlated with either of these variables . ) . ] we have introduced a new morphometric measure for characterising the micro - architecture of trabecular bone , _ long range node strength _ , which measures the degree to which a pixel in a 2-dimensional bone image has long - range connectivity in three or more directions , each at least 90 degrees from the others . in addition , we have calculated the mean node strength , ndstr , for each of the 26 bone samples considered in the study . we have found that ndstr has a strong positive correlation with trabecular bmd ( @xmath60 , after exponential transformation ) . furthermore , we have ascertained a strong correlation ( @xmath68 ) between ndstr and the established histomorphometric measure , node - terminus ratio ( nd / tm ) . moreover , qualitative comparison of images with similar trabecular bmd but different mean ndstr ( see figs . [ fig4 ] and [ fig5 ] ) suggest that ndstr successfully quantifies how `` lattice - like '' the micro - architecture is . further studies , including either clinical data or measured bone strength , are needed in order to determine the utility of this measure relative to other existing morphometric measures . such studies could also determine the most useful choice of the parameters that appear in the algorithm : the transmission constant , minimum strength constant , and bending coefficients . for example , these constants could be chosen to maximise correlations with bone strength . in the future , the sensitivity of the method on these parameters and on different ct settings ( e.g. , ct pixel size , slice thickness , mean ct value of bone ) should be systematically investigated . moreover , a future study on synthetic data could be used to verify that ndstr measures structure and not just bone volume or mass , and specifically that it finds nodes . also different skeletal sites should be analysed with our approach in order to test its potential to describe structural differences . we would like to note that our algorithm for computing node strength is dependent on pixel size , and is thus unsuitable for absolute comparisons between studies . however , this limitation is not unique to the node strength measure . for example , guggenbuhl _ et al . _ have showed using ct images with different thickness ( 1 mm , 3 mm , 5 mm , and 8 mm ) that the outcome of texture analysis depends substantially on the slice thickness@xcite . in the present study we have used pqct equipment with an in - plane pixel size of 200 @xmath3 m @xmath4 200 @xmath3 m and a slice thickness of 1 mm . we have not conducted a formal investigation into the influence of slice thickness on the long range node strength , but it is fair to assume that the long range node strength is similarly affected by the slice thickness . if further studies were to confirm the practical value of the measure , an implementation could be developed to produce node strengths that are broadly comparable between images with different pixel size . however , the technological development of high - resolution pqct equipment is progressing very quickly , and already high - resolution pqct scanners exist that can image a tibia at an isotropic pixel size of approximately 100 @xmath3 m , which is comparable to the trabecular thickness in the human proximal tibia@xcite , thus making the slice thickness less of an issue . some previous studies have investigated the trabecular bone micro - architecture using texture analysis applied on x - ray images of bone@xcite . however , these techniques are not comparable to the method presented in the present study as they are based on projections of a 3-dimensional trabecular network on a 2-dimensional plane , whereas our method is applied to 2-dimensional sections obtained through the 3-dimensional trabecular network . nevertheless , ranjanomennahary _ _ very recently showed some significant correlation did exist between radiograph based texture analysis and @xmath3ct based unbiased 3-dimensional measures of trabecular micro - architecture@xcite . _ compared 3-dimensional measures of micro - architecture based on 3-dimensional synchrotron radiation ct with more than 350 texture parameters obtained from simulated radiographs that were created from the synchrotron radiation ct data sets@xcite . they found using multiple regression analysis that a combination of a subset of texture analysis parameters correlated to the 3-dimensional measures of micro - architecture . however , they had to use at least three texture analysis parameters in the correlation with each of the 3-dimensional measures of micro - architecture . investigated ct images of the distal radius using texture analysis@xcite . the ct image used by cortet _ _ used similar pixel size to those used in the present study . they analysed the ct images using the traditional node - strut analysis and texture analysis including the grey level run length method@xcite . as illustrated in the present work there is a moderate correlation ( @xmath69 ) between the node - strut analysis and the ndstr measure , which we believe illustrates that the ndstr measure captures somewhat different information from that obtained with the node - strut analysis grey level run length is based on runs that have exactly the same grey level , and is therefore very sensitive to the choice of discretisation of the grey levels . in contrast , the ndstr measure treats the grey level as a continuous parameter , and thus it is able to detect runs with variable grey levels , and is consequently less sensitive to discretisation choices . another advantage of the method is that the proposed method utilises all of the ct value information in the pqct images , as no binarization of the images is performed . in the present study we have not applied the developed method to the histological sections directly . the reason for this is that the histological sections only cover a limited area of interest and thereby a limited trabecular length , which renders the ndstr less meaningful . however , this is a limitation of the histological examination procedure where it is only feasible to investigate a smaller sample ( biopsy ) of a larger structure ( the proximal tibia ) and not a limitation of the proposed method . in addition , the histological images are segmented into bone and marrow by their nature and it is thus not possible to assign an intensity value to pixels , analogous to a ct value , which is needed by the algorithm in its present form . however , it is possible to apply the concept of long range node - strut analysis to such binary images but this is outside the scope of the present investigation . in the present study the long range node - strut analysis was applied to pqct images of the proximal tibia . however , we would like to stress that the method is not limited to this skeletal site and thus can be applied to 2-dimensional ct images obtained from any skeletal site like e.g. the vertebral body or the calcaneous . finally , we note that the general method of long range node - strut analysis provides more than just the mean node strength . in the present study , we have focused on mean node strength for simplicity , but the intermediate measures of _ directional strand strength _ , used in the computation of node strength , may be useful in themselves as a measure of directional strength . in the present study the long range node - strut analysis has been applied to 2-dimensional pqct images obtained in the horizontal plane . however , the trabecular micro - architecture of the proximal tibia is mostly isotropic in the horizontal plane , whereas the micro - architecture in vertical direction is highly anisotropic compared with the horizontal plane@xcite . as @xmath3ct scanners become more and more prevalent and as the pixel size and imaging capacity of pqct scanners steadily improve it will be an important future task to generalise the long range node - strut analysis into three dimensions and to apply the technique to 3-dimensional data sets obtained from such equipment . in particular , the directional strand strength could be used to investigate anisotropic differences of the trabecular micro - architecture of such 3-dimensional data sets . the data acquisition parts of this project were made possible by grants from the microgravity application program / biotechnology from the manned spaceflight program of the european space agency ( esa ) ( esa project # 14592 , map ao-99 - 030 ) . the authors would like to thank professor g. bogusch and professor r. graf , center for anatomy , charit berlin , germany for kindly providing the bone specimens . wolfgang gowin , formerly at campus benjamin franklin , charit berlin , germany , is gratefully acknowledged for preparing the bone specimens and harvesting the bone biopsies . erika may and martina kratzsch , campus benjamin franklin , charit berlin and inger vang magnussen , university of aarhus , denmark are acknowledged for their excellent technical assistance scanning the ct images and preparing the histological sections . 52ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1016/j.jbiomech.2003.12.001 [ * * ( ) , 10.1016/j.jbiomech.2003.12.001 ] link:\doibase 10.1016/j.bone.2004.09.023 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1016/j.actaastro.2005.01.007 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1016/j.physleta.2006.08.058 [ * * , ( ) ] link:\doibase 10.1140/epjst / e2007 - 00078-x [ * * , ( ) ] link:\doibase 10.1103/physreve.79.021903 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1111/j.1365 - 2818.2005.01469.x [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ , ed . ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( )	* purpose * we present a new morphometric measure of trabecular bone micro - architecture , called _ mean node strength _ ( ndstr ) , which is part of a newly - developed approach called _ long range node - strut analysis_. our general aim is to describe and quantify the apparent `` lattice - like '' micro - architecture of the trabecular bone network . * methods * similar in some ways to the topological node - strut analysis introduced by garrahan et al . , our method is distinguished by an emphasis on long - range trabecular connectivity . thus , while the topological classification of a pixel ( after skeletonisation ) as a node , strut , or terminus , can be determined from the @xmath0 neighbourhood of that pixel , our method , which does not involve skeletonisation , takes into account a much larger neighbourhood . in addition , rather than giving a discrete classification of each pixel as a node , strut , or terminus , our method produces a continuous variable , _ node strength_. the node strength is averaged over a region of interest to produce the _ mean node strength _ ( ndstr ) of the region . * results * we have applied our long range node - strut analysis to a set of 26 high - resolution peripheral quantitative computed tomography ( pqct ) axial images of human proximal tibiae acquired 17 mm below the tibial plateau . we found that ndstr has a strong positive correlation with volumetric trabecular bone mineral density ( bmd ) . after an exponential transformation , we obtain a pearson s correlation coefficient of @xmath1 . qualitative comparison of images with similar bmd but with very different ndstr values suggests that the latter measure has successfully quantified the prevalence of the `` lattice - like '' micro - architecture apparent in the image . moreover , we found a strong correlation ( @xmath2 ) , between ndstr and the conventional node - terminus ratio ( nd / tm ) of garrahan et al . the nd / tm ratios were computed using traditional histomorphometry performed on bone biopsies obtained at the same location as the pqct scans . * conclusions * the newly introduced morphometric measure allows a quantitative assessment of the long - range connectivity of trabecular bone . one advantage of this method is that it is based on pqct images that can be obtained noninvasively from patients , i.e. without having to obtain a bone biopsy from the patient .
the density of states ( dos ) is a fundamental quantity from which many macroscopic quantities can be derived . indeed , once the dos is known , the partition function can be computed yielding the free energy of the system . it follows that the heat capacity , forces , etc can be derived . a well - known example of a macroscopic quantity that follows immediately from the knowledge of the electromagnetic dos is the casimir force@xcite . other examples are shear forces @xcite and heat transfer @xcite between two semi - infinite dielectrics . recently , it has been shown that unexpected coherence properties of thermal emission at short distances from an interface separating vacuum from a polar material are due to the contribution to the density of states of resonant surface waves @xcite . it has also been shown that the casimir force can be interpreted as essentially due to the surface waves contribution to the dos @xcite . calculating and measuring the local density of states ( ldos ) in the vicinity of an interface separating a real material from a vacuum is therefore necessary to understand many problems currently studied . the density of states is usually derived from the green function of the system by taking the imaginary part of the green s function@xcite . in solid - state physics , the electronic local density of states at the fermi energy at the surface of a metal can be measured with a scanning tunneling microscope ( stm ) @xcite . this has been proved by several experiments , in particular the so - called quantum corral experiments @xcite . although one can formally generalize the definition of the electromagnetic ldos by using the trace of the imaginary part of the green s tensor@xcite , it turns out that this definition does not yield the correct equilibrium electromagnetic energy density . recently , it has been shown theoretically that the stm and the scanning near - field optical microscope ( snom ) have strong analogies @xcite . more precisely , in the weak tip - sample coupling limit , it was demonstrated that a unified formalism can be used to relate the stm signal to the electronic ldos and the snom signal to electromagnetic ldos . snom instruments @xcite have been used to perform different kinds of emission spectroscopy , such as luminescence @xcite , raman spectroscopy @xcite or two - photon fluorescence @xcite . for detection of infrared light , apertureless techniques @xcite have shown their reliability for imaging @xcite as well as for vibrational spectroscopy on molecules @xcite . moreover , recent calculations and experiments have shown that an optical analog of the quantum corral could be designed , and that the measured snom images on such a structure present strong similarities with the calculated electromagnetic ldos @xcite . these results suggest that the electromagnetic ldos could be directly measured with a snom . the purpose of this article is to show how the electromagnetic ldos can be related to the electric green - function , and to discuss possible measurements of the ldos in snom . we first introduce a general definition of the electromagnetic ldos in a vacuum in presence of materials , possibly lossy objects . then , we show that under some well - defined circumstances , the ldos is proportional to the imaginary part of the trace of the electrical green function . the results are illustrated by calculating the ldos above a metal surface . we show next that the signal detected with a snom measuring the thermally emitted field near a heated body is closely related to the ldos and conclude that the natural experiment to detect the ldos is to perform a near - field thermal emission spectrum . we discuss the influence of the tip shape . we also discuss whether standard snom measurements using an external illumination can detect the electromagnetic ldos @xcite . as pointed out in the introduction , the ldos is often defined as being the imaginary part of the trace of the electric - field green dyadic . this approach seems to give a correct description in some cases @xcite , but to our knowledge this definition has never been derived properly for electromagnetic fields in a general system that includes an arbitrary distribution of matter with possible losses . the aim of this section is to propose an unambiguous definition of the ldos . let us consider a system at equilibrium temperature @xmath0 . using statistical physics , we write the electromagnetic energy @xmath1 at a given _ positive _ frequency @xmath2 , as the product of the dos by the mean energy of a state at temperature @xmath0 , so that @xmath3 where @xmath4 is planck s constant and @xmath5 is boltzmann s constant . we can now introduce@xcite a _ local _ density of states by starting with the local density of electromagnetic energy energy @xmath6 at a given point @xmath7 in space , and at a given frequency @xmath2 . this can be written by definition of the ldos @xmath8 as @xmath9 the density of electromagnetic energy is the sum of the electric energy and of the magnetic energy . at equilibrium , it can be calculated using the system green s function and the fluctuation - dissipation theorem . let us introduce the electric and magnetic field correlation functions for a stationnary system @xmath10 note that here , the integration over @xmath2 goes from @xmath11 to @xmath12 . if @xmath13 is the electric current density in the system , the electric field reads @xmath14 . in the same way , the magnetic field is related to the density of magnetic currents @xmath15 by @xmath16 . in these two expressions , @xmath17 and @xmath18 are the dyadic green functions of the electric and magnetic field , respectively . the fluctuation - dissipation theorem yields @xcite that @xmath19}\frac{\mu_0\omega}{2\pi}im g^{e}_{ij}({{\bf r}},{{\bf r}}',\omega ) \\ \mathcal{h}_{ij}({{\bf r}},{{\bf r}}',\omega ) & = & \frac{\hbar \omega}{\left[\exp(\hbar \omega / k_{b}t ) - 1\right]}\frac{\epsilon_0\omega}{2\pi}im g^{h}_{ij}({{\bf r}},{{\bf r}}',\omega ) \end{aligned}\ ] ] if one considers only the positive frequencies @xmath20 $ ] so that @xmath21}\frac{\omega}{\pi c^2}im tr\left[{{\stackrel{\leftrightarrow}{\bf g}}}^e({{\bf r}},{{\bf r}},{\omega})+{{\stackrel{\leftrightarrow}{\bf g}}}^h({{\bf r}},{{\bf r}},{\omega})\right ] \label{eq : imtr}\ ] ] it is important to note that the magnetic - field green function and the electric - green function are not independent . in fact , one has @xmath22.{{\stackrel{\leftrightarrow}{\bf g}}}^e({{\bf r}},{{\bf r}}',{\omega}).\left[\nabla_{{{\bf r}}'}\times\right]\ ] ] a comparison of eqs . ( [ eq : defldos ] ) and ( [ eq : imtr ] ) shows that the ldos of the electromagetic field reads @xmath23 = f({{\stackrel{\leftrightarrow}{\bf g}}}^e ) \label{eq : ldosimtr}\ ] ] in which @xmath24 is an operator which will be discussed more precisely in the next section . the goal of this section is to study the ldos behavior for some well - characterised physical situations , based on the result in eq . ( [ eq : ldosimtr ] ) . in a vacuum , the imaginary part of the trace of the electric and magnetic field green functions are equal . indeed , the electric and magnetic field green functions obeys the same equations and have the same boundary conditions in this case ( radiation condition at infinity ) . in a vacuum , the ldos is thus obtained by considering the electric field contribution only , and multiplying the result by a factor of two . one recovers the familiar result @xmath25 which shows in particular that the ldos is homogeneous and isotropic . let us consider a plane interface separating a vacuum ( medium 1 , corresponding to the upper half - space ) from a semi - infinite material ( medium 2 , corresponding to the lower half - space ) characterised by its complex dielectric constant @xmath26 ( the material is assumed to be linear , isotropic and non - magnetic ) . inserting the expressions of the electric and magnetic - field green functions for this geometry @xcite into equation ( [ eq : ldosimtr ] ) , one finds the expression of the ldos at a given frequency and at a given height @xmath27 above the interface in vacuum . in this situation , the magnetic and electric green functions are not the same . this is due to the boundary counditions at the interface which are different for the electric and magnetic fields . in order to discuss the origin of the different contributions to the ldos , we define and calculate an electric ldos ( @xmath28 ) due to the electric - field green function only , and a magnetic ldos ( @xmath29 ) due to the magnetic - field green function only . the total ldos @xmath30 has a clear physical meaning unlike @xmath28 and @xmath29 . note that @xmath28 is the quantity which is usually calculated and considered to be the true ldos . in the geometry considered here , the expression of the electric ldos is @xcite @xmath31\right.\nonumber\\ & + & \left.\int_{1}^{\infty}\frac{\kappa d\kappa}{\|p\| } \left[im(r_{12}^s)+\left(2\kappa^2 - 1\right)im(r_{12}^p)\right]e^{-2\|p\|{\omega}z / c}\right\ } \label{ldoselec}\end{aligned}\ ] ] this expression is actually a summation over all possible plane waves with wave number @xmath32 where @xmath33 if @xmath34 and @xmath35 if @xmath36 . @xmath37 and @xmath38 are the fresnel reflection factors between media 1 and 2 in @xmath39 and @xmath40 polarisations , respectively , for a parallel wave vector @xmath41 @xcite . @xmath42 corresponds to propagating waves whereas @xmath36 corresponds to evanescent waves . a similar expression for the magnetic ldos can be obtained : @xmath43\right.\nonumber\\ & + & \left.\int_{1}^{\infty}\frac{\kappa d\kappa}{\|p\| } \left[im(r_{12}^p)+\left(2\kappa^2 - 1\right)im(r_{12}^s)\right]e^{-2\|p\|{\omega}z / c}\right\ } \label{ldosmagn}\end{aligned}\ ] ] adding the electric and magnetic contributions yields the total ldos : @xmath44\right.\nonumber\\ & + & \left.\int_{1}^{\infty}\frac{\kappa^3 d\kappa}{\|p\| } \left[im(r_{12}^s)+im(r_{12}^p)\right]e^{-2\|p\|{\omega}z / c}\right\ } \label{ldos}\end{aligned}\ ] ] it is important to note that the electric and magnetic ldos have similar expressions , but are in general not equal . the expression of @xmath45 is obtained by exchanging the @xmath39 and @xmath40 polarisations in the expression of @xmath46 . as a result , the two polarisations have a symmetric role in the expression of the total ldos @xmath47 . the vacuum situation can be recovered from the previous expression by setting the values of the fresnel reflection factors to zero . the same result is also obtained by taking the ldos at large distance from the interface , i.e , for @xmath48 where @xmath49 is the wavelength . this means that at large distances , the interface does not perturb the density of electromagnetic states . in fact , @xmath50 becomes negligible for the evanescent waves and @xmath51 is a rapidly oscillating function for the propagating waves when integrating over @xmath52 . the result is that all the terms containing exponential do not contribute to the integral giving the ldos in the vacuum . conversely , at short distance from the interface , @xmath53 is drastically modified compared to its free - space value . equations ( [ ldoselec])-([ldos ] ) show that the fresnel coefficients and therefore the nature of the material play a crucial role in this modification . for example , as pointed out by agarwal @xcite , in the case of a perfectly conducting surface , the contribution of the electric and magnetic ldos vanish , except for their free - space contribution . in this particular case , one also retrieves the vacuum result . we now focus our attention to real materials like metals and dielectrics . we first calculate @xmath54 for aluminum at different heights . aluminum is a metal whose dielectric constant is well described by a drude model for near - uv , visible and near - ir frequencies @xcite : @xmath55 with @xmath56 10@xmath57 rad.s@xmath58 and @xmath59 10@xmath60 rad.s@xmath58 . we plotted in fig.[ldosz ] the ldos @xmath53 in the near uv - near ir frequency domain at four different heights . we first note that the ldos increases drastically when the distance to the material is reduced . as discussed in the previous paragraph , at large distance from the material , one retrieves the vacuum density of states . note that at a given distance , it is always possible to find a sufficiently high frequency for which the corresponding wavelength is small compared to the distance so that a far - field situation is retrieved . when the distance to the material is reduced , additional modes are present : these are the evanescent modes that are confined close to the interface and that can not be seen in the far field . moreover , aluminum exhibits a resonance around @xmath61 . below this frequency , the material supports resonant surface waves ( surface - plasmon polaritons ) . additional modes are therefore seen in the near field . this produces an increase of the ldos close to the interface . the enhancement is particularly important at the resonant frequency which corresponds to @xmath62=-1 $ ] . this behavior is analogous to that previously described in ref.@xcite for a sic surface supporting surface - phonon polaritons . also note that in the low frequency regime , the ldos increases . finally , fig.[ldosz ] shows that it is possible to have a ldos smaller than that of vacuum at some particular distances and frequencies . figure [ fig2 ] shows the propagating and evanescent waves contributions to the ldos above an aluminum sample at a distance of 10 nm . the propagating contribution is very similar to that of the vacuum ldos . as expected , the evanescent contribution dominates at low - frequency and around the surface - plasmon polariton resonance ) , where pure near - field contributions dominates . we now turn to the comparison of @xmath63 with the usual definition often encountered in the literature , which corresponds to @xmath64 . we plot in fig.[fig3 ] @xmath65 , @xmath66 and @xmath67 above an aluminum surface at a distance @xmath68 . in this figure , it is possible to identify three different domains for the ldos behaviour . we note again that in the far - field situation ( corresponding here to high frequencies i.e. @xmath69 ) , the ldos reduces to the vacuum situation . in this case @xmath70 . around the resonance , the ldos is dominated by the electric - field green contribution . conversely , at low frequencies , @xmath29 dominates . thus , fig.[fig3 ] shows that we have to be very careful when using the expression @xmath71 . above aluminum and at a distance @xmath72 , this approximation is only valid on a small range between @xmath73 and @xmath74 . in order to get more physical insight , we have calculated the asymptotic ldos behaviour in the three regimes mentioned above . as we have already seen , the far - field regime ( @xmath77 ) corresponds to the vacuum case . to study the near - field situation , we focus on the evanescent contribution as suggested by the results in fig.[fig2 ] . when @xmath78 , the exponential term @xmath79 is small only for @xmath80 . in this ( quasi - static ) limit , the fresnel reflection factors reduce to @xmath81 asymptotically , the expressions of @xmath64 and @xmath82 are latexmath:[\ ] ] where @xmath154 . comparing this expression and ( [ erecfourier]),then @xmath155 where @xmath156 . furthermore , using the fact that @xmath157 and defining @xmath158 @xmath159 let us denote @xmath160 the field radiated by the reciprocal current in @xmath161 . the dipole induced then writes @xmath162 and @xmath163 using the fact that for all dyadic @xmath164 and for all vector @xmath165 and @xmath166 @xmath167\cdot{{\bf b}}={{\bf a}}\cdot[{{\stackrel{\leftrightarrow}{\bf a}}}^t\cdot{{\bf b}}]\ ] ] that @xmath168 is a symetric dyadic ( @xmath169 ) , that @xmath143 is transverse to the direction @xmath170 and the definition of @xmath145 . @xmath171 d.w . pohl and d. courjon ( eds . ) , _ near - field optics _ ( kluwer , dordrecht , 1993 ) ; m.a . paesler and p.j . moyer , _ near field optics : theory , instrumentation and applications _ ( wiley - interscience , new - york , 1996 ) ; m. ohtsu ( ed . ) , _ near - field nano / atom optics and technology _ ( springer - verlag , tokyo , 1998 ) . f. zenhausern , m.p . oboyle and h.k . wickramasinghe , appl . * 65 * , 1623 ( 1994 ) ; y. inouye and s. kawata , opt . lett . * 19 * , 159 ( 1994 ) ; p. gleyzes , a.c . boccara and r. bachelot , ultramicroscopy * 57 * , 318 ( 1995 ) .	we propose in this article an unambiguous definition of the local density of electromagnetic states ( ldos ) in a vacuum near an interface in an equilibrium situation at temperature @xmath0 . we show that the ldos depends only on the electric field green function of the system but does not reduce in general to the trace of its imaginary part as often used in the literature . we illustrate this result by a study of the ldos variations with the distance to an interface and point out deviations from the standard definition . we show nevertheless that this definition remains correct at frequencies close to the material resonances such as surface polaritons . we also study the feasability of detecting such a ldos with apetureless snom techniques . we first show that a thermal near - field emission spectrum above a sample should be detectable and that this measurement could give access to the electromagnetic ldos . it is further shown that the apertureless snom is the optical analog of the scanning tunneling microscope which is known to detect the electronic ldos . we also discuss some recent snom experiments aimed at detecting the electromagnetic ldos .
in the 105 deg@xmath5 field of the _ kepler _ space telescope there are @xmath140 rr lyr stars brighter than 17th magnitude being observed at both long and short cadence ( _ i.e. _ , integrated exposure times of @xmath130-min and 1-min , respectively ) . with the notable exception of rr lyrae itself ( kolenberg _ et al . _ 2010 , 2011 ) little was known about the stars prior to the launch of telescope . several of the rr lyrae stars are c - type stars pulsating in the first overtone mode ( moskalik _ et al . _ 2011 , in preparation ) . the remainder are ab - type stars pulsating in the fundamental mode , approximately half of which exhibit the blazhko effect ( see benk _ et al . _ 2010 ) and show phase and amplitude variations of the light curves occurring on time scales of @xmath110 to hundreds of days . the other half of the rrab stars appear to oscillate in a remarkably constant manner and are the subject of our recent paper `` fourier analysis of non - blazhko ab - type rr lyrae stars observed with the _ kepler _ space telescope '' ( nemec _ et al . _ 2011 ) , the results of which are summarized here . elsewhere in these conference proceedings kinemuchi ( 2011 ) reviews the general properties of the _ kepler _ rr lyrae star observations , and kolenberg ( 2011 ) discusses the blazhko rr lyrae stars . lcclclr nr lyr & 3733346 & 12.683 & 0.6820264(2 ) & 54964.7381 & lc : q1-q5 & 18333 + v715 cyg & 3866709 & 16.265 & 0.47070494(4 ) & 54964.6037 & lc : q1-q5 & 18374 + v782 cyg & 5299596 & 15.392 & 0.5236377(1 ) & 54964.5059 & lc : q1-q5 & 18381 + v784 cyg & 6070714 & 15.370 & 0.5340941(1 ) & 54964.8067 & lc : q1-q5 & 18364 + kic 6100702 & 6100702 & 13.458 & 0.4881457(2 ) & 54953.8399 & lc : q0-q4 & 14404 + nq lyr & 6763132 & 13.075 & 0.5877887(1 ) & 54954.0702 & lc : q0-q5 & 18759 + fn lyr & 6936115 & 12.876 & 0.52739845(1 ) & 54953.2690 & sc : q0+q5 & 149925 + & - & - & 0.527398471(4 ) & 54953.2690 & lc : q1-q5 & 18338 + kic 7021124 & 7021124 & 13.550 & 0.6224926(7 ) & 54965.6471 & lc : q1 & 1595 + kic 7030715 & 7030715 & 13.452 & 0.6836137(2 ) & 54953.8434 & lc : q0-q5 & 18802 + v349 lyr & 7176080 & 17.433 & 0.5070740(2 ) & 54964.9555 & lc : q1-q5 & 18314 + v368 lyr & 7742534 & 16.002 & 0.4564851(1 ) & 54964.7828 & lc : q1-q5 & 18273 + v1510 cyg & 7988343 & 14.494 & 0.5811436(1 ) & 54964.6695 & lc : q1-q5 & 18394 + v346 lyr & 8344381 & 16.421 & 0.5768281(1 ) & 54964.9211 & lc : q1-q5 & 18362 + v350 lyr & 9508655 & 15.696 & 0.5942369(1 ) & 54964.7795 & lc : q1-q5 & 18326 + v894 cyg & 9591503 & 13.293 & 0.5713866(2 ) & 54953.5627 & lc : q1-q5 & 18362 + v2470 cyg & 9947026 & 13.300 & 0.5485894(1 ) & 54953.7808 & lc : q0-q5 & 18794 + v1107 cyg & 10136240 & 15.648 & 0.5657781(1 ) & 54964.7532 & lc : q1-q5 & 18373 + v838 cyg & 10789273 & 13.770 & 0.4802799(1 ) & 54964.5731 & lc : q1-q5 & 18241 + aw dra & 11802860 & 13.053 & 0.6872158(6 ) & 54954.2160 & sc : q0 & 14240 + & - & - & 0.687217(1 ) & 54954.2160 & lc : q1 & 1614 + & - & - & 0.6872158(2 ) & 54954.2160 & lc : q5 & 4474 + & - & - & 0.68721632(3 ) & 54954.2160 & sc : q5 & 135380 + after deriving pulsation periods , times of maximum light , and total amplitudes , and characterising the light curves using fourier decomposition methods , attention is turned to [ fe / h ] values and how the stars compare with other rr lyrae stars in our galaxy and in the large magellanic cloud . after establishing that there are systematic offsets between the @xmath6 and @xmath7 fourier parameters , the @xmath7 versions of the correlations established by kovcs , jurcsik , walker , sandage and others are used to derive physical characteristics for the stars . these include the dereddened colour @xmath8 , mean effective temperature , @xmath9 , surface gravity , log @xmath10 , and pulsational and evolutionary luminosities ( @xmath2 ) and masses ( @xmath3 ) . finally , the results are compared with new theoretical model results computed with the warsaw non - linear hydrodynamics code . for each of the 19 stars listed in * table 1 * the _ kepler _ raw - flux data acquired over the first 420 days of the telescope s operation ( q0-q5 ) was analysed . the mean apparent magnitudes of the stars ( from the _ kepler _ input catalog ) range from @xmath11 to 17.4 mag , with distances @xmath13 - 23 kpc in the direction @xmath12 . the last two columns of table 1 identify the particular @xmath13 data that was analyzed . * figure 1 * shows the raw - flux ( left panel ) and fully - transformed @xmath7-magnitude ( right panel ) as a function of time for one of the non - blazhko stars , v894 cyg ( graphs such as these are typical ) . the flux is in units of @xmath14/s / cadence ( log@xmath15 scale ) and 18911 long cadence ( lc ) aperture photometry measurements are plotted . since the 1.4-m telescope was rolled by 90@xmath16 every three months resulting in a given star being observed with different pixels every ` quarter ' , it was necessary to correct for sensitivity variations from quarter to quarter , and for drift within each quarter . the @xmath7 magnitudes were derived from the fully normalized and detrended photometry . the scalloped pattern seen at maximum and minimum light , and the moir pattern seen everywhere , result from the pulsation period for this star , 0.571 d , not being evenly divisible by the lc sampling period , @xmath130 min , giving rise to 27 observations per pulsation cycle . the constancy of the range in the total amplitude is remarkable , as is the continuous moir pattern . the pulsation period for each star ( fundamental mode ) was derived from the high - precision aperture photometry using the period04 period - finding program ( lenz & breger 2005 ) and a version of the clean program written by dr . seung - lee kim ( see nemec , walker & jeon 2009 ) . typically the lc data consisted of @xmath118000 points , and the short cadence ( sc ) observations of fn lyr and aw dra ( q0 and q5 ) amounted to another @xmath1150000 data points per star . the uncertainties in the derived periods are @xmath11 - 2@xmath17d , depending on the particular data set and pulsation period , @xmath18 . times of maximum light , @xmath19 , accurate to @xmath200.0005 d also were computed . the @xmath18 and @xmath19 values are summarized in columns 4 and 5 of table 1 . phased light curves for all 19 non - blazhko stars were plotted using the derived pulsation periods and times of maximum light . two such light curves are illustrated in the top panels of * figure 2. the light curve for fn lyr is based on 18338 lc data points acquired over @xmath1400 days in q1-q5 , and that on the right is for aw dra and is based on 135380 sc data points acquired over 90 days in q5 . when the light curves for all the stars are plotted one clearly sees variations in the ` risetimes ' ( _ i.e. _ , the time from minimum to maximum light , expressed in phase units ) , the stars with the shortest risetimes having the largest amplitudes and a tendency to have a bump on the rise to maximum light . also analyzed were new high precision ground - based @xmath6 photometry for three of the stars ( nr lyr , fn lyr and aw dra ) , and all sky automated survey ( asas ) @xmath21 photometry ( see pigulski _ et al . _ 2009 ) for nine of the brightest stars . the high - precision calculations were instrumental in establishing the @xmath6-@xmath7 offsets in the fourier parameters ( needed to permit use of the kovcs , jurcsik _ et al . _ relations ) ; and the asas - north data ( http://www.astrouw.edu.pl/asas/kepler ) , although not providing high - precision fourier parameters , were sufficient for deriving reliable loop diagrams in the hr - diagram , from which it is clear that in every case the non - blazhko stars are bluest when brightest ( as expected ) . by combining the historical data ( going back more than 100 years for fn lyr , nq lyr and aw dra ) with the _ kepler _ data we were also able to derive period change rates for some of the stars . in particular , the periods for fn lyr and aw dra were seen to be increasing . unfortunately the baseline of the _ kepler _ observations is still too short to derive significant period variations directly . finally , one of the non - blazhko stars , kic 7021124 , was discovered to be doubly - periodic and to have properties very similar to v350 lyr ( benk _ et al . both appear to be pulsating in the fundamental and second - overtone modes simultaneously . the period ratios are almost identical , @xmath22 for kic 7021124 , and 0.592 for v350 lyr . when masses are derived for these two stars using the petersen @xmath23 vs. @xmath24 diagram the best agreement with the models is obtained , in both cases , for a high luminosity and a high mass : @xmath25 and @xmath26 . fourier decomposition of the light curves was performed on the fully calibrated photometry for all the stars by fitting the following fourier series to the apparent magnitudes : @xmath27 , \\\ ] ] where @xmath28(@xmath29 ) is the apparent magnitude ( either @xmath7 for the _ kepler _ data , or @xmath30 for the ground based @xmath6-photometry ) , @xmath31 is the number of fitted terms , @xmath32 is the ( assumed single ) pulsation frequency of the star ( = @xmath33 ) , @xmath29 is the observed time of the observation ( bjd-54953 for the _ kepler _ data ) , @xmath19 is the time of maximum light used to phase the light curves ( so that maximum light occurs at zero phase ) , and the @xmath34 and @xmath35 are the amplitude and phase coefficients for the individual fourier terms . the fourier calculations were made with two fortran programs , one kindly provided by dr . gza kovcs , and the other by dr . pawel moskalik . the assumed pulsation periods were those derived either from the _ kepler _ data or from the period change rate analysis , and the @xmath19 values were calculated numerically . the fourier analyses were performed on various subsets of the data . the bottom panels of fig.2 show residual plots after fitting 38-term fourier descriptions to all the data . in both cases the standard deviations about the mean curve are below 1 mmag , specifically @xmath160 @xmath36magn , and the greatest differences occur where the slope is steepest . for the purposes of using extant fourier correlations to derive physical characteristics only the lower order fourier terms are needed . owing to the small risetimes as many as 100 or more terms could be needed for a normal distribution of the residuals . another way of looking at the data was to calculate fourier coefficients and parameters for each pulsation cycle . for the long cadence ( 30 min ) data there were @xmath124 observations per bin for an rrab star with a period of 12 hours , and fewer ( more ) observations per cycle for stars with shorter ( longer ) periods . none of the 19 stars exhibits the recently discovered `` period doubling '' effect , thus confirming the result found earlier by szab _ for the fourier calculations ` direct fourier fitting ' ( dff ) rather than ` template fourier fitting ' ( tff ) methods were used ( see kovcs & kupi 2007 ) , and for each star the resulting time series were plotted for four fourier parameters : @xmath37 , the first term in the fourier series ( see right panel of fig . 2 ) ; @xmath38 , the phase of the first term ; @xmath39 , the amplitude ratio @xmath40 ; and @xmath41 , the fourier parameter found by simon , kovcs and others to be one of the most significant variables for deriving physical characteristics . typical time series are illustrated in figure 3 * , on the left for the unmodulated rrab star nq lyr having intermediate amplitude ( @xmath37=0.279 mag ) and intermediate brightness , @xmath42 mag , and on the right for v783 cyg , a low - amplitude blazhko star with the shortest blazhko period ( @xmath43 d ) among the stars in the sample . for nq lyr linear trend lines were fit to each time series and in almost every case the slope is zero to within the systematic and random uncertainties . the mean values for the four parameters are as follows : @xmath44 , with residual standard error , @xmath45 ; @xmath46 , with @xmath47 ; @xmath48 , with @xmath49 ; and @xmath50 mag , with @xmath51 . since the other non - blazhko stars show random variations of the fourier parameters similar to those shown here for nq lyr these means and errors provide a measure of the typical uncertainties . inspection of the nq lyr time series in fig . 3 shows that the stability over the @xmath1420 d interval ( q0-q5 ) is exceptional , not only for each fourier parameter but for the ensemble of parameters . to better illustrate the remarkable stationarity of the light curves for all 19 sample stars a set of ` animated gif ' light curves was prepared and these are available in the electronic version of the mnras paper . in * table 2 * some of the resulting fourier coefficients and parameters are summarized . column 2 gives the standard deviation about the mean light curve ( see fig.2 ) , most of which are below one mmag . column 3 contains the first term in the fourier series , which correlates extremely well with the total amplitude . columns 4 and 5 contain amplitude ratios ( defined above ) and columns 6 and 7 contain phase parameters ( see simon 1988 and references therein ) . lccccccc nr lyr & 0.69 & 0.266 & 0.456 & 0.352 & 2.416 & 5.115 & -2.34 + v715 cyg & 1.74 & 0.338 & 0.479 & 0.358 & 2.314 & 4.901 & -1.44 + v782 cyg & 0.79 & 0.190 & 0.488 & 0.279 & 2.777 & 5.808 & -0.47 + v784 cyg & 0.96 & 0.234 & 0.487 & 0.253 & 2.904 & 6.084 & -0.14 + kic 6100702 & 0.66 & 0.209 & 0.493 & 0.279 & 2.743 & 5.747 & -0.35 + nq lyr & 0.65 & 0.280 & 0.471 & 0.356 & 2.389 & 5.096 & -1.83 + fn lyr & 0.64 & 0.380 & 0.444 & 0.350 & 2.322 & 4.818 & -1.90 + kic 7021124 & 1.10 & 0.283 & 0.512 & 0.351 & 2.372 & 5.060 & -2.08 + kic 7030715 & 0.71 & 0.231 & 0.494 & 0.303 & 2.683 & 5.606 & -1.66 + v349 lyr & 3.24 & 0.346 & 0.450 & 0.352 & 2.328 & 4.845 & -1.73 + v368 lyr & 1.57 & 0.405 & 0.464 & 0.341 & 2.272 & 4.784 & -1.53 + v1510 cyg & 0.82 & 0.345 & 0.473 & 0.355 & 2.389 & 5.068 & -1.83 + v346 lyr & 2.83 & 0.330 & 0.473 & 0.352 & 2.372 & 5.060 & -1.83 + v350 lyr & 1.67 & 0.340 & 0.485 & 0.342 & 2.389 & 5.124 & -1.84 + v894 cyg & 0.91 & 0.377 & 0.490 & 0.338 & 2.364 & 5.067 & -1.79 + v2470 cyg & 0.79 & 0.220 & 0.488 & 0.282 & 2.745 & 5.737 & -0.71 + v1107 cyg & 0.99 & 0.280 & 0.495 & 0.350 & 2.421 & 5.196 & -1.56 + v838 cyg & 1.22 & 0.393 & 0.465 & 0.349 & 2.300 & 4.853 & -1.56 + aw dra & 0.55 & 0.306 & 0.525 & 0.346 & 2.730 & 5.560 & -1.74 + it is of interest to compare the derived fourier parameters for the 19 non - blazhko stars in the _ kepler _ field ( derived using the @xmath7 photometry but transformed to @xmath6 values using the offsets ) with parameters derived for galactic globular clusters and for rr lyrae stars in the large magellanic cloud . * figure 4 * compares the non - blazhko stars with 177 rr lyr stars in several well - studied globular clusters ( from kovcs & walker 2001 ) . the same symbols and colour coding are used in all four panels . the * upper - left * panel is analogous to the @xmath18-@xmath52 diagram ( see fig.6 below ) and was used to derive [ fe / h ] values for the _ kepler _ stars . the two diagonal lines represent mean relations when the globular cluster data ( small black dots ) are sorted into two [ fe / h ] bins ( the ` bb ' subscript refers to the butler - blanco system , which approximates the carretta - gratton system ) : a metal - poor bin consisting of 19 stars ( surrounded by blue squares ) with metallicities between @xmath53 and @xmath54 dex and average @xmath55 dex ( bb - scale ) ; and an intermediate - metallicity bin consisting of 39 stars ( circled in red ) with [ fe / h]@xmath56 between @xmath57 and @xmath58 dex and average @xmath59 dex . the equations of the lines are : @xmath60 ( upper metal - poor bin ) and @xmath61 ( lower intermediate metallicity bin ) . four of the _ kepler _ non - blazhko stars are clearly richer than [ fe / h]@xmath56=@xmath59 dex , while the remainder are apparently more metal poor . the * upper - right * panel of fig . 4 shows that the _ kepler _ stars appear to be drawn from a distribution similar to that of the globular clusters . in particular , the _ kepler _ stars with low @xmath62 values are not unusual , except possibly that they all have relatively high @xmath39 values . since the globular cluster stars of higher metallicity ( red open circles ) all tend to reside on the right side of the diagram this separation is probably a metallicity effect , supporting our conclusion that v784 cyg , v782 cyg , kic 6100702 and v2470 cyg are metal rich . likewise , the * lower - left * panel shows that the majority of the _ kepler _ stars do not differ from the stars in globular clusters . note too that there is very little metallicity discrimination in this plane . the two stars located at the extreme upper edge of the envelope of the globular cluster distribution are aw dra and v784 cyg . finally , the * lower - right * panel shows close agreement between the phase parameters of the _ kepler _ and globular cluster rr lyr stars , which supports the conclusion drawn earlier ( lower - right panel in fig . 4 ) that there is a strong approximately - linear correlation between @xmath63 and @xmath41 . the diagram also shows very little dependence on metallicity . * figure 5 * compares the @xmath39 , @xmath62 , @xmath64 and @xmath65 fourier parameters for the 19 _ kepler _ non - blazhko rr lyr stars with those for the field rr lyr stars in the central regions of the lmc . the lmc data are from the massive ogle - iii catalog of variable stars by soszynski _ ( 2009 ) , which comprises almost 25000 rr lyr stars . all the parameters are @xmath6-band values , the lmc values having been transformed from @xmath66 to @xmath6 using the transformation equations given by morgan , simet & bargenquast ( 1998 ) , and the values for the _ kepler _ rr lyr stars transformed from @xmath7 to @xmath6 using the well - determined offsets from diagrams such as fig.6 below . the * upper - left * panel shows that the @xmath39 values for the _ kepler _ stars are typical for ab - type rr lyr stars and they all lie in the narrow range 0.45 to 0.51 . the period ranges of the rrab stars also are similar . where the lmc stars differ from the _ kepler _ stars is in the much larger @xmath39 range ; in particular , none of the _ kepler _ stars are among the lmc stars in the ` dropdown ' at @xmath67 . it is suspected that the majority of these are rrab blazhko variables . in the * upper - right * panel of fig . 5 the @xmath39 and @xmath62 amplitude ratios are compared . this graph puts into perspective the small ranges of @xmath39 and @xmath62 seen for the _ kepler _ non - blazhko rrab stars , both relative to that for the lmc rrab stars , and relative to the ranges seen for other types of rr lyr stars ( rrc , rrd and rre ) . the upper left panel of fig . 4 has already shown that the _ kepler _ stars sort into four metal - rich and 15 metal - poor stars , and that the metal - poor stars are similar to the rr lyr stars found in galactic globular clusters ( both ooi and ooii types ) ; however , the metal - rich stars have no counterparts , at least among the kovcs & walker ( 2001 ) sample of globular cluster rr lyr stars . in the * bottom * panels of fig . 5 ( both of which are metallicity diagnostics ) we see that both the metal - rich _ kepler _ rr lyr stars ( the clump of four stars at log@xmath68 and @xmath69 ) and the metal - poor _ kepler _ rr lyr stars ( the majority with lower @xmath64 values ) have counterparts in the lmc , and that the metal - rich stars appear to have relatively long periods for metal - rich rrab stars ( _ i.e. _ , the subgroup of rrab stars with shorter periods at a given @xmath65 ) . several different physical characteristics were established using equations derived by jurcsik ( 1998 , hereafter j98 ) , kovcs & walker ( 2001 ) , and sandage ( 2004 , 2006 ) . since these equations were originally established for @xmath6 data , and we have only @xmath7 data , it was necessary first to establish @xmath7 versions of the equations . the specific equations used to derive the physical quantities , and the detailed tables given in nemec _ et al . _ ( 2011 ) are not repeated here , but the results of their application are summarized in several graphs and discussed below . * figure 6 * shows four different metallicity diagnostic diagrams . the upper - left panel is the classical period - amplitude diagram , often called the ` bailey @xmath18-@xmath70 diagram ' but coined the ` oosterhoff - arp - preston @xmath18-@xmath70 diagram ' by sandage ( 2004 , hereafter s04 ) . at a given total - amplitude lower [ fe / h ] stars tend to have longer periods . spectroscopically calibrated diagonal lines , such as that shown in the lower right panel for the messier 3 rrab stars ( @xmath71 ) , can be constructed , from which offsets give [ fe / h ] values . since the pattern of the stars seen in all four panels is very similar all four variables can , in principle , be used to derive [ fe / h ] values . four different estimates of [ fe / h ] were calculated by nemec _ ( 2011 ) , based on the @xmath41 relation of j98 ( her equation 1 ) , the @xmath65 relation of s04 ( his equation 3 ) , the @xmath52 relation of s04 ( his equation 6 ) , and the ` risetime ' relation of s04 ( his equation 7 ) . the @xmath72 relations seem to be the most reliable and the [ fe / h ] values based on the j98 formula , after conversion to the zinn - west scale , are given in the last column of table 2 . recently acquired cfht 3.6-m espadons spectra confirm that kic 6100702 is indeed metal - rich , at [ fe / h ] = [email protected] dex , while aw dra is of intermediate metallicity , with [ fe / h ] = [email protected] ( nemec , ripepi , chadid _ et al . _ 2011 , in preparation ) . * figure 7 * summarizes the derived de - reddened colour information in the form of correlation diagrams . in all four panels the four more metal rich stars stand apart from the more metal - poor stars . for this reason they have been plotted with open squares . the linear correlations show that the reddest metal - poor stars ( those nearest the red edge of the instability strip ) are more metal poor , have longer periods and smaller total amplitudes , and are more luminous than the bluest metal - poor stars . the diagrams also show that at a given colour the four metal - rich stars ( plotted with open squares ) have shorter periods , smaller amplitudes and are less luminous than the metal - poor stars . not represented in these diagrams but computed and given in the full paper are absolute magnitudes ( @xmath6-passband ) , surface gravities , mean effective temperatures , distances , and luminosities and masses . both pulsational and evolutionary @xmath2 and @xmath3 values were computed . the @xmath2(puls ) were calculated using equations 16 and 17 of j98 , and in both cases the lower metallicity stars have the higher @xmath2(puls ) . the @xmath3(puls ) were calculated using equations 14 and 22 of j98 . the average mass for the four metal - rich stars is @xmath10.50 @xmath73 compared with the average mass for the more metal - poor stars @xmath10.60 @xmath73 . three different @xmath2(evol ) were calculated , using equations 8 , 10 and 12 of sandage ( 2006 ) . for the metal - poor stars all three @xmath2(evol ) are systematically larger than the @xmath2(puls ) values . for the four metal - rich stars the agreement is better but there is a wide range of @xmath2(evol ) owing to the linear or non - linear [ fe / h ] dependencies . regardless of which formula was used the most luminous stars have the lowest metallicities and there is internal consistence . it is not clear whether the @xmath2(puls ) or @xmath2(evol ) are correct . as was the case for the luminosities , the @xmath3(evol ) are all larger than the corresponding @xmath3(puls ) values . * figure 8 * compares the results with the locations of model zero - age horizontal branches and evolutionary tracks . the top two panels show hr diagrams with @xmath74 as abscissa and @xmath75 as ordinate , and two sets of zahb loci from dorman ( 1992 ) computed for different masses along the zahb . the more luminous horizontal branch assumes [ fe / h]@xmath76 dex , and the less luminous branch [ fe / h]@xmath77 dex . the numbers next to the symbols are the masses for the individual models , which are seen to be higher for the low-[fe / h ] tracks than for the high-[fe / h ] tracks . for both assumed metallicities oxygen enhanced and non - enhanced zahbs have been plotted the effect of increasing the oxygen to iron ratio from [ o / fe]=0 to 0.66 for the low - metallicity zahbs is to lower the luminosity and reduce the mass at a given temperature . for the high - metallicity tracks an oxygen enhancement from [ o / fe]=0 to 0.23 has little effect on the derived @xmath2 or @xmath78 . also plotted in the low-[fe / h ] case are the evolutionary paths away from the zahb for two masses , 0.66 and 0.68 @xmath79 . in both panels the non - blazhko rr lyr stars with low metallicities are represented by large black squares , the four high - metallicity stars are plotted with open squares , and the @xmath9 are the average of the kovcs & walker ( 2001 , hereafter kw01 ) and sandage ( 2006 , hereafter s06 ) values . in the top left panel of fig . 8 the luminosities and masses ( labelled in red ) of the _ kepler _ non - blazhko rr lyr stars were calculated with eqs . 17 and 22 of j98 and thus are based on pulsation theory . the @xmath2(puls ) and @xmath78(puls ) are seen to be systematically smaller than values derived from the zahb tracks ( for the appropriate metal abundance ) . the reddest non - blazhko rr lyr stars lie close to the fundamental mode red - edge and have the smallest amplitudes . this graph also shows blue and red edges of the instability strip for the fundamental mode ( red and blue solid lines ) and first - overtone mode ( red and blue dashed lines ) . the edges were calculated with the warsaw pulsation code assuming a mass of 0.65 @xmath80 . the _ kepler _ non - blazhko stars all lie in the fundamental mode region of the variability strip , and the smallest amplitude rr lyr stars ( the four metal - rich stars , kic 7030715 and nr lyr ) have locations near the fundamental red edge ( fre ) of the instability strip . as expected , all the stars near the fre have low @xmath62 values . in the top right panel of fig . 8 the luminosities were calculated with eq . 12 of s06 , and the masses ( labelled in red ) with eq . 15 of s06 ; thus they are evolutionary @xmath2 and @xmath78 values . in this case there is very good agreement with the stellar evolution models , as one expects since they are based on stellar evolution models . enhancing the oxygen to iron ratio by the plotted amounts makes little difference . it is unclear which are correct , the @xmath2(puls ) and @xmath78(puls ) , or the @xmath2(evol ) and @xmath78(evol ) . the mass and luminosity discrepancies go in the same direction as seen for cepheids . ( 2010 ) recently derived a dynamical mass @xmath78(dynam ) for a classical cepheid in a well detached , double - lined eclipsing binary in the lmc . the mass they derive is very accurate and favours @xmath78(puls ) . the reason for the discrepancies may be the same , as suggested by pietrzynski _ et al_. not enough mass loss has been taken into account in the evolution models . finally , the nemec _ ( 2011 ) paper concludes with a section on hydrodynamic models computed with the smolec and moskalik ` warsaw convective pulsation program ' . perhaps most relevant for the discussion above is the conclusion that in the @xmath18-@xmath70 diagram when [ fe / h ] is varied from @xmath81 to @xmath82 dex there is very little motion at a given amplitude . instead , motion to longer periods at a given amplitude occurs for stars of higher @xmath2 for a given mass , or for stars of lower mass for a given luminosity . this suggests that the main factors causing period shifts in the @xmath18-@xmath70 diagram seem to be the mass and luminosity , the latter being caused either by higher zahb mass or by post - zahb evolution . the main conclusions reached by nemec _ et al . _ ( 2011 ) concerning the _ kepler _ non - blazhko rr lyrae stars are : ( 1 ) fourier - based [ fe / h ] values have been derived for the stars ( using four different correlations ) , and four of the 19 stars ( v782 cyg , v784 cyg , kic 6100702 and v2470 cyg ) appear to be much more metal - rich than the others recent cfht spectra confirm that kic 6100702 is metal - rich with [ fe / h]@xmath83 dex , and aw dra is of intermediate metallicity with [ fe / h]=@xmath84 dex . ( 2 ) the star kic 7021124 is found to be doubly - periodic ( @xmath22 ) with characteristics similar to v350 lyr ( see benk _ 2010 ) ; ( 3 ) all the stars show remarkably constant light curves ( @xmath0 mmag ) over the 420-day interval of the q0-q5 observations ; ( 4 ) fn lyr and aw dra are found to have increasing periods ; ( 5 ) the asas - north photometry of nine of the brightest stars shows that their colours are bluest when brightest , as expected for rr lyrae stars ; ( 6 ) fourier - based physical characteristics for all the stars have been derived ; ( 7 ) distances range from 3 to 23 kpc ; ( 8) pulsational and evolutionary masses and luminosities are calculated for the stars , the latter tending to be higher in both cases . when the observed periods are compared with periods computed with the warsaw non - linear convective pulsation code better agreement is achieved assuming _ pulsational _ @xmath2 and @xmath3 values rather than the ( higher ) _ evolutionary _ @xmath2 and @xmath3 values . ( 9 ) finally , the warsaw models show that varying [ fe / h ] has a relatively small effect on the @xmath18-@xmath70 relation and that the main factors causing the period shifts must be luminosity and mass . funding for the _ kepler _ discovery mission is provided by nasa s science mission directorate . we acknowledge the entire _ kepler _ team for their outstanding efforts that have made these results possible . jmn thanks dr . andrew mcwilliam for the invitation to present these results in pasadena at such a stimulating conference .	this paper summarizes the main results of our recent study of the non - blazhko rr lyrae stars observed with the _ kepler _ space telescope . these stars offer the opportunity for studying the stability of the pulsations of rr lyrae stars and for providing a reference against which the blazhko rr lyrae stars can be compared . of particular interest is the stationarity of the low - dispersion ( @xmath0mmag ) light curves constructed from @xmath118000 long - cadence ( 30-min ) and ( for fn lyr and aw dra ) the @xmath1150000 short - cadence ( 1-min ) photometric data points . fourier - based [ fe / h ] values and other physical characteristics are also derived . when the observed periods are compared with periods computed with the warsaw non - linear convective pulsation code better agreement is achieved assuming _ pulsational _ @xmath2 and @xmath3 values rather than the ( higher ) _ evolutionary _ @xmath2 and @xmath3 values . [ 1996/06/01 ] 2@xmath4 2c ii 4c iv 2fe ii 3fe iii 1 mg i 2 mg ii 2si ii 4si iv 2al ii 3al iii 1o i 1n i 1h i =
the top quark reaches great interest since it breaks the @xmath3 symmetry maximally and it has rich decay products due to its large mass . the rare decays of the top quark have been studied in the literature , in the framework of the standard model ( sm ) and beyond @xcite-@xcite ; the one - loop flavor changing transitions @xmath4 in @xcite and @xmath0 in @xcite . in the sm , these decays are suppressed as a result of glashow - iliopoulos- miaiani ( gim ) mechanism @xcite . the branching ratios ( @xmath5 ) of the decays @xmath4 have been predicted in the sm as @xmath6 in @xcite . @xmath0 , which can give strong clues about the nature of electroweak symmetry breaking , has been calculated at the order of the magnitude of @xmath7 in the sm , in @xcite . these are small rates for the measurement , even at the highest luminosity accelerators and therefore , there is a need to analyse these rare decays in a new physics beyond the sm . in the present work , we study the flavor changing @xmath0 decay in the framework of the general two higgs doublet model ( model iii),where @xmath8 is the sm higgs boson . here , we take the yukawa couplings complex and switch on the cp violating effects . since the @xmath5 is at the order of @xmath9 in the sm , we neglect this contribution and calculate the new physics effects in the model iii . in the calculations , we take into account the interactions due to the internal mediating charged higgs boson @xmath10 and neglect the ones including internal neutral higgs bosons , @xmath11 and @xmath12 ( see discussion part ) . the numerical results show that the @xmath5 of this process can reach to the values of order @xmath13 , playing with the free parameters of the model iii , respecting the existing experimental restrictions . this prediction is almost seven orders larger compared to the one in the sm and it is a measurable quantity in the accelerators . furthermore , we predict the possible cp asymmetry @xmath14 at the order of @xmath15 for the intermediate values of the cp parameter , due to the complex yukawa coupling @xmath16 ( see section 2 for its definition ) . this is purely a new physics effect and the measurement of @xmath14 for the process under consideration may open a new window to go beyond the sm and test the new physics . the paper is organized as follows : in section 2 , we present the @xmath5 and @xmath14 of the decay @xmath17 in the framework of model iii . section 3 is devoted to discussion and our conclusions . the flavor changing transition @xmath0 is quite suppressed in the sm due to the glashow - iliopoulos - maiani ( gim ) mechanism @xcite . the extended higgs sector could bring large contributions to this decay and make cp violation possible , in general . this section is devoted to the calculation of the @xmath5 and the cp violating asymmetry of the decay under consideration , in the general two higgs doublet model , so called model iii . in the model iii , the flavor changing neutral currents in the tree level are permitted and various new parameters , such as yukawa couplings , masses of new higgs bosons , exist . the @xmath0 process is controlled by the yukawa interaction and in the model iii , it reads as @xmath18 where @xmath19 and @xmath20 denote chiral projections @xmath21 , @xmath22 for @xmath23 , are the two scalar doublets , @xmath24 are left handed quark doublets , @xmath25 are right handed up ( down ) quark singlets , with family indices @xmath26 . the yukawa matrices @xmath27 and @xmath28 have in general complex entries . it is possible to collect sm particles in the first doublet and new particles in the second one by choosing the parametrization for @xmath29 and @xmath30 as @xmath31\ , ; \phi_{2}=\frac{1}{\sqrt{2}}\left(\begin{array}{c c } \sqrt{2 } h^{+}\\ h_1+i h_2 \end{array}\right ) \,\ , . \label{choice}\end{aligned}\ ] ] with the vacuum expectation values , @xmath32 here , @xmath33 and @xmath34 are the mass eigenstates @xmath11 and @xmath12 respectively since no mixing occurs between two cp - even neutral bosons @xmath8 and @xmath11 at tree level , for our choice . the flavor changing ( fc ) interaction can be obtained as @xmath35 where the couplings @xmath36 for the fc charged interactions are @xmath37 and @xmath38 is defined by the expression @xmath39 here the index `` n '' in @xmath38 denotes the word `` neutral '' . the sm contribution to the @xmath5 of the process @xmath0 is negligibly small , which is at the order of the magnitude @xmath40 . therefore , we take into account only the new effects beyond the sm . the relevant diagrams are given in fig [ fig1 ] . at this stage , we would like to discuss the possibilities not to take into account the tree level contribution to the decay underconsideration , in the model iii . first , it can be assumed that all the off diagonal neutral yukawa couplings vanish and therefore the coupling @xmath41 vanishes . another possibility is to take the mixing between two neutral higgs bosons @xmath8 and @xmath11 is small and the tree level interaction @xmath42 is negligible . in our case , we consider the gauge and @xmath43 invariant higgs potential which spontaneously breaks @xmath3 down to @xmath44 as : @xmath45 ^ 2 + c_4 [ ( \phi_1^+ \phi_1 ) ( \phi_2^+ \phi_2)-(\phi_1^+ \phi_2)(\phi_2^+ \phi_1 ) ] \nonumber \\ & + & c_5 [ re(\phi_1^+ \phi_2)]^2 + c_{6 } [ im(\phi_1^+ \phi_2)]^2 + c_{7}\,\ , . \label{potential}\end{aligned}\ ] ] since we assume that only @xmath46 has vacuum expectation value , no mixing occurs between two neutral higgs bosons and the tree level interaction vanishes . furthermore , since we take the c - quark mass zero and the coupling @xmath41 negligible compared to the couplings @xmath47 and @xmath48 ( see @xcite ) , the main contribution comes from the diagrams with internal charged higgs boson and for the matrix element , we get @xmath49 where @xmath50 and the functions @xmath51 , @xmath52 , @xmath53 are defined as @xmath54 with @xmath55 . here @xmath56 and @xmath57 . using the eq . ( [ m1 ] ) , it is straightforward to obtain the decay width as @xmath58 now we would like to give the expression for @xmath14 of the above process . here , we take @xmath16 and @xmath47 complex with the parametrizations @xmath59 and assume that the complexity of @xmath47 is small . using the definition of the cp violating asymmetry @xmath14 @xmath60 we get @xmath61 where @xmath62 and the functions @xmath51 , @xmath52 and @xmath53 are given in eq . ( [ f1f2f3 ] ) . here the symbol @xmath63 denotes the complex conjugation . as it is shown from the eq . ( [ acp ] ) , @xmath14 vanishes when two cp violating angles @xmath64 and @xmath65 are equal . this is interesting , since @xmath14 can vanish even in the existence of complex phases , in the model iii . in this section , we study the @xmath66 , @xmath67 and @xmath68 dependencies of the @xmath5 and @xmath14 for the process @xmath0 , in the model iii . for the calculation of the @xmath5 we take the value of the total decay width @xmath69 as @xmath70 . notice that the coupling @xmath68 is defined as @xmath71 . the process @xmath0 exists also in the sm and model ii ( or i ) version of the 2hdm . in both models this process appears at least at loop level . in the sm model internal mediating @xmath72 bosons are responsible for this decay and its br is very small , at the order of the magnitude of @xmath40 . in the model ii , additional contribution comes from the charged higgs boson @xmath10 and can be enhanced compared to the sm result by playing with the free parameter @xmath73 , respecting the experimental measurements . however , no @xmath14 occurs in the model ii ( i ) and also in the sm . this forces one to go into the model iii , with complex and possibly large yukawa couplings . model iii contains large number of free parameters such as yukawa couplings , @xmath74 , the masses of new higgs bosons , @xmath10 , @xmath11 and @xmath12 , and they should be restricted by using experimental measurements . at this stage we summarize these restrictions : * we neglect all the yukawa couplings except @xmath75 and @xmath76 since they are negligible due to their light flavor contents , by our assumption . notice that we also neglect the off diogonal coupling @xmath77 , since it is smaller compared to @xmath78 ( see @xcite ) . therefore , the new neutral higgs bosons do not have any contribution to the @xmath5 of the decay under consideration . * we take @xmath75 and @xmath76 complex and respect the constraint for the angle @xmath65 and @xmath79 , due to the experimental upper limit of neutron electric dipole moment , @xmath80 , which leads to @xmath81 for @xmath82 gev @xcite . * we find a constraint region for these free parameters by restricting the wilson coefficient @xmath83 , which is the effective coefficient of the operator @xmath84 ( see @xcite and references therein ) , in the region @xmath85 . here upper and lower limits were calculated using the cleo measurement @xcite @xmath86 and all possible uncertainities in the calculation of @xmath83 @xcite . for completeness we present the wilson coefficient @xmath83 . denoting the contribution for the sm with @xmath87 and the additional charged higgs contribution with @xmath88 , we have the initial values @xmath89 the lo corrected coefficient @xmath90 is given as @xmath91 and @xmath92 , @xmath93 and @xmath94 are the numbers which appear during the evaluation @xcite . the explicit forms of the functions @xmath95 are , @xmath96 the discussion given above allows us to obtain a constraint region for the couplings @xmath75 , @xmath76 and the cp violating parameters , @xmath97 and @xmath66 . here , we assume that the coupling @xmath75 has a small imaginary part and we choose @xmath98 . notice that , in figures , the @xmath5 is restricted in the region between solid ( dashed ) lines for @xmath99 ( @xmath100 ) . here , there are two possible solutions for @xmath83 due to the cases where @xmath101 and @xmath102 . in the case of complex yukawa couplings , only the solutions obeying @xmath101 exist . in fig . [ brsintb ] , we plot the @xmath5 with respect to @xmath66 for @xmath103 , @xmath104 , @xmath105 , @xmath106 . as shown in this figure , the @xmath5 can reach to the values at the order of the magnitude of @xmath107 and it is not so much sensitive to the parameter @xmath66 . its magnitude is almost @xmath108 larger for @xmath99 compared to the one for @xmath100 . fig . [ brksibb ] represents the @xmath5 with respect to @xmath109 for @xmath103 , @xmath110 , @xmath104 and @xmath105 . this figure shows that the @xmath5 is strongly sensitive to the the coupling @xmath109 and it can get the values at the order of @xmath13 even for @xmath111 . this observation is an important clue about the upper limit of the coupling @xmath109 , with the possible future measurement of the @xmath5 . for the small coupling @xmath109 , the restricted region for the @xmath5 becomes narrow , for both @xmath99 and @xmath100 . finally , we show the @xmath67 dependence of the @xmath5 in fig . [ brmh0 ] for @xmath103 , @xmath110 , @xmath104 and @xmath112 . with the increasing values of @xmath67 , the @xmath5 decreases and the restriction region becomes narrower . at this stage , we would like to analyse the cp asymmetry @xmath14 of the decay @xmath0 and show @xmath66 and @xmath67 dependencies of @xmath14 in the figures [ acpsintb ] and [ acpmh0 ] , respectively . @xmath14 is at the order of the magnitude of @xmath15 for the intermeditate values of @xmath66 and it can reach to @xmath113 for @xmath99 . notice that @xmath97 is taken small and @xmath14 vanishes when two cp parameters have the same values , namely @xmath114 . if @xmath14 is positive ( negative ) , @xmath115 can have both signs . however , if it is negative , @xmath115 must be negative . this observation is useful in the determination of the sign of @xmath115 . the same behaviour is observed in fig . [ acpmh0 ] which represents the mass @xmath67 dependence of @xmath14 . when the higgs mass @xmath67 increases , an enhancement in @xmath14 is detected , especially for @xmath116 . now we will summarize our results : * the @xmath5 of the flavor changing process @xmath0 is at the order of @xmath40 in the sm and the extended higgs sector brings large contributions , at the order of @xmath117 , which can be measured in the future experiments . this ensures a crucial test for the new physics beyond the sm . * the @xmath5 is sensitive to @xmath109 and its measurement makes it possible to predict an upper limit for this coupling . furthermore , the measurement of the @xmath5 of the decay under consideration can give important information about the mass of higgs boson @xmath8 . * @xmath14 is at the order of the magnitude of @xmath15 for the intermeditate values of @xmath66 and it rises up to the values @xmath113 for @xmath99 . the measurement of @xmath14 can ensure a hint for the determination the sign of @xmath83 . therefore , the experimental investigation of the process @xmath118 will be effective for understanding the physics beyond the sm . this work was supported by turkish academy of sciences ( tuba / gebip ) . 1 g.mahlon and s.parke , _ phys . lett . _ * b347 * ( 1995 ) 394 . e. jenkins , _ phys . _ * d56 * ( 1997 ) 458 ; d. atwood and m. sher , _ phys . lett . _ * b411 * ( 1997 ) 306 . g.eilam , j.l.hewett and a.soni , _ phys . * d44 * ( 1991 ) 1473 , erratum _ phys . * d59 * 039901 ( 1999 ) . c.s.li , r.j.oakes and j.m.yang , _ phys . rev . _ * d49 * ( 1994 ) 293 ; j.m.yang and c.s.li , _ phys . _ * d49 * ( 1994 ) 3412 ; g.couture , c.hamzaoui and h.knig , _ phys . * d52 * ( 1995 ) 171 ; j.l.lopez , d.v.nanopoulos and r.rangarajan , _ phys . rev . _ * d56 * ( 1997 ) 3100 ; g.couture , m.frank and h.knig , _ phys . * d56 * ( 1997 ) 4213 ; g.m.de divitiis , r.petronzio , l.silvestrini , _ nucl . phys . _ * b504 * ( 1997 ) 45 . t.han , r.d.peccei , and x.zhang , _ nucl . _ * b454 * ( 1995 ) 527 ; t.han , k.whisnant , b .- l.young , and x.zhang , _ phys . _ * d55 * ( 1997 ) 7241 ; _ phys . _ * b385 * ( 1996 ) 311 ; m.hosch , k.whisnant , and b .- l.young , _ phys . rev . _ * d56 * ( 1997 ) 5725 ; e. malkawi and t. tait , _ phys . _ * d54 * ( 1996 ) 5758 ; t. tait and c.p . yuan , _ phys . rev . _ * d55 * ( 1997 ) 7300 ; t. han , m. hosch , k. whisnant , b .- young and x. zhang , _ phys . _ * d58 * ( 1998 ) 073008 ; f. del aguila and j.a . aguilar - saavedra , _ phys . _ * b462 * ( 1999 ) 310 ; f. del aguila and j.a . aguilar - saavedra , _ nucl . phys . _ * b576 * ( 2000 ) 56 . w.s.hou , _ phys . lett . _ * b296 * ( 1992 ) 179 ; k.agashe , m.graesser , _ phys . _ * d54 * ( 1996 ) 4445 ; m.hosch , k.whisnant , b.l.young , _ phys . rev . _ * d56 * ( 1997 ) 5725 ; j.guasch , hep - ph/9710267 . l.t.handoko and j.hashida , _ phys . * d58 * ( 1998 ) 094008 . g.eilam , b.haeri and a.soni , phys . d41 ( 1990 ) 875 . b. mele , s. petrarca and a. soddu , _ phys . * b435 * , ( 1998 ) 401 . t. han and jing jiang , b. mele , _ phys . lett . _ * b516 * , ( 2001 ) 337 . s. l. glashow , j. iliopoulos and l. maiani , _ phys . _ * d2 * ( 1970 ) 875 t. m.aliev , e. iltan , _ j. phys . _ * g25 * ( 1999 ) 989 . d. b. chao , k. cheung and w. y. keung , _ phys . * 59 * ( 1999 ) 115006 . s. chen , et . al . , cleo collaboration , _ phys . * 87 * ( 2001 ) 251807 . a. j. buras and m. mnz , _ phys . _ * d52 * ( 1995 ) 186 .	we study the flavor changing @xmath0 decay in the framework of the general two higgs doublet model , so called model iii . here , we take the yukawa couplings complex and switch on the cp violating effects . we predict the branching ratio six orders larger compared to the one calculated in the sm , namely @xmath1 , and observe a measurable cp asymmetry , at the order of @xmath2 .
current cosmological observations are very precise to give us a lot of information on the evolution and present state of the universe . we usually extract the information by constraining cosmological parameters such as energy densities of baryon , dark matter and dark energy , the hubble constant , reionization optical depth , spectral index of primordial fluctuation and so on . among them , in this paper , we focus on the primordial helium abundance @xmath0 , which has been of great interest in cosmology . one of the reasons why the primordial helium abundance has been considered to be interesting and important is that , in the context of the standard big bang nucleosynthesis ( sbbn ) , we can know the baryon density once @xmath0 is determined from observations . however , it has been discussed that a significant systematic error dominates when one infers the value of @xmath0 from measurements in low - metallicity extragalactic hii region @xcite . furthermore , there have been some discussions that there may be a large uncertainty in the neutron lifetime @xcite , which results in uncertainties in the predictions for the abundances of light elements . in this respect , the study of other independent measurements of the helium abundance would be interesting . recent precise measurements of cosmic microwave background ( cmb ) such as wmap can now enable us to constrain cosmological parameters with great accuracies . however , the helium abundance has not been discussed much in the literature when one study of cosmological constraints from cmb since @xmath0 has been considered to have little effect on cmb power spectrum . in most of analyses , @xmath0 is fixed to be @xmath6 which is probably motivated from a somewhat old value of the observed primordial helium abundance of @xmath7 @xcite @xcite . ] . but , in fact , there have been some works which discuss the effects of @xmath0 on cmb and some constraints were given @xcite . since the helium abundance affects the recombination history , the cmb power spectrum can be affected mainly through the diffusion damping . although the constraint is not so severe , it is important to notice that they are obtained independently from bbn , which can be used to cross - check our understanding of the helium abundance . furthermore , the value of @xmath0 at the time of bbn may be different from that at late time when cmb observations can probe - balls , the decay of @xmath8-balls produce extra baryon after bbn has completed @xcite . in a model of this kind , the value of @xmath0 can vary at different epochs . ] . after we studied the constraint on @xmath0 in @xcite , the data from wmap has been updated @xcite . in addition , the data at higher multipoles where the effects of @xmath0 become significant have been updated by acbar @xcite and cbi @xcite . thus it is a good time to investigate the constraint on the helium abundance using these cmb data , which is one of the aims of this paper . furthermore , we expect a more precise measurement of cmb from the future planck satellite @xcite , which can give us a much better constraint on @xmath0 . in fact , future constraint on @xmath0 has already been studied using the fisher matrix formalism @xcite . although this method is fast and usually adopted to predict the precision of the future measurements of cosmological parameters , it can give some inaccurate predictions in some cases , for example , when the likelihoods does not respect a gaussian form . in addition , the fisher matrix formalism predicts only the uncertainty for the parameter estimation since it just concerns with the derivatives with respect to parameters around the fiducial values . however , since some parameters are correlated in general , fixing the values of some parameters can bias the estimation of other parameters . namely , priors we assume on some parameters can cause the estimated central values to deviate from the input fiducial values , but such effects can not be quantified by the fisher matrix approach . thus , in this paper , we use the markov chain monte carlo ( mcmc ) approach to extract reliable future constraints on @xmath0 and other cosmological parameters . in particular , when we forecast the sensitivity for other cosmological parameters , we assume some different priors on @xmath0 and investigate to what extent the information of @xmath0 is important to determine other cosmological parameters . the organization of this paper is as follows . in the next section , the effects of @xmath0 on cmb power spectrum are briefly discussed . then in section [ sec : current ] , we study the constraint on @xmath0 using cmb data currently available including acbar , boomerang and cbi as well as wmap5 . in section [ sec : planck ] , we investigate a constraint on @xmath0 and other cosmological parameters from the future planck experiment . in addition , we also study how the prior on @xmath0 affects the determination of other cosmological parameters . a brief discussion on the significance of the uncertainties of the recombination theory in deriving cosmological constraint is given too . the final section is devoted to the summary of this paper . here we briefly discuss the effects of the primordial helium abundance @xmath0 on cmb power spectrum where @xmath9 with @xmath10 and @xmath11 being the number density of hydrogen and helium-4 respectively . as has been discussed in @xcite , the value of @xmath0 can affect the recombination history , which changes the structure of acoustic peaks . the effects of @xmath0 on acoustic peaks mainly come from the diffusion damping which suppresses the power on small scales and the shift of the position of acoustic peaks due to the change of the recombination epoch . before recombination , the number density of electron @xmath12 can be given by @xmath13 , where @xmath14 is the baryon number density . thus the increase of @xmath0 indicates the decrease of the number of electrons . when the number of electrons is reduced , the mean free path of the compton scattering becomes larger , which means that fluctuations on larger scales can be more affected by the diffusive mixing and rescattering . thus the damping scale below which fluctuation of photon is exponentially suppressed becomes larger . furthermore , due to the change of the number density of electrons , the epoch of recombination is also affected even though its effect is not so significant . this effect shifts the position of acoustic peaks slightly . in fig . [ fig : cl ] , we show the cmb @xmath15 spectrum with several values of @xmath0 . for reference , we also plot the current data ( left panel ) and the expected data from the future planck experiment ( right panel ) . as mentioned above , by increasing the value of @xmath0 , the power on small scales is damped more . in addition , it is noticeable that the position of acoustic peaks also shifts slightly . to characterize the effects of the change in @xmath0 and other cosmological parameters on cmb @xmath15 power spectrum @xmath16 , we consider some useful quantities @xcite . first of all , to see how cosmological parameters affect the position of acoustic peaks , we investigated the response of the position of the first peak by the change of the parameters , which we denote @xmath17 . in addition , to see the effects of the diffusion damping and some other effects by cosmological parameters , we study the height of the first peak relative to that at @xmath18 and the height of the second peak ( and higher peaks up to 5th peak ) relative to the first peak , which are denoted as @xmath19 , @xmath20 , @xmath21 , @xmath22 and @xmath23 , respectively . for clarity , we give the definitions of these quantities . the definition of @xmath19 is @xmath24 and the height of the @xmath25-th peak relative to the first peak is defined as @xmath26 where @xmath27 . we varied cosmological parameters including @xmath0 around a fiducial model and obtained partial derivatives by fitting linearly around the fiducial value . for the fiducial cosmological values , we assumed the mean values of the wmap 5-yr result ( wmap5 ) for a power - law @xmath28cdm model . regarding @xmath0 , we take @xmath29 which corresponds to the value obtained in the sbbn for the wmap5 baryon density . the resulting derivatives are : @xmath30 where @xmath31 and @xmath32 are energy densities of baryon and matter , @xmath33 is the scalar spectral index of primordial fluctuation , @xmath34 is the hubble constant in units of @xmath35 . in the formula for @xmath19 , we do not show the dependence on @xmath36 since its effect on @xmath19 is very small compared to that of the other parameters . as seen from the negative signs of @xmath37 for @xmath38 - 5 , the diffusion damping becomes more efficient as @xmath0 increases . we can also see the correlation of @xmath0 with other cosmological parameters , which can be useful when we interpret the results , in particular , for a planck forecast . the prior ranges for the parameters used in the analysis . priors shown in the 1st and 2nd columns are adopted for current and expected planck data , respectively . note that we adopt the prior range for @xmath0 shown above only in the cases with @xmath0 being treated as a free parameter whereas @xmath0 is a derived parameter in the case where we assume the sbbn relation . for the analysis with the current data , we also vary the amplitude of the sunyaev - zeldovich effect @xmath39 , which is omitted in the planck data analysis . we also include two additional parameters @xmath40 and @xmath41 which represent uncertainties in the theory of recombination ( see section [ sec : planck ] for more details ) . [ cols="<,^,^ " , ] we studied the constraint on @xmath0 and the effects of the priors for @xmath0 on constraining other cosmological parameters using current cmb data from wmap5 , acbar , boomerang and cbi , and also from the future planck experiment . after briefly reviewing the effects of @xmath0 on cmb , we studied current constraints on the primordial helium abundance . we obtained the current limit on @xmath0 from wmap5 alone as @xmath42 at 95%c.l . , which is improved to be @xmath43 at 68% ( 95% ) c.l . by adding the data of acbar , boomerang and cbi around the damping tail . we have also considered how the prior of @xmath0 can affects the constraints on other cosmological parameters using currently available data . we found that , at the present precision level of cmb measurements , the prior on @xmath0 has little effect for determinations of other cosmological parameters . we have also investigated the future constraint from the planck experiment . by performing a mcmc analysis , we derived an expected error for the helium abundance from the future planck experiment and found that it will be well measured with the accuracy of @xmath44 ( 68% c.l . ) in the planck experiment , which is 10 times smaller values compared with current data . furthermore , it may be interesting to notice that this precision is comparable to that obtained by hii region observations . as for the effects of the prior on @xmath0 on the determination of other cosmological parameters , we found that , with the precision of planck , the assumption on @xmath0 can affect the constraints on other cosmological parameters such as @xmath31 and @xmath33 . in this respect , the prior on @xmath0 can be important for determining the other parameters . in addition , the constraint on @xmath0 from cmb itself can be an independent test from other methods such as using hii region observations . in the near future , we can have more precise measurements of cmb . such upcoming data would give us more precise information of the primordial helium abundance . at the same time , it is necessary to study the effects of the helium abundance more rigorously in order to extract information of other cosmological parameters . * acknowledgments : * this work is supported in part by the sumitomo foundation ( t.t . ) and the grant - in - aid for scientific research from the ministry of education , science , sports , and culture of japan , no.18840010 ( k.i . ) and no.19740145 ( t.t . ) . t.s . would like to thank the japan society for the promotion of science for financial support . k. a. olive and e. d. skillman , astrophys . j. * 617 * , 29 ( 2004 ) [ arxiv : astro - ph/0405588 ] ; r. h. cyburt , b. d. fields , k. a. olive and e. skillman , astropart . * 23 * , 313 ( 2005 ) [ arxiv : astro - ph/0408033 ] . m. fukugita and m. kawasaki , astrophys . j. * 646 * , 691 ( 2006 ) . m. peimbert , v. luridiana and a. peimbert , arxiv : astro - ph/0701580 . y. i. izotov , t. x. thuan and g. stasinska , astrophys . j. * 662 * , 15 ( 2007 ) [ arxiv : astro - ph/0702072 ] . a. serebrov _ et al . _ , phys . b * 605 * , 72 ( 2005 ) [ arxiv : nucl - ex/0408009 ] . g. j. mathews , t. kajino and t. shima , phys . d * 71 * , 021302 ( 2005 ) [ arxiv : astro - ph/0408523 ] . a. p. serebrov , phys . lett . b * 650 * , 321 ( 2007 ) [ arxiv : nucl - ex/0611038 ] . b. d. fields and k. a. olive , astrophys . j. * 506 * , 177 ( 1998 ) [ arxiv : astro - ph/9803297 ] . r. trotta and s. h. hansen , phys . rev . d * 69 * , 023509 ( 2004 ) [ arxiv : astro - ph/0306588 ] . g. huey , r. h. cyburt and b. d. wandelt , phys . d * 69 * , 103503 ( 2004 ) [ arxiv : astro - ph/0307080 ] . k. ichikawa and t. takahashi , phys . d * 73 * , 063528 ( 2006 ) [ arxiv : astro - ph/0601099 ] . k. ichikawa , m. kawasaki and f. takahashi , phys . lett . b * 597 * , 1 ( 2004 ) [ arxiv : astro - ph/0402522 ] . e. komatsu _ et al . _ [ wmap collaboration ] , arxiv:0803.0547 [ astro - ph ] . j. dunkley _ et al . _ [ wmap collaboration ] , arxiv:0803.0586 [ astro - ph ] . g. hinshaw _ et al . _ [ wmap collaboration ] , arxiv:0803.0732 [ astro - ph ] . r. s. hill _ [ wmap collaboration ] , arxiv:0803.0570 [ astro - ph ] . m. r. nolta _ et al . _ [ wmap collaboration ] , arxiv:0803.0593 [ astro - ph ] . c. l. reichardt _ et al . _ , arxiv:0801.1491 [ astro - ph ] . j. l. sievers _ et al . _ , arxiv : astro - ph/0509203 . [ planck collaboration ] , arxiv : astro - ph/0604069 . w. hu , m. fukugita , m. zaldarriaga and m. tegmark , astrophys . j. * 549 * , 669 ( 2001 ) [ arxiv : astro - ph/0006436 ] . w. c. jones _ et al . _ , astrophys . j. * 647 * , 823 ( 2006 ) [ arxiv : astro - ph/0507494 ] . f. piacentini _ et al . _ , astrophys . j. * 647 * , 833 ( 2006 ) [ arxiv : astro - ph/0507507 ] . t. e. montroy _ et al . _ , astrophys . j. * 647 * , 813 ( 2006 ) [ arxiv : astro - ph/0507514 ] . a. lewis and s. bridle , phys . d * 66 * , 103511 ( 2002 ) [ arxiv : astro - ph/0205436 ] . a. kosowsky , m. milosavljevic and r. jimenez , phys . d * 66 * , 063007 ( 2002 ) [ arxiv : astro - ph/0206014 ] . e. komatsu and u. seljak , mon . not . astron . soc . * 336 * , 1256 ( 2002 ) [ arxiv : astro - ph/0205468 ] . l. perotto , j. lesgourgues , s. hannestad , h. tu and y. y. y. wong , jcap * 0610 * , 013 ( 2006 ) [ arxiv : astro - ph/0606227 ] . p. j. zhang , u. l. pen and h. trac , mon . not . soc . * 347 * , 1224 ( 2004 ) [ arxiv : astro - ph/0304534 ] . m. g. santos , a. cooray , z. haiman , l. knox and c. p. ma , astrophys . j. * 598 * , 756 ( 2003 ) [ arxiv : astro - ph/0305471 ] . j. hamann , j. lesgourgues and g. mangano , jcap * 0803 * , 004 ( 2008 ) [ arxiv:0712.2826 [ astro - ph ] ] . s. burles , k. m. nollett and m. s. turner , astrophys . j. * 552 * , l1 ( 2001 ) [ arxiv : astro - ph/0010171 ] . j. chluba , j. a. rubino - martin and r. a. sunyaev , mon . not . * 374 * , 1310 ( 2007 ) [ arxiv : astro - ph/0608242 ] . j. chluba and r. a. sunyaev , arxiv : astro - ph/0702531 . e. r. switzer and c. m. hirata , phys . d * 77 * , 083006 ( 2008 ) [ arxiv : astro - ph/0702143 ] . c. m. hirata and e. r. switzer , phys . d * 77 * , 083007 ( 2008 ) [ arxiv : astro - ph/0702144 ] . e. r. switzer and c. m. hirata , arxiv : astro - ph/0702145 . s. seager , d. d. sasselov and d. scott , astrophys . j. suppl . * 128 * , 407 ( 2000 ) [ arxiv : astro - ph/9912182 ] . s. seager , d. d. sasselov and d. scott , astrophys . j. lett . * 523 * , l1 ( 1999 ) arxiv : astro - ph/9909275 .	we studied a constraint on the primordial helium abundance @xmath0 from current and future observations of cmb . using the currently available data from wmap , acbar , cbi and boomerang , we obtained the constraint as @xmath1 at 68% c.l . we also provide a forecast for the planck experiment using the markov chain monte carlo approach . in addition to forecasting the constraint on @xmath0 , we investigate how assumptions for @xmath0 affect constraints on the other cosmological parameters . may 2008 + .5 in * primordial helium abundance from cmb : + a constraint from recent observations and a forecast * .45 in kazuhide ichikawa@xmath2 , toyokazu sekiguchi@xmath3 , and tomo takahashi@xmath4 .45 in _ @xmath3 institute for cosmic ray research , university of tokyo , kashiwa 277 - 8582 , japan + @xmath5 department of physics and astronomy , university college london , gower street , london , wc1e 6bt , u.k . + @xmath4 department of physics , saga university , saga 840 - 8502 , japan _ .4 in
weak interaction processes play a decisive role in the early stage of the core collapse of a massive star @xcite . first , electron capture on nuclei in the iron mass region , starting after the core mass exceeds the appropriate chandrasekhar mass limit , reduces the electron pressure , thus accelerating the collapse , and lowers the electron - to - baryon ratio , @xmath3 , thus shifting the distribution of nuclei present in the core to more neutron - rich material . second , many of the nuclei present can also @xmath0 decay . while this process is quite unimportant compared to electron capture for initial @xmath3 values around 0.5 , it becomes increasingly competative for neutron - rich nuclei due to an increase in phase space related to larger @xmath4 values . however , @xmath0 decay on nuclei with masses @xmath5 have not yet been considered in core collapse studies @xcite . this is surprising since gerry brown pointed out nearly a decade ago @xcite that certain nuclei heavier than @xmath6 , like @xmath7co and @xmath8co , have very strong @xmath0-decay matrix elements making it conceivable that they can actually compete with electron capture . brown argued that this might have quite interesting consequences for the collapse . during this early stage of the collapse , neutrinos produced in both electron capture and @xmath0 decay still leave the star . therefore , a strong @xmath0-decay rate will cool the star without lowering the @xmath3 value . as a consequence , the @xmath3 value at the formation of the homologous core ( after neutrino trapping ) might be larger than assumed . this results in a smaller envelope , and less energy is required for the shock to travel through the material . following brown s suggestion , @xmath0 decay rates for nuclei in the mass range @xmath9 were investigated @xcite . these studies were based on the same strategy and formalism as already employed by the pioneering work in this field by fuller , fowler and newman ( commonly abbreviated by ffn ) @xcite . the important idea in ffn was to recognize the role played by the gamow - teller resonance in @xmath0 decay . other than in the laboratory , @xmath0 decay rates under stellar conditions are significantly increased due to thermal population of the gamow - teller back resonance in the parent nucleus ( the gt back resonance are the states reached by the strong gt transitions in the inverse process ( electron capture ) built on the ground and excited states , see @xcite ) allowing for a transition with a large nuclear matrix element and increased phase space . indeed , fuller et al . concluded that the @xmath0 decay rates under collapse conditions is dominated by the decay of the back resonance . in a more recent work , aufderheide et al . came to the same conclusion . inspired by the independent particle model , the authors of ref . @xcite estimated the @xmath0 decay rates in a similar fashion as the electron capture rates , and phenomenologically parametrized the position and the strength of the back resonance . this estimate was supplemented by an empirical contribution , placed at zero excitation energy , which simulates low - lying transition strength missed by the gt resonance . when extending the ffn rates to nuclei with @xmath5 , aufderheide et al . found indeed that the @xmath0 decay rates are strong enough to balance the electron capture rates for @xmath10 . nevertheless , these results have never been explored in details in core collapse calculations . in recent years the parametrization of the electron capture and @xmath0 decay rates as adopted by ffn and aufderheide et al . have become questionable due to experimental data @xcite and have been critizised on the basis of more elaborate theoretical models @xcite . we will show in this paper , that although the previous weak interaction rates for core collapse are systematically incorrect , the important observation that electron and @xmath0 decay rates balance each other for a certain range of @xmath3 values is indeed correct . our conclusions will be based on large - scale shell model calculations for several key nuclei which , due to aufderheide et al . @xcite , contribute most significantly to the electron capture and @xmath0 decay rates at various stages of the collapse . these shell model calculations reproduce the measured gt strength distributions for nuclei in the mass range @xmath11 very well @xcite . furthermore modern large - scale shell model calculations also agree with measured half - lifes very well . thus for the first time one has a tool in hand which allows for a reliable calculation of presupernova electron capture and @xmath0 decay rates . modern shell model calculations come in two varieties : large - scale diagonalization approaches @xcite and shell model monte carlo ( smmc ) techniques @xcite . the latter can treat the even larger model spaces , but has limitations in its applicability to odd - a and odd - odd nuclei at low temperatures , which does not apply to the former . more importantly the diagonalization approach allows for detailed spectroscopy , while the smmc model yields only an `` averaged '' gt strength distribution which introduces some inaccuracy into the calculation of the capture and decay rates . we will consistenly use in the following the shell model diagonalization approach to study these rates . due to the very large m - scheme dimensions involved , the gt strength distributions have been calculated in truncated model spaces which fulfill the ikeda sum rule . however , at the chosen level of truncation involving typically 10 million configurations or more , the gt strength distribution is virtually converged . as residual interaction we adopted the recently modified version of the kb3 interaction which corrects the slight inefficiencies in the kb3 interaction around the @xmath12 subshell closure @xcite . in fact the modified kb3 interaction i ) reproduces all measured gt strength distributions very well and ii ) describes the experimental level spectrum of the nuclei studied here quite accurately @xcite . as @xmath13 shell model calculations , i.e. calculations performed in one major shell , overestimate the experimental gt strength by a universal factor @xcite , we have scaled our gt strength distribution by this factor , @xmath14 . large - scale shell model calculations of the electron capture rates for key nuclei in the presupernova collapse have been reported already elsewhere @xcite ; see also the smmc results in @xcite . in these studies it became apparent that the phenomenological parametrization of the gt contribution to the electron capture and @xmath0 decay rates , as introduced in ffn @xcite and subsequently used by aufderheide et al . @xcite , is systematically incorrect . these authors have placed the centroid of the gt strength distribution at too high an excitation energy in the daughter nucleus for electron capture on even - even nuclei , while for capture on odd - a and odd - odd nuclei they underestimated the energy of the gt centroid noticeably . for capture on even - even nuclei this has comparably little effect as ffn overcompensate the misplacement of the gt centroid by a too large empirical contribution at zero excitation energy . typically the ffn rates are roughly a factor of 5 larger than the shell model rates for capture on even - even nuclei like @xmath15ni or @xmath16fe . for capture on odd - odd and odd - a nuclei the misplacement of the gt centroid makes the ffn rates about 1 - 2 orders of magnitude too large compared to the shell model rates . as a consequence , ffn and aufderheide et al . have noticeably overestimated the electron capture rates for the early stage of the supernova collapse . which consequences do the misplacement of the gt centroids have for the competing @xmath0 decays ? in odd - a and even - even nuclei ( the daughters of electron capture on odd - odd nuclei ) , experimental data and shell model studies place the back - resonance at higher excitation energies than assumed by ffn and aufderheide et al . correspondingly , its population becomes less likely at the temperatures available during the early stage of the collapse ( @xmath17 , where @xmath18 measures the temperature in @xmath19 k ) and hence the contribution of the back - resonance to the @xmath0 decay rates for even - even and odd - a nuclei decreases . due to aufderheide et al . @xcite , some of the most important @xmath0 decay nuclei ( defined by the product of abundance and @xmath0 decay rate and listed in tables 18 - 22 in @xcite ) are odd - odd nuclei . for these nuclei , all available data , stemming from ( n , p ) reaction cross section measurements on even - even nuclei like @xmath20fe or @xmath21ni , and all shell model calculations indicate that the back - resonance resides actually at lower excitation energies than previously parametrized . consequently , the contribution of the back - resonance to the @xmath0 decay rate of odd - odd parent nuclei should be larger than assumed in the compilations . we note that this general expectation has already been conjectured in ref . @xcite on the basis of ( n , p ) data available at that time . these authors have attempted to fit the data within a strongly truncated shell model calculation which then in turn has been used to predict a corresponding @xmath0 decay rate . this procedure is viewed as rather uncertain as i ) the large energy resolution in the data made its convolution into a @xmath0 decay rate imprecise and ii ) the shell model truncation level was too inaccurate in order to estimate reliably the contribution of other states than the back - resonance to the decay rate . these shortcomings can be overcome in recent state - of - the - art large scale shell model calculations . we have calculated the @xmath0 decay rates for several nuclei under relevant core collapse conditions ( @xmath22 , where @xmath23 measures the density in @xmath24 g/@xmath25 and temperatures @xmath26 ) . these nuclei include even - even ones ( @xmath27ti,@xmath28cr , @xmath29fe ) , odd - a nuclei ( @xmath30mn , @xmath31fe , @xmath32co ) and odd - odd nuclei ( @xmath33sc,@xmath34mn , @xmath35co ) . the selection has been made to include those nuclei which have been ranked as most important for core collapse simulations by aufderheide et al . in fact , with the rates of ref . @xcite these 15 nuclei contribute between @xmath36 and @xmath37 to the change of @xmath3 due to @xmath0 decay in the range @xmath38 . although the formula for the presupernova @xmath0 decay rate @xmath39 is well known ( e.g. @xcite ) , we have chosen to quote the basic result here as this allows for the easiest discussion of the improvement incorporated in our calculation compared to previous work . thus , where @xmath41 , @xmath42 , and @xmath43 are the electron energy , momentum , and chemical potential , and @xmath44 ; @xmath45 is the nuclear energy difference between the initial and final states , while @xmath46 is their gt transition strength . @xmath47 is the charge number of the parent nucleus , @xmath48 is the partition function , @xmath49}$ ] , while @xmath50 is the fermi function which accounts for the distortion of the electron s wave function due to the coulomb field of the nucleus . the values for the chemical potential are taken from @xcite . to estimate the rates at finite temperatures , the compilations employed the so - called brink hypothesis @xcite which assumes that the gt strength distribution on excited states is the same as for the ground state , only shifted by the excitation energy of the state . we have not used this approximation , but have performed shell model calculations for the individual transitions . our sum over initial states includes i ) explicitly the ground state and several excited states in the parent nucleus ( usually at least all levels below 1 mev excitation energy ) and ii ) all back - resonances which can be reached from the levels in the daughter nucleus below 1 mev excitation energy . as these back - resonances also include parent states below 1 mev , special care has been taken in avoiding double - counting . the partition function is consistenly summed over the same initial states . here a word of caution is in order . we have calculated the gt strength distributions using 33 lanczos iterations in all allowed angular momentum and isospin channels . this is usually sufficient to converge in the states at excitation energies below @xmath51 mev . at higher excitation energies , @xmath52 mev , the calculated gt strengths represent centroids of strengths , which in reality are split over many states . while this does not introduce uncertainties in the summing over the gt strengths ( the numerator in ( 1 ) ) , it might be inconsistent for the calculation of the partition function . however , this is practically not the case , as at the rather low temperatures of concern here the partition function is given by those states which have already converged in our model space . nevertheless there might be states outside of our model space ( intruder states ) which will be missed in our evaluation of the @xmath0 decay rates . but their statistical weight in both numerator and denominator in the rate equation ( 1 ) is small . although our calculations agree well with the experimental informations available ( excitation energies and gt transition strengths ) , we have replaced the shell model results by data whenever available . in fig . 1 we compare our shell model @xmath0 decay rates with those of ffn for selected nuclei representing the three distinct classes : even - even ( @xmath28cr , @xmath53fe ) , odd - a ( @xmath30mn , @xmath54fe ) and odd - odd ( @xmath34mn , @xmath16co ) . we note again that @xmath0 decay of the nuclei studied here is important at temperatures @xmath55 @xcite . for the odd - odd nuclei we calculate rates similar to those of ffn . this approximate agreement is , however , somewhat fortunate . in ffn the misplacement of the gt back - resonances has been compensated by too large values for the total gt strengths ( ffn adopted the unquenched single particle estimate ) and the low - lying strengths . at higher temperatures ( @xmath56 ) , the ffn rates for odd - odd nuclei are larger than our shell model rates . for odd - a and even - even nuclei our shell model rates are significantly smaller than the ffn rates as the back resonance , for @xmath57 , is less populated thermally than in the ffn parametrization . using the ffn rates , even - even nuclei were found to be unimportant for @xmath0 decay in the core collapse ; our lower rates make them even less important . this situation is somewhat different for odd - a nuclei which ( @xmath54fe , @xmath30mn ) have been identified as important in ref . @xcite adopting the ffn rate . aufderheide et al . have added several odd - a nuclei with masses @xmath5 ( which are not calculated in ffn ) to the list of those nuclei which significantly change the @xmath3 value during the collapse by @xmath0 decays . these nuclei include @xmath58fe and @xmath32co ; we will show below that their rates have also been overestimated significantly in ref . @xcite . our shell model rates indicate that the importance of odd - a nuclei is significantly overestimated when the previously compiled values are adopted . in ref . @xcite @xmath0 decay rates for several nuclei have been estimated in strongly truncated shell model calculations , in which these authors allowed a maximum of 1 nucleon to be excited from the @xmath59 shell to the rest of the pf - shell in the daughter nucleus , and fitted the single particle energy spectra to reproduce measured ( n , p ) data for @xmath34fe , @xmath16ni and @xmath30co ; the ( n , p ) data constrain the back - resonance transition to the ground states in the @xmath0 decays of @xmath34mn , @xmath16co and @xmath30fe . our shell model rates are compared to the estimates of ref . @xcite in fig . 2 . for the 3 odd - odd nuclei , the agreement is usually better than a factor 2 . this is due to the fact that these rates are dominated by the back - resonances , i.e. the ( n , p ) data of the daughter nucleus , which are reproduced in our large - scale shell model approach and have been fitted in ref . for the odd - a nucleus @xmath30fe our @xmath0 decay rate is about an order of magnitude lower than the estimate of ref . @xcite at @xmath60 , while the rates agree for @xmath61 , where it is dominated by the back - resonances . at the lower temperatures , @xmath0 decays of low - lying states are important which might be overestimated in the truncated calculation . what might the revised @xmath0 decay rates mean for the core collapse ? to investigate this question we study the change of the electron - to - baryon ratio , @xmath62 , along a stellar trajectory . following ref . @xcite , we define @xmath63 where the sum runs over all nuclear species present in the core . @xmath47 , @xmath64 , and @xmath65 are the charge , mass number , and mass fraction of the nucleus , respectively . the mass fraction is given by nuclear statistical equilibrium @xcite ; we will use the values as given in tables 14 - 24 of ref . noting that @xmath0 decay ( @xmath0 ) increases the charge by one unit , while electron capture ( ec ) reduces it by one unit , we have @xmath66 where @xmath67 and @xmath68 are the electron capture and @xmath0 decay rates of nucleus @xmath69 . for several key nuclei we have calculated these rates within large - scale shell model studies . some of the results are listed in tables 1 and 2 , where they are also compared to the ffn rates and the ones of ref . this comparison also includes the @xmath0 decay rates for @xmath58fe and @xmath32co , which , due to @xcite and earlier suggested by brown @xcite , are important when the stellar trajectory reaches electron - to - baryon values @xmath70 . our shell model rates agree for @xmath7co with the rate of aufderheide et al . , but are smaller than the estimates of these authors by factors 2 and 5 for @xmath58fe and @xmath58co , respectively . we note that the strong ground state decay of @xmath7co contributes about @xmath71 to the total rate at the condition listed in table 2 . some of the electron capture rates are taken from @xcite , while several other shell model rates are presented here for the first time ( e.g. for @xmath34fe , @xmath16ni and the odd - a nuclei ) . although the nuclei , for which reliable shell model rates are now available , include the dominant ones at the various stages of the early collapse ( due to the ratings in ref . @xcite ) , there are upto 250 nuclei present in nse at higher densities @xcite . although we are currently working at a revised compilation for @xmath0 decay and electron capture rates for nuclei in the mass range @xmath72 , its completion is computer - intensive and tedious . nevertheless some important conclusions can already be drawn from the currently available data . at first we will follow the stellar trajectory as given in ref . @xcite , although some comments about this choice are given below . we estimated @xmath73 and @xmath74 separately on the basis of the 25 most important nuclei listed in tables 14 - 24 in @xcite . we used shell model rates for the nuclei listed in table 1 and 2 . for the other nuclei we scaled the ffn rates using the following scheme which corrects for the systematic misplacement of the gt centroid and is derived by the comparison of ffn and shell model rates for the nuclei listed in tables 1 and 2 . the ffn electron capture rates have been multiplied by 0.2 ( even - even ) , 0.1 ( odd - a ) and 0.04 ( odd - odd ) , while the ffn @xmath0 decay rates have been scaled by 0.05 ( even - even ) , 0.025 ( odd - a ) and 1.5 ( odd - odd ) . the results for @xmath75 are plotted in fig . 3 , where they are also compared to the values obtained for the ffn rates . one observes that the shell model rates reduce @xmath76 significantly , by more than an order of magnitude for @xmath77 . this is due to the fact , that , except for @xmath78ni , all shell model electron capture rates are smaller than the recommendations given in the ffn and aufderheide et al . compilations @xcite . in particular , this is drastic for capture on odd - odd nuclei , which due to these compilations , dominate @xmath76 at densities @xmath79 . the shell model @xmath0 decay rates also reduce @xmath74 , however , by a smaller amount than for electron capture . this is mainly caused by the fact that the shell model @xmath0 decay rates of odd - odd nuclei are about the same as the ffn rates or even slightly larger , for reasons discussed above . it is interesting to note that ffn typically give higher @xmath0-decay rates for odd - a nuclei than aufderheide et al . @xcite , while it is vice versa for odd - odd nuclei . as a consequence @xmath80 is dominated by odd - a nuclei for @xmath81 if the ffn rates are used , while odd - odd nuclei contribute significantly if the rates of @xcite are adopted . in either case , both compilations yield rather similar profiles for @xmath75 ( see fig . 14 in @xcite ) . the important feature in fig . 3 is the fact that the @xmath0 decay rates are larger than the electron capture rates for @xmath82 , which is also already true for the ffn rates @xcite . so far we have used the same stellar trajectory as in ref . this allowed a comparison with the conclusions obtained in that reference . however , this assumption is inconsistent , and , in fact , was already inconsistent in @xcite . the chosen stellar trajectory is based on runs performed with the stellar evolution code kepler @xcite which uses the ffn electron capture rates , but quite outdated @xmath0 decay rates @xcite , following the old belief that @xmath0 decay rates are unimportant @xcite . the outdated @xmath0 decay rates were derived basically from a statistical model approach @xcite and are orders of magnitude too small @xcite . what are the consequences and will electron capture and @xmath0 decay rate also balance in a consistent model ? at the beginning of the collapse , electron capture is significantly faster than @xmath0 decay ( see fig . the shell model rates make @xmath78ni the most important contributor , but it can not quite compensate for the reduction of the @xmath83co rate . thus , at @xmath84 the total electron capture rate @xmath73 drops slightly . this reduction is more severe for smaller @xmath3 values , until at @xmath85 electron capture and @xmath0 decay balance . the consequence is that , due to the slower electron capture , the star radiates less energy away in form of neutrinos until @xmath85 is reached . thus one expects that in the early stage the stellar trajectory is , for a given density , at a higher temperature . this , of course , increases both the @xmath0 decay and electron capture rates . importantly both rates have roughly the same temperature dependence in the relevant temperature range : typically electron capture rates are enhanced by an order of magnitude if the temperature raises from @xmath86 to @xmath87 . but this increase is the same order of magnitude by which the @xmath0 decay rates grow in the same temperature interval . consequently the two rates will also be balanced at around @xmath88 if a consistent stellar trajectory is used . as stated above , the dominance of @xmath0 decay over electron capture during a certain stage of the core collapse of a massive star has been suggested or noted before @xcite . however , previous argumentation has been based on rates for these two processes which had been empirically and intuitively parametrized , rather than derived from a reliable many - body model . moreover , it was shown in recent years that the assumed parametrizations , mainly with respect to the energy of the gamow - teller centroid , were systematically incorrect . shell model calculations are now at hand which allow , for the first time , the reliable calculation of these rates under stellar conditions . given the fact that the large - scale shell model studies reproduce all important ingredients ( spectra , half - lives , gt strength distributions ) very well , the shell model rates are rather reliable . we stress an important point , that the shell model @xmath0 decay rates are larger than the electron capture rates for @xmath89 . this might have important consequences for the core collapse . a first investigation into these consequences has been performed by aufderheide et al . @xcite , however , using the ffn values for both rates . they find that the competition of @xmath0 decay and electron capture leads to cooler cores and larger @xmath3 values at the formation of the homologuous core . these results are important motivation enough to derive a complete set of shell model rates and then use them in core collapse calculations . this work was supported in part by the danish research council and through grant de - fg02 - 96er40963 from the u.s . department of energy . oak ridge national laboratory is managed by lockheed martin energy research corp . for the u.s . department of energy under contract number de - ac05 - 96or22464 .. grants of computational resources were provided by the center for advanced computational research at caltech . kl and gmp thank michael strayer and the oak ridge physics division for their kind hospitalty during which parts of the manuscript have been written . .electron capture rates for selected even - even , odd - a and odd - odd nuclei . the chosen stellar conditions reflect those at which the nuclei are considered to be most important due to the ranking given by aufderheide et al . @xcite . the shell model rates ( labelled sm ) are compared to those recommended by ffn @xcite and ref . the last column , named importance ratio , gives the percentage of the total change in @xmath62 ( for the definition see text ) assigned to the respective nucleus by aufderheide et al . at the respective stellar conditions . all rates are in s@xmath90 . exponents are given in parentheses . [ cols="^,^,^,^,^,^,^",options="header " , ]	we calculate supernova electron capture and @xmath0 decay rates for various @xmath1-shell nuclei using large - scale shell model techniques . we show that the centroid of the gamow - teller strength distribution has been systematically misplaced in previous rate estimates . our total electron capture rates are significantly smaller than currently adopted in core collapse calculations , while the total @xmath0 decay rates change less . our calculation shows that for electron - to - baryon ratios @xmath2 - 0.46 @xmath0 decay rates are larger than the competing electron capture rates . 2
the tully - fisher relation ( tully & fisher , 1977 ) is a direct indication of a close relationship between the detected baryons and the total mass in spiral galaxies . the detected baryons consist of the stellar and gaseous content , i.e. the visible mass , and this sets the luminosity profile of the galaxy while the total gravitational mass , which includes the dark matter content ( and possibly a component of baryonic dark matter ) , sets its rotation velocity . numerous studies have been carried out to investigate this relation , crucial in determining extragalactic distances ( e. g. pierce & tully 1988 , tully & pierce 2000 ) , in the study of evolution of galaxies ( e.g. puech et al . 2008 ) and also in giving constraints on cosmological galaxy formation models ( e. g. portinari & sommer - larsen 2007 ) . the tully - fisher relation is undoubtedly a crucial test for galaxy evolution models and although it has been the focus of a number of studies , its origin is still being debated . a few authors argue a cosmological origin ( e. g. avila - reese , firmani & hernndez 1998 ) while others suggest that this relation is regulated by star forming processes ( e. g. silk 1997 ) . on the other hand , the tully - fisher relation is related to the stellar populations of galaxies as it is suggested by its steeper slope when the luminosity is measured in the near infrared ( nir ) bands , when compared to the slope measured in the optical ( e.g. tully & pierce , 2000 ) . the use of nir bands in the tully - fisher relation has shown to be extremely useful , because nir bands present lower internal extinction than optical bands ( verheijen 2001 ) and the mass - to - light ratio is less contaminated by younger stellar populations , giving a better reflection of the stellar mass of the galaxies . luminosities can be converted into stellar masses by scaling them by a given mass - to - light ratio . thus , instead of linking rotation velocities to luminosities , a few authors have chosen to show the correlation of rotation velocities to stellar masses . this relation is called stellar tully - fisher relation ( e.g. bell & de jong 2001 ) . in order to estimate the whole baryonic mass , the gaseous component should be added to the stellar mass . the relation between luminosities and baryonic masses is called the baryonic tully - fisher relation . this relation has been studied by several authors ( e. g. mcgaugh 2000 , verheijen 2001 , geha et al . 2006 , bell & de jong 2001 ) . the slope of the baryonic tf relation is an important test for galaxy formation models . a steeper slope indicates that the baryonic mass of massive galaxies tends to approximately match the total mass of the galaxy . to date , most of the works devoted to study the tully - fisher relation have used compilations of several galaxy surveys , observed in different ways ( hi line profiles and rotation curves ) adding factors of uncertainties due to sample non - homogeneity . beside the problems of using different samples , the use of certain observational techniques may add other uncertainties in the study of the tully - fisher relation . for example , an over or under prediction of the position angle in long - slit observation could produce an erroneous estimate of the maximum velocity of a galaxy , which will be reflected in the tully - fisher relation . in order to avoid the problems listed in the previous paragraph , we have made use of the homogeneous galaxy survey gassendi halpha survey of spirals ( ghasp ) to study the nir , stellar and baryonic tully - fisher relations and their implications for the total mass of galaxies . the ghasp survey represents a large effort to constitute a sample of field galaxies in an homogeneous way . first , a strict isolation criterion has been used to insure the isolation of the galaxies . rather close galaxies have been chosen in order to guarantee a high spatial resolution , compared to hi surveys . high and low inclination objects have been excluded in order to minimize uncertainties in the de - projected rotation curve . second , all ghasp galaxies have been observed using the same instrument ; a scanning fabry - perot attached to a focal reducer at the 193 cm at observatoire de haute provence ( ohp ) . in addition to obtaining data for the whole sample with the same instrument , which is a great advantage to insure homogeneity , the scanning fabry - perot is certainly the most adapted instrument to obtain rotation curves in the most proper way . because it gives a 2d velocity field , we can obtain the rotation curve ( and the maximum velocity ) without any previous assumptions about the position angle or the inclination like it is the case for long - slit observations . this technique avoids , in this way , a great factor of uncertainty that is common to not be taken into account using other techniques . third , the data reduction has been performed in an homogeneous way , in order to derive the rotation curves from the velocity fields in the cleanest and most rigorous possible way ( see epinat et al . 2008a , b for details ) . these three points allowed eliminating ( or at least greatly minimizing ) several problems that previous studies have encountered . using rotation curves to obtain the maximum rotation velocity is a more precise way compared to the hi line width profile technique , used by several others studies . the higher spatial resolution of optical velocity maps , compared to hi , avoids the problem of missing the maximum velocity because of lack of resolution ( beam smearing ) . together with the kinematic information , we have used h and k - band photometry from 2mass survey , mass - to - light ratios derived from stellar population models , hi fluxes and h@xmath2 masses from the literature to perform , for the first time , the nir , stellar and baryonic tully - fisher relations for the ghasp sample . in 2 we describe the data , including the method used to compute the stellar , gaseous and baryonic masses . in 3 , we present the results . in 4 we discuss and compare our results with previous works . finally , we summarize our main findings in 5 . ghasp is the largest sample of spiral and irregular galaxies observed to date using fabry - perot techniques . it consists of 3d h@xmath3 data cubes for 203 galaxies , covering a large range in morphological types and absolute magnitudes . all the ghasp galaxies have been recently reanalyzed in a homogeneous way in epinat et al . ( 2008a , b ) . these authors published velocity fields , monochromatic h@xmath3 images , dispersion velocity maps , rotation curves and maximum rotation velocities ( v@xmath4 ) for each galaxy . a sub - sample of 93 galaxies has been selected by removing from the sample : 1 ) galaxies with radial systemic velocities lower than 3000 km s@xmath5 ( to avoid the effect of the local group infall ) for which no other individual measurements of distances were available ( the references are indicated in epinat et al . 2008b ) and 2 ) galaxies with inclinations lower than 25 degrees for which the uncertainties on the rotational velocity is comparatively high . rotation curves shown in epinat et al . ( 2008a , b ) present a large variety of shapes ( from falling to rising ) and degrees of asymmetry . in order to study the influence of the shape of the rotation curves in the tully - fisher relation , we have made a classification of our sample in three subsamples , i. e. rising , flat and decreasing rotation curves , which will be described with the letters `` r '' , `` f '' and `` d '' , respectively . an `` + '' or ` ' sign has been added if the rotation curve is respectively more or less extended than the optical radius of the galaxy ( r@xmath6 ) . no symbol is added if the radius of the observed rotation curve is barely equal to r@xmath6 . a decreasing or flat rotation curve displays a clear v@xmath4 . this is not the case for a rotation curve which rises up to the very last observed radius , for which @xmath7 is possibly underestimated . this is even worse if the rising rotation curve does not reach the optical radius . in this case , v@xmath4 was computed at r@xmath6 by extrapolating an _ arctan _ function ( @xmath8 , where @xmath9 is the core radius ) to the rotation curves . if observed rotational velocities at r@xmath10r@xmath6 are higher than the modeled value , the observed values have been used . this could be the case if large scale bumps in the inner parts of rotation curves are present . table [ appendix1 ] ( column 8) presents the results of the rotation curve quality assessments . the main assumption necessary to derive a rotation curve from the observed velocity field is that rotation motions are dominant and non circular motions are not part of a large - scale pattern . thus , by construction , a rotation curve provides a measurement , for each radius , of the axi - symmetric component of the gravitational potential well of the galaxy . by consequence , if the motions in the galaxy disk are purely circular , the receding and the approaching sides of a rotation curve should match and the residual velocity field should not display any structure . once the parameters of the rotation curves are properly computed by minimizing the velocity dispersion in the residual velocity field ( epinat et al . 2008 ) , the remaining residuals are the signature of non - circular motions in or out the plane of the disk ( e. g. bars , oval distortions , spiral arms , local inflows and outflows , warps ) , including the intrinsic turbulences of the gas . to quantify the effects of these non - circular motions on the tully - fisher relations we have computed , for each galaxy , two indicators . the first one is based on the asymmetries between both sides of the rotation curves , it quantifies the mean difference of amplitude between the receding and approaching sides of the rotation curve . for each ring centered on the galaxy center , the weighted ( absolute ) velocity difference between both sides is computed . the weight is provided by the number of bins in each ring . each bin is an independent velocity measurement on the velocity field , it may be constituted by @xmath1150 pixels for low signal - to - noise region of the galaxy . depending of the spatial resolution , each ring contains from two to several hundreds bins . due to the fact that their radius are smaller , the central rings contains a number of pixels lower than the outer ones , their weights is thus smaller . the second indicator is based on the mean velocity dispersion extracted from the residual velocity field . this parameter is quantified by computing the average local velocity dispersion on the whole residual velocity field . to not overestimate the weight of non circular motion in slowly rotating systems with respect to high rotators , both indicators have been normalized by the maximum rotation velocity . we found that both indicators show the same trend on the tully - fisher relation , thus we will only illustrate the results using the indicator related to the asymmetries in the rotation curve . we computed the near - infrared magnitudes using 2mass data ( skrutskie et al . 2mass h and k - band data were available for 83 galaxies of the ghasp sub - sample defined in section 2.1 . absolute magnitudes were obtained using : @xmath12 distances ( _ d _ ) were taken from epinat et al . ( 2008b ) . they are computed from the systemic velocities ( from the ned database ) corrected from virgo infall and assuming h@xmath13=75 km s@xmath5 mpc@xmath5 , except when accurate distance measurements where available ( references are listed in epinat et al . the magnitudes m@xmath14 have been computed using the flux within the isophote of 20 mag arcsec@xmath5 ( where uncertainties were taken from 2mass ) . we corrected the magnitude for galactic extinction using the schlegel maps ( schlegel et al . k - corrections ( k@xmath14 ) , extinction due to the inclinations ( a@xmath14 ) and seeing were applied using the method given in masters et al . columns 1 , 2 and 3 in table [ appendix1 ] list the name , h and k - band absolute magnitudes for the sample . k - band luminosities were estimated using l@xmath15=10@xmath16 , where the k - band absolute solar magnitude of 3.41 was taken from allen ( 1973 ) . given the homogeneity of the _ sloan digital sky ( _ sdss _ ) , g and r - band optical magnitudes were extracted from this database . moreover , most of the mass - to - light ratio recipes use b - v colors as an input parameter . for this reason , we have converted g and r - band data into b and v - band magnitudes by using the recipes given in lupton ( 2000 ) . we have extracted the optical size for each galaxy from the _ sdss_. in this case , we used the parameter isoa ( in the r - band ) , which corresponds to the diameter of the isophote where the disk surface brightness profile drops to 25 mag arcsec@xmath17 . these values were converted in radii ( in kpc ) by using the distance to each galaxy . in order to compute the mass - to - light ratio for ghasp galaxies , g - r colors were corrected by galactic extinction using the values given in the _ sdss _ database and then converted into b - v colors . _ sdss _ colors were available for 45 galaxies from which we removed five objects for which their magnitudes and optical radius are obviously incorrect ( compared with the optical radius , and magnitudes , given in hyperleda ) . for other six galaxies , there were no radius measurement . all the analysis including the radius of the galaxies were thus performed with 34 objects . b - band luminosities were estimated by using l@xmath18=10@xmath19 , where the b - band absolute solar magnitude of 5.48 was taken from binney & merrifield ( 1998 ) . the main uncertainty in the study of the stellar and baryonic tully - fisher relations states in the stellar mass - to - light disk ratio @xmath20 . two main methods to estimate this parameter are used . spano et al . ( 2008 ) have modeled the stellar mass distribution of rotation curves , by scaling the r - band surface brightness profile to the rotation curve , obtaining an estimation of @xmath20 . bell & de jong ( 2001 ) have used stellar populations synthesis models to predict a relation between the colors of galaxies and their @xmath20 . although both approaches attempt to compute the same physical parameter , several authors have shown that surprisingly there seems to be no clear correlation between the @xmath20 obtained from these two methods ( e.g. barnes et al . other authors have invoked the modified newton dynamics ( mond ) to obtain the @xmath20 and study its implication on the baryonic tully - fisher relation ( mcgaugh 2005 ) . in this work , we have estimated @xmath20 and stellar masses using stellar population synthesis models recipes . one of them consists of simply fixing the value of @xmath20 . we have compared our results with other works available in the literature following the stellar populations synthesis models given by bell & de jong ( 2001 , b&j ) , bell et al . ( 2003 , be ) and portinari et al . ( 2004 , po ) ( equations [ eq1 ] and [ eq1b ] , [ eq2 ] and [ eq2b ] and [ eq3 ] respectively ) . @xmath21 @xmath22 @xmath23 @xmath24 @xmath25 - 0.115 } l_{k } \label{eq3}\ ] ] b&j and be suggested an uncertainty of 0.1 dex in the @xmath20 estimation . the adopted uncertainty in @xmath20 is larger than the uncertainties of the optical colors . we have adopted the same uncertainty for the po relation . b&j used a scaled salpeter initial mass function ( imf ) . po used a salpeter imf , with masses ranging between 0.1 and 100 m@xmath26 . these models are available for several colors ( to estimate @xmath20 ) and also for the luminosity in several bands ( to estimate the stellar mass ) . we have converted _ sdss _ g - r colors into b - v colors to obtain the @xmath20 parameter by using the recipes listed above . stellar masses were calculated using the k - band luminosities . this band is more likely to be reflective of the stellar mass of galaxies . in order to compare stellar masses derived from the k - band luminosities , we have also used the b - band luminosity , despite the fact that this photometric band could be contaminated with the emission of young stars . we removed from this analysis galaxies for which no _ sdss _ colors were available in the literature . therefore , in the stellar and baryonic analysis , we were left with 45 galaxies in total . k - band luminosities were estimated using l@xmath15=10@xmath16 ( see section 2.2 ) . stellar masses were also estimated using a fixed mass - to - light ratio following mcgaugh et al . ( 2000 , _ mg _ ) . these authors defined the mass - to - light ratio in the k - band as @xmath27 . in this case , the stellar mass is simply : @xmath28 it is interesting to note that gurovich et al . ( 2010 ) estimated the @xmath20 for a sample of local galaxies by modeling their stellar population histories . these authors did not find differences in the tully - fisher relation when the stellar masses were computed by using the modeled @xmath20 or when this parameter was fixed to @xmath20=0.8 ( mcgaugh et al . 2000 ) . the mass of a galaxy is constituted of its content in stars and stellar remnants , gas ( neutral , molecular and a negligible component of ionized gas ) , dust ( usually negligible ) and dark matter . the baryonic mass is the sum of the stellar and gas contents . the total mass in gas , m@xmath29 , is : @xmath30 where m@xmath31 is the neutral gas , m@xmath32 is the mass in helium and metals , m@xmath33 is the mass in molecular hydrogen and m@xmath34 is the ( negligible ) mass in ionized hydrogen . in order to obtain the baryonic mass for the ghasp sample we have calculated the observed hi mass for each galaxy using the corrected 21-cm line flux taken from hyperleda ( paturel et al . fluxes were converted into mass using the relation : @xmath35 where d is the distance to the galaxy in mpc , and f@xmath36 is the hi flux in jy km@xmath5 . to take into account the correction for helium and metals in the gas content ( e. g. mcgaugh et al . 2000 ) , m@xmath32 is related to the hi mass through : @xmath37 the h@xmath2 mass has been computed following the formula given by mcgaugh & de blok ( 1997 ) , using the morphological type of the galaxies ( young & knesek , 1989 ) . @xmath38 nevertheless , to avoid uncertainties linked to our bad knowledge in the h@xmath2 content , the baryonic mass studied in this paper does not include h@xmath2 , except when it is explicitly mentioned . the baryonic mass , m@xmath39 , is defined as : @xmath40 where m@xmath41 is the total stellar mass . uncertainties in the baryonic mass are the results of the quadratic sum of the uncertainties in stellar masses and uncertainties in the h i masses , which were taken from hyperleda . we have compared the baryonic mass to the total dynamical mass for each galaxy of our sample . although almost the whole baryonic mass is approximately within the optical radius , the total dark matter content is not , thus we compute the dynamical total mass only within the optical radius . to estimate the total dynamical mass , we assumed the mass has a spherical distribution , which is likely the case for the dark halo component , as supported by observations ( e.g. ibata et al . 2001 ) and n - body simulations ( e.g. kazantzidis et al . 2010 ) by using : @xmath42 where @xmath3 is a parameter depending on the mass profile distribution ( equal to one for an uniform distribution ) . to compute m(r ) we have used the r - band optical radius from _ sdss _ , as tabulated in the appendix . in order to obtain an estimation of the dark matter content at the optical radius , we have subtracted the baryonic mass ( as estimated in 2.5 ) from the dynamical total mass within the optical radius . galaxies having the same rotational velocity do not necessarily have the same luminosity ( or reciprocally ) , thus the observed tully - fisher relation presents a dispersion which may be produced by intrinsic properties of galaxies . beside this dispersion , uncertainties in magnitudes / masses and rotational velocities should be taken into account when the slope and zeropoint of this relation are computed ( see hogg et al . 2010 for details about fitting straight lines ) . several efforts have be performed to fit straight lines to fundamental relations , taken into account together the uncertainties in both axis and the intrinsic dispersion of the relation ( e. g. tremaine et al . 2002 , weiner et al . 2006 ) . in this paper , we have followed the prescription given by tremaine et al . ( 2002 ) , by adding ( in quadrature ) a dispersion factor to the uncertainties estimation of the nir magnitudes , stellar and baryonic masses . the value of the dispersion factor is chosen in order to reach a @xmath43 of unity per degree of freedom . to fit the tully - fisher relation , we used linear relation of the form : @xmath44 where , @xmath45 and @xmath46 ) , for the nir and stellar / baryonic tully - fisher relation , respectively , and @xmath47 . to obtain the slope and the zerpoint of this relation , we used the task fitexy ( press et al . in table [ appendix1 ] we list the nir magnitudes and the different rotational velocities that we obtained , from the observations and from the arctan model ( see 2.1 ) . columns 1 , 2 and 3 list the name , h and k - band magnitudes . column 4 shows the observed maximum rotational velocity for each galaxy . column 5 gives the modeled velocity at r@xmath6 . column 6 corresponds to the rotational velocity used in the tf relation . finally , in column 7 we classify the shape of rotation curves of the ghasp sample as shown in 2.1 . in table [ appendix2 ] we list the different determinations of mass - to - light ratios ( @xmath20 ) and masses for each galaxy . columns 1 , 2 and 3 indicate respectively the name , the radius of the galaxies ( taken from _ sdss _ ) and the b - v colors ( transformed from g - r colors ) . in columns 4 , 5 and 6 we list the mass - to - light ratios calculated from equations [ eq1 ] , [ eq2 ] and [ eq3 ] , respectively . values for the stellar masses are shown in columns 7 , 8 and 9 , following b&j , be and po , respectively . in column 10 , we list the m@xmath48+m@xmath32 , where m@xmath48 is calculated using the observed hi mass for the ghasp sample . column 11 corresponds to the baryonic masses excluding the h@xmath2 content ( m@xmath49+m@xmath48+m@xmath32 ) , m@xmath49 is here calculated following be given in column 8 . in fig . [ btf_ghasp_fig2a ] ( upper panels ) we plot the tully - fisher relations for the h and k - band ( left and right panels , respectively ) . in both panels , flat , decreasing and rising rotation curves are indicated by black dots , green triangles and red stars , respectively . galaxies having a rising rotation curve show a large dispersion on the tully - fisher relation , while most of the flat rotation curves lie on the relation . this may simply reflect the scatter in the determination of v@xmath4 for rising rotation curves , for which v@xmath4 may be uncertain . alternatively , this might indicated that rising rotation curves , that are usually dark matter dominated galaxies , show an intrinsic scatter in the tully - fisher relation . in the bottom panel of fig . [ btf_ghasp_fig2a ] we plot the k - band tully - fisher relation in which we divided the sample by their asymmetries in the rotation curve . galaxies displaying the largest non - circular motions ( red stars ) lie preferentially in the low velocity / luminosity region of the plot and present a larger scatter than galaxies less affected by non circular motion ( black dots ) . the conclusion is that non - circular motions contribute to the scatter in the nir tully - fisher relation , at least in the low luminosity ( and mass ) region of the plot . inspecting fig . [ btf_ghasp_fig2a ] , we observe a break in the tully - fisher relation at m@xmath1 - 20 . this effect has already been noted by mcgaugh et al . ( 2000 ) , gurovich et al . ( 2004 ) and amorn et al . ( 2009 ) . in order to quantify this break , we have applied a fit ( as discusses in 2.5 ) to all galaxies ( black dashed line ) and to the galaxies with m@xmath50 - 20 ( red dotted - dashed lines ) . for the first case , we have derived the followings equations : in the equations above we have included the one - sigma uncertainties in the slope and zero point . three galaxies ( @xmath114@xmath55 of the sample ) are 1@xmath56 off the tully - fisher relation . these galaxies could have slightly too high magnitudes for their rotational velocities or too low rotational velocities for their magnitudes ( upper left region in each panel of fig . [ btf_ghasp_fig2a ] ) . in spite of the rotation curves of these objects reach the optical radius r@xmath6 , two of them have rising rotation curves ( red stars in the left panels of fig . [ btf_ghasp_fig2a ] ) . in this sense , we can not exclude that both galaxies could have higher rotational velocities than the observed values , placing both galaxies on the tf relation . on the other hand , one galaxy has a flat rotation curve ( black dot in fig . [ btf_ghasp_fig2a ] ) , meaning that this object already reached its maximum rotational velocity . two possible scenarios could explain the high near - infrared magnitude of this galaxy . agn activity , that could enhance the near - infrared magnitude of seyfert 1 galaxies ( riffel et al . 2009 ) and/or the contribution of tp - agb stars into the k - band luminosity ( maraston 1998 ) . tp - agb stars are often not taken into account in the models and they are quite important specially for stellar populations with ages below 1 gyr . a detailed study of this galaxy is out of the scope of this paper . taking into account the dispersions , the slopes of the h and k - band tf relations are similar , being slightly steeper in the k - band . table [ table2 ] summarizes the fit parameters ( denoted by @xmath3 for slopes and @xmath57 for the zero points ) found for the tf relation in h and k - bands . in the same table we list values of @xmath3 and @xmath57 found in the literature . in order to compare different estimators of the stellar mass available in the literature , we have calculated the slope and zero points of the stellar tf relations of our ghasp sample using different prescriptions for mass determination given by b&j , be , po and mg ( see equations [ eq1 ] , [ eq2 ] , [ eq3 ] and [ eq3b ] ) . table [ table3 ] summarizes the fit parameters using the different stellar mass estimators . slopes from different estimators are consistent within 1@xmath56 . in the following , we used the be results to study the stellar and baryonic tf relations ( the be models are updated versions of the b&j models ) . in fig . [ btf_ghasp_fig2b ] ( left panel ) , we show the stellar tf relation for the ghasp sample . the central dashed line follows equation [ stellar ] ( and one-@xmath56 in the zeropoint ) and represents the best fit on the data . in this case , we use a dispersion factor of 0.31 dex in the stellar mass . we found a slope of [email protected] ( equation [ stellar ] ) when we calculated the stellar masses following be . b&j found a slope of [email protected] for the stellar tf relation , for a sample of galaxies in the ursa major cluster . as previously found in the h and k - band tf relation , low - mass galaxies in fig . [ btf_ghasp_fig2b ] ( with stellar masses lower than 10@xmath59 m@xmath26 ) lie below the relation defined by high - mass spirals . by using cosmological hydrodynamical simulations , de rossi et al . ( 2010 ) suggested that sn feedback is the main responsible for this behavior in low - mass systems . in fig . [ btf_ghasp_fig6 ] we plot stellar masses derived from the k and b - band luminosities , following the recipes of be . we found that the b - band luminosities slightly overestimate ( underestimate ) the stellar mass for the low - mass ( high - mass ) galaxies , when it is compared with the value derived from the k - band luminosity . such an overestimation ( and underestimation ) for the stellar masses of low - mass ( and high - mass ) galaxies could strongly affect the slope of the baryonic tully - fisher relation , adding a band - dependence effect " . the same trend was found when we use the recipes given in bj and po . in fig . [ btf_ghasp_fig2b ] ( right panel ) , we plot the baryonic tf relation for ghasp . the central dashed line represents the best fit for the data and one-@xmath56 in the zeropoint ( see equation [ bary ] ) . in this case , we use a dispersion factor of 0.21 dex in the baryonic mass . the shapes of the rotation curves are represented by different symbols ( circles , triangles and stars indicated flat , decreasing and rising rotation curves ) . it is interesting to note that low - mass galaxies now lie on the same relation defined by high - mass galaxies . this fact is attributed to the inclusion of the gaseous mass into the stellar budget . in this plot , the stellar mass was estimated from be . the slope of the baryonic tf that we derive for the ghasp sample is [email protected] ( which is in agreement with the slope obtained from an unweighted bisector fit , i.e. [email protected] ) . as done for the stellar mass tf relation , we obtained the slope and zero point of the baryonic tf relation when the stellar mass was calculated using b&j , be , po and mg . these values are listed in table [ tablehenri ] , where @xmath3 and @xmath57 correspond to the zero points and slopes , respectively . in table [ tablehenri ] we have included the resulting slopes and zero points obtained when h@xmath2 masses for the galaxies are included in the baryonic budget . we found that the slope of the baryonic tf relation , when h@xmath2 is included , does not change significantly . note , in particular , that the use of the bell & de jong ( 2001 ) or bell et al . ( 2003 ) mass - to - light recipes , on the ghasp sample , results in a very similar slope of the baryonic tf relation ( lines 1 and 9 of table [ tablehenri ] ) . in table [ tablehenri ] we also list the fit parameters for the baryonic tully - fisher relation when the stellar masses were derived from the b - band luminosities ( and the gaseous mass was corrected by the h@xmath2 gas mass ) . in this case , the slope is shallower ( @xmath61 ) than in the case when stellar masses are computed from the k - band luminosities . in fig . [ r25bary ] , we show that , even with some scatter in the relation , the baryonic mass @xmath62 grows almost linearly with the optical galactic radius r@xmath6 in log units ( @xmath63 ) . we used a weighted bisector least square fit to obtain the dependence between these parameters ( we used a bisector fit given that the _ sdss _ database does not quote errors in the radius ) . we found @xmath64 , where @xmath65 . these results suggest that the baryonic mass density , defined by @xmath66=@xmath67 , depends weakly on the sizes of the galaxies . more precisely , @xmath68 .	we studied , for the first time , the near infrared , stellar and baryonic tully - fisher relations for a sample of field galaxies taken from an homogeneous fabry - perot sample of galaxies ( the ghasp survey ) . the main advantage of ghasp over other samples is that maximum rotational velocities were estimated from 2d velocity fields , avoiding assumptions about the inclination and position angle of the galaxies . by combining these data with 2mass photometry , optical colors , hi masses and different mass - to - light ratio estimators , we found a slope of [email protected] and [email protected] for the stellar and baryonic tully - fisher relation , respectively . we found that these values do not change significantly when different mass - to - light ratios recipes were used . we also point out , for the first time , that rising rotation curves as well as asymmetric rotation curves show a larger dispersion in the tully - fisher relation than flat ones or than symmetric ones . using the baryonic mass and the optical radius of galaxies , we found that the surface baryonic mass density is almost constant for all the galaxies of this sample . in this study we also emphasize the presence of a break in the nir tully - fisher relation at m@xmath120 and we confirm that late - type galaxies present higher total - to - baryonic mass ratios than early - type spirals , suggesting that supernova feedback is actually an important issue in late - type spirals . due to the well defined sample selection criteria and the homogeneity of the data analysis , the tully - fisher relation for ghasp galaxies can be used as a reference for the study of this relation in other environments and at higher redshifts . [ firstpage ] galaxies : evolution galaxies : kinematics and dynamics
appearance of the geometric phases in physical systems has been , since their first systematic treatment by @xcite , among the most fascinating physical phenomena . generalized to systems with degenerate spectrum,@xcite they give the possibility to devise systems in which the dynamical contribution is only the overall phase factor , and the actual transformations are of purely geometric origin . + quantum geometric transformations ( below referred to as _ holonomies _ ) have attracted even more attention after @xcite proved their potential use in quantum computing . as compared to the ordinary , dynamical computation , quantum gates are here realized by cyclic evolution of parameters , and the result depends only on the geometry of the traversed path . construction of holonomies has been discussed for various physical realizations , among them are also superconducting nanocircuits . @xcite in these realizations the necessary number of independent tunable parameters results in high complexity of the considered systems . here , motivated also by the recent experiment with josephson - junction system composed of two coupled charge qubits , @xcite we consider a similar design as a potential candidate for observation of quantum holonomies . as compared to the proposals in refs . @xcite , where the simplest two - dimensional holonomies are constructed using four coupled charge qubits , we achieve substantial simplification using only two qubits . also , we realize the transformations within a twofold degenerate ground state , rather than excited state ( as in @xcite ) . in this way we avoid the problem of depopulation of the subspace in which the holonomies are realized . we begin with describing the system and defining the operational subspace . in the four - dimensional space of two charge qubits we find a configuration of parameters for which the ground state is degenerate . we perform the holonomies by selecting two of the parameters and varying them adiabatically along a certain loop . during the transformation the degeneracy of the ground state is maintained by adjusting the remaining parameters as a function of the others . the scheme presented here may be realized in a system without strong constraints on its parameters provided that the noise level is low . since the system of charge qubits is usually affected by the charge fluctuations , in the following we optimize the design in order to suppress the noise ( for certain models of errors we find a decoherence - free subspace @xcite ) . finally we discuss possible extensions of the scheme . the specific system considered here is described by the hamiltonian of two coupled qubits . in the discussion we also comment on the possible application of our scheme to an arbitrary system with this model hamiltonian . the system we consider consists of two `` charge qubits''.@xcite they are coupled to each other via tunable josephson junction a symmetric dc - squid ( superconducting quantum interference device ) ( see fig . [ fi : system ] ) . as discussed later , we need to have the possibility to switch off all the josephson energies completely ( and we use dc - squids instead of simple junctions ) . on the other hand , to construct nontrivial transformations , the amplitude and phase ( or equivalently the real and imaginary parts ) of the couplings @xmath0 and @xmath1 should be controllable during the operations . to achieve this we replace one of the junctions in the left and right squids with a further dc - squids ( the same technique has been used in ref . @xcite ) . if the dimensionless gate charges @xmath2 are close to @xmath3 , the only relevant charge states for each island are @xmath4 and @xmath5 ( zero or one extra cooper pair on the island ) . the hamiltonian may be then reduced to the four - dimensional space and in the charge basis @xmath6 has the form @xmath7 here the diagonal elements are the electrostatic terms @xmath8 [ @xmath9 is the number of excess cooper pairs on the first ( second ) island ] with @xmath10 , and similarly for @xmath11 . here @xmath12 is the sum of all capacitances connected to the first ( second ) island . the electrostatic coupling term due to finite capacitance of the middle dc - squid equals @xmath13 ( we do not assume weak electrostatic coupling and the capacitance @xmath14 does not need to be small ) . the josephson coupling @xmath15 $ ] ( and similarly for @xmath16 ) , where @xmath17.@xcite here @xmath18 is the josephson energy of the junctions in the dc - squids . the term @xmath19 is the tunable josephson coupling of the small squid . finally @xmath20 is the josephson energy of the middle squid . here the fluxes are in units of @xmath21 , the superconducting flux quantum . we assume that the system is operated in the charge regime ( @xmath22 ) . a similar system has been used already in the context of geometric phases . @xcite used two coupled charge qubits to devise an experiment in which the berry phase could be detected . also , as already mentioned , the experiment by @xcite shows that a system of two coupled charge qubits can be well controlled and manipulated with available experimental techniques . the difference between our system and those used in refs . @xcite is that we couple the islands using dc - squid instead of a capacitor , and we require controllable josephson couplings @xmath23 . the set of controllable parameters defines the _ control manifold_. the hamiltonian is characterized by the tunable terms : @xmath24 , @xmath11 , @xmath0 , @xmath1 , @xmath25 . we will use in the discussion rather the parameters that can be controlled directly , _ i.e. _ , @xmath26 . the fluxes @xmath27 are used only to control the symmetry of the squids , and we will not consider them to be used as the actual tunable parameters in the following discussion ( between the operations they are kept at the values for which the couplings @xmath23 are real , and may be tuned to zero . just before the operations we can switch them to different values for which the dc - squids are asymmetric . during the operations they are kept constant ) . the operational subspace for the simplest non - abelian holonomies is spanned by two lowest - energy eigenstates . to make the transformations of purely geometric nature we need to make them degenerate . this property in our system does not follow from the symmetry arguments and we need to control it by making the parameters not fully independent . in the following we will compare the energies of the lowest states , and from this constraint calculate one of the parameters as the function of the remaining ones . in this way we define the degeneracy domain in the control manifold . for the following configuration of the parameters @xmath28 the hamiltonian eq.([eq:7ra ] ) is diagonal in the charge basis , and the ground state is twofold degenerate and spanned by the states @xmath29 we will refer to this particular point in the parameter space as the _ starting point _ , or the `` no - op '' regime ( as this configuration will be kept between the sequences of transformations ) , and to the states @xmath30 and @xmath31 as the logical basis . before we perform any transformations the system is initialized to the state @xmath30 by turning off all the josephson couplings , and tuning the electrostatic energy @xmath32 to be the lowest one . after long enough time the system will relax , and we switch to the starting point . we perform a holonomic transformation by varying the parameters adiabatically along a suitable loop . during the operations the other charge states may be involved as well . the system , however , at any instant of time remains in the twofold degenerate ground state , and the resulting unitary transformation ( holonomy ) is limited to the logical basis only . the correspondence between the loop , and the performed unitary transformation is given by @xmath33 where @xmath34 is the path - ordering operator , @xmath35 the path , @xmath36 is the set of parameters ( control manifold ) , and the @xmath37 matrix @xmath38 is the _ wilczek - zee connection_,@xcite @xmath39 @xmath40 and @xmath41 are parameter - dependent states spanning the operational subspace , and for the no - op regime they can be either @xmath30 or @xmath31 . to construct the holonomies we need to find a loop corresponding to the desired transformation . any unitary transformation can be factorized as a product of two noncommuting rotations in the qubit space . here we find two such holonomies , namely , @xmath42 and @xmath43 . for our purposes it is convenient to interpret these rotations as phase shifts between @xmath30 and @xmath44 , and between @xmath45 and @xmath46 , respectively . for given rotation , at each point of the constructed paths the wilczek - zee connection @xmath47 is then diagonal [ in the basis @xmath48 and @xmath49 respectively ] , and we omit the path - ordering operation . the holonomies we find are in our case the aforementioned rotations , @xmath50 and @xmath51 . the first transformation we consider is the phase shift @xmath52 . if we keep @xmath53 ( and change the flux @xmath54 to make the coupling @xmath1 complex ) , the hamiltonian simplifies to @xmath55 the obtained block form may be solved for each part independently . we calculate the desired phase shift as the difference between the berry phases for each block . at the same time , to make the phase shift of purely geometric origin , we compare the energies of the lower states within each block to find the degeneracy subspace in the control manifold . since the energies of the lowest states are functions of @xmath56 , and @xmath57 , from the condition @xmath58 we find the degeneracy condition in the functional form @xmath59 . this dependence is shown in the upper part of fig . [ fi : plotsphase ] the parameter - dependent spectrum ( for the degeneracy domain in the control manifold ) is shown in the middle plot of fig . [ fi : plotsphase ] . the effectively four - level system has only three distinct energy values due to the degeneracy of the ground state . finally the phase shift may be presented conveniently as the integral , @xmath60 , \label{}\ ] ] where @xmath61 is the domain in the @xmath62-@xmath57 plane limited by the path @xmath63 , and the `` berry field '' is given by @xcite @xmath64 here @xmath65 and @xmath66 are the excited states from the upper - left and the lower - right block of @xmath67 respectively , and @xmath68 their energies . the field difference @xmath69 is shown in the lower plot of fig . [ fi : plotsphase ] . similarly , we can find the second generic holonomy , namely , @xmath70 . for @xmath71 ( the coupling should be complex during the operations , so the fluxes @xmath27 should have different values than at the starting point , but chosen in this way that the josephson energies may be equal ) , @xmath72 , the hamiltonian may be again presented in the block form , this time in the basis @xmath73 : @xmath74 comparison of the lowest energy of the upper - left block with @xmath75 gives the degeneracy subspace , as shown in the upper plot of fig . [ fi : plotsflip ] . the middle plot shows the parameter - dependent spectrum , again the ground state being twofold degenerate . since the state @xmath76 does not vary with the parameters , the desired phase is simply the berry phase acquired by the ground state in the upper - left block . the phase here equals @xmath77 the field @xmath78 is shown in the lower plot of fig . [ fi : plotsflip ] . the scheme described above is valid for a system without any strong constraints on its parameters the only idealization so far is the assumption of perfect symmetry of the small dc - squids . as already noted , the operations are performed within a ground state , and the relaxation does not depopulate the operational subspace . however , as the degeneracy is maintained by external parameters , we should expect that the system will suffer from their fluctuations . now we will give qualitative analysis of the noise in the considered system , and describe how the design can be optimized in order to suppress the dissipation . in josephson charge qubits the ohmic noise and the low - frequency charge fluctuations coming from the substrate are usually relatively strong . since the system is operated in the charge regime , and both sources couple to the charge on the islands we apply here the spin - boson model with the system hamiltonian approximated by its electrostatic part only @xmath79 here @xmath80 are the bath operators . we divided the interaction term into the part describing the common environment for the two islands , and part with independent environments . if now the two islands have equal energy splitting ( @xmath81 ) , and the dominant interaction with environment can be modeled by common environment , the subspace @xmath82 is well protected ( in the limiting case of completely correlated noise @xmath83 the states span the decoherence - free subspace ) . our goal is then to minimize the uncorrelated noise in the islands as much as possible . to achieve this for the noise coming from the substrate we can , as discussed in ref . @xcite place the islands close to each other . to suppress the ohmic noise we could use a common voltage source controlling the bias of the islands ( as in the experiment by @xcite ) . as we present here , this may work well also for our purposes , provided that the islands are nearly identical . we use again two voltage sources , but now their roles are not symmetric . instead , the main , common source ( the fluctuations of which give correlated noise ) will be used to tune the system , while the second , auxiliary , is connected to one island via very small capacitance ( compared to @xmath84 ) as it is used to tune a small ( as explained below ) difference between the gate charges @xmath85 and @xmath62 . the resulting noncorrelated noise will be then much lower than the correlated noise of the main source . to be more specific , the fluctuations in @xmath86 should be correlated ( rather than in @xmath87 ) and we require that the electrostatic properties of both islands are nearly identical ( @xmath88 , and @xmath89 ) . then the common voltage source @xmath90 gives first of all nearly equal energy splitting of the islands [ @xmath91 in eq.([eq:8ra ] ) ] , and also relates the parameters @xmath92 . we note that for the starting point as well as during the operation @xmath93 we have @xmath94 . this for identical islands gives @xmath95 . also during the operation @xmath96 there is a region in the parameter space in which the relation @xmath97 nearly coincides with @xmath95 [ this region is determined by @xmath98 , where the states @xmath30 and @xmath44 have equal and lowest electrostatic energies . the small difference between @xmath62 and @xmath99 is in this region a result of small , but nonzero josephson energies ] . we then can use an additional voltage source that is coupled only to the first island . this voltage will give an additional term @xmath100 which compensates the effect of small differences in the properties of the islands , and during the @xmath96 operation the small difference @xmath101 . since each gate can be performed using various paths , we may even further improve the fidelity of the operations by examining the sensitivity of the degeneracy splitting to the fluctuations of each parameter , and choose paths from the region least affected by the fluctuations . for instance , close to the starting point [ defined in eq.([eq : qra ] ) ] the main voltage @xmath87 fluctuations to the first order do not shift the system out of the degeneracy subspace . the same applies to @xmath57 for the @xmath96 operation and to @xmath102 for @xmath51 . this can be easily seen in the upper plots of figs . [ fi : plotsphase ] and [ fi : plotsflip ] . those fluctuations lie within a plane tangential to the degeneracy subspace . to protect the system also during the operation we should construct the loops relatively close to the starting point . to summarize , we have found that the system properties give us the opportunity to realize any unitary transformation within a qubit space using gates of purely geometric origin . assuming the perfect performance not affected by the noise we may rely on the simplest design without strong constraints on the system parameters . however , such a system composed of two solid - state qubits is usually strongly probed by the environment which results in fast dephasing . our qualitative analysis made for this particular system specifies the conditions under which the system should be well protected even in the presence of noise . we have not discussed here the coupling between such elementary blocks , as the question of applicability of holonomies in quantum computing remains open and requires detailed analysis of resulting fault tolerance for individual proposals . however , as the logical basis here is defined by states differing by the charge configuration of the islands , such extension based on the method described in ref . @xcite should be adjustable to our system . the scheme presented here may be applied to any system described by the hamiltonian of two interacting qubits of the form @xmath103 provided the parameters are tunable . since there are until now many well - developed schemes concerning qubit control as well as the controlled qubit - qubit coupling ( usually for the sake of quantum computation ) , realization of quantum holonomies may appear very natural and straightforward in such systems . for the same reason the model of a four - dimensional quantum system as a candidate to construct and observe holonomies may turn out to be more applicable than the simplest , and widely used in implementation - independent discussions three - level model ( see for example @xcite ) . as far as the josephson - junction systems are considered , we may also use the device to study the process of the adiabatic charge transport:@xcite the logical basis is spanned by the states with one excess cooper - pair on one island and the second with zero . quite naturally the second discussed operation , generated by @xmath104 , corresponds for @xmath105 to the charge pumping cycle.@xcite the author thanks yu . makhlin for inspiring discussions and numerous comments on the manuscript , r. fazio , j. siewert , and g. falci for discussions and useful hints . this work was supported by the dfg - schwerpunktprogramm `` quanten - informationsverarbeitung '' , eu ist project squbit , and ec research training network . 99 m. v. berry , proc . london , ser . a * 392 * , 45 ( 1984 ) . f. wilczek and a. zee , phys . lett . * 52 * , 2111 ( 1984 ) . j. pachos , p. zanardi , and m. rasetti , phys . a * 61 * , 010305(r ) ( 2000 ) ; p. zanardi and m. rasetti , phys . a * 264 * , 94 ( 1999 ) . l. faoro , j. siewert , and r. fazio , phys . * 90 * , 028301 ( 2003 ) . m. -s . choi , j. phys . : condens . 15(46 ) , 7823 ( 2003 ) . yu . a. pashkin , t. yamamoto , o. astafiev , y. nakamura , d. v. averin , and j. s. tsai , nature ( london ) * 421 * , 823 ( 2003 ) . duan and g .- c . guo , phys . rev . lett . * 79 * , 1953 ( 1997 ) . p. zanardi and m. rasetti , phys . lett . * 79 * , 3306 ( 1997 ) . d. a. lidar , i. l. chuang , and k. b. whaley , phys . * 81 * , 2594 ( 1998 ) . makhlin , g. schn , and a. shnirman , nature ( london ) * 398 * , 305 ( 1999 ) . m. tinkham , _ introduction to superconductivity _ , 2nd ed . ( mcgraw - hill , new york , 1996 ) . g. falci , r. fazio , g. m. palma , j. siewert , and v. vedral , nature ( london ) * 407 * , 355 ( 2000 ) . a. b. zorin , f .- j . ahlers , j. niemeyer , t. weimann , and h. wolf , phys . b * 53 * , 13 682 ( 1996 ) . j. p. pekola , j. j. toppari , m. aunola , m. t. savolainen , and d. v. averin , phys . b * 60 * , 9931 ( 1999 ) .	we examined properties of a josephson - junction system composed of two coupled cooper - pair boxes ( charge qubits ) as a candidate for observation of quantum holonomies . we construct a universal set of transformations in a twofold degenerate ground state , and discuss the effects of noise in the system .
during the past decade , we have learned ( at last ) how to make rigorous contact between experimental low energy hadronic physics and qcd which is presumed to underlie such processes . this contact is provided by chiral perturbation theory ( chpt ) @xcite , which exploits the ( broken ) chiral invariance of the light quark component of the qcd lagrangian and provides a representation for interaction amplitudes as an expansion in energy - momentum divided by the chiral scale parameter @xmath2 gev @xcite . a review of chpt techniques will not be attempted here , but has been presented in a number of sources , wherein it is shown that , at least in the goldstone boson sector , such a chiral approach provides a remarkably successful and predictive representation of a wide variety of experimental processes @xcite . the chiral technology begins by writing down an effective chiral lagrangian , the simplest ( two - derivative ) form of which is , in the goldstone sector , @xcite @xmath3 where @xmath4 is a nonlinear function of the pseudoscalar fields , @xmath5 is the quark mass matrix , @xmath6 is a phenomenological constant , @xmath7 is the covariant derivative , and @xmath8 is the pion decay constant in the limit of chiral symmetry . although these are only two of an infinite number of terms , already at this level there exists predictive power e.g . , tree level evaluation of @xmath9 yields the familiar weinberg predictions ( at @xmath10 ) for s - wave @xmath11 scattering lengths @xcite which are approximately borne out experimentally . loop diagrams required by unitarity produce terms of @xmath12 and contain divergences . however , just as in qed such infinities can be absorbed into renormalizing phenomenological chiral couplings , and the most general `` four - derivative '' lagrangian has been given by gasser and leutwyler @xcite @xmath13 here the bare @xmath14 coefficients are themselves unphysical and are related to empirical quantities @xmath15 measured at scale @xmath16 via @xmath17 where @xmath18 are constants defined in ref . @xcite and @xmath19 is the usual parameter arising in dimensional regularization , with @xmath20 being the number of dimensions . gasser and leutwyler have obtained empirical values for the phenomenological constants @xmath21 . a wide range of electroweak and strong interactions of these goldstone bosons have been successfully treated within this formalism , but there is at least one recurring problem whenever the s - wave i=0 @xmath11 interaction is involved the simple one loop predictions have in general been found to be wanting @xcite . this is perhaps not surprising , as the associated phase shift @xmath22 passes through @xmath23 somewhere in the vicinity of @xmath24 mev , which has sometimes been associated with the existence of a broad resonance in this region . such resonant behavior can certainly not be treated in any perturbative fashion and thus in this channel the chiral expansion must break down well before this energy is reached . there are a number of ways which have been used in order to avoid this difficulty . one is simply to confine predictions to a low enough energy that one - loop corrections should be sufficient . however , it is also possible to treat such effects nonperturbatively either by inclusion of s - wave i=0 ( @xmath25 or @xmath26 ) pole effects or by use of dispersion relations which effectively treat the @xmath11 interaction to all loop orders but at the price of introduction of model dependence @xcite . this cost is usually considered worth paying , however , as by inclusion of such well - motivated model - dependence one can often reliably extend the usual region of validity of chiral predictions ( @xmath27 mev ) up to much higher values ( @xmath28 gev ) . a particularly useful example of this procedure has recently been provided for the process @xmath29 . because only neutral particles are involved there exists no tree level two - derivative or four - derivative contribution to this reaction , which guarantees that the one loop chiral perturbative prediction must itself be finite . this has been calculated as @xcite @xmath30 here @xmath31 ^ 2=1+{y\over x}\ln^2\left({\beta(s)+1\over \beta(s)-1}\right)\nonumber\\ \qquad{\rm with}\quad\beta ( x)=\sqrt{x-4y\over x}.\end{aligned}\ ] ] however , the associated @xmath32 cross section is given by @xmath33 and is found to bear little resemblance to its recently measured experimental analog , as shown in figure 1 . the solution to this problem has recently been explored by a number of authors and has been found to be related to the inadequacy of the one - loop approach in the i=0 s - wave @xmath11 channel @xcite . the solution is most clearly presented in terms of a dispersion relation approach . we assume , consistent with the chiral expansion , that when we are in the near - threshold region the only relevant higher order effects are in the helicity conserving s - wave amplitude , which we write as @xmath34\nonumber\\ \gamma\gamma\rightarrow\pi^{0}\pi^{0}:\qquad f^{n}(s)&=&\frac{2}{3}\left[f_{0}(s ) - f_{2}(s)\right],\end{aligned}\ ] ] where i = 0 , 2 refers to the isospin of the final @xmath35 state . for neutral pion production and working in the gauge wherein @xmath36 the transition amplitude is @xmath37 in the charged pion case the born and seagull contributions to this multipole must also be included , so that the full amplitude becomes @xmath38.\end{aligned}\ ] ] here @xmath39 where @xmath40 is the born approximation value for the helicity conserving s - wave multipole . in the threshold region the phase of @xmath41 is required by unitarity to be equal to the corresponding @xmath35 phase shift @xmath42 . when @xmath43 , inelastic reactions involving four pions are allowed . however , the inelasticity is small , being of order @xmath44 in the chiral expansion and will be neglected . the functions @xmath41 are analytic functions of s except for cuts along the positive and negative real axis . for positive @xmath45 , the right hand cut extends from @xmath46 and is due to the @xmath45 channel @xmath35 state . for negative @xmath45 , the left hand cut is due to @xmath47-channel intermediate states such as @xmath48 or @xmath49 , and extends from @xmath50 . the single - channel final state unitarization problem has a simple solution in terms of the omns function @xcite @xmath51 the result must have the form @xmath52 where @xmath53 is an analytic function with no cuts along the positive real axis . morgan and pennington consider a function @xmath54 which has the same left hand singularity structure as @xmath41 , but which is real for @xmath55 . they then write a twice subtracted dispersion relation for the difference @xmath56 , with the result @xcite @xmath57,\ ] ] where @xmath58 are subtraction constants . picking @xmath59 to be given by its born value and matching onto the known form of the low energy amplitude required by chiral symmetry determines these constants unambiguously to be @xmath60 one can now address the origin of the large corrections found in the @xmath61 amplitude . do they arise simply from the unitarization of the amplitude ( _ i.e. _ @xmath62 ) or are new inputs needed in the amplitude ? it turns out that the rescattering physics in @xmath63 is most important , and that the main corrections are due to well - known ingredients . in our subsequent discussion , we will use a full phenomenological treatment but it is useful here to explore the case with a simple analytic form for @xmath63 . the condition @xmath64 defines the [ 0,1 ] pad approximation for the omns function @xcite , _ i.e. _ @xmath65 and allows one an approximate but simple analytic representation for the @xmath32 amplitude . the constant @xmath66 is chosen to match the small @xmath45 behavior of the experimental @xmath67 function , and @xmath68 is chosen from a fit to @xmath69 @xmath35 scattering . the resulting form for the @xmath61 amplitude is @xmath70\nonumber\\ & + & { 4\over 3f_\pi^2}(l_9^r+l_{10}^r)s ( d_0^{-1}(s)-d_2^{-1}(s))\end{aligned}\ ] ] which , when the pad forms of @xmath71 are used , provides a consistent analytic solution to the dispersion relation while also displaying the correct chiral properties to @xmath72 . in figure 1 , we plot the resulting cross section , in comparison with the data and the lowest order result . it can be seen that the omns functions produce a substantial modification even near threshold . of these , the most important is @xmath67 which reflects the strong attractive @xmath35 scattering in the @xmath73 channel @xcite . a much more satisfactory fit is found by use of an omns function @xmath74 determined via the use of _ experimental _ values of the pion - pion phase shifts @xcite as well as including contributions to the left hand cut due to @xmath75 exchange diagrams , which leads to the very good fit given in figure 1 . details of this calculation can be found in ref . @xcite . the lesson to be learned from this example is the importance of inclusion of i=0 s - wave @xmath11 rescattering corrections especially in processes which have no counterterm contributions and are generated from a simple one - loop chiral calculation . we shall see in the next sections two additional examples of this type , when we extend our formalism to cover the nonleptonic weak - radiative decays @xmath76 and @xmath77 . a good deal of work has been done extending the chiral formalism into the regime of nonleptonic weak processes . to two - derivative order the form of the effective su(3 ) octet chiral lagrangian is unique @xmath78 where @xmath79 is a constant whose value can be determined empirically . the corresponding four - derivative weak effective lagrangian has also been written down and involves thirty - seven additional terms @xcite , whose explicit form will not be needed here , since as in the case of @xmath29 there is no tree level contributions to the weak - radiative decay process @xmath80 . nevertheless there does exist a finite one - loop piece . defining @xcite @xmath81 the one - loop chiral prediction is found to be @xmath82\right ) . \label{chpt : b}\ ] ] with the value @xmath83 determined from the tree level prediction for @xmath84 @xmath85 we find the one - loop chiral prediction to be in good agreement with the recently determined experimental number @xmath86 an alternative derivation of this result is provided by the use of a dispersion relation . we note that the function b(s ) has a cut along the line element @xmath87 and by unitarity has the imaginary part @xcite @xmath88 where @xmath89 is the amplitude for the weak decay @xmath90 and @xmath91 is the amplitude for s - wave radiative pion annihilation @xmath92 discussed in the previous section . note that since we are using an su(3 ) octet assumption for the weak transition , both di - pions are required to be in an isoscalar configuration . to lowest order in chiral perturbation theory we have @xmath93 we now write a doubly subtracted dispersion relation for the function @xmath94 using as subtraction constants the requirements that the amplitude vanish at both @xmath95 and @xmath96 , as given in the lowest order chiral analysis , yielding @xmath97 if the lowest order chiral values eq . [ loword]are used to determine i m @xmath98 then the integration can be performed analytically to yield @xmath99 which is precisely the ( pion contribution to the ) one - loop chiral result eq . [ chpt : b ] . at one level then the dispersive technique represents merely an alternative ( and simpler ! ) way by which to perform the one - loop calculation . however , at a deeper level the dispersion relation provides a means to undertake a much more complete calculation of the radiative decay process by using for i m @xmath98 not just the lowest order chiral forms for these amplitudes but instead values which have more experimental validity . of course , the subtraction constants must be fixed by some other means we assume that chiral perturbation theory to one loop is accurate enough to describe the amplitude at very low energies . the @xmath100 decay amplitude @xmath89 itself is an analytic function with a discontinuity along a cut @xmath87 given in terms of the s - wave i=0 @xmath11 scattering phase shift @xmath101 the general solution of such an equation is given @xmath102 where @xmath103 is an arbitrary polynomial and @xmath74 is the omns function discussed in section 1 . in order to determine the polynomial @xmath103 we demand that the full @xmath84 amplitude given in eq . [ a0omn ] match the simple chiral form eq . [ loword ] in the absence of rescattering ( _ i.e. _ when @xmath104 ) . we have then @xmath105 while the corresponding solution for @xmath91 was derived in ref . @xcite and has been outlined in section 1 of this paper . for @xmath89 is used , the value @xmath106 must be employed in order that the proper normalization to the @xmath84 decay rate be preserved . ] substitution into eq . [ imb ] and numerical evaluation of the dispersive integral then provides a prediction for the @xmath107 amplitude which is much more complete and founded in experiment than is its simple one - loop chiral analog . the result of the numerical integration is shown in table 1 , where values are given for the radiative decay amplitude @xmath108/branching ratio for four scenarios a ) simple one - loop chiral perturbation theory ; b ) use of the full unitarized values and with the pad form for the omns function ; c ) use of the full unitarized values and with a numerical representation of the omns function based on experiment and given by gasser @xcite ; d ) same as c ) but with vector and axial - vector exchange contributions included also in @xmath91 , which provides the best fit to the @xmath32 system . .shown are the amplitudes and predicted branching ratios for the process @xmath107 using various theoretical inputs to the dispersion relation eq . 28 as described in the text . [ cols="<,<,<,<",options="header " , ] the observed enhancement is clearly seen to come from two effects : . ] * the use of a corrected @xmath109 amplitude in agreement with data gives an enhancement factor of @xmath110 in the branching ratio . this has been noted earlier @xcite ; in a very simple approach one could just scale @xmath79 in eq . [ k3pica ] to reproduce the experimental @xmath109 amplitude , leading to a similar enhancement factor for @xmath111 . however , a consistent calculation clearly should explain all relevant processes , @xmath112 , @xmath109 , @xmath113 , @xmath80 and @xmath77 by the same method . we have explained in detail above how this can be achieved . * the inclusion of the neutral two pion intermediate state in the unitarity relation brings an additional enhancement factor of @xmath114 . as mentioned before , this contribution is missed in all approaches where the @xmath115 vertex is treated only at the born level . although this effect is moderate , it goes in the right direction . moreover , it is critical to include the @xmath116 intermediate state in order to properly implement the constraints of unitarity . finally we note that using a twice subtracted dispersion relation , the contributions from the high energy region to the dispersive integral are very much suppressed . cutting the integral at @xmath117 instead of 1 gev changes the branching ratio by less than one percent . also , the calculated spectrum in the invariant mass of the two photons is plotted in figure 3 . it deviates insignificantly from the spectrum obtained in one loop chiral perturbation theory . if only the charged intermediate state is included , the maximum of @xmath118 is shifted toward higher @xmath119 values , contrary to the experimental trend . inclusion of the neutral intermediate state in the unitarity relation restores the maximum to its original location . the process @xmath120 has traditionally been a difficult one to understand within the context of chiral perturbation theory . indeed , from other successes one might have expected the one - loop chiral prediction to be accurate to @xmath121(20% ) or so , whereas in fact the predicted branching ratio is nearly a factor of three too small . previous work in this area has attempted to explain this discrepancy in terms of vector meson pole contributions and/or in terms of higher loop final state interaction effects . we have here noted , however , that previous dispersion - based final state interaction calculations have included only the intermediate @xmath122 intermediate state , omitting its potentially important @xmath116 analog . above we have given a mutually consistent analysis of both @xmath123 _ and _ @xmath120 processes using dispersion relations and have shown that by including this previously neglected @xmath116 intermediate state piece the branching ratio for @xmath77 is significantly enhanced , although it remains too low to fully explain the data . this should by no means be considered to be complete analysis indeed many effects such as i=1,2 @xmath35 scattering effects as well as higher order contributions to the @xmath124 process have been neglected . nevertheless , we believe that this calculation opens up interesting questions for future study about the importance of effects beyond the one loop approximation in chiral perturbation theory . * indicated are experimental data points for @xmath32 compared to the one - loop chiral perturbative prediction ( dashed ) and dispersive calculations using pad ( dotted ) and gasser ( solid ) @xmath74 functions . * shown are the results of one - loop chiral perturbation theory ( dashed ) and dispersive analysis scenario d ) ( solid ) for i m @xmath94 ( a ) and re @xmath94 ( b ) respectively . * normalized spectra @xmath125 in the invariant mass of the two photons ( @xmath126 ) . plotted are chiral perturbation theory ( dashed ) and dispersive analysis using gasser @xmath74 function ( solid ) . * i m @xmath127 ( a ) and re @xmath127 ( b ) in the physical decay region . shown are the results of one - loop chiral perturbation theory ( dashed ) and dispersive analysis using gasser @xmath74 function ( solid ) . also shown is the result of the dispersive analysis including the charged intermediate state only ( dot - dashed ) .	agreement between the experimental value @xmath0 and the number predicted via a one - loop chiral perturbation theory calculation has been cited as a success for the latter . on the other hand the one - loop prediction for the closely related process @xmath1 has been found to be a factor three below the experimental value . using the inputs of unitarity and dispersion relations , we demonstrate the importance of higher order loop effects to both of these processes .
in condense matter physics , quantum phase transitions imply fluctuations , which happened at the zero temperature @xcite . when a controlling parameter changes across critical point , some properties of the many - body system will change dramatically . many results show that entanglement existed naturally in the spin chain when the temperature is at zero . the quantum entanglement of a many - body system has been paid much attention since the entanglement is considered as the heart in quantum information and computation @xcite . as the bipartite entanglement measurement in a pure state , the von neumann entropy @xcite in the antiferromagnetic anisotropic spin chain @xcite and isotropic spin chain @xcite are investigated respectively . by using the cross fields of quantum many - body theory and quantum - information theory , von neumann entropy is applied to detect quantum critical behaviors @xcite . a typical example is that osborne solved exactly one - dimensional infinite - lattice transverse - field ising model to obtain entropy by the jordan - wigner transform . the entropy predicts the quantum phase transition successfully @xcite . moreover , another concept from quantum information science , the ground state fidelity has been used to qualify quantum phase transitions in the last few years @xcite . because the fidelity is a measure of similarity between states , the fidelity should drop abruptly at critical points as a consequence of the dramatic changes in the structure of the ground states , regardless of what type of internal order is present in quantum many - body states . this result is the orthogonality of different ground states due to state distinguishability . many results shown that the fidelity and the entanglement entropy have similar predictive power for identifying quantum phase transitions in the most systems @xcite . however , the ground state fidelity is a model - dependent indicator for quantum phase transitions . it can not be used to characterize the berezinskii - kosterlitz - thouless ( bkt ) transition @xcite . the bkt - like transition occurs at heisenebeg isotropic spin chain with next - nearest - neighbor and in antiferromagnetic anisotropic heisenberg model with @xmath0 @xcite . entanglement , fidelity and their relations with quantum phase transition in high spin chain , just like spin-1 chain need to be further investigated . in this paper , the fidelity and entanglement in the easy - axis spin-1 chain are numerically investigated by using the density - matrix renormalization - group ( dmrg ) technique . in section ii , the hamiltonian of easy - axis spin-1 chain and its critical property are presented . in section iii , the effects of anisotropic interaction on fidelity is investigated and its relation with quantum phase transition is analyzed . the effects of anisotropic interaction on entanglement entropy is calculated and its relation with quantum phase transition is analyzed in section iv . at last , a discussion concludes the paper . it is known that there is a gap in the spectrum of isotropic heisenberg spin-1 chain in the thermodynamic limit . the gap between the ground state and the first excited state energy is usually called haldane gap @xcite . the hamiltonian of an anisotropic heisenberg antiferromagnetic spin-1 chain of @xmath1 sites can be given by @xmath2 where @xmath3 are spin operators on the @xmath4-th site and @xmath1 is the length of the spin chain . the parameter @xmath5 denotes the antiferromagnetic coupling and @xmath6 is considered in the paper . the parameter @xmath7 is anisotropic interaction , @xmath8 is the easy - plane anisotropy and @xmath9 is the easy - axis anisotropy . it is predicted that a novel phase @xmath10 < \delta < \delta_{c2}=1.17 $ ] appears between the xy phase and the nel phase @xcite . many results suggested that a bkt transition occurs at @xmath11 , while that the transition at @xmath12 belongs to the 2d ising universality class @xcite . since the fidelity can not detect a bkt - like phase transition @xcite , we only concentrate on quantum phase transition which happens at easy - axis anisotropy . it is found that the hamiltonian in eq . ( 1 ) can be transformed conveniently to give @xmath13,\\\ ] ] where @xmath14 . the critical point of transition from haldane spin liquid to nel spin solid is around @xmath15 . it is noted that the model of eq . ( 2 ) is invariant for @xmath16 , so that the results in the paper can be equally applied to a system with a ferromagnetic coupling on the @xmath17 plane . the ground state fidelity can be applied to detect the existence of the quantum phase transitions . a general hamiltonian of quantum many - body system can be written as @xmath18 where @xmath19 is the driving hamiltonian and @xmath20 denotes its strength . if @xmath21 represents a state of the system , the ground state fidelity between @xmath21 and @xmath22 can be defined as @xmath23.\ ] ] if the state can be written as @xmath24 , the eq . ( 3 ) can be rewritten as @xmath25 . because @xmath26 reaches its maximum value @xmath27 for @xmath28 , on expanding the fidelity in powers of @xmath29 , the first derivative @xmath30 . by using the property , the fidelity can written by @xmath31 therefore , the average fidelity susceptibility @xmath32 can be given by @xcite @xmath33}{n\delta^2}.\ ] ] it is well known that it is hard to calculate the ground state fidelity because of the lack of knowledge of the ground state function . for models that are not exactly solvable , most of researchers resort to exact diagonalization to obtain the ground state for small size . this method can not precisely quantify the quantum phase transition because the size of the system is too small . recently , the method of density - matrix renormalization - group ( dmrg ) @xcite can be applied to obtain the ground state of the model . moreover , the technique of calculating the overlap of two different ground states by dmrg has been used @xcite . for high precision , the method in @xcite is used to calculate the ground state fidelity susceptibility . we calculate @xmath1 up to @xmath34 when @xmath35 @xcite . the total number of density matrix eigenstates held in system block is @xmath36 in the basis truncation procedure . the matlab codes of finite size density - matrix renormalization - group have double precision . they are performed with three sweep in private computer , and the truncation error is smaller than @xmath37 . the ground state fidelity susceptibility @xmath38 is plotted as a function of anisotropic parameter @xmath39 for different sizes in fig . 1 _ * in dimensionless units_. it is shown that there is a peak in @xmath38 . the peak of @xmath38 increases when the size increases . the location of the peak decreases slightly to small value of @xmath39 as @xmath1 increases . the ground state fidelity @xmath40 is plotted as a function of @xmath39 for different sizes in the inset of fig . 1 _ in dimensionless units_. it is seen that there is a sharp valley corresponding to the peak of @xmath38 . the critical point of fidelity susceptibility @xmath38 and fidelity @xmath40 is about @xmath41 . as we known , the fidelity measures the similarity between two states , while quantum phase transitions are intuitively accompanied by an abrupt change in the structure of the ground state wave - function . this primary observation motivates researchers to use fidelity to predict quantum phase transitions . up to now , the ground state fidelity and fidelity susceptibility have been applied in various many - body systems to detect quantum phase transitions successfully @xcite . it confirms further that the haldane spin liquid@xmath42nel spin solid transition occurs at the point in our system . for comparison , the ground state entanglement entropy is also used to detect the quantum phase transition . for density matrix @xmath43 of any pure state , the entanglement entropy @xmath44 between subsystem @xmath45 and @xmath46 can be defined as @xmath47 where @xmath48is the reduced density matrix obtained from @xmath43 by taking the partial trace over the state space of subsystem @xmath49 . in order to avoid boundary effects , the entanglement entropy of two neighboring central sites of large system needs to be calculated by using the method of dmrg . the entropy of two neighboring central sites labeled @xmath50 is plotted as a function of anisotropy @xmath39 with sizes of @xmath51 in fig . 2 _ in dimensionless units*_. the entanglement entropy of the two neighboring central sites increases monotonously when the anisotropy @xmath39 increases . it is known that the peak or discontinuity in entanglement entropy indicates the quantum phase transition . the peak does not occur at the quantum critical point . it seems that this is due to the monogamy property @xcite . we also calculate the entanglement entropy between the right - hand @xmath52 contiguous qubits and the left - hand @xmath52 contiguous qubits . the result is similar with that shown in fig . 2 . for avoiding repetition , the result is not plotted again . this means that the entropy can not be used to predict quantum phase transition here . however , the critical properties can be captured by the derivatives of the entropy as a function of the anisotropy @xmath39 @xcite . it is noted that the transition from haldane spin liquid to nel spin solid is continuous and belongs to a second order quantum phase transition . it has been shown that the density matrix may be continuous but its first - order derivative is singular at a second order quantum phase transition point @xcite . consequently , we calculate the first - order derivative of entropy @xmath53 at the ground state with @xmath35 . the first derivative of the central two sites entanglement entropy @xmath53 is plotted as a function of anisotropy @xmath39 for different sizes in fig . it is seen that there is a peak in @xmath53 . the location of the peak in @xmath53 moves slightly to small value of @xmath39 as @xmath1 increases . the critical value is about @xmath41 . it represents indeed the quantum phase transition from haldane spin liquid to nel spin solid . in the paper , the fidelity susceptibility and entropy in an easy - axis antiferromagnetic heisenberg spin-1 chain are studied . by using the density - matrix renormalization - group for the model , the effect of anisotropic interaction on fidelity susceptibility and entropy of large size is presented . their relations with quantum phase transition are investigated . it is shown that the quantum phase transition from haldane spin liquid to nel spin solid is clearly marked by the peak ( or valley ) of the fidelity susceptibility ( or fidelity ) . however , the entanglement entropy can not have similar predictive power for revealing quantum phase transition in the system due to the monogamy property , while the first - order derivation of entropy can be used to successfully detect the quantum phase transition . it is a pleasure to thank yinsheng ling , jianxing fang , and xiang hao for their many helpful discussions . the financial support from the national natural science foundation of china ( grant no . 10774108 ) is gratefully acknowledged . of the ground state is plotted as a function of anisotropy @xmath39 for different size @xmath1 . the fidelity @xmath40 is plotted in the inset . the symbols are for @xmath54{0.00,0.00,0.00}{\blacksquare } ) , n=60(\textcolor[rgb]{0.98,0.00,0.00}{{\bullet } } ) , n=80(\textcolor[rgb]{0.00,1.00,0.00}{{\blacktriangle}})$ ] . ]	the fidelity and entropy in an easy - axis antiferromagnetic heisenberg spin-1 chain are studied numerically . by using the method of density - matrix renormalization - group , the effects of anisotropy on fidelity and entanglement entropy are investigated . their relations with quantum phase transition are analyzed . it is found that the quantum phase transition from haldane spin liquid to nel spin solid can be well characterized by the fidelity . the phase transition can be hardly detected by the entropy . but it can be successfully detected by the first deviation of the entropy .
the phenomena of flux creep@xcite and thermally activated flux flow@xcite were initially discovered in conventional hard type - ii superconductors . the irreversibility line " in the @xmath4 plane , which separates the reversible ( or flux flow ) and irreversible magnetization ( or flux pinned ) regions , was thought to be exclusive to the high - temperature superconductors and was thus interpreted as a quasi de almeida - thouless line " in analogy to the irreversibility observed between the zero - field - cooled ( zfc ) and field - cooled ( fc ) magnetization in spin glasses and the possibility of a superconducting glass phase was conjectured.@xcite the vortex - glass phase was suggested to be a new thermodynamic phase within the mixed state of a type - ii superconductor.@xcite the term vortex glass was used to the signify the destruction of long range translational order in the abrikosov lattice due to flux pinning by the underlying disorder in the superconducting material . the irreversibility line was shown to be a sharp boundary across a thermodynamic phase transition between the disordered vortex solid ( or glass ) and the vortex liquid where thermally activated flux flow is possible.@xcite an alternate interpretation of the irreversibility line , without invoking the glass phase in the mixed state , was later offered in terms of the more conventional phenomenon like the flux creep.@xcite these paradigms of high-@xmath5 superconductors were applied to conventional type - ii materials like niobium thin films where it was found that the irreversibility line represents the vortex - melting transition.@xcite the model of the giant flux creep"@xcite was extended to explain the _ width _ and _ shape _ of the resistive transition as a function of applied magnetic field.@xcite the key ingredient to the flux creep picture is the estimate of the activation energy @xmath6 which must be surmounted for the flux movement to occur.@xcite the elemental volume of the flux line lattice which moves during the creep is governed by the activation energy and thus gives the temperature width of the resistive transition as a function of applied field.@xcite the superconducting glass picture@xcite and the giant flux creep picture@xcite were thus shown to be _ qualitatively _ similar by accounting for the field dependence of width of transition ( and in - turn the shape of irreversibility line ) without the need for invoking sample inhomogeneities.@xcite the combination of quenched disorder , higher temperatures , shorter coherence length and large magnetic penetration depth in case of high t@xmath7 superconductors lead to a modification of the mean field phase diagram of the conventional low t@xmath7 superconductors.@xcite an alternate way of studying the vortex matter is to consider the collection of vortices as an elastic object,@xcite instead of treating the problem within the framework of microscopic ginzburg - landau phenomenology.@xcite the rich variety of vortex phases within the superconducting state could be well explained by an elastic theory of vortex matter.@xcite by supplementing the traditional theory of collective pinning@xcite with the concepts of elastic media in random potential , the various properties of the vortex matter could be described in an unified manner.@xcite the application of the theory of elasticity to vortex matter has found similarities in case of other systems like magnetic domain walls , charge density waves and wigner crystals.@xcite the vortex matter has thus provided a model system to study the properties of moving glasses or driven elastic media in presence of quenched disorder.@xcite in the work presented here , we experimentally attempt to find similarities between the flux flow in vortex matter with one more class of a driven elastic system , namely the flow of a molecular glass or viscous liquid . we try to address one of the fundamental questions arising in the studies of vortex matter , which is , how much distance does the fluxon ( or the elemental volume of flux - line lattice ) move for a finite resistance to appear across the resistive transition in a hard type - ii superconductor ? " while this question can be of technological importance for tuning the nature of quenched disorder to achieve desirable critical current densities , it can also lead to fundamental understanding of elastic systems influenced by disorder and driven by external force . we have chosen the nb@xmath0zr@xmath1 alloy for this study . the nb - zr alloy system had generated considerable interest in the past for its perceived use in superconducting wires for high field magnets.@xcite the @xmath8 phase alloy with the b.c.c . structure was well studied due to its high critical current carrying ability.@xcite the same composition in this alloy system was used to investigate fundamental phenomenon like the flux creep in hard superconductors@xcite or the scaling laws for flux pinning in hard type - ii superconductors,@xcite which are now known popularly as kramer scaling . our approach to solve the above posed question is as follows . we first characterize the microsctructure at various length scales of the nb@xmath0zr@xmath1 superconducting alloy under consideration to know about the nature of disorder which can act as pinning potential for the flux - line lattice . the estimate of the activation energy required for flux - line movement is obtained from electrical resistivity measurements across the transition as a function of both temperature and magnetic field . the non - arrhenius shape of the resistive transition draws our attention to a model of viscous flow which has been used earlier to explain a non - arrhenius temperature dependent viscosity across a molecular glass transition . the model is based on purely elastic energy considerations where the activation energy is related to the shear modulus through some sort of a _ correlation _ volume and is popularly known as the _ shoving _ model.@xcite the field - dependent pinning force in the mixed state of our sample can be explained only if more than one pinning mechanisms are considered . such a possibility is explored within the framework of collective pinning,@xcite where the estimates of correlation lengths over which the flux bundles are pinned can be obtained . the elastic constants of vortex matter like the shear modulus ( @xmath9 ) and the tilt modulus ( @xmath10 ) are obtained from magnetization measurements using this framework . the activation energy for flux - line movement and the elastic constants of the vortex matter are then correlated through the correlation volume of flux bundle to get the estimate of the distance over which the flux - line lattice moves before a finite resistance appears across the transition . interestingly , we find that the displacement of vortex matter estimated from the shoving model is of the same order of magnitude as that of the flux - bundle hopping length when a finite resistance appears across the resistive transition . the viscous flow of disordered solids and the flux - line movement in vortex matter appear to be just different cases of the same general phenomenon . these results could possibly lead to newer directions for encompassing diverse phenomenon arising in periodic systems influenced by quenched disorder with vortex matter as a model system . nb@xmath0zr@xmath1 alloy ( @xmath11 1gm mass ) was prepared by arc melting the constituent elements of 99.99% purity in a water cooled copper hearth placed in an inert argon atmosphere . the resulting sample button was remelted six times to ensure homogeneity and was not subjected to any further heat treatment . the sample was then characterized by x - ray diffraction ( xrd ) using a commercial diffractometer ( panalytical x - pert pro mrd ) with cu - k@xmath12 radiation . @xmath13 - 2@xmath13 scans were recorded in the bragg - brentano geometry . a monochromatized ( cu - k@xmath14 ) and collimated ( about 20 arc - sec in the plane of scattering ) x - ray beam was obtained using a hybrid monochromator . the obtained x - ray peak widths , which were significantly larger than the instrumental broadening , were used for the estimation of particle size and lattice strain . a small piece of sample was used for optical metallography which was subjected to slow grinding using sic paper and polishing using diamond compound on a soft cloth . after obtaining a mirror quality finish on the surface , the sample was chemically etched using a mixture of 70 ml c@xmath15h@xmath16o@xmath15 ( lactic acid ) with 30 ml hno@xmath15 ( nitric acid ) and 2 ml hf ( hydrofluoric acid ) for about 15 seconds . the microstructure was observed with a commercial inverted metallurgical optical microscope ( leica dmi5000 m ) . a commercial scanning electron microscope ( philips , xl30cp ) equipped with a energy dispersive spectrometer ( bruker , xflash@xmath17 silicon drift detector ) was used for observing the microstructure on sub - micron length scales and determining the chemical composition at various locations of the sample . for visualizing the disorder / grain structure at even smaller length scales , a small piece of the sample was cut from the same parent button for transmission electron microscopy ( tem ) studies . a commercial tem ( phillips , cm200 ) with a tungsten filament as cathode , was used at an accelerating anode voltage of 200 kv . electrical resistivity was measured using the standard four - probe technique with a home - made variable temperature insert in a commercial superconducting magnet and cryostat system ( american magnetics inc . , usa ) magnetization ( @xmath18 ) measurements were performed as a function of field ( @xmath19 ) and temperature ( @xmath20 ) using a commercial vibrating sample magnetometer ( vsm quantum design , usa ) . we consistently use the si units for all the physical properties reported here . figure [ xrd ] shows the xrd pattern of the sample . the peaks which are quite broad in nature could be indexed with the @xmath8 phase b.c.c . structure with a lattice constant of 3.368 . to estimate the grain size and lattice strain , we use the williamson - hall ( w - h ) plot,@xcite which is shown in the inset of fig . the plot is between @xmath21 and @xmath22 , where @xmath23 is the full width at half maximum for each of the bragg peaks . the ( 110 ) peak appears to be anomalously sharp and does not belong to the straight line w - h fit obtained from the other peaks . the reciprocal of the y - intersept gives the average grain size which is estimated to be around 98 nm . the slope of the w - h plot gives the lattice strain which is found to be about 1.33% . the microstructure ( see fig . [ micro ] ) shows a dendritic growth during the solidification of the alloy from melt , which is quite commonly seen in many materials when a thermal gradient is present between the melt and the solidified portion of the material.@xcite the combined results of xrd and optical metallography show that the alloy has probably solidified with substantial disorder due to the rapid cooling it had to experience because of the small size of the sample button placed in the water cooled copper hearth . the rapid cooling probably did not allow sufficient time for the phase to grow homogeneously . the composition variation across this dendritic pattern was determined using energy dispersive spectroscopy . figure [ sem - eds ] shows the microstructure at two different length scales at a random location on the sample surface . when the composition is determined over a larger area ( fig . [ comp ] ) , it is quite close to the nominal composition of nb@xmath0zr@xmath1 within the error bar of eds measurements , which is nearly 3% for nb and 1% for zr . at certain locations on the protrusions of the dendritic arms ( fig . [ zr - rich ] ) the zr concentration is much higher than the target composition ( 25% ) and has reached up to 40 % at certain locations of the sample . these results imply that the sample is a mixture of zr - rich and zr - deficient regions on the local scale even though the 3:1 ratio of nb to zr is maintained on a larger length scale . figure [ tem ] shows the results of the high resolution tem studies which bring out the nature of disorder in a more clear manner . in fig . [ mismatch ] the major defects are lattice plane bending , mismatch in interplanar spacing and edge dislocations . the inset shows the selected area diffraction ( sad ) pattern which shows the elongation of laue spots . it is known that the lattice constant in the nb - zr alloy system varies with zr concentration.@xcite thus the mismatch in interplanar spacing from 2.61 to 2.42 at a particular location shown in fig . [ mismatch ] is probably due to composition variation across that region . this observation is consistent with the results of eds measurements mentioned earlier . the formation of dislocations ( shown in encircled region ) is most probably due to a large lattice mismatch at certain locations due to compositional variation . when the strain increases beyond a certain limit in heterostructures , the interface accommodates a part of the lattice mismatch through the introduction of dislocations.@xcite figure [ dislocation ] shows another location of the sample which contains lattice plane bending and possibly a jog at one of the edge dislocation . figure [ bending ] shows a location of the sample within the dendritic growth where the lattice is heavily deformed due to the composition variation . it should be noted that no well defined grains ( or grain boundaries ) were observed on length scales estimated from x - ray diffraction . the sample is mostly a disordered alloy with intermittent disturbances in the periodicity of lattice at very short length scales without any sharp grain boundaries . as we see later , the typical length scale of these disturbances in the periodicity of lattice is smaller than the coherence length , which makes the collective pinning of the flux line lattice@xcite more probable than single particle pinning . the applicability of the collective pinning theory will also be shown during the course of this article by estimating certain superconducting parameters . the microstructure of the sample at various length scales shows that the material is highly disordered and this inhomogeneity is expected to manifest in the superconducting properties as well . it is known that the @xmath5 in nb - zr system varies quite substantially with zr concentration.@xcite moreover the lack of long range ordering in the underlying lattice could also result in a response of the sample which is quite different from the bulk of same alloy system . contrary to this expectation in the case of the present sample , we see that the material behaves like a typical bulk type - ii hard superconductor . we highlight this observation before discussing the results of resistivity measurements . figure [ mt ] shows the temperature dependent magnetization of the nb@xmath0zr@xmath1 sample in a low applied magnetic field of @xmath24 = 2mt . the measurement was performed under three different thermomagnetic histories . in the zero - field - cooled ( zfc ) protocol , the sample is first cooled in zero field down to the lowest temperature . the magnetic field is applied at the lowest temperature and magnetization is measured during warming of the sample . the superconducting transition temperature t@xmath25 is estimated to be slightly below 11k . in the field - cooled - cooling ( fcc ) protocol , the sample is cooled in presence of field from above @xmath5 . the temperature at which the fcc curve deviates from the zfc curve is termed as the irreversibility temperature t@xmath26 . after reaching the lowest temperature in presence of field , the measurements are performed while warming the sample in field and this protocol is named as the field - cooled - warming ( fcw ) protocol . the fcw curve deviates from the fcc curve and meets the zfc curve at a temperature t@xmath27 which is slightly lower than t@xmath26 . these thermal history effects , especially the hysteresis between the fcc and fcw curves , are indications that the sample behaves like a typical bulk hard type - ii superconductor.@xcite we had shown earlier that in case of nanocrystalline superconductors,@xcite the fcc and fcw curves may coincide because the grain size is much smaller than the flux trapping depth@xcite inside the sample . figure [ width ] shows the resistive transition from the normal state to the superconducting state in the of the nb@xmath0zr@xmath1 sample in various applied fields . the superconducting transition width is less than 0.1k in zero field which is quite unusual for the sample with such composition variation shown earlier this further confirms that though there is substantial amount of disorder in the sample , it behaves like a bulk superconducting sample of single composition . the width of the transition as a function of field is shown in the inset to fig . [ width ] . the observed width can be described by the equation , @xmath28 with @xmath29 = 1.82 . the value of the exponent is not 2/3 as calculated theoretically for a giant flux creep model or superconducting glass.@xcite however , it should be noted that the quantitative dependence of the transition width on the applied field depends on the nature of activation energy.@xcite the quantitative predictions of the theory may actually depend on the fluxon jumping length.@xcite a deviation from the 2/3 behaviour has been observed in other systems as well.@xcite it is the central aim of this work to find out the distance over which the flux moves before a finite resistance appears across the transition . it is therefore important to know the exact form of the activation energy as we see next . figure [ rt - non - arr ] shows resistivity as a function of inverse temperature . the resistivity can not be explained by a simple arrhenius relation where the activation energy is independent of temperature . rather , it takes a non - arrhenius form which can be expressed as,@xcite @xmath30 where @xmath31 substituting equation [ u - h ] in equation [ rho ] , we get the temperature dependent resistivity at a particular field as , @xmath32 \label{non - arr - rt - eqn}\ ] ] the best fit is obtained for @xmath33 = 4 with a regression better than 0.996 . figure [ rh - non - arr ] shows the resistive transition as a function of magnetic field at various temperatures . the magnetic field dependent isothermal resistivity can be expressed with a similar non - arrhenius form as , @xmath34 \label{non - arr - rh - eqn}\ ] ] with the exponent @xmath8 once again = 4 . the activation energies @xmath35 determined from both temperature dependent resistivity in constant magnetic field and isothermal field dependent resistivity match quite closely as can be seen from table [ tab : activation ] . the prefactor @xmath6 is quite large and is not the actual activation energy . rather , it is the tangential slope of the arrhenius plot and is thus termed as the apparent activation energy.@xcite [ cols="^,^,^,^",options="header " , ] [ tab : disp ] in conclusion , we have studied the resistive transition of vortex matter in highly strained sample of nb@xmath0zr@xmath1 . the detailed study of microstructure of the sample enabled us to apply the theory of collective pinning to understand the pinning properties of flux line lattice in our sample and determine the elastic constants of vortex matter . the non - arrhenius shape of the resistive transition , both as a function of temperature and magnetic field , showed that the models of viscous flow of disordered solids can be indeed applied to vortex matter . the arguments were based on purely elastic energy considerations instead of assuming any particular model or shape of the irreversibility line in the field - temperature phase space . our results show that the viscous flow of disordered solids and the flux - creep phenomenon in hard type - ii superconductors could be the manifestation of same underlying physical principles . these studies should provide sufficient interesting inputs for further experiments which can image the actual flow of flux line lattice and those theories which treat the flux flow in type - ii superconductors as special case of the general phenomenon of plastic depinning of driven systems . the authors wish to acknowledge dr . l. s. sharath chandra for help with magnetization measurements , dr . tapas ganguli for help with the x - ray diffraction measurements and dr . k. srivastava for guidance on the transmission electron microscopy measurements .	we present the results of magnetization and magneto - transport measurements in the superconducting state of an as - cast nb@xmath0zr@xmath1 alloy . we also report the careful investigation of the microstructure of our sample at various length scales by using optical , scanning electron and transmission electron microscopies . the information of microstructure is used to understand the flux pinning properties in the superconducting state within the framework of collective pinning . the magneto - transport measurements show a non - arrhenius behaviour of the temperature and field dependent resistivity in the flux flow region . this non - arrhenius behaviour is understood in terms of a model , which was originally proposed for viscous flow of disordered solids and is popularly known in the literature as the ` shoving ' model . the activation energy for flux flow is obtained from magneto - transport measurements and is assumed to be mainly the elastic energy stored in the flux - line lattice . the critical current density estimated from magnetization measurements is moderately high of the order of 10@xmath2 am@xmath3 at a temperature of 2k . the scaling of pinning force density with respect to reduced field indicates the presence of two pinning mechanisms of different origins . the elastic constants of the flux - line lattice are estimated from magnetization measurements and are used to estimate the length scale of vortex lattice movement , or the volume displaced by the flux - line lattice , during flux flow . it appears that the vortex lattice displacement estimated from elastic energy considerations is of the same order of magnitude as that of the flux - bundle hopping length when a finite resistance appears during flux flow . our results could provide possible directions for establishing a framework where vortex matter and glass forming liquids or amorphous solids can be treated in a similar manner for understanding the phenomenon of viscous flow in disordered solids or more generally the pinning and depinning properties of elastic manifolds in random media .
the inverse scattering problem , non - destructive evaluation , is one of the intriguing research topics since it is closely related to human life . it is because it applies not only to physics , engineering , or image medical science but also to identifying the cracks of the structures such as concrete walls , machines , or buildings therefore , it has been considerably investigated by many researchers to suggest the algorithm regarding this problem or to experiment and analyze previously suggested algorithms . related works can be found in @xcite and references therein . however , the inverse scattering problem is such a difficult problem that not many methods have been studied other than the reconstruction method based on the iterative method such as newton - type method , refer to @xcite . regarding the algorithms of using a newton - type method , in case the initial shape is quite different from the unknown target , the reconstruction of material leads to failure with the non - convergence or yielding faulty shapes . even though the reconstruction ends up with a successful result , it could take a great deal of time . therefore , several non - iterative algorithms have been suggested because they can reconstruct the shape that is quite similar to the target , and thus it can be used as a good initial guess , which also takes a short time and efficient in iterative methods ( see @xcite and references therein ) . among them , the non - iterative reconstruction algorithm such as kirchhoff and subspace migration has been consistently studied thanks to its better imaging product . however , the existing research on this algorithm has been applied heuristically . that is , it mostly relied on experimental results . although research on several structures was conducted , it was based on experiments or statistical approach , not revealing the mathematical structures explicitly , refer to @xcite and references therein . it resulted in the difficulties of explaining the results theoretically . recently , an analysis of mathematical structure of single- and multi - frequency subspace migration for imaging of small electromagnetic materials has been conducted by establishing a relationship with bessel function of integer order in full - view inverse scattering problems , refer to @xcite . this remarkable research has shown the reason why subspace migration is effective and why application of multi - frequency guarantees better imaging performance than application of single - frequency . motivated by this work , this analysis is successfully extended to the limited - view inverse scattering problems ( see @xcite ) . afterwards , a multi - frequency subspace migration weighted by applied frequency has been suggested in @xcite to obtain more precise results . furthermore , the reason why the suggested algorithm presents better imaging products was demonstrated mathematically and an analysis of multi - frequency subspace migration weighted by the power of applied frequency has been considered in @xcite . this research concludes that increasing the power of applied frequency is meaningless , so that multi - frequency subspace migration weighted by applied frequency is a good algorithm for imaging . recently , it has been confirmed that a multi - frequency subspace migration weighted by the logarithmic function of applied frequency can yield more appropriate imaging result than the one suggested in @xcite . however , one can face difficulties in identifying the reason why it shows the better performance through the mathematical analysis . motivated by this difficulty , we derive an indefinite integration of square of bessel function of order zero of the first kind multiplied by the natural logarithmic function . based on this integration , we discover the structure of multi - frequency subspace migration weighted by the logarithmic function of applied frequency by establishing a relationship with bessel function of integer order of the first kind , and provide the reason of better imaging performance . the organization of this study is as follows . in section [ sec2 ] , we briefly introduce two - dimensional direct scattering problem and subspace migration . section [ sec3 ] provides a survey on the structures of single- , multi- , and weighted multi - frequency subspace migrations , the derivation of indefinite integration of square of bessel function multiplied by the natural logarithmic function , and the the mathematical analysis on why multi - frequency subspace migration weighted by natural logarithmic function shows better imaging performance than the traditional one . in section [ sec4 ] , several results of numerical experiments with noisy data are presented in order to support our analysis . finally , a short conclusion is mentioned in section [ sec5 ] . let us consider two - dimensional electromagnetic scattering from a thin , curve - like homogeneous inclusion within a homogeneous space @xmath0 . the latter contains an inclusion denoted as @xmath1 which is localized in the neighborhood of a curve @xmath2 . that is , @xmath3 where the supporting @xmath2 is a simple , smooth curve in @xmath0 , @xmath4 is the unit normal to @xmath2 at @xmath5 , and @xmath6 is a strictly positive constant which specifies the thickness of the inclusion ( small with respect to the wavelength ) , refer to figure [ thininclusion ] . throughout this paper , we denote @xmath7 be the unit tangent vector at @xmath8 . sketch of the thin inclusion @xmath1 in two - dimensional space @xmath0.,scaledwidth=35.0% ] in this paper , we assume that every material is characterized by its dielectric permittivity and magnetic permeability at a given frequency . let @xmath9 and @xmath10 denote the permittivity and permeability of the embedding space @xmath0 , and @xmath11 and @xmath12 the ones of the inclusion @xmath1 . then , we can define the following piecewise constant dielectric permittivity @xmath13 and magnetic permeability @xmath14 respectively . note that if there is no inclusion , i.e. , in the homogeneous space , @xmath15 and @xmath16 are equal to @xmath17 and @xmath18 respectively . in this paper , we set @xmath19 and @xmath20 for convenience . at strictly positive operation frequency @xmath21 ( wavenumber @xmath22 ) , let @xmath23 be the time - harmonic total field which satisfies the helmholtz equation @xmath24 similarly , the incident field @xmath25 satisfies the homogeneous helmholtz equation @xmath26 as is usual , the total field @xmath23 divides itself into the incident field @xmath25 and the scattered field @xmath27 , @xmath28 . notice that this unknown scattered field @xmath29 satisfies the sommerfeld radiation condition @xmath30 uniformly in all directions @xmath31 . in this paper , we consider the illumination of plane waves @xmath32 and far fields in free space where @xmath33 is a two - dimensional vector on the unit circle @xmath34 in @xmath0 . for convenience , we denote @xmath35 to be a discrete finite set of observation directions and @xmath36 be the set of incident directions . the far - field pattern is defined as a function @xmath37 which satisfies @xmath38 as @xmath39 uniformly on @xmath40 and @xmath41 . then , based on @xcite , @xmath37 can be written as an asymptotic expansion formula . for @xmath42 and @xmath43 , the far - field pattern @xmath44 can be represented as @xmath45 where @xmath46 is uniform in @xmath47 , @xmath42 , and @xmath48 is a @xmath49 symmetric matrix defined as follows : let @xmath50 and @xmath51 denote unit tangent and normal vectors to @xmath2 at @xmath52 , respectively . then * @xmath48 has eigenvectors @xmath50 and @xmath51 . * the eigenvalue corresponding to @xmath50 is @xmath53 . * the eigenvalue corresponding to @xmath51 is @xmath54 . now , we introduce subspace migration for imaging of thin inclusion @xmath1 . detailed description can be found in @xcite . in order to introduce , we generate a multi - static response ( msr ) matrix @xmath55_{j , l=1}^{n}\in\mathbb{c}^{n\times n}$ ] whose element @xmath56 is the collected far - field at observation number @xmath57 for the incident number @xmath58 . in this paper , we assume that @xmath59 , i.e. , we have the same incident and observation directions configuration . it is worth emphasizing that for a given frequency @xmath60 , based on the resolution limit , any detail less than one - half of the wavelength can not be retrieved . hence , if we divide thin inclusion @xmath1 into @xmath61 different segments of size of order @xmath62 , only one point , say , @xmath63 , @xmath64 , at each segment will affect the imaging ( see @xcite ) . if @xmath65 , the elements of msr matrix can be represented as follows : @xmath66e^{i\omega({\boldsymbol{\theta}}_j+{\boldsymbol{\theta}}_l)\cdot{\mathbf{y}}_m } , \end{aligned}\end{aligned}\ ] ] where @xmath67 denotes the length of @xmath2 . now , let us perform the singular value decomposition ( svd ) of @xmath68 @xmath69 where @xmath70 , @xmath71 nonzero singular values such that @xmath72 and @xmath73 and @xmath74 are left- and right - singular vectors of @xmath68 , respectively . based on the structure of ( [ elementofmsr ] ) , define a vector @xmath75 as @xmath76e^{i\omega{\boldsymbol{\theta}}_1\cdot{\mathbf{z}}},{\mathbf{c}}\cdot[1,{\boldsymbol{\theta}}_2]e^{i\omega{\boldsymbol{\theta}}_2\cdot{\mathbf{z}}},\cdots , { \mathbf{c}}\cdot[1,{\boldsymbol{\theta}}_n]e^{i\omega{\boldsymbol{\theta}}_n\cdot{\mathbf{z}}}\bigg]^t,\ ] ] where the selection of @xmath77 depends on the shape of the supporting curve @xmath78 ( see ( * ? ? ? * section 4.3.1 ) for a detailed discussion ) . then , in accordance with @xcite , @xmath79 for some @xmath80 and @xmath81 , @xmath64 . based on the orthonormal property of singular vectors , the first @xmath61 columns of @xmath82 and @xmath83 are orthonormal , it follows that @xmath84 where @xmath85 . hence , we can introduce subspace migration for imaging of thin inclusion at a given frequency @xmath21 as @xmath86 based on the properties ( [ orthonormal ] ) , map of @xmath87 should exhibit peaks of magnitude @xmath88 at @xmath89 , and of small magnitude at @xmath90 . in recent work @xcite , the structure of @xmath87 is derived as follows . throughout this paper , we assume that the set of incident ( and also observation ) directions @xmath91 spans @xmath34 , and @xmath92 be the _ bessel function of integer order @xmath93 of the first kind_. if the total number of incident and observation directions @xmath94 is sufficiently large and satisfies @xmath95 . then single - frequency subspace migration ( [ singlesm ] ) can be represented as follows : @xmath96 this result tells us that although the shape of @xmath1 can be recognized via the map of @xmath87 , some unexpected artifacts should appear in the map of @xmath87 due to the oscillating property of bessel function . in order to obtain better result , a normalized multi - frequency subspace migration is considered . this is introduced as follows ; for multi - frequency @xmath97 , by an assumption of @xmath98 for all @xmath99 , a normalized multi - frequency subspace migration @xmath100 is given by @xmath101 where @xmath102 is the number of nonzero singular values of msr matrix @xmath103 , @xmath104 . then , the structure of ( [ multism ] ) can be represented as follows . if the total number of incident and observation directions @xmath94 is sufficiently large and satisfies @xmath105 for @xmath104 . then , multi - frequency subspace migration ( [ multism ] ) can be represented as follows : @xmath106 this shows that ( [ multism ] ) yields better images owing to less oscillation than ( [ singlesm ] ) does so that unexpected artifacts in the plot of @xmath100 are mitigated when @xmath107 is sufficiently large . this result indicates why a multi - frequency subspace migration offers images with good resolution . on the other hand , it is possible to examine the same conclusion throughout the basis of statistical hypothesis testing in statistical theory , refer to @xcite . in order to improve multi - frequency subspace migration ( [ multism ] ) , one must eliminate or control the last term of ( [ structuremultism ] ) . for this purpose , a weighted multi - frequency subspace migration has been introduced @xmath108 and following result is obtained . if the total number of incident and observation directions @xmath94 is sufficiently large and satisfies @xmath105 for @xmath104 . then , multi - frequency subspace migration ( [ weightedmultism ] ) can be represented as follows : @xmath109 based on recent work @xcite , the term @xmath110 , which is disturbing the imaging performance , is eliminated when @xmath111 and remained when @xmath112 , respectively . hence , it can be said that @xmath113 is an improved version of @xmath100 . based on the improved procedure presented above , it is natural to consider the following multi - frequency subspace migration weighted by some function @xmath114 : @xmath115 then , based on the results in section [ sec3 - 1 ] , @xmath116 should be represented as the following form : @xmath117 where @xmath118 is a positive definite function . based on several recent works @xcite , the term @xmath119 is disturbing the imaging performance since it generates some unnecessary artifacts . however , if we can find a suitable function @xmath120 , which makes @xmath119 as a negative valued function in the neighborhood of @xmath1 , and small ( or negative ) valued one at the outside of neighborhood of @xmath1 , @xmath116 will be an improved subspace migration . recently , it has been confirmed that @xmath116 is an improved version of @xmath113 when @xmath121 . however , this fact has been examined via some results of numerical simulations . thus , identification of mathematical structure of ( [ wmsm ] ) is still remaining . in order to identify this , we derive the following indefinite integration . this plays an important role in exploring the structure of ( [ wmsm ] ) . note that the derivation of following integration is easy but we have been unable to find such a derivation . [ theoremlogintegral ] for every positive real number @xmath122 , following identity holds @xmath123 since @xmath124 an integration by parts yields @xmath125 applying following identity ( see @xcite ) @xmath126 leads us @xmath127 combining ( [ term1 ] ) and ( [ term2 ] ) , we can obtain the desired result ( [ integralformulalog ] ) . this completes the proof . applying theorem [ theoremlogintegral ] , we can explore the structure of @xmath116 . [ structurewmsm ] assume that total number of incident and observation directions @xmath94 is sufficiently large and satisfies @xmath105 for @xmath104 . then , multi - frequency subspace migration ( [ wmsm ] ) weighted by @xmath121 can be represented as follows : @xmath128 for the sake of simplicity , we assume that @xmath98 for all @xmath104 , and denote @xmath129 then by virtue of @xcite , ( [ structureweightedmultism ] ) can be written as follows @xmath130 then , the change of variable @xmath131 yields @xmath132 since @xmath133_{\omega_1\|{\mathbf{z}}-{\mathbf{y}}_m\|}^{\omega_f\|{\mathbf{z}}-{\mathbf{y}}_m\| } -\int_{\omega_1\|{\mathbf{z}}-{\mathbf{y}}_m\|}^{\omega_f\|{\mathbf{z}}-{\mathbf{y}}_m\|}(2-\ln(x))j_1(x)^2dx\right\ } , \end{gathered}\ ] ] applying theorem [ theoremlogintegral ] yields @xmath134 finally , let us apply following indefinite integral formula of the bessel function ( see @xcite ) : @xmath135 then , an elementary calculus leads us to @xmath136 by combining ( [ term3 ] ) and ( [ term4 ] ) , we can obtain @xmath137 with this , we can obtain the desired result by applying ( [ besselformula ] ) again . this completes the proof . the above result leads us to the following result of improvement . [ improvementtheorem ] under the same assumption of theorem [ structurewmsm ] , multi - frequency subspace migration ( [ wmsm ] ) weighted by @xmath121 improves @xmath113 of ( [ weightedmultism ] ) . throughout the proof , we assume that all @xmath138 are sufficiently large enough and @xmath139 are small enough such that , we set @xmath140 and @xmath141 . therefore , the value @xmath142 is smaller than @xmath88 . ] @xmath143 in order to show the improvement of ( [ wmsm ] ) , we recall the structure ( [ structureweightedmultism ] ) @xmath144 where @xmath145 based on the structure , it is enough to show that @xmath146 , i.e , the term @xmath147 is negative . 1 . suppose that @xmath148 is far away from @xmath63 , then since @xmath21 is sufficiently large , the following asymptotic form holds for any integer @xmath93 , @xmath149 hence , @xmath150 . moreover , applying lhpital s rule , it is easy to observe that @xmath151 hence , we can conclude that the values @xmath152 and @xmath153 are negligible , i.e. , @xmath154 . assume that @xmath148 is close enough to @xmath63 such that @xmath155 for all @xmath104 . let @xmath156 denotes the gamma function . then , since the following asymptotic property of bessel function holds for any integer @xmath93 , @xmath157 we can observe that @xmath158 and @xmath159_{\omega_1}^{\omega_f}\\ & = \frac{1}{2}\bigg(\omega_f\ln(\omega_f)-2\omega_f-\omega_1\ln(\omega_1)+2\omega_1\bigg)\\ & = \frac{1}{2}\left\{\omega_f\ln\left(\frac{\omega_f}{\omega_1}\right)+(\omega_f-\omega_1)\ln(\omega_1)-2(\omega_f-\omega_1)\right\}\\ & < \frac{1}{2}\bigg(c\omega_f+\omega_f\ln(\omega_1)+2\omega_1\bigg)=\mathcal{o}(\omega_f ) . \end{aligned}\ ] ] therefore , the term @xmath153 is dominated by @xmath152 , i.e. , @xmath147 is of negative value . based on above observations , it is true that the term @xmath147 is negative in the neighborhood of @xmath1 , and close to zero at the outside of the neighborhood of @xmath1 ( see figure [ negativeregion ] ) . this means that the results via ( [ wmsm ] ) with @xmath160 will be better owing to less oscillation than ( [ weightedmultism ] ) with @xmath111 does . this completes the proof . description of the result in theorem [ improvementtheorem].,scaledwidth=50.0% ] note that @xmath116 has its maximum value at @xmath161 . hence , we can immediately obtain following result of unique determination . let the applied frequency @xmath21 be sufficiently high . if the total number @xmath94 of incident and observation directions and total number @xmath107 of applied frequencies are sufficiently large , then the shape of supporting curve @xmath2 of thin inclusion @xmath1 can be obtained uniquely via the map of @xmath116 . some numerical simulation experiments are performed in order to support theorems [ structurewmsm ] and [ improvementtheorem ] . throughout this section , the search vector @xmath148 is included in the square @xmath162\times[-1,1]$ ] . in order to describe thin inclusions @xmath163 , two smooth curves are selected as follows : @xmath164^t~:~-0.5\leq s\leq0.5\right\}}\\ \sigma_2&={\left\{[s+0.2,s^3+s^2 - 0.6]^t~:~-0.5\leq s\leq0.5\right\}}.\end{aligned}\ ] ] the thickness @xmath6 of thin inclusions @xmath163 is equally set to @xmath165 . we denote @xmath166 and @xmath167 be the permittivity and permeability of @xmath163 , respectively , and set parameters @xmath167 , @xmath17 , @xmath166 and @xmath18 are @xmath168 and @xmath88 , respectively . since @xmath17 and @xmath18 are set to unity , the applied frequencies reads as @xmath138 at wavelength @xmath169 for @xmath170 , which will be varied in the numerical examples between @xmath140 and @xmath171 . for the incident directions @xmath33 , they are selected as @xmath172^t\quad\mbox{for}\quad l=1,2,\cdots , n,\ ] ] and @xmath173 total number of directions has chosen . since @xmath174 , the dominant eigenvectors of @xmath103 are @xmath175 , @xmath176^t$ ] will be the best choice . however , we have no _ a priori _ information of @xmath175 , @xmath177^t$ ] has selected for generating @xmath178 of ( [ vecw ] ) . for a more detailed discussion , we recommend a recent work ( * ? ? ? * section 4.3 ) . in order to show the robustness , a white gaussian noise with @xmath179db signal - to - noise ratio ( snr ) added to the unperturbed far - field pattern data @xmath180 via a standard matlab command ` awgn ` included in the ` communications system toolbox ` package . for discriminating non - zero singular values of msr matrix @xmath103 , a @xmath181threshold scheme is applied for each @xmath99 ( see @xcite for instance ) . throughout this section , only both permittivity and permeability contrast case is considered . in figure [ figure1 ] , some imaging results via @xmath182 , @xmath183 , and @xmath184 are exhibited when the thin inclusion is @xmath185 . by comparing these results , we can immediately observe that unexpected artifacts can be examined in the results via @xmath182 and @xmath183 but , as we expected , they dramatically disappeared in the map of @xmath184 . maps of @xmath182 ( left ) , @xmath183 ( center ) , and @xmath184 ( right ) when the thin inclusion is @xmath185.,title="fig:",scaledwidth=32.5% ] maps of @xmath182 ( left ) , @xmath183 ( center ) , and @xmath184 ( right ) when the thin inclusion is @xmath185.,title="fig:",scaledwidth=32.5% ] maps of @xmath182 ( left ) , @xmath183 ( center ) , and @xmath184 ( right ) when the thin inclusion is @xmath185.,title="fig:",scaledwidth=32.5% ] now , let us consider the imaging results of @xmath186 in figure [ figure2 ] . based on this , we can observe the same phenomenon as in figure [ figure1 ] . although proposed algorithm successfully eliminates artifacts , some part of @xmath186 ca nt be visible . this is due to the selection of @xmath177^t$ ] . based on the shape of @xmath187 , the unit normal direction @xmath188 is similar to @xmath189^t$ ] for @xmath190 . when @xmath191 , @xmath188 is immensely different from @xmath189^t$ ] . hence , finding an optimal @xmath192 is still remaining as an interesting subject . same as figure [ figure1 ] except the thin inclusion is @xmath186.,title="fig:",scaledwidth=32.5% ] same as figure [ figure1 ] except the thin inclusion is @xmath186.,title="fig:",scaledwidth=32.5% ] same as figure [ figure1 ] except the thin inclusion is @xmath186.,title="fig:",scaledwidth=32.5% ] now , let us extend the proposed algorithm for imaging of two-(or more ) different thin inclusions @xmath185 and @xmath186 . for the sake of simplicity , we denote @xmath193 . figure [ figurem1 ] shows maps of @xmath182 , @xmath183 , and @xmath184 for @xmath194 with the same permittivities @xmath195 and permeabilities @xmath196 . it is interesting to observe that unlike the case of single inclusion , where almost unexpected artifacts are eliminated , proposed imaging functional @xmath116 does not improve the traditional ones @xmath100 and @xmath113 . hence , further analysis is needed to identify the reason . maps of @xmath182 ( left ) , @xmath183 ( center ) , and @xmath184 ( right ) when the thin inclusion is @xmath194 with same material property.,title="fig:",scaledwidth=32.5% ] maps of @xmath182 ( left ) , @xmath183 ( center ) , and @xmath184 ( right ) when the thin inclusion is @xmath194 with same material property.,title="fig:",scaledwidth=32.5% ] maps of @xmath182 ( left ) , @xmath183 ( center ) , and @xmath184 ( right ) when the thin inclusion is @xmath194 with same material property.,title="fig:",scaledwidth=32.5% ] for the final example , we consider the imaging of multiple inclusions @xmath194 with different properties @xmath197 and @xmath198 . results in numerical simulations are exhibited in figure [ figurem2 ] . by comparing the results in figure [ figurem1 ] , we can observe that same as the imaging of single inclusion , almost every artifact has disappeared in the map of @xmath184 while some of they are still remaining in the maps of @xmath182 , @xmath183 . same as figure [ figurem2 ] except different material properties.,title="fig:",scaledwidth=32.5% ] same as figure [ figurem2 ] except different material properties.,title="fig:",scaledwidth=32.5% ] same as figure [ figurem2 ] except different material properties.,title="fig:",scaledwidth=32.5% ] to conclude this section , let us present some remarks . from the derivation of theorem [ structurewmsm ] , it follows that the number @xmath94 of incident and observation directions and the number @xmath107 of applied frequencies have to be large enough . furthermore , the selection of @xmath192 of ( [ vecw ] ) must follow the unit normal direction on the supporting curve @xmath2 . on the other hand , we point out that if there are two inclusions with the same material property , our analysis and observation are not valid anymore . in fact , the derivation of asymptotic expansion formula in the existence of multiple inclusions , one must assume that they are well - separated to each other . hence , we expect that if the distance between two inclusions are sufficiently large , the result will be nice . but , this fact does not guarantee the improvement of proposed algorithm . thus , finding a method of improvement is still required . finally , we believe that although the result in this paper does not guarantee the complex shape of thin inclusions due to the intrinsic rayleigh resolution limit , they can be good initial guesses of a level - set method or of a newton - type reconstruction algorithm , refer to @xcite and references therein . in the present study , we have proposed a multi - frequency subspace migration weighted by the natural logarithmic function for imaging of thin , crack - like electromagnetic inclusions . this is based on the asymptotic expansion formula of far - field pattern in the existence of such inclusions and the structure of constructed msr matrix operated at multiple frequencies . throughout a careful analysis and numerical experiment , it is confirmed that proposed method successfully improves traditional approaches . however , a counter example was discovered when one tries to find the shape of multiple inclusions with the same material properties . hence , investigating the reason will be an interesting work . furthermore , for achieving the best imaging of inclusions , finding _ a priori _ information of supporting curve , e.g. , unit outward normal vector , should be a remarkable research . in this paper , we considered the imaging of inclusions located in the homogeneous space but based some recent works @xcite , subspace migration can be applicable for imaging of targets buried in the half - space . hence , an extension to the half - space problem is expected . and , similarly to @xcite , the improvement considered herein can be extended to the limited - view inverse scattering problems . 00 r. acharya , r. wasserman , j. stevens , and c. hinojosa , biomedical imaging modalities : a tutorial , comput . medical imag . graph . , 19 ( 1995 ) , 325 . d. lvarez , o. dorn , n. irishina and m. moscoso , crack reconstruction using a level - set strategy , j. comput . ( 2009 ) , 57105721 . h. ammari , g. bao , and j. flemming , an inverse source problem for maxwell s equations in magnetoencephalography , siam j. appl . , 62 ( 2002 ) , 13691382 . h. ammari , e. bonnetier and y. capdeboscq , enhanced resolution in structured media , siam j. appl . , 70 ( 2009 ) , 14281452 . h. ammari , j. garnier , h. kang , w .- k . park and k. slna , imaging schemes for perfectly conducting cracks , siam j. appl . math , 71 ( 2011 ) , 6891 . h. ammari , e. iakovleva and d. lesselier , a music algorithm for locating small inclusions buried in a half - space from the scattering amplitude at a fixed frequency , multiscale model . simul . , 3 ( 2005 ) , 597628 . h. ammari , h. kang , e. kim , k. louati , m. vogelius , a music - type algorithm for detecting internal corrosion from electrostatic boundary measurements , numer . ( 2008 ) , 501528 . h. ammari , h. kang , h. lee and w .- k . park , asymptotic imaging of perfectly conducting cracks , siam j. sci . comput . , 32 ( 2010 ) , 894922 . s. r. arridge , optical tomography in medical imaging , inverse problems , 15 ( 1999 ) , r41r93 . e. beretta and e. francini , asymptotic formulas for perturbations of the electromagnetic fields in the presence of thin imperfections , contemp . , 333 ( 2003 ) , 4963 . g. bao , s. hou and p. li , inverse scattering by a continuation method with initial guesses from a direct imaging algorithm , j. comput . phys . , 227 ( 2007 ) , 755762 . l. borcea , g. papanicolaou and c. tsogka , subspace projection filters for imaging in random media , c. r. mecanique , 338 ( 2010 ) , 390401 . m. burger , b. hackl and w. ring , incorporating topological derivatives into level set methods , j. comput . phys . , 194 ( 2004 ) , 344362 . x. chen and k. agarwal music algorithm for two - dimensional inverse problems with special characteristics of cylinders , ieee trans . antennas propag . , 56 ( 2008 ) , 18081812 . x. chen and y. zhong , music electromagnetic imaging with enhanced resolution for small inclusions , inverse problems , 25 ( 2009 ) , 015008 . m. cheney and d. isaacson , distinguishability in impedance imaging , ieee trans . biomed . , 39 ( 1992 ) , 852860 . m. cheney , d. isaacson , and j. c. newell , electrical impedance tomography , siam rev . , 41 ( 1999 ) , 85101 . d. colton and a. kirsch , a simple method for solving inverse scattering problems in the resonance region , inverse problems , 12 ( 1996 ) , 383 . a. j. devaney , super - resolution processing of multi - static data using time - reversal and music , available at http://www.ece.neu.edu/faculty/devaney/ajd/preprints.htm . o. dorn and d. lesselier , level set methods for inverse scattering , inverse problems , 22 ( 2006 ) , r67r131 . t. d. dorney , j. l. johnson , j. v. rudd , r. g. baraniuk , w. w. symes and d. m. mittleman , terahertz reflection imaging using kirchhoff migration , opt . ( 2001 ) , 15131515 . a. s. fokas , y. kurylev , and v. marinakis , the unique determination of neural currents in the brain via magnetoencephalography , inverse problems , 20 ( 2004 ) , 10671082 . r. griesmaier , reciprocity gap music imaging for an inverse scattering problem in two - layered media , inverse probl . imag . , 3 ( 2009 ) , 389403 . y. a. gryazin , m. v. klibanov , t. r. lucas , two numerical methods for an inverse problem for the 2-d helmholtz equation , j. comput . phys . , 184 ( 2003 ) , 122 - 148 . s. hou , k. huang , k. slna and h. zhao , a phase and space coherent direct imaging method , j. acoust . , 125 ( 2009 ) , 227238 . s. hou , k. slna and h. zhao , imaging of location and geometry for extended targets using the response matrix , j. comput . phys . , 199 ( 2004 ) , 317338 . d. isaacson , distinguishability of conductivities by electric current computed tomography , ieee trans . medical imag . , 5 ( 1986 ) , 9195 . d. isaacson and m. cheney , effects of measurements precision and finite numbers of electrodes on linear impedance imaging algorithms , siam j. appl . , 51 ( 1991 ) , 17051731 . a. ishimaru , t .- k . chan and y. kuga , an imaging technique using confocal circular synthetic aperture radar , ieee trans . , 36 ( 1998 ) , 15241530 . a. ishimaru , c. zhang , m. stoneback and y. kuga , time - reversal imaging of objects near rough surfaces based on surface flattening transform , waves random complex media , 23 ( 2013 ) , 306 - 317 y .- d . joh , y. m. kwon , j. y. huh and w .- k . park , structure analysis of single- and multi - frequency subspace migrations in inverse scattering problems , prog . , 136 ( 2013 ) , 607622 . y .- d . joh , y. m. kwon , j. y. huh and w .- k . park , weighted multi - frequency imaging of thin , crack - like electromagnetic inhomogeneities , poster session , proceeding of progress in electromagnetics research symposium in taipei , ( 2013 ) , 631635 . d . joh and w .- k . park , an optimized weighted multi - frequency subspace migration for imaging perfectly conducting , arc - like cracks , proceedings of the 6th international congress on image and signal processing , 1 ( 2013 ) , 250255 . o. kwon , e. j. woo , j. r. yoon , and j.k . seo , magnetic resonance electrical impedance tomography ( mreit ) : simulation study of @xmath199substitution algorithm , ieee trans . biomed . , 49 ( 2002 ) , 160167 . o. kwon , j. r. yoon , j. k. seo , e. j.woo , and y. g. cho , estimation of anomaly location and size using impedance tomography , ieee trans . biomed . engr . , 50 ( 2003 ) , 8996 . y. m. kwon and w .- k . park , analysis of subspace migration in limited - view inverse scattering problems , appl . lett . , 26 ( 2013 ) , 11071113 . m. l. moran , r. j. greenfield , s. a. arcone and a. j. delaney , multidimensional gpr array processing using kirchhoff migration , j. appl . geophys . , 43 ( 2000 ) , 281295 . park , analysis of a multi - frequency electromagnetic imaging functional for thin , crack - like electromagnetic inclusions , appl . , 77 ( 2014 ) , 3142 . w .- k . park , non - iterative imaging of thin electromagnetic inclusions from multi - frequency response matrix , prog . res . , 106 ( 2010 ) , 225241 . w .- k . park , on the imaging of thin dielectric inclusions buried within a half - space , inverse problems , 26 ( 2010 ) , 074008 . w .- k . park and d. lesselier , electromagnetic music - type imaging of perfectly conducting , arc - like cracks at single frequency , j. comput . 228 ( 2009 ) , 80938111 . w .- k . park and d. lesselier , fast electromagnetic imaging of thin inclusions in half - space affected by random scatterers , waves random complex media , 22 ( 2012 ) , 323 . k . park and d. lesselier , music - type imaging of a thin penetrable inclusion from its far - field multi - static response matrix , inverse problems , 25 ( 2009 ) , 075002 . w .- k . park and d. lesselier , reconstruction of thin electromagnetic inclusions by a level set method , inverse problems , 25 ( 2009 ) , 085010 . w .- k . park and t. park , multi - frequency based direct location search of small electromagnetic inhomogeneities embedded in two - layered medium , comput . , 184 ( 2013 ) , 16491659 . w. rosenheinrich , tables of some indefinite integrals of bessel functions , 2011 , available at http://www.fh-jena.de/~rsh/forschung/stoer/besint.pdf f. santosa , a level set approach for inverse problems involving obstacles , esaim contr . optim . var . , 1 ( 1996 ) , 1733 . v. sen , m. k. sen and p. l. stoffa , pvm based 3-d kirchhoff depth migration using dynamically computed travel - times : an application in seismic data processing , parallel comput . , 25 ( 1999 ) , 231248 . j. k. seo , o. kwon , h. ammari , and e. j. woo , mathematical framework and anomaly estimation algorithm for breast cancer detection using ts2000 configuration , ieee trans . biomedical engineering , 51 ( 2004 ) , 18981906 . e. taillet , j. f. lataste , p. rivard , and a. denis , non - destructive evaluation of cracks in massive concrete using normal dc resistivity logging , ndt & e int . , 63 ( 2014 ) , 1120 . g. ventura , j. x. xu and t. belytschko , a vector level set method and new discontinuity approximations for crack growth by efg , int . j. numer . meth . engng , 54 ( 2002 ) , 923944 . y. zhong and x. chen , music imaging and electromagnetic inverse scattering of multiply scattering small anisotropic spheres , ieee trans . antennas propag . , 55 ( 2007 ) , 35423549 .	the present study seeks to investigate mathematical structures of a multi - frequency subspace migration weighted by the natural logarithmic function for imaging of thin electromagnetic inhomogeneities from measured far - field pattern . to this end , we designed the algorithm and calculated the indefinite integration of square of bessel function of order zero of the first kind multiplied by the natural logarithmic function . this is needed for mathematical analysis of improved algorithm to demonstrate the reason why proposed multi - frequency subspace migration contributes to yielding better imaging performance , compared to previously suggested subspace migration algorithm . this analysis is based on the fact that the singular vectors of the collected multi - static response ( msr ) matrix whose elements are the measured far - field pattern can be represented as an asymptotic expansion formula in the presence of such inhomogeneities . to support the main research results , several numerical experiments with noisy data are illustrated . multi - frequency subspace migration , weighted by natural logarithmic function , thin electromagnetic inhomogeneities , multi - static response ( msr ) matrix , numerical experiments .
in the last decade , the ly@xmath0 forests of qso s absorption spectra at redshifts @xmath8 have played an important role in understanding the diffuse cosmic baryon gas and the uv ionizing photon background , and constraining cosmological models and parameters ( e.g. rauch et al . 1997 ; croft et al . 2002 ; bolton et al . 2005 ; seljak et al . 2005 ; jena et al . 2005 ; viel et al . 2006 ) . recently , more and more uv photon sources , including qsos , grb , lyman - break galaxies , and ly@xmath0-emitters at redshifts @xmath9 have been observed ( see ellis 2007 and reference therein ) . due to the rapidly increase of gunn - peterson ( gp ) optical depth at @xmath10 , their absorption spectra show long dark gaps on scales of tens of mpc separated by tiny transmitted leaks . it has been suggested that we are observing the end stage of reionization ( fan et al . 2006 ) . it has been known that the dark gaps are from the igm where the density of neutral hydrogen are high enough to produce almost complete absorptions , however , where the transmitted leaks come from is still unclear . in photoionization equilibrium , the density of neutral hydrogen @xmath11 , here @xmath0 is the recombination rate and @xmath12 is the density of igm ; therefore , even when most of the igm are neutral enough to produce complete ly@xmath0 absorptions , it is still possible for the lowest density voids to provide prominent transmitted fluxes . on the other hand , the leaks can also come from ionized patches around ionizing sources where the intensity of uv radiation are higher than average . the origin of leaking is crucial to understanding the history of reionization . according to commonly accepted scenario of the reionization , in the early stage , the ionized regions are isolated patches in the neutral hydrogen background ( e.g. , ciardi et al . 2003 ; sokasian et al . 2003 ; mellema et al . 2006 ; gnedin 2006 ; trac & cen 2006 ) . and the subsequent growing and overlapping of the ionized patches lead to the ending of reionization ( e.g. , ciardi et al . 2003 ; sokasian et al . 2003 ; mellema et al . 2006 ; gnedin 2004 ) . thus , if leaks mostly come from ionized patches , reionization should happen in the early stage . in contrast , if they were produced in lowest density voids , the the uv ionizing background might has already underwent an evolution from highly inhomogeneous to uniform distribution . a variety of statistics has been used to study the evolution of reionization , such as the mean and dispersion of gp optical depth , the probability distribution function ( pdf ) of the flux , and the size of dark gaps ( e.g. , fan et al . 2002 , 2006 ; songaila & cowie 2002 ; paschos & norman 2005 ; kohler et al . 2007 ; gallerani et al . 2006 ; becker et al . 2006 ) , but all of them seem to be ineffective to provide the information of leak s origin and the inhomogeneity of uv ionizing background . the gp optical depth is an average , and not sensitive to details of reionization . the statistical properties ( mean and variance ) of the gp optical depth at @xmath5 can be well explained by either the fluctuation of ionizing background ( fan et al . 2006 ) or models with uniform ionizing background ( lidz et al . 2006 liu et al . 2006 , hereafter paperi ) . the pdf of the flux is also insensitive to the geometry of reionization . in addition , the pdf is heavily contaminated by noise and distorted by resolution . dark gaps are defined to be continuous regions with optical depth above a threshold in spectra . intuitively , the statistics of dark gap should contain the same information as leaks . however , the size of dark gaps are sensitive to the instrumental resolution , because higher resolution data contain more small leaks ( e.g. , paschos & norman 2005 ; paperi ) , and they also are contaminated by observational noise . moreover , dark gaps are from saturated absorptions , they are featureless and contains generally less information of non - saturated absorption . in this paper we made a statistical approach to ly@xmath0 leaks . the purpose is to show that the statistical features of ly@xmath0 leaks would be effective tool to reveal the origin of ly@xmath0 leaks , and to probe the evolution of reionization . similar to ly@xmath0 absorption lines , ly@xmath0 leaks have a rich set of statistical properties , such as the width of leak profile . unlike dark gaps , the properties of ly@xmath0 leaks can be defined through integrated quantities , which are less contaminated by resolution and noise . moreover , ly@xmath0 leaks are from regions of non - saturated absorptions and encode more information of reionization ; therefore , the ly@xmath0 statistics would provide more underlying physics of reionization than all the above - mentioned statistics . the paper is organized as follows . 2 describes the method to produce ly@xmath0 absorption samples . 3 presents the statistical properties of ly@xmath0 leaks with a uniform ionizing background . 4 analyzes the effect of inhomogeneous ionizing background . conclusion and discussion are given in 5 . we simulate lyman series absorption spectra of qsos between @xmath13 and 6.5 using the same lognormal method as those for low redshifts @xmath14 ( e.g. , bi et al . 1995 ; bi & davidsen 1997 ) . in this model , the density field @xmath15 of the igm is given by an exponential mapping of the linear density field @xmath16 as @xmath17,\ ] ] where @xmath18 is the variance of the linear density field on scale of the jeans length . obviously , the 1-point pdf of @xmath15 is lognormal . in this model , the velocity field of baryon gas is produced by considering the statistical relation between density and velocity field ( bi & davidsen , 1997 ; choudhury et al . 2001 ; veil et al . 2002 ) . the dynamical bases of the lognormal model have gradually been settled in recent years . first , although the evolution of cosmic baryon fluid is governed by the naiver - stokes equation , the dynamics of growth modes of the fluid can be sketched by a stochastic force driven burgers equation ( berera & fang 1994 ) . on the other hand , the lognormal field is found to be a good approximation of the solution of the burgers equation ( jones 1999 ) . the one - point distribution of the cosmic density and velocity fields on nonlinear regime are consistent with lognormal distribution ( e.g. yang et al . 2001 , pando et al . especially , it has been shown recently that the velocity and density fields of the baryon matter of the standard @xmath19cdm model is well described by the so - called she - lvque s universal scaling formula , which is given by a hierarchical process with log - poisson probability distribution ( he et al . 2006 , liu & fang 2007 ) . the simulation is performed in the concordance @xmath19cdm cosmological model with parameters @xmath20 , @xmath21 , @xmath22 , and @xmath23 . the thermodynamic evolution in the igm is actually a rather complex process , because the nonlinear clustering leads to a multi - phased igm . as shown in cosmological hydrodynamic simulations ( e.g. , he et al . 2004 ) , for a given mass density , the temperature of the igm could have large scatters with differences up to two orders . nevertheless , the equation of state in ly@xmath0 clouds can be well approximated by a polytropic relation with @xmath24 ( hui & gnedin ; he et al . the neutral fraction @xmath25 is obtained by solving the photoionization equilibrium equation . the photoionization rate @xmath26 will be given in 2.2 . we then construct synthetic absorption spectra by convoluting the neutral hydrogen density field with voigt profiles . for each given @xmath27 , the size of the simulation samples is @xmath28 , and there are @xmath29 pixels in each simulation box . if the distribution of the igm is uniform and the uv ionizing background is independent of redshift , the mean gp optical depth of ly@xmath0 absorption should approximately increase with redshift as @xmath30 . the observations of dark gaps directly show that the gp optical depth undergoes a stronger evolution at @xmath31 , and consequently , the uv ionizing background would decrease rapidly with redshift at @xmath5 . the strong evolution scenario of the uv ionizing background is supported by a number of simulations or semi - analytical models of reionization ( e.g. , razoumov et al . 2002 ; gnedin 2004 ; oh & furlanetto 2005 ; pascho & norman 2005 ; wyithe & loeb 2005 ; kohler et al . 2007 ; gallerani et al . 2006 ) . it has been found that an evolution of photoionization rate as follows can fit the strong redshift evolution of the gp optical depth ( paper i ) : @xmath32^{2.4 } \},\ ] ] which is in units of 10@xmath33 s@xmath34 . note that the power index 2.4 in equation ( 2 ) is little different from the one used in paper i because we use a different @xmath35 in this paper ( also see below ) . with eq.(2 ) , we calculate the redshift dependencies of neutral hydrogen fraction @xmath36 and the effective optical depth , @xmath37 , where @xmath38 is the mean of transmitted flux . the results are plotted in figure 1 . the data points for @xmath39 are taken from ly@xmath0 observations of fan et al . ( 2006 ) . for best fitting , the amplitude @xmath40 is in the range 5 - 7 , which can be considered as the allowed range of @xmath41 . in this paper , we will use @xmath42 as the fiducial photoionization rate . it is interesting to note that @xmath36 approaches to @xmath43 at redshift @xmath44 , which is consistent with the electron scattering optical depth given by the data of cmb polarization of wmap iii ( page et al . 2007 ) . it should be pointed out that the assumption of @xmath45 k ( 2.1 ) is well reasonable at @xmath46 ( e.g. hui & gnedin 1997 ; he et al 2004 ) and may still be applicable at @xmath1 if the mass averaged neutral fraction of hydrogen is not larger than 10@xmath47 , and the photon heating rate is small . however , at higher redshift , say @xmath48 , the temperature @xmath35 might be redshift - dependent . yet , no proper information on @xmath49 is available at high redshift , and this leads to uncertainty of the model . fortunately , in photoionization equilibrium , the neutral fraction @xmath36 depends mainly on a degenerate factor @xmath50 . thus , the problem with the uncertainty of @xmath49 can be overcame if we use the combined parameter of @xmath51^{0.75}$ ] to fit the data . in the range @xmath52 , this parameter actually is @xmath53 ; in the range of @xmath54 , it is different from @xmath53 by a factor of @xmath55^{0.75}$ ] . thus , the redshift - evolution of @xmath41 would be slower than eq.(2 ) if @xmath49 is less than @xmath56 k at higher redshifts . as an example of ly@xmath0 leaks , we plot a simulated sample of ly@xmath0 absorption spectrum at @xmath57 with a uniform ionizing background in figure 2 , which shows the transmitted flux @xmath58 , the density @xmath12 of baryon gas , the bulk velocity @xmath59 , and the neutral hydrogen fraction @xmath36 . as expected , the mean of transmitted flux is very small , about 0.004 , and corresponds to an effective optical depth 5.5 . nevertheless , we see spiky features with the transmitted flux @xmath58 as large as @xmath60 . they are leaks . at low redshifts @xmath61 , the ly@xmath0 forests in qso s spectra have a spectral filling factor significantly less than one and can be decomposed into individual ly@xmath0 absorption lines . at redshifts @xmath9 , it is meaningless to decompose the spectra into individual lines since almost the whole spectra are absorbed completely . we note , however , the transmitted leaks look like emission features upon the dark background , and the absorption spectra can be decomposed into individual `` emission lines '' , i.e. , ly@xmath0 leaks . comparing the top , the second and bottom panels , we see that all the leaks comes from the regions with mass density less than 0.3 of the mean mass density of baryon gas . the neutral fraction for leaks is @xmath62 , which yields a gp optical depth @xmath63 for overdensity 0.25 and a @xmath64 , while the mean neutral fraction of the entire example is about @xmath65 , which is high enough to produce dark gaps . the column density of neutral hydrogen of the leaking features is mainly in the range of 10@xmath66 - 10@xmath67 @xmath68 , which are the non - saturated absorption regions , and therefore , leaks can come from regions where no enough neutral hydrogen to produce complete absorptions . similar to very high density clouds , @xmath69 , the regions with very low density @xmath70 are rare events in the cosmic clustering . therefore , leaks may provide valuable test on models of clustering . for instance , the lognormal distributions are long tailed in both high and low density sides , and it contains more high density events as well as low density events than gaussian model . it should be emphasized that , once the photoionization rate @xmath26 is determined from the gp optical depth , the statistical property of the sample shown in figure 2 does nt contain free parameters . these samples have been successfully used to explain the following observations : 1 . ) the large dispersion of the gp optical depth ; 2.)the pdf of the flux , and 3 . ) the evolution of the size of dark gaps ( paper i ) . now we use them to study the statistical properties of ly@xmath0 leaks . the ly@xmath0 absorption lines at low redshifts are described by well defined quantities , such as the equivalent width , fwhm ( full width half maximum ) , and the voigt profile , all of which are easily related to physical interpretations . since the transmitted leaks look like emission features , one can decompose the spectra into individual `` emission lines '' and describes it by quantities similar to the ly@xmath0 absorption lines . in this section we study the statistical distribution of ly@xmath0 leaks in the model with a uniform ionizing background . the effects of inhomogeneous ionizing background will be discussed in next section . the center of a leak is identified as the maximum of transmitted flux , and the boundaries are two nearby positions around the center where the flux falls down to zero or to noise level ( we take it to be @xmath71 in this paper ) . the properties of leaks can be measured by two quantities : an equivalent width @xmath2 , which is the total area under the profile of a leak , and a half width @xmath72 , which is defined as the width around center within which the area under the profile of the leak is equal to half of the total area of the leak . the equivalent width @xmath2 and the half width @xmath3 are two independent measurements of ly@xmath0 leaks . this point can be seen in figure 3 , which gives the value of @xmath2 and @xmath3 for each leak at @xmath57 . it does not indicate a significant correlation between @xmath2 and @xmath3 , especially in the region of @xmath73 and @xmath74 . figure 3 also shows that the distribution of @xmath72 has a lower limit @xmath75 , which is due to the jeans scale used for smoothing the sample . on the other hand , the equivalent width @xmath2 distributes in the range from @xmath76 to @xmath77 . obviously , one is unable to introduce two independent quantities for characterizing dark gaps . the profiles of 100 randomly sampled leaks with @xmath2 = 0.01 , 0.1 , and 1 are displayed in fig.[profile ] . the tails of the profiles in the three panels of fig.[profile ] look like the lorentz profile , and of course , they do not have the meaning of the natural width of an absorption line . the profiles of leaks for a given equivalent width @xmath2 have very large dispersions ; for example , for @xmath2=0.1 , the flux covers a range from @xmath78 to 0.15 , and @xmath3 can be 0.4 to 4 . as the leaks are formed out of the two neighboring complete absorption troughs , which depend on inhomogeneities of density , velocity , and temperature fields , the large scatter of the profiles is expected . figure 4 also shows that some leaks may have multiple local maximums above the noise level . for clarity and easy - operating , we treat them as one leak . the current observational resolution is of the order of 10km / s , which corresponds to @xmath79 in observer s frame , or about 13 pixels of simulation . therefore , the observed ly@xmath0 leaks would have a resolvable width . similar to ly@xmath0 forests , we can define the cumulative number densities @xmath80 and @xmath81 of leaks as the number of leaks with widths larger than a given @xmath2 and @xmath3 at @xmath27 per unit @xmath27 , respectively . the differential number densities are @xmath82 and @xmath83 . it should be pointed out that statistics of @xmath2 and @xmath84 are not the same as the largest peak width statistics proposed by gallerani et al . ( 2007 ) , which considered only the largest peak width . a peak may contains more than one leaks , i.e. leak statistics describe the details of the leaking area . the mean transmitted flux at @xmath27 within @xmath85 is @xmath86 . we calculate the number densities of leaks in redshift range @xmath87 . in each redshift region we produce 100 light - of - sight samples to calculate the density functions . the results are shown in figure 5 . the errors are estimated by jackknife method , i.e. , the variance over 5 subsamples , each of which contains 20 light - of - sight samples . the number density @xmath88 shown in the up - left panel of figure 5 are similar to a schechter function : they follow a power law at small @xmath2 and have a cut - off at large @xmath2 . the distribution of @xmath89 at @xmath90 declines with decreasing @xmath91 , this is because of the jeans length smoothing . the slope of @xmath88 and @xmath89 are smaller for small redshift . it means the lack of low density voids with small size . that is , the increase of voids of small size is less than voids of large size . this trend can also be seen from the flattening of the cumulative density distributions @xmath92 and @xmath80 . the redshift - evolution of the number densities @xmath80 and @xmath93 of leaks for @xmath2 , @xmath3 = 0.5 , 1.0 , and 3.0 are shown in figure [ zdis ] . as has been seen in figure [ ldis ] , the number densities dramatically decrease at higher redshifts . the evolution is more rapidly for large leaks : the number density @xmath94,@xmath95 drops by a factor of @xmath96 5 when redshift increasing from 5 to 5.8 , while @xmath97,@xmath98 drops by a factor of @xmath96 60 . the evolution trend is the same for the number density @xmath89 . from the error bars of figures [ ldis ] and [ zdis ] we see that the predicted features of leaks would be able used to compare with observed data set containing 20 or more light - of - sight samples with the similar quality as simulation . in this section , we study the observational and instrumental effects on the statistics of leaks . since both @xmath2 and @xmath3 are defined through the area under the profile of leaks , the effects of resolution and noise would be small . to simulate the observational effects , we bin the original data to a coarse grid corresponding to a resolution of 20000 , and we add gaussian noises with signal - to - noise ratio s / n=3 on binned pixels . the number densities of leaks for the noisy binned samples are shown in figures [ noise ] . the effect of binning and noise is very small for @xmath2 : the original plot of @xmath2 actually is the same as the plot of binned @xmath2 , and the distribution of noisy @xmath2 is affected only when @xmath99 . the effect of binning for @xmath91 is also small on scales larger than the binning length . as expected , the noise effects for @xmath91 are significant for @xmath100 . the noise effects for @xmath91 are even smaller if we smooth the noisy sample . this is very different from the pdf of the flux and the size of dark gaps , both of which are heavily contaminated by instrumental resolution and observational noises . one can compare the uncertainty of @xmath40 with the effect of noises . figures [ gamma ] shows the number densities for the uv background amplitude @xmath40 = 5 , 6 , and 7 ( eq.2 ) , which represent the allowed range of @xmath40 . different from noises , the difference of @xmath40 will cause uncertainty in the whole ranges of @xmath2 and @xmath3 . the uncertainties of number densities are within a factor of 2 when the amplitude @xmath40 changes from 7 to 5 . these uncertainties essentially are from the mass density perturbations with long wavelengths ( paper i ) . the error bars from the scattering of light - of - sight samples is also shown in figure [ gamma ] . therefore , the scattering of @xmath101 light - of - sight samples actually is less than the uncertainty of @xmath40 . in the early stage of reionization , ionizing photons are mainly in ionized patches around uv ionizing sources , and therefore , the spatial distribution of ionizing background is highly inhomogeneous and has patchy structures . when the ionized patches spread over the whole space , the ionizing background become uniform or quasi - uniform , and so the ionizing background underwent an inhomogeneous - to - uniform transition during the reionization . in this section we study the effects of inhomogeneous ionizing background on ly@xmath0 leak statistics . the first question is how to model the inhomogeneous spatial distribution of the ionizing photon field . to our problem , the most important property is the relation between the fields of mass density and ionizing photon background or photoionization rate . if the spatial fluctuation of photoionization rate @xmath102 is statistically uncorrelated with the density field @xmath103 , the reionization of igm will statistically be the same as a uniform ionizing background , regardless of the details of the fluctuation of @xmath102 . in this case , the only effect of the fluctuations would be to yield a larger variance in relevant statistics . however , as shown in last sections , the uncertainty of the leak statistics is already large even when the ionizing background is uniform , and therefore , one would not be able to distinguish the fluctuating photon field from inhomogeneous density field if they are uncorrelated . although many simulations on the uv photon field at the epoch of reionization have been done , there still lack the results of the correlation between photon and density fields . in this context , we will consider the following four models on the statistical relation between the inhomogeneous fields of photon and density , which are mainly based on physical consideration of different stage of reionization . _ model 0_. the photoionization rate is spatially uniform . it corresponds to the post - overlapping stage of reionization . this model has been used in last two sections . _ model i_. the photoionization rate at a give redshift is assumed to be proportional to the density field of igm , @xmath104 , @xmath105 being a constant . this model is motivated by the so - called inside - out scenario : high density regions around uv sources are ionized first , and is most probable at the early stage of reionization . _ model ii_. just opposite to model i , the photoionization rate at a give redshift is assumed to be inversely proportional to the density field of igm , @xmath106 , @xmath107 being a constant . this model comes from the so - called outside - in scenario : under - dense regions are ionized first ( miralda - escude et al . 2000 ) , which is applicable at the late stage of reionization ( furlanetto & oh 2005 ) . in order to fit the observed effective optical depth @xmath39 at redshift @xmath57 , we take the following parameters : @xmath108 , @xmath109 where @xmath40 is the photoionization rate at @xmath57 for uniform ionizing background [ eq.(2 ) ] . _ model iii or patch model_. this model corresponds to the stage before the overlapping of ionized patches . the igm is almost fully neutral except for some isolated ionizing patches . we model the ionized patch as a stromgren sphere : the neutral hydrogen fraction is equal to 0 within the patch and equal to 1 outside the patch . the mean radius of the stromgren sphere is assumed to be @xmath110 comoving mpc , which corresponding to an uv photon source with luminosity @xmath111 erg s@xmath34 with active time @xmath112 year and a @xmath113 spectrum . to fit the observed optical depth , we found there should be 2 patches every simulated box ( @xmath114 ) . this number is actually consistent with the following estimation @xmath115 where @xmath116 the comoving luminosity function , i.e. the 3-d number density of sources with luminosity @xmath117 . using @xmath118 mpc@xmath47 ( bouwens et al . 2006 ) , we have @xmath119 and @xmath120 . for the four models described in the last section , we recalculate the transmitted flux with the same underlying density and velocity fields of fig2 . the results are shown in figure 9 . as mentioned above , the mean of transmitted flux for all models takes approximately the same value of 0.004 , which corresponds to an effective optical depth @xmath121 . however , figure 9 shows clearly different behaviors of ly@xmath0 leaks in various models . for models i and ii , the leaks appear exactly at the same positions as model 0 except some small leaks , which are more prominent for model i. in other words , all the leaks of models 0 , i and ii are from lowest density voids . therefore , the three models have the same distribution of dark gaps , and it is impossible to discriminate among these models with the dark gap statistics . however , the profiles of the leaks of model 0 , i and ii are statistically different from each other . for model i , the leaks generally have larger width and lower height than model 0 , while for model ii , the width of leaks generally is narrower than model 0 , and the height of leaks is larger than model 0 . the reason is straightforward . comparing with model 0 , model i gives a higher @xmath36 at low density and lower @xmath36 at high density this leads to lower amplitude and broader width . for model ii , the effect is just opposite to model i and yields higher amplitude and narrower width . in the patch model , all the leaks come from ionized patches , within which ionizing sources are enclosed , and so the internal information of ionized patches can be inferred from the leak statistics . the size of dark gaps is actually given by the distance between ionized patches . generally , ly@xmath0 leaks in the patch model have a maximum flux @xmath122 , which is higher than the maximum flux of other models , 0.2 . this is because the neutral fraction within the stromgren sphere is much less that than other models . this behavior is similar to the so - called proximity effect of qsos at low redshift ( e.g. , rauch 1998 ) . thus , the statistics of the maximum flux of ly@xmath0 leaks may be used to reveal the patchy origin of ly@xmath0 leaks . as we sample the size of patches along the sight of light according to the impact probability , the size of patches is smaller than @xmath123 . if the size is too small , the ionized patch will be opaque to ly@xmath0 photons due to the damping wing of the surrounding neutral hydrogen absorption ( miralda - escude 1998 ) . obviously , it explains why only one leak is apparent in the patch model as displayed in figure 9 . we now calculate the number densities @xmath124 , @xmath80 , @xmath125 and @xmath126 for three inhomogeneous ionizing background models at @xmath57 , and the results are shown in figures 10 . we can see from figure 10 that the effects of different models on the number densities of @xmath2 and @xmath91 are different . that is , although all the models of 0 , i , ii and iii give the same mean effective optical depth , their leak distributions are different . for clarifying , in figures 10 we show only error bars for the curves of model 0 . it would be enough to show that these results can be tested with data set containing about 20 light - of - sight data . first , for model i , both of the number densities @xmath88 and @xmath80 have only small deviations from model 0 . however , @xmath125 are significantly different from their counterparts of model 0 . the number density of leaks with @xmath127 is much more than model 0 . the basic feature of model i is to increase the number of leaks with large @xmath3 as shown in figure 9 . second , for model ii , we see once again that the number densities @xmath88 and @xmath80 have only small deviations from model 0 . yet , the number densities @xmath89 and @xmath81 of the model ii are systematically lower than model 0 . therefore , the basic feature of model ii is to keep the total area under the profile of leaks almost unchanged , but the widths of leaks are significantly narrowed . finally , the behavior of patch model is very different from models 0 , i and ii . as expected , the patch model yields more leaks with large @xmath2 , and @xmath88 shows a bump around @xmath128 , which characterize the area of the ionized patches . on the other hand , the number densities of @xmath89 and @xmath81 are lower than model 0 because the characteristic size of ionized patches is less than that of low density voids . it is interesting to note that for the patch model , the tails of @xmath88 , @xmath80 and @xmath129 , @xmath126 are quite different from each other . generally , the tails of @xmath88 and @xmath80 can extend to as large as @xmath130 , while for @xmath125 and @xmath81 , there are no tails higher than 3 . it results partially from the damping wing effect ( miralda - escude 1998 ) . in summary , the statistical properties of leaks with respective to @xmath2 and @xmath91 are sensitive to the details of ionizing photon field . combining the distribution of @xmath2 and @xmath91 , the ly@xmath0 leaks would be able to probe the origin of themselves , and thus reveal the ionization state of igm , the inhomogeneity of ionizing background , and the evolution stage of reionization . it should be pointed out that we considered only the patches of the hi regions around high redshift galax ies . the hi regions around quasars or the proximity effect would also be the patches of leaking . the mean luminosity of quasars probably is higher than galaxies , and therefore , the above - mentioned feature of @xmath89 and @xmath81 would be more prominent for the patches of quasars . the damped ly@xmath0 absorption system is important for modeling low redshift ly@xmath0 absorption . since these systems have high column density of neutral hydrogen , it will not contribute to leaking either in density void models or patch model . we show that the ly@xmath0 absorption spectra of uv photon emitters at redshifts around @xmath131 can be decomposed into ly@xmath0 leaks , which come from either lowest density voids or ionized patches containing ionizing sources . the ly@xmath0 leaks are well defined and described by the equivalent width @xmath2 and the width of half area @xmath91 . since both @xmath2 and @xmath91 are defined through integrated quantities , the distributions of ly@xmath0 leaks in terms of @xmath2 and @xmath91 are stable with respect to observational noises and instrumental resolution . although the number densities @xmath88 , @xmath80 , @xmath125 , and @xmath81 evolve very rapidly at redshift @xmath5 , these statistics are measurable up to @xmath132 . if the ly@xmath0 leaks come from lowest density voids , the igm should be still highly ionized and the ionizing background is almost uniform ; in contrast , if the leaks come from isolated ionized patches , the ionizing background should be inhomogeneous , and the reionization is still in the overlapping stage . therefore , the origin of ly@xmath0 leaks is crucial to understand the history of reionization . we show that the statistics with @xmath2 and @xmath91 are sensitive to the origin of ly@xmath0 leaks , because the ly@xmath0 leaks are sensitive to the correlation between photon and density fields . based on physical consideration of reionization , we studied four phenomenological models of the photon field : the uniform ionizing background ( model 0 ) , the photoionization rate @xmath4 is proportional to the density @xmath12 ( model i ) , @xmath4 and @xmath12 are anti - correlated ( mode ii ) , and ionized regions only given by stromgren sphere around ionizing photo sources ( patch model ) . we found that , although all the four models can fit the observed mean and variance of optical depth at @xmath5 , the width distribution of ly@xmath0 leaks show very different behaviors . for model 0 , i , and ii , most of ly@xmath0 leaks are from lowest density voids , and the distribution of dark gaps are similar . however , the properties of individual ly@xmath0 leaks are different . model i gives broader width @xmath91 than model 0 ; model ii gives narrower width @xmath91 than model 0 . for patch model , the ly@xmath0 leaks are mainly from ionized sphere , and they generally have higher maximum of transmitted flux than other models . there is a bump in the distribution of the equivalent width @xmath2 , which characterize the intensity of uv photons within ionized patches . ly@xmath0 leaks from ionized patches will provide the information of ionized patches , such as their intensity and size , and constrain the properties of uv sources that contribute most of ionizing photons of the reionization . finally , we point out that similar analysis is also applicable to the reionization of heii . the optical depth of heii ly@xmath0 evolves rapidly at @xmath133 . it reaches @xmath134 at @xmath135 . that is , the evolution of heii reionization at @xmath135 would be similar to that of h at @xmath5 . observed sample of heii ly@xmath0 absorption indeed show structures similar to the leaks of hydrogen absorption spectrum ( zheng et al . one can also draw the information of background photons at the energy band of heii ionization . this work is supported in part by the us nsf under the grant ast-0507340 . llf acknowledges support from the national science foundation of china ( nsfc ) .	ly@xmath0 absorption spectra of qsos at redshifts @xmath1 show complete gunn - peterson absorption troughs ( dark gaps ) separated by tiny leaks . the dark gaps are from the intergalactic medium ( igm ) where the density of neutral hydrogen are high enough to produce almost saturated absorptions , however , where the transmitted leaks come from is still unclear so far . we demonstrate that leaking can originate from the lowest density voids in the igm as well as the ionized patches around ionizing sources using semi - analytical simulations . if leaks were produced in lowest density voids , the igm might already be highly ionized , and the ionizing background should be almost uniform ; in contrast , if leaks come from ionized patches , the neutral fraction of igm would be still high , and the ionizing background is significantly inhomogeneous . therefore , the origin of leaking is crucial to determining the epoch of inhomogeneous - to - uniform transition of the the ionizing photon background . we show that the origin could be studied with the statistical features of leaks . actually , ly@xmath0 leaks can be well defined and described by the equivalent width @xmath2 and the full width of half area @xmath3 , both of which are less contaminated by instrumental resolution and noise . it is found that the distribution of @xmath2 and @xmath3 of ly@xmath0 leaks are sensitive to the modeling of the ionizing background . we consider four representative models : uniform ionizing background ( model 0 ) , the photoionization rate of neutral hydrogen @xmath4 and the density of igm are either linearly correlated ( model i ) , or anti - correlated ( model ii ) , and @xmath4 is correlated with high density peaks containing ionizing sources ( model iii ) . although all of these models can match to the mean and variance of the observed effective optical depth of the igm at @xmath5 , their distribution of @xmath2 and @xmath3 are very different from each other . consequently , the leak statistics provides an effective tool to probe the evolutionary history of reionization at @xmath6 . similar statistics would also be applicable to the reionization of he ii at @xmath7 = 7.4truecm = 17.7truecm = 6.0truecm [ firstpage ] cosmology : theory - intergalactic medium - large - scale structure of the universe
the possibility of muon colliders was introduced by skrinsky et al.@xcite , neuffer@xcite , and others . more recently , several workshops and collaboration meetings have greatly increased the level of discussion@xcite,@xcite . a detailed feasibility study@xcite was presented at snowmass 96 . hadron collider energies are limited by their size , and technical constraints on bending magnetic fields . lepton ( or ) colliders , because they undergo simple , single - particle interactions , can reach higher energy final states than an equivalent hadron machine . however , extension of @xmath1 colliders to multi - tev energies is severely performance - constrained by beamstrahlung , the luminosity @xmath2 of a lepton collider can be written : [ lumeq ] 1 4e n_2 r_o p_beam _ y n_collisions where @xmath3 is the average vertical ( assumed smaller ) beam spot size , @xmath4 is the beam energy , @xmath5 is the total beam power , @xmath6 is the electromagnetic constant , @xmath7 is the classical radius , and @xmath8 is the number of photons emitted by one bunch as it passes through the opposite one . if this number is too large then the beamstrahlung background of electron pairs and other products becomes unacceptable . as the energy rises , the luminosity , for the same event rate , must rise as the square of the energy . for an electron collider , @xmath9 , and , for a fixed background , we have the severe requirement : e^3 in a muon collider there are two significant changes : 1 ) the classical radius @xmath7 is now that for the muon and is 200 times smaller ; and 2 ) the number of collisions a bunch can make @xmath10 is no longer 1 , but is now related to the average bending field in the muon collider ring , for 6 t , it is 900 . in addition , with muons , synchrotron radiation is negligible , and the collider is circular . in practice this means that it can be much smaller than a linear electron machine . the linacs for the 0.5 tev nlc will be 20 km long . the ring for a muon collider of the same energy would be only about 1.2 km circumference . there are , of course , technical difficulties in making sufficient muons , cooling and accelerating them before they decay and dealing with the decay products in the collider ring . despite these difficulties , it appears possible that high energy muon colliders might have luminosities comparable to or , at energies of several tev , even higher than those in colliders . the basic parameters of a 4 tev collider are shown schematically in fig.[schematic ] and given in tb.[sum ] together with those for a 0.5 tev demonstration machine based on the ags as an injector . it is assumed that a demonstration version based on upgrades of the fermilab machines would also be possible . .parameters of collider rings [ cols="<,<,^,^ " , ] the proton driver is a high - intensity ( four bunches of @xmath11 protons per pulse ) 30 gev proton synchrotron , operating at a repetition rate of 15 hz . two of the bunches are used to make @xmath12 s and two to make @xmath13 s . prior to targeting the bunches are compressed to an rms length of 1 ns . for a demonstration machine using the ags@xcite , two bunches of @xmath14 at a repetition rate of 2.5 hz at @xmath15gev could be used . predictions of nuclear monte - carlo programs@xcite@xcite@xcite suggest that @xmath16 production is maximized by the use of heavy target materials , and that the production is peaked at a relatively low pion energy ( @xmath17mev ) , substantially independent of the initial proton energy . cooling requirements dictate that the target be liquid : liquid lead and gallium are under consideration . in order to avoid shock damage to a container , the liquid could be in the form of a jet . pions are captured from the target by a high - field ( @xmath18 t , 15 cm aperture ) hybrid magnet : superconducting on the outside , and a water cooled bitter solenoid on the inside . a preliminary design@xcite has a bitter magnet with an inside coil diameter of 24 cm ( space is allowed for a 4 cm heavy metal shield inside the coil ) and an outside diameter of 60 cm ; it provides half ( 10 t ) of the total field , and would consume approximately 8 mw . the superconducting magnet has a set of three coils , all with inside diameters of 70 cm and is designed to give 10 t at the target and provide the required tapered field to match into the decay channel . the decay channel consists of a periodic superconducting solenoidal ( @xmath19 t and radius @xmath20 cm ) . a linac is introduced along the decay channel , with frequencies and phases chosen to deaccelerate the fast particles and accelerate the slow ones ; i.e. to phase rotate the muon bunch . [ evsctpol2 ] shows the energy vs ct at the end of the decay channel . the selected muons have a mean energy 150 mev , rms bunch length @xmath21 m , and rms momentum spread @xmath18% ( @xmath22% , @xmath23 ) . the number of muons per initial proton in this selected bunch is @xmath24 0.3 . if nothing is done then the average muon polarization is about 0.19 . if higher polarization is desired , some selection of muons from forward pion decays @xmath25 is required . this can be done by momentum selecting the muons at the end of the decay and phase rotation channel . a snake@xcite is used to generate the required dispersion . varying the selected minimum momentum of the muons yields polarization as a function of luminosity loss as shown in fig.[polvscut ] . dilutions introduced in the cooling have been calculated@xcite and are included . a siberian snake is also required in the final collider ring . for the required collider luminosity , the phase - space volume must be greatly reduced ; and this must be done within the @xmath26 lifetime . cooling by synchrotron radiation , conventional stochastic cooling and conventional electron cooling are all too slow . optical stochastic cooling@xcite , electron cooling in a plasma discharge@xcite and cooling in a crystal lattice@xcite are being studied , but appear very difficult . ionization cooling@xcite of muons seems relatively straightforward . in ionization cooling , the beam loses both transverse and longitudinal momentum as it passes through a material medium . subsequently , the longitudinal momentum can be restored by coherent reacceleration , leaving a net loss of transverse momentum . the equation for transverse cooling ( with energies in gev ) is : @xmath27 where @xmath28 is the normalized emittance , @xmath29 is the betatron function at the absorber , @xmath30 is the energy loss , and @xmath31 is the radiation length of the material . the first term in this equation is the coherent cooling term , and the second is the heating due to multiple scattering . this heating term is minimized if @xmath29 is small ( strong - focusing ) and @xmath31 is large ( a low - z absorber ) . energy spread is reduced by placing a transverse variation in absorber density or thickness at a location where position is energy dependent , i.e. where there is dispersion . the use of such wedges can reduce energy spread , but it simultaneously increases transverse emittance in the direction of the dispersion . it thus allows the exchange of emittance between the longitudinal and transverse directions . the cooling is obtained in a series of cooling stages . in general , each stage consists of three components with matching sections between them : 1 . a fofo lattice consisting of spaced axial solenoids with alternating field directions and lithium hydride absorbers placed at the centers of the spaces between them , where the @xmath29 s are minimum . a lattice consisting of more widely separated alternating solenoids , and bending magnets between them to generate dispersion . at the location of maximum dispersion , wedges of lithium hydride are introduced to interchange longitudinal and transverse emittance . 3 . a linac to restore the energy lost in the absorbers . in a few of the later stages , current carrying lithium rods replace item ( 1 ) above . in this case the rod serves simultaneously to maintain the low @xmath29 , and attenuate the beam momenta . similar lithium rods , with surface fields of @xmath32 t , were developed at novosibirsk and have been used as focusing elements at fnal and cern@xcite . the emittances , transverse and longitudinal , as a function of stage number , are shown in fig.[cooling ] . in the first 10 stages , relatively strong wedges are used to rapidly reduce the longitudinal emittance , while the transverse emittance is reduced relatively slowly . the object is to reduce the bunch length , thus allowing the use of higher frequency and higher gradient rf in the reacceleration linacs . in the next 7 stages , the emittances are reduced close to their asymptotic limits . in the last 3 stages , using lithium rods , there are no wedges and the energy is allowed to fall to about 15 mev . transverse cooling continues , and the momentum spread is allowed to rise . the total length of the system is 750 m , and the total acceleration used is 5 gev . the fraction of muons remaining at the end of the cooling system is calculated to be @xmath33% . following cooling and initial bunch compression the beams must be rapidly accelerated to full energy ( 2 tev , or 250 gev ) . a sequence of recirculating accelerators ( similar to that used at cebaf)could be used but would be relatively expensive . a more economical solution would be to use fast pulsed magnets in synchrotrons with rf systems consisting of significant lengths of superconducting linac . for the final acceleration to 2 tev in the high energy machine , the power consumed by a ring using only pulsed magnets would be excessive and alternating pulsed and superconducting magnets@xcite are used instead . after acceleration , the @xmath12 and @xmath13 bunches are injected into a separate storage ring . the highest possible average bending field is desirable to maximize the number of revolutions before decay , and thus maximize the luminosity . collisions occur in one , or perhaps two , very low-@xmath34 interaction areas . the magnet design is complicated by the fact that the @xmath26 s decay within the rings ( @xmath35 ) , producing electrons whose mean energy is approximately 0.35 that of the muons . these electrons travel toward the inside of the ring dipoles , radiating a fraction of their energy as synchrotron radiation towards the outside of the ring , and depositing the rest on the inside . the total average power deposited , in the ring , in the 4 tev machine is 13 mw . the beam must thus be surrounded by a @xmath24 6 cm thick warm shield@xcite , which is located inside a large aperture conventional superconducting magnet . the quadrupoles can use warm iron poles placed as close to the beam as practical , with coils either superconducting or warm , as dictated by cost considerations . in order to maintain a bunch with rms length 3 mm , without excessive rf , an isochronous lattice , of the dispersion wave type@xcite is used . for the 3 mm beta at the intersection point , the maximum beta s in both x and y are of the order of 400 km ( 14 km in the 0.5 tev machine ) . local chromatic correction is essential . two lattices have been generated@xcite@xcite , one of which@xcite , after the application of octupole and decapole correctors , has been shown to have an adequate calculated dynamic aperture . studies of the resistive wall impedance instabilities indicate that the required muon bunches would be unstable if uncorrected . in any case , the rf requirements to maintain such bunches would be excessive . bns@xcite damping , applied by rf quadrupoles@xcite , is one possible solution , but needs more careful study . monte carlo study@xcite,@xcite indicated that the background , though serious , should not be impossible . further reductions are expected as the shielding is optimized , and it should be possible to design detectors that are less sensitive to the neutrons and photons present . there would also be a background from the presence of a halo of near full energy muons in the circulating beam . the beam will need careful preparation before injection into the collider , and a collimation system will have to be designed to be located on the opposite side of the ring from the detector . there is a small background from incoherent ( i.e. @xmath36 ) pair production in the 4 tev collider case . the cross section is estimated to be @xmath37 , which would give rise to a background of @xmath38 electron pairs per bunch crossing . approximately @xmath39 of these , will be trapped inside the tungsten nose cone , but those with energy between 30 and @xmath40mev will enter the detector region . * considerable progress has been made on a scenario for a 2 + 2 tev , high luminosity collider . much work remains to be done , but no obvious show stopper has yet been found . * the two areas that could present serious problems are : 1 ) unforeseen losses during the 25 stages of cooling ( a 3% loss per stage would be very serious ) ; and 2 ) the excessive detector background from muon beam halo . * many technical components require development : a large high field solenoid for capture , low frequency rf linacs , multi - beam pulsed and/or rotating magnets for acceleration , warm bore shielding inside high field dipoles for the collider , muon collimators and background shields , etc . but : * none of the required components may be described as _ exotic _ , and their specifications are not far beyond what has been demonstrated . * if the components can be developed and the problems overcome , then a muon - muon collider could be a useful complement to colliders , and , at higher energies could be a viable alternative . this research was supported by the u.s . department of energy under contract no . de - aco2 - 76-ch00016 and de - ac03 - 76sf00515 . 99 e. a. perevedentsev and a. n. skrinsky , proc . 12th int . conf . on high energy accelerators , f. t. cole and r. donaldson , eds . , ( 1983 ) 485 ; _ early concepts for @xmath41 colliders and high energy @xmath26 storage rings _ , _ physics potential & development of @xmath41 colliders . 2@xmath42 workshop _ , sausalito , d. cline , aip press , woodbury , new york , ( 1995 ) . d. neuffer , ieee trans . * ns-28 * , ( 1981 ) 2034 . , napa ca , nucl inst . and meth . , * a350 * ( 1994 ) ; proceedings of the muon collider workshop , february 22 , 1993 , los alamos national laboratory report la- ur-93 - 866 ( 1993 ) and _ physics potential & development of @xmath41 colliders 2@xmath42 workshop _ , sausalito , ca , ed . d. cline , aip press , woodbury , new york , ( 1995 ) . transparencies at the _ 2 + 2 tev @xmath41 collider collaboration meeting _ , feb 6 - 8 , 1995 , bnl , compiled by juan c. gallardo ; transparencies at the _ 2 + 2 tev @xmath41 collider collaboration meeting _ , july 11 - 13 , 1995 , fermilab , compiled by robert noble ; proceedings of the 9th advanced icfa beam dynamics workshop , ed . j. c. gallardo , aip press , conference proceedings 372 ( 1996 ) . , bnl-52503 , fermilab - conf.-96/092 , lbnl-38946 , proceedings of the snowmass workshop 96 , to be published . t. roser , _ ags performance and upgrades : a possible proton driver for a muon collider _ , proceedings of the 9th advanced icfa beam dynamics workshop , ed . j. c. gallardo , aip press , conference proceedings 372 ( 1996 ) . d. kahana , et al . , _ proceedings of heavy ion physics at the ags - hipags 93 _ , ed . g. s. stephans , s. g. steadman and w. e. kehoe ( 1993 ) ; d. kahana and y. torun , _ analysis of pion production data from e-802 at 14.6 gev / c using arc _ , bnl report # 61983 ( 1995 ) . n. v. mokhov , _ the mars code system user s guide _ , version 13(95 ) , fermilab - fn-628 ( 1995 ) . j. ranft , dpmjet code system ( 1995 ) . r. weggel , private communication ; physics today , pp . 21 - 22 , dec . f. chen , _ introduction to plasma physics _ , plenum , new york , pp . 23 - 26 ( 9174 ) ; t. tajima , _ computational plasma physics : with applications to fusion and astrophysics _ , addison - wesley publishing co. , new york , pp . 281 - 282 ( 1989 ) . b. norum and r. rossmanith , _ polarized beams in a muon collider _ , in physics potential & development of @xmath41 colliders , proc . , 3rd int . conf . , san francisco , dec . 1995 , elsevier , in press . a. a. mikhailichenko and m. s. zolotorev , phys . lett . * 71 * , ( 1993 ) 4146 ; m. s. zolotorev and a. a. zholents , slac - pub-6476 ( 1994 ) . a. hershcovitch , brookhaven national report ags / ad / tech . note no . 413 ( 1995 ) . z. huang , p. chen and r. ruth , slac - pub-6745 , _ proc . workshop on advanced accelerator concepts _ , lake geneva , wi , june ( 1994 ) ; p. sandler , a. bogacz and d. cline , _ muon cooling and acceleration experiment using muon sources at triumf _ , _ physics potential & development of @xmath41 colliders 2@xmath42 workshop _ , sausalito , ca , ed . d. cline , aip press , woodbury , new york , pp . 146 a. n. skrinsky and v.v . parkhomchuk , sov . j. of nucl . physics * 12 * , ( 1981 ) 3 ; d. neuffer , particle accelerators , * 14 * , ( 1983 ) 75 ; d. neuffer , proc . 12th int . conf . on high energy accelerators , f. t. cole and r. donaldson , eds . , 481 ( 1983 ) ; d. neuffer , in advanced accelerator concepts , aip conf . 156 , 201 ( 1987 ) . g. silvestrov , proceedings of the muon collider workshop , february 22 , 1993 , los alamos national laboratory report la - ur-93 - 866 ( 1993 ) ; b. bayanov , j. petrov , g. silvestrov , j. maclachlan , and g. nicholls , nucl . inst . and meth . * 190 * , ( 1981 ) 9 ; colin d. johnson , hyperfine interactions , * 44 * ( 1988 ) 21 ; m. d. church and j. p. marriner , annu . * 43 * ( 1993 ) 253 . d. summers , presentation at the 9th advanced icfa beam dynamics workshop , unpublished . i. stumer , presentation at the bnl - lbl - fnal collaboration meeting , feb 1996 , bnl , unpublished . s.y . lee , k .- y . ng and d. trbojevic , fnal report fn595 ( 1992 ) ; phys * e48 * , ( 1993 ) 3040 ; d. trbojevic , et al . , _ design of the muon collider isochronous storage ring lattice _ , _ micro - bunches workshop _ , bnl oct . ( 1995 ) , to be published . a. garren , et al . , _ design of the muon collider lattice : present status _ , in physics potential & development of @xmath41 colliders , proc . , 3rd int . , san francisco , dec . 1995 , elsevier , in press . k. oide , private communication . v. balakin , a. novokhatski and v. smirnov , proc . 12_th _ int . conf . on high energy accel . , batavia , il , 1983 , ed . f.t . cole , batavia : fermi natl . accel . lab . ( 1983 ) , p. 119 . a. chao , _ physics of collective beam instabilities in high energy accelerators _ , john wiley & sons , inc , new york ( 1993 ) . g. w. foster and n. v. mokhov , _ backgrounds and detector performance at 2 + 2 tev @xmath41 collider _ , _ physics potential & development of @xmath41 colliders 2@xmath42 workshop _ , sausalito , ca , ed . d. cline , aip press , woodbury , new york , pp . 178 ( 1995 ) . presentation at the unpublished .	muon colliders have unique technical and physics advantages and disadvantages when compared with both hadron and electron machines . they should be regarded as complementary . parameters are given of 4 tev high luminosity collider , and of a 0.5 tev lower luminosity demonstration machine . we discuss the various systems in such muon colliders . @xmath0 ) = = 10000
in this note we consider three reactions @xmath0 + h + [ lrg ] + p \,\ , ; \label{r1}\\ p + p & \longrightarrow & x_1 + [ lrg ] + h + [ lrg ] + x_2 \,\ , ; \label{r2}\\ p + p & \longrightarrow & b + \bar b + x + [ lrg ] + p \,\,;\label{r3}\end{aligned}\ ] ] where lrg denotes the large rapidity gap between produced particles and @xmath1 corresponds to a system of hadrons with masses much smaller than the total energy . the first two reactions are so called double pomeron production of higgs meson while the third is the single diffraction production of bottom - antibottom pair . the goals of this note are the following : 1 . to give the highest from reasonable estimates for the cross sections of reactions and ; 2 . to summarize all uncertainties which we see in doing these estimates ; 3 . to show that there is a new mechanism of diffractive heavy quark production ( ) which is suppressed in dis and dominates in hadron - hadron collision at the tevatron ; 4 . to estimate the value of the cross section of reaction due to this new mechanism and to show that all attempts to compare the diffraction dissociation in hadron - hadron collisions and dis@xcite look unreliable without a detail experimental study of this process at fermilab . inclusive higgs production has been studied in many details @xcite for the tevatron energies . the main source for higgs is gluon - gluon fusion which gives @xmath2 for higgs with mass @xmath3 @xcite . the reference point for our estimates is the cross section of higgs production due to w and z fusion which is equal to @xmath4 @xcite . in this process we also expect the two lrg @xcite and in some sense this is a competing process for reactions of and . let us estimate the simplest digram for the dp higgs production , namely , fig.1 without any of s - channel gluons . this diagram leads tpo the amplitude [ m1 ] m ( q q q h q ) = @xmath5 for reaction of , @xmath6 and therefore , [ m2 ] m ( q + q q + h + q ) . has an infrared divergency that is regularized by the size of the colliding hadrons . in other words , one can see that the simplest diagrams shows that dp higgs production is a typical soft " process . in fig.1 one can see that we have two sets of gluon which play a different role . the first one is the gluons that connect @xmath7-channel lines . their contribution increases the value of cross section @xcite [ g1 ] \|_y = 0 = @xmath8 can be rewritten in the form [ g2 ] \|_y = 0 = @xmath9 which is convenient for numeric estimates . however , first we need to find the value of @xmath10 . in inclusive production the value of @xmath11 has been calculated @xcite [ hx ] g^2_h= g_f ^2_s(m^2_h ) n^2/9 ^2 . however , i think that the scale of @xmath12 for our process is not the mass of higgs but the soft " scale ( @xmath13 with @xmath14 ) . indeed , using blm procedure @xcite we can include the bubbles with large number of light quarks only in @xmath7-channel gluon line which carry the `` soft '' transverse momenta . this gives a sizable effect in numbers , since @xmath15 for @xmath16 is equal to 1.16 pb ( @xmath17 ) and to 20 pb ( @xmath18 ) . taking the last value we have [ n1 ] to gluon emission suppressed the value of the cross section . actually , we have to multiply the cross section of by two factors to obtain the estimate for the experimental cross section [ sp1 ] d ( pp pp h)d y\|_y = 0 @xmath19 the first factor is the probability that there is no inelastic interaction of the spectators in our process . i the situation with calculation of this factor has been reported in this workshop @xcite and the conclusion is that this factor @xmath20 at the tevatron energies . the discussion for double pomeron processes you can find in ref . @xcite the second factor in describe the probability that there is no parasite emission in fig.1 which leads to a process with hadrons in central rapidity region which do not come from the higgs decay . the generic formula for @xmath21 is [ sp2 ] s^2_spect = e^- < n_g(y = ln(m^2_h / s_0 ) > where @xmath22 is the mean number of gluon in interval @xmath23 . in pqcd this number is large @xcite @xmath24 which leads to very small cross section for higgs production . for `` soft '' double pomeron production we can estimate the value of @xmath25 assuming that the hadron production is two stage process : ( i ) production of mini jet with @xmath26 and ( ii ) minijet decay in hadrons which can be taken from @xmath27 process . finally , [ sp3 ] < n_g(y ) > = 2 3 , which gives @xmath28 . finally , we have [ god1 ] \|_y = 0 = 0.02 pb we can increase the cross section , measuring reaction of . its cross section is equal to [ god2 ] ( pp \rightarrow pp h)}{d y}\|_{y = 0 } \left ( \frac{\sigma^{sd}\cdot b_{el}(\sqrt{s}/m_h)}{4\,\sigma_{el}\cdot b_{dd}(\sqrt{s}/m_h ) } \right)^2 = \ ] ] @xmath29 and are our results . i firmly believe that they give the maximum values of the cross sections which we could obtain from reasonable estimates . however , i would like to summarize the most sensitive points in our estimates : 1 . the scale for running coupling qcd constant in cross section of higgs production . we took the `` soft '' scale for our estimates . however , it is a point which needs more discussion and even more it looks in contradiction with our feeling , as i have realised during our last meeting . my argument is the blm procedure but more discussions are needed ; 2 . we took @xmath30 for double pomeron processes the same as for hard " lrg process . the justification for this is eikonal type model @xcite , but it could be different opinions as well as direct experimental data ; 3 . the estimates for @xmath21 is very approximate and we need to work out better theory for this suppression . we would like also to mention that the new ideas on high energy interaction such as the saturation of the gluon density at high energy@xcite , will give a more optomistic estimates for the process of interest . the main observation is that there are two contributions for heavy quark diffractive production ( see ): ( i ) the first is so called ingelman - schlein mechanism @xcite which described by fig.2-a and ( ii ) the second one is closely related to coherent diffraction suggested in ref . @xcite and which corresponds to fig . the estimates of both of them have been discussed in ref . the main conclusion is that the main contribution for the tevatron energies stems from cd ( see also @xcite while the is mechanism leads to the value of the cross section in one order @xcite less than cd one . on the other hand in dis the cd contribution belongs to the high twisdt and because of that it is rather small @xcite . our conclusion is very simple . at the tevatrom we has a good chance to measure a new contribution to hard " diffraction which is small in dis . the typical values of the cross section is @xmath31 [ hq1 ] 10 ^ -4 10 ^ -10 forp_t , min = 550gev one can find all details in ref . i am very grateful to a. gotsman and u. maor for encouraging optimism and their permanent discussions on the subject . my special thanks goes to larry mclerran and his mob at the bnl for very creative atmosphere and fruitful discussions . this research was supported in part by the israel science foundation , founded by the israeli academy of science and humanities , and bsf @xmath32 9800276 . this manuscript has been authorized under contract no . de - ac02 - 98ch10886 with the u.s . department of energy . 99 l. alvero , j.c . collins and j.j . whitmore , psu - th-200,hep - ph/9806340 ; + l. alvero , j.c . collins , j. terron and j.j . whitmore , phys . d59 ( 1999 ) 074022 . z. kunszt and w.j . stirling , aachen ecfa ws , p.428 , 1990 . dokshitzer , v.a . khoze and t. sjostrand , phys . b274 ( 1992 ) 116 . marciano and s. willenbrock , phys . d49 ( 1994 ) 1354 . a. bialas and p.v . landshoff , phys . b256 ( 1991 ) 540 . b. muller and alec j. schramm , nucl a523 ( 1991 ) 677 . j - r cudell and oscar f. hernandez , nucl . b471 ( 1996 ) 471 . v. barger , r.j.n.phillips and d.zeppenfeld , phys . b346 ( 1995 ) 106 . s. dawson , nucl . b359 ( 1991 ) 283 ; a. djouadi , m. spira and p. zerwas , phys . lett . b264 ( 1991 ) 440 . brodsky , p. lepage and p. b. mackenzie , phys.rev . d28 ( 1983 ) 228 . e. levin , talk at this ws . e. gotsman , e. levin and u. maor , phys.lett . b353 ( 1995 ) 526 . martin , m.g . ryskin and v.a . khoze , phys.rev . d56 ( 1997 ) 5867 ; phys.lett . b401 ( 1997 ) 330 . l. v. gribov , e. m. levin , and m. g. ryskin , phys . rep . * 100 * , 1 ( 1983 ) ; + mclerran and r. venugopalan , phys . rev . * d 49 * ( 1994 ) 2233 ; * 49 * ( 1994 ) 3352 ; * 50 * ( 1994 ) 2225 ; + yu . v. kovchegov , phys * d 54 * 5463 ( 1996 ) ; * d 55 * 5445 ( 1997 ) ; + kovner , l. mclerran , and h. weigert , phys . rev . * d52 * , 3809 ( 1995 ) ; 6231 ( 1995 ) ; yu . v. kovchegov and dirk h. rischke , phys * c56 * , 1084 ( 1997 ) . g. alves , e. levin and a. santoro , phys . dd55 ( 1997 ) 2683 . g. ingelman and p. schlein , phys . b152 ( 1985 ) 256 . j. collins , l. frankfurt and m. strikman , phys . b307 ( 1993 ) 161 . d. e. soper , talk at dis97 , hep - ph/9707384 . m. wusthoff and a.d . martin , hep - ph/9909362 . f. yuan and k .- chao , hep - ph/9810340 ; hep - ph/9811285 .	in this note we give the highest of reasonable estimates for the value of cross section of the double pomeron higgs meson production and suggest a new mechanism for heavy quark diffractive production which will dominate at the tevatron energies . # 1eq . ( [ # 1 ] ) bnl - nt-99/9 + taup-2615 - 99 +
since the early 1960 s , determinantal ( and closely related pfaffian ) random point processes have served as a key tool in asymptotic analysis of exactly solvable probabilistic systems in mathematics and physics . in the late 1990 s , the domain of their applicability was extended to random growth models and interacting particle systems in ( 1 + 1 ) dimensions , see , e.g. , the surveys of johansson @xcite , ferrari - spohn @xcite , borodin - gorin @xcite , and references therein . about ten years ago , the work of tracy - widom @xcite on asymptotics of the partially asymmetric simple exclusion process ( asep , for short ) started a new wave of developments . the asep is one of the most basic interacting particle systems whose large time asymptotics did not seem to be susceptible to standard determinantal or pfaffian methods . tracy - widom employed a different approach ( coordinate bethe ansatz ) , and a flurry of activity followed , see , e.g. , a survey of corwin @xcite and references therein . tracy - widom showed that the large time fluctuations of current for the asep with step initial data were described by the gue tracy - widom distribution , which originated from the airy determinantal point process @xcite . however , before the large time limit the determinantal processes were nowhere to be seen . some hope for a greater involvement of the determinantal point processes appeared with the work on asymptotics of directed polymers in random media . for the point - to - point continuum brownian polymer ( equivalently , the kardar - parisi - zhang stochastic partial differential equation with so - called narrow wedge initial data ) , that is known to be a limiting object for the asep due to bertini - giacomin @xcite , it was shown by amir - corwin - quastel @xcite , calabrese - le doussal - rosso @xcite , dotsenko @xcite , and sasamoto - spohn @xcite , that the laplace transform of the distribution of its partition function is equal to an average of a simple multiplicative functional on the airy determinantal process ( cf . a recent note of borodin - gorin @xcite for this form of the result ) . for another , semi - discrete brownian polymer model known as the oconnell - yor polymer @xcite , it was shown that the laplace transform of the partition function can be realized via averages over signed determinantal point processes , see oconnell @xcite , imamura - sasamoto @xcite . unfortunately , signed ( i.e. , non - positive ) measures are often of limited probabilistic use , although imamura - sasamoto were able to take the limit of their result to see the kpz - airy connection of the previous paragraph . one goal of this work is to make an explicit connection between the asep and determinantal point processses , and to show how this connection can be used for analyzing the large time asymptotics . the asep can be realized as a limit of another random growth system known as the stochastic six vertex model ; its definition goes back to gwa - spohn @xcite . very recently , borodin @xcite noticed that certain averages over the stochastic six vertex model coincide with other averages for the so - called schur measures ( introduced by okounkov in @xcite ) ; the latter can be thought of as prototypical examples of the determinantal point processes ; see , e.g. , @xcite for detailed explanations . taking the asep limit of this coincidence is not entirely straightforward , and this is the first main result of the present paper . the family of determinantal processes that corresponds to the asep ( with step initial data ) turns out to be a novel one . we call them the discrete laguerre ensembles ; they live on @xmath0 , and their correlation kernels are expressed through the classical laguerre orthogonal polynomials . we prove that the q - laplace transform of the asep height function is equal to the average of a multiplicative functional on the corresponding discrete laguerre ensemble . we then show how this result implies three different asymptotic regimes for the asep . they correspond to two limit regimes of the discrete laguerre ensemble . in the first limit , that deals with finitely many first asep particles , the discrete laguerre ensemble converges to the discrete hermite ensemble ( that goes back to borodin - olshanski @xcite ) . in the second and third limits , that correspond to the asep height function convergence to the gue tracy - widom distribution and the solution of the kpz equation mentioned above , the discrete laguerre ensemble converges to the airy process . the difference between these two limits on the side of the discrete laguerre ensemble is provided solely by different asymptotic behavior of the multiplicative functional . we also explain what the corresponding limits mean for the stochastic six vertex model ( the convergence to the gue tracy - widom distribution was previously obtained in @xcite and @xcite ) . in another direction , we introduce the discrete jacobi ensemble , explain the operator - theoretic mechanism of how all our discrete ensembles arise from the theory of classical orthogonal polynomials , and exhibit numerous limit transitions between more classical orthogonal polynomial ensembles and the new ones . let us now explain our results in more detail . in what follows , we assume the reader s familiarity with the basic definitions and properties of the determinantal point processes ; cf . section [ sect2.1 ] and references therein . rather than introducing the discrete laguerre ensemble ( denoted by @xmath1 below ) by a formula , let us explain a general construction , of which @xmath1 is a particular case . let @xmath2 be a measure on @xmath3 such that ( a ) it is absolutely continuous with respect to the lebesgue measure @xmath4 ; ( b ) it has finite moments of any order ; ( c ) the moment problem for @xmath5 is determinate . let @xmath6 denote the orthonormal polynomials with respect to @xmath5 with positive highest coefficients ; they form a basis in @xmath7 . given a point @xmath8 inside the support of @xmath5 , consider the orthogonal decomposition @xmath9 , where @xmath10 and @xmath11 are the subspaces of functions supported by @xmath12 and @xmath13 , respectively . we have an isomorphism of hilbert spaces @xmath14 by means of the correspondence @xmath15 , @xmath16 . under this isomorphism , the decomposition @xmath9 induces an orthogonal decomposition @xmath17 . denote by @xmath18 and @xmath19 the orthogonal projections onto @xmath20 and @xmath21 , respectively . these operators define determinantal point processes @xmath22 on @xmath23 . the correlation kernels of these point processes ( i.e. , the matrices of @xmath24 ) have the form @xmath25 choosing @xmath26 to be one of the three classical weights @xmath27 we arrive at the discrete hermite , laguerre , and jacobi ensembles , respectively . they are very different from the orthogonal polynomial ensembles associated with these weights ; those live on @xmath3 and have finitely many particles almost surely ; cf . section [ sc : orth - poly - ens ] . note that the roles of the index and the independent variable of the orthogonal polynomials are swapped when one moves from one type of ensembles to the other , which bears certain similarity to the idea of _ bispectrality _ , cf . grnbaum @xcite . one relation between the two types of ensembles is discussed in section [ sc : duality ] . the discrete hermite ensembles previously appeared in borodin - olshanski @xcite , two special cases of the discrete jacobi kernel previously appeared in borodin - kuan @xcite , the discrete laguerre and the general discrete jacobi ensembles appear to be new . a key feauture of @xmath22 that is important to us , is that for the classical weights , the correlation kernels are spectral projections for rather simple second order difference operators on @xmath23 . these are noting else but the _ jacobi matrices _ associated with the corresponding systems of orthogonal polynomials . in other words , they are tridiagonal @xmath28 matrices that represent operators of the form @xmath29 in the basis of the corresponding orthogonal polynomials , where @xmath30 is the operator of multiplication by the independent variable @xmath31 in @xmath32 . the concrete form of these matrices for the classical weights can be found in section [ sect5 ] below . let us explain why this is important to us . the classical hypergeometric polynomials of the askey scheme can always be viewed as eigenfunctions of a suitable second order differential or difference operator ( with polynomial coefficients ) . hence , the correlation kernels of the corresponding orthogonal polynomial ensembles can be viewed as spectral projections for the associated operators . there are several asymptotic regimes , in which the orthogonal polynomials ensembles cease to be such ( for example , the number of particles may tend to infinity ) , but the associated differential / difference operators are easily seen to have a limit . in the case when the state spaces before and after the limit are discrete , this is actually sufficient to claim the convergence of the spectral projections for such operators , thus the convergence of the corresponding determinantal point processes . moreover , the computations are significantly simpler than in more traditional approaches to such asymptotics that typically involve steepest descent arguments or the riemann - hilbert problem methods . however , if some of the involved state spaces are continuous , we are currently not able to produce a rigorous justification of this method , and it should be viewed as heuristic . however , in all the examples we know , it does yield a correct statement with rather modest computations , and it also provides possibly the simplest way of guessing the correct scaling . the difference operator approach to asymptotics of the correlation kernels was first used by borodin - olshanski in @xcite . it was also applied by borodin - gorin @xcite , gorin @xcite , olshanski @xcite , and ( independently ) discussed in the continuous setup by tao @xcite ; see also bornemann @xcite . it is also somewhat similar in spirit to the idea of edelman - sutton @xcite that lead to spectacular progress in understanding limits of the general @xmath33-ensembles in random matrix theory , with the key difference that one has to deal with _ random _ tridiagonal matrices in that case ( which is much harder ) . another related idea can be found in the work of breuer duits @xcite , who used the asymptotics of jacobi matrices to prove gaussian fluctuations for the corresponding orthogonal polynomial ensembles with growing number of particles . in section [ sect6 ] below , we use our approach to prove several limiting statements , realizing the discrete hermite , laguerre , and jacobi ensembles as limits in different ways , cf . figure [ fig ] in that section . one of those statements deals with convergence of the meixner orthogonal polynomial ensembles to the discrete laguerre ensemble . it is this limit that , coupled with a result of @xcite on matching observables of the stochastic six vertex model and the schur measures ( of which meixner ensembles are a special case ) , leads to our first main result : consider the asep on @xmath34 with particles occupying all negative integers at time @xmath35 , and with the left jump rate @xmath36 and right jump rate @xmath37 . let @xmath38 denote the number of the asep particles weakly to the right of the position @xmath39 . then at any time moment @xmath40 , and for any @xmath41 , @xmath42 , we have @xmath43 where @xmath44 is the determinantal point process of the form @xmath45 with the laguerre weight @xmath46 . a similar relation holds for @xmath47 as well . a further limit transition from the discrete laguerre ensemble to the discrete hermite one , which we also prove with the operator method , leads to the asymptotics of the first asep particles at large times . another limit transition , from the discrete laguerre ensemble to the airy one , involves a continuous state space ( for airy ) , so we provide heuristics with our approach , which in particular gives us the ( not so trivial ) correct normalization for the limit . the only piece of this limit transition that is needed for the gue tracy - widom asymptotics proof for the asep is the convergence of the distribution of the first particle , and that is actually equivalent ( thanks to the results of section [ sc : duality ] ) to the convergence of the first particle of the laguerre orthogonal polynomial ensemble to the first airy particle a well - known fact that goes back to @xcite . on the other hand , the convergence of the weakly asymmetric asep to kpz requires the convergence of the full discrete laguerre ensemble to airy , and we sketch the argument , assuming that convergence . a rigorous proof would require a verification of the trace - class convergence of the kernels by a different method ( e.g. , by classical steepest descent arguments ) , which is standard but technical , and we do not pursue it . finally , we utilize the convergences of the meixner and charlier orthogonal polynomial ensembles to the discrete hermite and airy ensembles to give similar large - scale asymptotic statements for two instances of the ( higher spin ) stochastic six vertex model in a quadrant . * acknowldegements . * we are very grateful to vadim gorin for very helpful remarks . the work of a. b. was partially supported by the nsf grants dms-1056390 and dms-1607901 . let @xmath48 be a locally compact polish space . by a _ point configuration _ in @xmath48 we mean an arbitrary subset @xmath49 without accumulation points ; it is either finite or countable . with any @xmath50 one associates the atomic measure @xmath51 ( the sum of delta - measures at the points of @xmath52 ) . the space of all point configurations is denoted by @xmath53 ; it can be endowed with a natural structure of measurable space . given a probability measure @xmath54 on @xmath53 , one can speak of a _ random _ point configuration @xmath52 . likewise , @xmath55 becomes a random radon measure on @xmath48 . averaging @xmath55 with respect to @xmath54 one obtains a ( non - random ) measure @xmath56 on @xmath48 , called the _ density measure _ of @xmath54 . it is the simplest invariant of @xmath54 . one can generalize this construction and obtain an infinite sequence of invariants @xmath57 , where @xmath58 is a symmetric measure on @xmath59 ( the @xmath60-fold product @xmath61 ) called the _ @xmath60-point correlation measure_. @xmath54 is said to be a _ determinantal measure _ , or a _ determinantal point process _ , if one can exhibit a radon measure @xmath62 on @xmath48 ( called the _ reference measure _ ) and a function @xmath63 on @xmath64 ( called the _ correlation kernel _ ) such that for each @xmath65 , @xmath58 has a density @xmath66 with respect to @xmath67 , and this density ( called the _ @xmath60-point correlation function _ ) is given by symmetric minors of the kernel : @xmath68_{i , j=1}^k.\ ] ] if such a pair @xmath69 , @xmath70 exists , then it determines @xmath54 uniquely . on the other hand , different pairs may produce the same determinantal measure . for instance , one can replace @xmath62 by an equivalent measure @xmath71 and at the same time replace the kernel @xmath63 by the new kernel @xmath72 ; then the correlation measures do not change . another possibility is to keep @xmath62 fixed but replace @xmath63 with @xmath73 , where @xmath74 is a nonvanishing function on @xmath48 ; this transformation also does not affect the correlation measures . which pairs @xmath69 , @xmath70 give rise to determinantal measures is a difficult question if one does not impose additional assumptions . usually one wants @xmath62 to be some natural measure ( say , the lebesgue measure on @xmath75 , or the counting measure when @xmath48 is a discrete space ) , but the problem is related to a description of admissible kernels @xmath63 . however , in the present paper we are dealing with a very special class of correlation kernels , and for our purposes the following abstract existence theorem is sufficient . let @xmath76 be a closed subspace of the hilbert space @xmath77 , and @xmath78 be the orthogonal projection onto @xmath76 . the kernel @xmath63 of the operator @xmath78 is the _ reproducing kernel _ of @xmath76 , which is defined by @xmath79 where @xmath80 is an arbitrary orthonormal basis of @xmath76 ; the definition does not depend on the choice of the basis . assume that the function @xmath81 is locally @xmath62-integrable ; then there exists a ( unique ) determinantal measure @xmath82 for which @xmath63 serves as a correlation kernel . for a more accurate formulation of this result , see soshnikov @xcite . note that one can change @xmath76 and hence @xmath78 without changing the corresponding measure @xmath83 . namely , let @xmath74 be a function on @xmath48 with the values on the unit circle in @xmath84 , and let @xmath85 consist of the functions of the form @xmath86 , where @xmath87 ranges over @xmath76 . then replacing @xmath76 with @xmath85 does not affect the determinantal measure . even if one wants @xmath74 to be real - valued , it may take values @xmath88 . concrete examples will be given below . if @xmath89 , the range of @xmath78 , has finite dimension , then the measure @xmath83 always exists , and it is concentrated on the subspace @xmath90 of @xmath91-point configurations , where @xmath92 ( the converse is also true ) . if @xmath76 has infinite dimension and @xmath83 exists , then the @xmath83-random configuration @xmath52 is infinite , with probability 1 . in this section we assume that @xmath48 is a finite set or a countable set with discrete topology ( this is what we mean by the discrete case ) . then @xmath53 is simply the set @xmath93 of all subsets of @xmath48 . if @xmath48 is finite , then @xmath53 is a finite set , too . if @xmath48 is countable , then @xmath53 is a compact , totally disconnected space with respect to the natural topology the base of the topology is formed by the cylinder sets @xmath94 we are mainly interested in the case when @xmath48 is countable , but occasionally we will need finite sets @xmath48 as well . we take as @xmath62 the counting measure on @xmath48 . the correlation functions of a probability measure @xmath54 on @xmath53 admit a simple interpretation : for a @xmath60-tuple of distinct points @xmath95 , the value @xmath66 is equal to @xmath96 ; in other words , @xmath66 is the probability that the random configuration contains all the points @xmath97 . for the counting measure @xmath62 , the hilbert space @xmath77 turns into the coordinate space @xmath98 with its distinguished basis @xmath99 indexed by the points of @xmath48 . in the discrete case , the subtleties related to an accurate definition of a reproducing kernel disappear , and any closed subspace @xmath100 gives rise to a determinantal measure . in the discrete case , there exists a special operation called the _ particle / hole involution_. this is the involutive self - map @xmath101 assigning to a subset @xmath52 its complement @xmath102 . the correspondence @xmath103 induces , in a natural way , an involutive map @xmath104 on the set of probability measures on @xmath53 . on the subset of determinantal measures , the latter map takes the form @xmath105 . here we define a class of determinantal point processes associated with orthogonal polynomials . consider a system @xmath106 of orthogonal polynomials with a weight measure @xmath107 on @xmath3 , and let @xmath108 be the support of @xmath107 . the system @xmath109 may be fairly general , but , to slightly simplify things , let us assume that it is taken from the askey scheme ( see koekoek swarttouw ( * ? ? ? * chapter 1 ) ) . then @xmath110 is either a discrete subset or a closed interval ( possibly , with infinite ends ) . in the latter case we denote by @xmath111 the density of @xmath107 with respect to the lebesgue measure . ) we set @xmath112 and @xmath113 , and for @xmath114 we denote by @xmath115 the @xmath91-dimensional subspace of @xmath116 formed by the polynomial functions of degree @xmath117 ( if the set @xmath48 is finite , then we suppose that @xmath91 is smaller that its size ) . let @xmath118 be the projection onto @xmath115 , and @xmath119 be the corresponding determinantal measure . the probability space @xmath120 is called the _ @xmath91-particle orthogonal polynomial ensemble _ , see koenig @xcite for a survey . its correlation kernel , taken with respect to the reference measure @xmath121 , is the _ christoffel darboux kernel _ @xmath122 where the norm is that of the weighted hilbert space @xmath116 . as is well known , the christoffel darboux kernel can also be written in the form @xmath123 where @xmath124 denotes the leading coefficient in @xmath125 . if instead we take as @xmath62 the counting measure ( in the discrete case ) or the lebesgue measure ( in the continuous case ) , then the kernel should be multiplied by the factor @xmath126 ; we write it as @xmath127 let @xmath2 be an arbitrary measure on @xmath3 with the following properties : \(i ) @xmath5 is absolutely continuous with respect to the lebesgue measure @xmath4 ; \(ii ) @xmath5 has finite moments of any order ; \(iii ) the moment problem for @xmath5 is determinate . condition ( i ) is not strictly necessary but it simplifies things . condition ( ii ) ensures the existence of an infinite system @xmath128 of orthogonal polynomials ; we denote the corresponding orthonormal system by @xmath129 . condition ( iii ) implies that the space @xmath130 $ ] of polynomials is dense in the weighted hilbert space @xmath7 , so that @xmath131 is a basis in @xmath132 ( see akhiezer ( * ? ? ? * corollary 2.3.3 ) ) . in what follows , we assume that for each @xmath133 , the leading coefficient of @xmath134 is strictly positive . this means that the polynomials @xmath134 are defined up to positive numeric factors , while the polynomials @xmath135 are uniquely defined by the weight measure . given a point @xmath8 , we consider the orthogonal decomposition @xmath9 , where @xmath10 and @xmath11 are the subspaces of functions supported by @xmath12 and @xmath13 , respectively . next , we assume that @xmath136 is inside the support of @xmath5 , so that both these semi - infinite intervals have strictly positive mass relative to @xmath5 . this means that both @xmath137 and @xmath138 have infinite dimension . finally , we define an isomorphism of hilbert spaces @xmath14 by means of the correspondence @xmath139 , @xmath16 . under this isomorphism , the decomposition @xmath9 induces an orthogonal decomposition @xmath17 . then we denote by @xmath18 and @xmath19 the projections onto @xmath20 and @xmath21 , respectively . these projection operators are the objects of interest for us . [ def3.a ] let @xmath5 and @xmath140 be as above , @xmath136 be a point inside the support of @xmath5 , and @xmath24 be the corresponding self - adjoint projection operators on @xmath141 as defined above . by the general theory ( section [ sect2.1 ] ) , @xmath24 gives rise to a determinantal measure @xmath22 on @xmath142 ( in other words , a determinantal point process on @xmath23 ) ; we call @xmath22 the _ discrete ensemble associated with the system @xmath140_. note that @xmath143 . note also that the @xmath22-random configuration contains almost surely infinitely many points : as pointed out in section [ sect2.1 ] , this follows from the fact that @xmath24 has infinite rank . [ def3.b ] let @xmath136 be a point inside the support of @xmath5 . introduce kernels @xmath144 on @xmath28 by setting @xmath145 we call @xmath144 the _ discrete kernel associated with the system @xmath140_. obviously , @xmath144 is the kernel of the projection @xmath24 . hence , @xmath144 serves as a correlation kernel for @xmath22 . observe that conditions ( i)(iii ) hold true for the hermite , laguerre , and jacobi orthogonal polynomials . indeed , conditions ( i ) and ( ii ) are obvious . as for condition ( iii ) , it is obvious in the jacobi case ( because then the weight measure has bounded support ) , and it is well known in the hermite and laguerre cases ( the moments of the weight measure do not grow too fast ) . therefore , the general construction described above is applicable to these three systems of polynomials . this leads us to three families of determinantal point processes on @xmath23 , which we call the _ discrete hermite ensemble _ , _ discrete laguerre ensemble _ , and _ discrete jacobi ensemble_. the same names are used for the corresponding kernels given by . each ensemble depends on the additional continuous parameter @xmath136 and has two variants , `` plus '' and `` minus '' , corresponding to the intervals @xmath13 and @xmath12 . below we examine these kernels in more detail . in particular , we explain how to write them in the _ integrable form _ ( see @xcite ) by making use of the forward and backward shift operators related to the corresponding systems of orthogonal polynomials . all necessary formulas can be found in koekoek swarttouw @xcite . the two variants of this ensemble are denoted by @xmath146 and @xmath147 . they correspond to the intervals @xmath13 and @xmath12 , respectively . here parameter @xmath136 ranges over @xmath148 . for the _ hermite polynomials _ we use the standardization and notation of @xcite , which are the most common ones . the weight measure of the hermite polynomials is @xmath149 , where @xmath150 . the @xmath151th hermite polynomial , denoted as @xmath152 , is specified by the property that its leading coefficient equals @xmath153 . in this standardization , @xmath154 using this formula we write down the integral representation of the discrete hermite kernel ( below @xmath155 ) : @xmath156 since @xmath157 , we have @xmath158 this yields the symmetry relation @xmath159 because the factor @xmath160 does not affect the determinantal measure . [ prop3.a ] for @xmath161 , the discrete hermite kernel can be written in the form @xmath162 we write @xmath163 next , we integrate by parts in the first integral using the formulas @xmath164 which are obtained from ( * ? ? ? * ( 1.13.6 ) and ( 1.13.8 ) ) ( the forward and backward shifts ) . then we do the same with the second integral . the resulting two integral terms are cancelled out , and we obtain the desired formula for the kernel @xmath165 . the case of @xmath165 is handled in exactly the same way . note that the sign @xmath166 in the right - hand side agrees with the fact that @xmath167 we define the _ laguerre polynomials _ as the orthogonal polynomials on @xmath168 with the weight measures @xmath169 , where @xmath170 . we denote by @xmath171 the @xmath151th laguerre polynomial ; in our standardization , its leading coefficient equals @xmath172 . this slightly differs from the conventional definition : the connection with the notation of @xcite is the following : @xmath173 where @xmath174 is the @xmath151th laguerre polynomial in the standardization of @xcite . here is the formula for the norm : @xmath175 the two variants of the discrete laguerre ensemble are denoted by @xmath176 . the corresponding discrete laguerre kernel has the form @xmath177 in contrast to the case of the hermite polynomials ( and that of the jacobi polynomials , see below ) , there is no symmetry relation which would reduce the first integral to the second one . [ prop3.b ] for @xmath161 , the discrete laguerre kernel can be written in the form @xmath178 with the convention that @xmath179 . as in the proof of proposition [ prop3.a ] , we apply the same trick with integration by parts . but now we use the backward shift first and the forward shift next . more precisely , these are the formulas @xmath180 which are derived from ( * ? ? ? * ( 1.11.8 ) and ( 1.11.6 ) ) . the jacobi polynomials are defined as in @xcite . they depend on two parameters @xmath181 , @xmath182 , and are denoted by @xmath183 . the argument @xmath31 and the additional parameter @xmath136 range over @xmath184 . the two variants of the discrete jacobi ensembles are denoted by @xmath185 . using the formula @xmath186 we write the integral representation as @xmath187 where the explicit expression for the norm is @xmath188 as in the case of the discrete hermite ensemble , we have a symmetry relation ; now it takes the form ( note the @xmath189 swap ) @xmath190 [ prop3.c ] for @xmath161 , the discrete jacobi kernel can be written in the form @xmath191 where @xmath192 the same trick as in propositions [ prop3.a ] and [ prop3.b ] . in the case of @xmath193 we write @xmath194 next , we integrate by parts in the first integral using the formulas @xmath195 and @xmath196 which are obtained from ( * ? ? ? * ( 1.8.8 ) ) ( backward shift ) and ( * ? ? ? * ( 1.8.6 ) ) ( forward shift ) , respectively . then we do the same with the second integral . again , the resulting integral terms are cancelled out , and we obtain the desired expression for the kernel @xmath197 . the case of @xmath198 is handled in exactly the same way . the fact that we obtain ( in the denominator ) the difference @xmath199 instead of @xmath200 , is caused by the very structure of the forward and backward shift operators in the jacobi case . trying @xmath200 , as before , we could not kill the integral terms . the transformation @xmath201 of the discrete index @xmath202 agrees with the limit transition discussed in section [ racah - djacobi ] below . let us return to the general setting of section [ sect3.1 ] and consider the discrete ensembles @xmath22 linked to a system @xmath203 of orthonormal polynomials with a weight measure @xmath26 satisfying the three conditions ( i)(iii ) . to handle the two variants together , let us use the common notation @xmath204 , where @xmath205 denotes either the interval @xmath13 ( in the case of @xmath206 ) or the interval @xmath207 ( in the case of @xmath208 ) . next , let us fix a natural number @xmath91 and denote by @xmath209 the continuous @xmath91-particle orthogonal polynomial ensemble coming from the same system of polynomials . we are going to show that certain two gap probabilities , which are related to @xmath210 and @xmath209 , respectively , are the same . [ thm3.a ] let @xmath52 denote the @xmath210-random infinite configuration on @xmath23 , and @xmath211 denote the @xmath209-random @xmath91-particle configuration on @xmath3 . then the following gap probabilities are the same : @xmath212 note that the probabilities in the left - hand side determine the distribution of the leftmost particle in @xmath22 , while the probabilities in the right - hand side do the same for the leftmost or the rightmost particle in @xmath209 , depending on whether @xmath213 or @xmath214 . as is seen from the proof , the special form of the interval @xmath205 is inessential here : it could be replaced an arbitrary borel subset of positive @xmath5-mass . we apply the well known fact that gap probabilities for determinantal processes are given by fredholm determinants ( see , e.g. , ( * ? ? ? * theorem 2 ) , where one has to specialize @xmath215 ) . then the equality in question reduces to the following one : @xmath216 where @xmath217 is the @xmath218 matrix with the entries @xmath219 and @xmath220 is the rank @xmath91 operator on the hilbert space @xmath221 given by the kernel @xmath222 comparing these two expressions we see that the desired equality is a continual analogue of the identity @xmath223 for two rectangular matrices @xmath224 of the same format . the proof of is a routine exercise . indeed , the left - hand side is equals to @xmath225_{i , j=1}^m,\ ] ] while the right - hand side is equal to @xmath226_{i , j=1}^m { \mathcal w}(dt_1)\dots { \mathcal w}(dt_m).\ ] ] then we have to show that for each fixed @xmath227 , the @xmath227-fold sum over @xmath228 is equal to the @xmath227-fold integral over @xmath229 . this is verified directly , by expanding each minor into a sum over permutations and making use of the definition of @xmath217 and @xmath220 . here we describe , in general form , our approach to studying limit transitions for certain determinantal point processes . concrete examples are given in the subsequent sections . we are dealing with determinantal measures of the form @xmath83 , where @xmath78 is a projection operator acting on @xmath230 and @xmath48 is a countable set ( see section [ sect2 ] above ) . [ prop4.a ] let @xmath231 be an infinite sequence of projection operators acting on @xmath230 , and @xmath78 be one more projection . if the kernels @xmath232 converge to the kernel @xmath63 pointwise on @xmath64 , then the determinantal measures @xmath119 weakly converge to the determinantal measure @xmath83 . from the interpretation of the correlation functions given in section [ sect2.2 ] one sees that if @xmath233 on @xmath64 , then @xmath234 for an arbitrary cylinder set @xmath235 . by the definition of the topology in @xmath53 , this implies the weak convergence @xmath236 . note that the pointwise convergence @xmath233 is equivalent to the weak convergence of projection operators @xmath237 , which in turn is equivalent to their strong convergence , because on the set of projections , the weak and strong operator topologies coincide . recall ( see @xcite ) that a sequence @xmath238 of self - adjoint operators on a hilbert space converges to a self - adjoint operator @xmath239 in _ strong resolvent sense _ if their resolvents converge strongly : @xmath240 given a self - adjoint operator @xmath239 and an open interval @xmath241 , possibly semi - infinite , we will denote by @xmath242_{({r}_1,{r}_2)}$ ] the spectral projection of @xmath239 corresponding to @xmath243 . we also abbreviate @xmath242_+:=[a]_{(0,+\infty)}$ ] . [ prop4.b ] let @xmath244 , @xmath114 , and @xmath239 be self - adjoint operators on a hilbert space and suppose that @xmath245 in the strong resolvent sense . next , let @xmath241 be an arbitrary open interval , possibly semi - infinite and such that its finite ends are not in the point spectrum of @xmath239 . then the spectral projections @xmath246_{({r}_1,{r}_2)}$ ] strongly converge to the spectral projection @xmath242_{({r}_1,{r}_2)}$ ] . for finite intervals , this assertion is contained in ( * ? ? ? * theorem viii.24 , claim ( b ) ) . in the case of a semi - infinite interval the argument is the same . recall one more general definition . let @xmath239 be a closed operator on a banach space ( in particular , on a hilbert space ) . a dense subspace @xmath247 contained in the domain of @xmath239 is said to be a _ core _ of @xmath239 if the closure of the operator @xmath248 ( the restriction of @xmath239 to @xmath247 ) coincides with @xmath239 . in particular , if @xmath239 is a self - adjoint operator on a hilbert space and @xmath247 is a core of @xmath239 , then the operator @xmath248 is essentially self - adjoint . the next result provides an effective tool for checking the strong resolvent convergence of self - adjoint operators . [ prop4.c ] let @xmath244 , @xmath114 , and @xmath239 be self - adjoint operators on a hilbert space @xmath132 , and suppose that there exists a dense subspace @xmath249 such that @xmath247 is a common core for all these operators and @xmath250 as @xmath251 , for any vector @xmath252 . then @xmath245 in the strong resolvent sense . see ( * ? ? ? * theorem viii.25 ) . a matrix @xmath253 with the entries @xmath254 is called _ tridiagonal _ if @xmath255 for @xmath256 . jacobi matrix _ is a real symmetric tridiagonal matrix whose off - diagonal entries are strictly positive . we will deal with finite and semi - infinite jacobi matrices ; in the former case we assume that the indices @xmath257 range over @xmath258 , where @xmath259 is a positive integer , and in the latter case we assume that @xmath257 range over @xmath23 . thus , a finite jacobi matrix is determined by the diagonal entries @xmath260 ( where @xmath261 ) and the off - diagonal entries @xmath262 ( where @xmath263 ) . likewise , a semi - infinite jacobi matrix is determined by two infinite sequences @xmath264 , @xmath265 indexed by @xmath266 . following @xcite , we call a semi - infinite jacobi matrix a _ @xmath267-matrix_. we also abbreviate @xmath268 . let @xmath269 denote the dense subspace of @xmath270 formed by the vectors with finitely many nonzero coordinates . every @xmath267-matrix @xmath253 determines a symmetric operator on @xmath270 with domain @xmath269 ; we denote this operator again by @xmath253 . as is well known ( see @xcite ) , the deficiency indices of @xmath253 are either @xmath271 or @xmath272 , and the question of which of these two cases holds is closely related to the classical moment problem . if the deficiency indices are @xmath272 , then @xmath253 is essentially self - adjoint , so that its closure @xmath273 is a self - adjoint operator and @xmath269 is a core for @xmath273 . here is a simple condition of essential self - adjointness which suffices for our purposes : [ prop4.d ] let @xmath253 be a @xmath267-matrix such that @xmath274 then its deficiency indices are @xmath272 , and hence @xmath253 is essentially self - adjoint . see @xcite , chapter i , addenda and problems , item 1 . for a @xmath267-matrix @xmath253 with deficiency indices @xmath272 , the domain of @xmath273 is easily described : [ prop4.e ] in the case of deficiency indices @xmath272 , the domain of the self - adjoint operator @xmath273 consists of those vectors @xmath275 for which the infinite vector @xmath276 with the coordinates @xmath277 is still in @xmath270 , and then @xmath278 . the fact that @xmath253 has deficiency indices @xmath272 just means that @xmath279 , and then one can apply a simple argument , see , e.g. , ( * ? ? ? * chapter iv , 1.1 ) . now we are in a position to explain how our method works . we consider certain projection operators @xmath280 , @xmath78 acting on @xmath270 , and we want to show that @xmath237 in the weak = strong operator topology . in the concrete cases under consideration we are able to exhibit essentially self - adjoint operators @xmath281 , @xmath253 with domain @xmath269 , such that @xmath282_+ , \quad k=[a]_+ , \quad \textrm{where } \quad n=1,2,\dots , \quad a_n:=\overline{\mathcal a}_n , \ ; a:=\overline{\mathcal a}\ ] ] ( we recall that the symbol @xmath283_+$ ] denotes the spectral projection corresponding to the interval @xmath284 ) . the operator @xmath253 is given by a @xmath267-matrix with deficiency indices @xmath272 , and each @xmath285 is given by a tridiagonal matrix which is either a @xmath267-matrix with deficiency indices @xmath272 or the direct sum of a finite jacobi matrix and a scalar matrix of infinite size . we verify that , as @xmath251 , @xmath286 this exactly means that @xmath245 on @xmath269 . since @xmath269 is a common core , proposition [ prop4.c ] shows that @xmath245 in the strong resolvent sense . in our examples , the spectrum of @xmath239 is purely continuous . therefore , we may apply proposition [ prop4.b ] and conclude that @xmath237 , as desired . the point is that ( in our examples ) the kernels of the projection operators are expressed through transcendental functions while the entries of the tridiagonal matrices are given by simple elementary formulas . for this reason , working with tridiagonal matrices turns out to be much easier than with kernels : asymptotic analysis reduces to elementary computations . in section [ sect6 ] we investigate five systems of discrete orthogonal polynomials : charlier , meixner , krawtchouk , hahn , and racah . the necessary information about these polynomials is contained in koekoek swarttouw @xcite . in all cases , large-@xmath91 limit transitions are computed by the same algorithm ; we proceed to its description . let @xmath287 denote any of these systems of polynomials and @xmath107 denote the corresponding weight function . there are slight differences between two variants : _ variant _ 1 : charlier and meixner . _ variant _ 2 : krawtchouk , hahn , and racah . the reason is that in variant 1 , the support of @xmath107 is the whole set @xmath23 and there are infinitely many polynomials , while in variant 2 , the support of @xmath107 is a finite set of the form @xmath288 and the system comprises finitely many polynomials @xmath289 only . let us examine variant 1 first . _ step _ 1 . we consider the weighted hilbert space @xmath290 and observe that the space of polynomials @xmath291 $ ] is its dense subspace . indeed , this claim is well known in the case of charlier and meixner polynomials . to verify it one can use the following general result : the space of polynomials is dense if and only if the moment problem related to the weight function @xmath107 is determinate ( see , e.g. , ( * ? ? ? * , prop . 4.15 ) ) , which in turn is guaranteed if the exponential generating series for the moments of @xmath107 has a nonzero radius of convergence ( ( * ? ? ? * , prop . 1.5 ) ) . the fact that the latter property holds in the charlier or meixner case is easy to check . _ step _ 2 . associated with @xmath292 is a second order difference operator @xmath293 with the following properties : @xmath294 the @xmath125 s are its eigenfunctions , @xmath295 , and one has @xmath296 @xmath294 the action @xmath293 on a test function @xmath297 on @xmath23 is given by a tridiagonal matrix : @xmath298 @xmath294 the off - diagonal coefficients @xmath299 are strictly positive ( with the only exception of @xmath300 ) and satisfy the relation @xmath301 ( it means that @xmath293 is symmetric with respect to the inner product of @xmath302 ; @xmath294 the diagonal coefficients are negative and given by @xmath303 the operator of multiplication by the function @xmath304 establishes an isomorphism of hilbert spaces @xmath305 . it transforms @xmath293 into another second order difference operator @xmath306 . its action is written as @xmath307 where @xmath308 note that @xmath309 so that @xmath310 is given by a jacobi matrix . we check that the assumption of proposition [ prop4.d ] is satisfied , and hence the operator @xmath311 , the restriction of @xmath310 to @xmath269 , is essentially self - adjoint . let @xmath312 denote its closure ; this is a self - adjoint operator on @xmath270 . on the other hand , since the polynomials @xmath125 are eigenfunctions of the difference operator @xmath293 with eigenvalues @xmath313 , we have @xmath314 we claim that the same holds with @xmath310 replaced by @xmath312 , that is , @xmath315 at first glance this looks evident , but actually is not , because @xmath312 is defined as the closure of the operator @xmath311 , while the functions @xmath316 do not belong to @xmath269 . but this difficulty is resolved with the help of proposition [ prop4.e ] : it tells us that the functions @xmath316 lie in the domain of @xmath312 , and on all these functions , the action of @xmath312 is implemented by @xmath310 . we conclude that the self - adjoint operator @xmath312 is diagonalized in the orthogonal basis of @xmath270 formed by the functions @xmath316 , @xmath133 . in particular , it has purely point , multiplicity free spectrum @xmath317 . _ step _ 3 . given @xmath114 , we set @xmath318 where @xmath319 is an appropriate constant . the results of the previous steps show that the spectral projection @xmath320_+=[d+\mu_n]_{(0,+\infty)}$ ] coincides with @xmath118 , the @xmath91-dimensional projection operator introduced in section [ sc : orth - poly - ens ] . ( recall that the range of @xmath118 is the @xmath91-dimensional subspace of @xmath270 spanned by the first @xmath91 functions @xmath316 , @xmath321 . ) division by a positive constant factor does not affect the spectral projection corresponding to the ray @xmath284 , so that we have @xmath322_+=k_n.\ ] ] on the other hand , we know that @xmath244 is determined by the jacobi matrix @xmath285 with the coefficients @xmath323 note that @xmath111 and @xmath299 depend on the parameters entering the definition of the polynomials @xmath125 . now our task is to find an appropriate limit regime : we let @xmath251 and tune these parameters ( which become depending on @xmath91 ) together with @xmath324 in such a way that there exist limits @xmath325 note that such a limit regime is not unique . in our examples , it depends on the deformation parameter @xmath136 . moreover , as is seen from the results of section [ sect6 ] , various limit regimes can differ in a more substantial way . _ step _ 4 . let @xmath253 denote the limit jacobi matrix . in our examples , it comes from the three - term relation for one of the three systems of orthogonal polynomials investigated in section [ sect3 ] . namely , let @xmath326 be the common notation for the hermite , laguerre , and jacobi polynomials . we denote by @xmath327 the corresponding orthonormal polynomials and by @xmath26 the weight measure . next , let @xmath30 denote the operator of multiplication by the coordinate function @xmath31 acting on the hilbert space @xmath328 . the jacobi matrix @xmath253 turns out to be the matrix of the operator @xmath329 in the basis @xmath140 with an appropriate choice of @xmath330 . in section [ sect5 ] below we write down the matrices @xmath253 explicitly in each of the three cases . their off - diagonal entries satisfy the assumption of proposition [ prop4.d ] and hence the deficiency indices of @xmath253 are @xmath272 . we denote by @xmath239 the self - adjoint operator on @xmath269 obtained by taking the closure of @xmath331 . we claim that @xmath239 is precisely the image of @xmath329 under the isomorphism @xmath332 taking the basis @xmath140 to the canonical basis of @xmath270 . indeed , to see this we argue as in the end of step 2 , with appeal to proposition [ prop4.e ] . because the spectrum of operator @xmath329 is purely continuous , the same holds for @xmath239 . then the argument of section [ sect4.3 ] shows that the spectral projections @xmath246_+$ ] converge to the spectral projection @xmath242_+$ ] . this gives the final result : the convergence of the @xmath91-point ensemble under consideration to one of the ensembles from section [ sect3 ] . the algorithm of section [ sect4.4 ] remains essentially the same , but due to finiteness of the support of the weight function , the situation is simplified and some of the arguments can be omitted . we only indicate necessary modifications . _ step _ 1 . instead of @xmath302 we have to deal with the finite - dimensional space @xmath333 . it coincides with the linear span of @xmath289 . no appeal to the moment problem is needed . _ step _ 2 . the infinite system of orthogonal polynomial is replaced by a finite one , @xmath334 , and the hilbert space @xmath270 is replaced by the @xmath335-dimensional hilbert space @xmath336 . because of this the situation is simplified ( no need to take the closure of a densely defined operator ) . _ step _ 3 . we define the jacobi matrix @xmath285 as before , but now it has finite format @xmath337 . the number @xmath259 is an additional parameter which will vary together with @xmath91 ; we assume @xmath338 , so @xmath259 will grow with @xmath91 . since the spaces @xmath336 will vary , it is convenient to extend the pre - limit operators @xmath244 to the space @xmath270 . to do this we consider the natural direct sum decomposition @xmath339 and set @xmath340 on the second component ; this does not affect the spectral projection @xmath246_+$ ] . _ step _ 4 . here nothing changes . the purpose of the present section is to exhibit @xmath267-matrices which will appear in various concrete instances of limit transitions . we begin with a general definition , where we adopt the notation and assumptions of section [ sect3.1 ] . thus , @xmath342 is a system of orthogonal polynomials with a weight measure @xmath26 satisfying conditions ( i)(iii ) , @xmath343 are the corresponding orthonormal polynomials , and @xmath136 is a fixed point inside the support of @xmath5 . recall that from these data we can construct determinantal point processes @xmath206 and @xmath208 . let @xmath30 denote the operator on @xmath32 consisting in multiplication by the coordinate function @xmath31 . this is a self - adjoint operator . its action in the basis @xmath140 is given by a semi - infinite tridiagonal matrix : @xmath344 this is a reformulation of the classical three - term relation which holds for any system of orthogonal polynomials . in the orthonormal basis @xmath345 , the action of @xmath30 can be written as @xmath346 where @xmath347 note that the matrix @xmath348 is symmetric : @xmath349 . [ def5.a ] we define the symmetric tridiagonal matrix @xmath350 via @xmath351 in other words , @xmath352 is the matrix of the operator @xmath353 in the basis @xmath345 , and @xmath354 is the matrix of the operator @xmath355 in the basis @xmath356 . [ prop5.a ] let @xmath350 be the matrix just defined . \(i ) @xmath357 for all @xmath266 , so @xmath253 is a @xmath267-matrix . \(ii ) @xmath253 has deficiency indices @xmath272 , whence it determines an essentially self - adjoint operator with domain @xmath269 . \(iii ) the self - adjoint operator @xmath358 on @xmath270 has simple , purely continuous spectrum filling the support of @xmath5 . \(iv ) the determinantal measure corresponding to the spectral projection @xmath242_+$ ] coincides with @xmath22 . note that of these four assertions , only ( ii ) is nonevident . \(i ) recall that in our standardization , the leading coefficients of polynomials @xmath134 are strictly positive . this implies that @xmath359 for all @xmath266 , which in turn means that @xmath360 and so @xmath361 . thus , @xmath253 is a @xmath267-matrix . \(ii ) by the definition of @xmath253 , ( ii ) can be rephrased as follows : the restriction of the operator @xmath30 to the subspace @xmath130\subset l^2({\mathbb r},{\mathcal w}(dt))$ ] is essentially self - adjoint . it is known ( see , e.g. , ( * ? ? ? * , theorem 2 ) ) that the latter fact holds if and only if the moment problem for @xmath5 is determinate . but this property has been postulated . alternatively , in the case of hermite , laguerre or jacobi polynomials , when we dispose of explicit expressions for the entries @xmath262 ( see below ) , we can deduce ( ii ) from proposition [ prop4.d ] . \(iii ) evident , because @xmath239 is equivalent to the operator @xmath329 . \(iv ) the determinantal measure in question has @xmath254 as a correlation kernel . on the other hand , according to definition [ def3.b ] , the correlation kernel @xmath144 of the measure @xmath22 is the matrix of @xmath329 in the basis @xmath345 . hence , @xmath362 thus , in the case @xmath363 , our assertion is trivial . in the case @xmath364 , we use the fact that the factor @xmath160 does not affect the determinantal measure ( see the end of section [ sect2.1 ] ) . below we write down the matrix @xmath253 for three systems of orthogonal polynomials : hermite , laguerre , and jacobi . this is a simple exercise : we use the formulas above and the explicit expression for the three - term relation , which we take from @xcite . the three - term relation for the hermite polynomials has the form ( see ( * ? ? ? * ( 1.13.3 ) ) ) @xmath366 together with the expression for the norm , see section [ sect3.2 ] , this implies @xmath367 the three - term relation for the laguerre polynomials in our standardization has the form ( see ( * ? ? ? * ( 1.11.3 ) ) ) @xmath369 together with the expression for the norm ( see section [ sect3.3 ] ) , we obtain @xmath370 the three - term relation for the jacobi polynomials has the form ( see ( * ? ? ? * ( 1.8.3 ) ) ) @xmath372 together with the expression for the norm , see section [ sect3.4 ] , this implies @xmath373 in this section we establish several limit transitions between determinantal point processes that can be seen in figure [ fig ] ; see sections [ sect4.4 ] and [ sect4.5 ] for a general description of how the computations proceed . ] the _ charlier polynomials _ are orthogonal polynomials on @xmath23 with the weight function @xmath375 depending on a parameter @xmath376 , see @xcite . we denote by @xmath377 the corresponding @xmath91-particle ensemble . we need the discrete hermite ensemble @xmath146 , which is defined in section [ sect3.2 ] . fix @xmath8 , let @xmath251 , and let the parameter @xmath378 vary together with @xmath91 in such a way that @xmath379 in this limit regime , @xmath380 . we apply the algorithm described in section [ sect4.4 ] , where we use the above expression for the weight function and the following formulas taken from ( * ? ? ? * ( 1.12.5 ) ) ) : @xmath381 where @xmath202 and @xmath151 range over @xmath23 . from these data we compute the entries of the jacobi matrix @xmath285 according to : @xmath382 let us take @xmath383 . then , as @xmath251 , @xmath384 the limit values coincide with the entries @xmath262 , @xmath260 of the jacobi matrix corresponding to the discrete hermite ensemble @xmath146 , see . this completes the proof . the _ meixner polynomials _ are orthogonal polynomials on @xmath23 with the weight function @xmath385 which depends on two parameters , @xmath170 and @xmath386 , see @xcite ( in @xcite , the parameter @xmath387 is denoted by @xmath388 ) . we denote by @xmath389 the corresponding @xmath91-particle ensemble . we also need the discrete laguerre ensemble @xmath390 , which is defined in section [ sect3.3 ] . [ thm4.a ] fix @xmath391 and @xmath170 , let @xmath251 , and let the parameter @xmath387 vary together with @xmath91 in such a way that @xmath392 , that is , @xmath393 in this limit regime , @xmath394 . we apply the algorithm of section [ sect4.4 ] , for which we use the above expression for the weight function and the following formulas taken from ( * ? ? ? * ( 1.9.5 ) ) : @xmath395 where @xmath202 and @xmath151 range over @xmath23 . from these data we compute the entries of the jacobi matrix @xmath285 according to : @xmath396 let us take @xmath397 . then , as @xmath251 , @xmath398 and the limit values coincide with the entries @xmath262 , @xmath260 of the jacobi matrix corresponding to the discrete laguerre ensemble @xmath399 , cf . . here we exhibit another limit regime for the meixner ensembles , which leads to the discrete hermite ensemble @xmath146 defined in section [ sect3.1 ] . [ th:6.3 ] fix @xmath8 , let @xmath400 , and let the parameters @xmath401 and @xmath387 vary together with @xmath91 in such a way that @xmath402 in this limit regime , @xmath403 . we assume that @xmath401 and @xmath387 depend on @xmath91 as indicated above , and we set in @xmath404 it follows that , as @xmath405 , @xmath406 and the limit values coincide with the entries of the jacobi matrix corresponding to @xmath146 , cf . . the _ krawtchouk polynomials _ are orthogonal polynomials on @xmath258 with the weight function @xmath407 where @xmath408 is a parameter , see @xcite . we denote by @xmath409 the corresponding @xmath91-particle ensemble , where we assume that @xmath338 ( see section [ sc : orth - poly - ens ] ) . [ th:6.4 ] fix @xmath8 , let @xmath400 , and let the parameters @xmath410 and @xmath259 vary together with @xmath91 in such a way that @xmath411 in this limit regime , @xmath412 . note that the restriction @xmath338 is not violated : indeed , @xmath413 and @xmath408 , so that @xmath259 grows faster than @xmath91 . we apply the algorithm described in section [ sect4.3 ] with the modifications indicated in section [ sect4.4 ] , and we use the above expression for the weight function and the following expressions taken from ( * ? ? ? * ( 1.10.5 ) ) ) : @xmath414 where both @xmath202 and @xmath151 range over @xmath258 . from these data we compute the entries of the jacobi matrix @xmath285 according to : @xmath415 let us take @xmath416 . then , as @xmath251 , @xmath417 and the limit values coincide with the entries @xmath262 , @xmath260 of the jacobi matrix corresponding to @xmath146 , see . note a similarity with the above computation for the meixner polynomials . this is not surprising because of a well - known relation between the meixner and krawtchouk polynomials , see the last formula in @xcite . the _ hahn polynomials _ are orthogonal polynomials on @xmath258 with the weight function @xmath418 depending on parameters @xmath181 and @xmath182 , see @xcite ( warning : in @xcite , our triple of parameters @xmath419 is denoted as @xmath420 ) . we denote by @xmath421 the corresponding @xmath91-particle ensemble , where we assume that @xmath338 . fix @xmath422 , @xmath181 , and @xmath182 . let @xmath400 and let the parameter @xmath259 vary together with @xmath91 in such a way that @xmath423 . in this limit regime , @xmath424 . we apply the algorithm described in section [ sect4.3 ] with the modifications indicated in section [ sect4.4 ] , and we use the above expression for the weight function and the following expressions taken from ( * ? ? ? * ( 1.5.5 ) ) : @xmath425 where both @xmath202 and @xmath151 range over @xmath258 . note that @xmath426 , as required . from these data we compute the entries of the jacobi matrix @xmath285 according to : @xmath427 let us take @xmath428 . then , as @xmath251 , @xmath429 and the limit values coincide with the entries @xmath262 , @xmath260 of the jacobi matrix corresponding to @xmath430 , cf . . the _ racah polynomials _ are at the top of the askey scheme , see @xcite . they depend on a quadruple of parameters @xmath431 , whose range splits into several pieces . for our purposes it suffices to choose one of them ; namely , in what follows we assume that @xmath432 where @xmath259 is a positive integer . we regard the racah polynomials as functions of the variable @xmath202 ranging over the set @xmath258 , but it should be noted that they are actually orthogonal polynomials on the _ quadratic grid _ @xmath433 thus , the @xmath151th racah polynomial is a polynomial of degree @xmath151 with respect to the variable @xmath434 , and not the variable @xmath202 , as before . however , this does not affect our algorithm , because the map @xmath435 defines a bijection between @xmath258 and @xmath436 . as in the case of krawtchouk or hahn polynomials , there are @xmath437 racah polynomials , they are linearly independent as functions on @xmath258 , and their linear span is precisely the space of all functions on this set . the weight function is given by @xmath438 see ( * ? ? ? * ( 1.2.2 ) ) ; here @xmath439 is the standard notation for the pochhammer symbol . the corresponding @xmath91-particle orthogonal polynomial ensemble is denoted by @xmath440 , where @xmath338 . recall that @xmath441 and @xmath259 are related by @xmath442 ; below we alternately use one of these two parameters . fix the parameters @xmath181 , @xmath182 , and @xmath443 , and consider the following limit regime : @xmath444 in this regime , @xmath445 we use the following formulas ( see ( * ? ? ? * ( 1.2.5 ) ) ) : @xmath446 we claim that @xmath447 for @xmath448 . indeed , due to the assumptions on the parameters , in the expression for @xmath449 , all the factors are strictly positive for @xmath450 . next , if @xmath451 , then the factor @xmath452 in the numerator is cancelled with the factor @xmath453 in the denominator , and the remaining factors are again stirctly positive . likewise , @xmath454 for @xmath455 , because all factors are strictly positive . we also have @xmath456 due to the factors @xmath457 and @xmath202 , respectively . finally , we have @xmath426 . thus , all necessary conditions that are required in our algorithm are satisfied . let @xmath458 and @xmath459 be defined according to , and set @xmath460 . then , in the limit regime specified in the formulation of the theorem , @xmath461 and @xmath462 \(i ) from the expression for the weight function we obtain ( recall that @xmath463 ) @xmath464 as @xmath465 , the first quantity in the curly brackets tends to 1 and hence @xmath466 next , from the expression for @xmath449 we obtain @xmath467 since @xmath460 , this implies the first formula . \(ii ) using the relation @xmath442 we obtain @xmath468 therefore , @xmath469 one readily checks that the right - hand side is equal to @xmath470 next , @xmath471 adding up the two expressions gives the second formula . the limit values computed in the lemma coincide with the entries of the jacobi matrix associated with @xmath193 , see and . this completes the proof of the theorem . here we find conditions under which the discrete laguerre ensemble @xmath472 converges to the discrete hermite ensemble @xmath341 . recall that both parameters of the discrete laguerre ( here denoted as @xmath473 , @xmath401 ) should be positive reals . the setting of the problem differs from that of the preceding sections , because this is not a large - n limit transition . but the method is the same , with obvious simplifications . [ thm6.lh ] we fix @xmath8 and consider the limit regime for @xmath472 in which @xmath474 and @xmath475 varies in such a way that @xmath476 then @xmath477 . recall that @xmath478 , so that we can equally well achieve the convergence @xmath479 . let @xmath480 and @xmath253 denote the jacobi matrices associated with @xmath472 and @xmath341 , respectively . it suffices to show that @xmath481 with an appropriate choice of quantities @xmath482 depending on @xmath473 and @xmath401 . recall ( see and ) that @xmath483 it follows that the desired limits hold provided we set @xmath484 . consider the airy differential operator @xmath485 where the variable @xmath486 ranges over @xmath3 . this differential operator is essentially self - adjoint on @xmath487 , so that its closure is a self - adjoint operator . the latter operator will be denoted by @xmath239 . it has simple , purely continuous spectrum filling the whole axis @xmath3 . the spectral projection @xmath488_+$ ] gives rise to a determinantal point process , called the _ airy ensemble_. we will denote it as @xmath489 . below we discuss limit transitions from discrete ensembles on @xmath23 to the airy ensemble . this inevitably assumes a scaling of the grid @xmath23 . we will consider the scaling of the form @xmath490 where @xmath266 is the discrete variable , @xmath491 is the continuous variable , and @xmath492 and @xmath493 are large scaling parameters such that @xmath494 . according to this , given a point process @xmath54 on @xmath23 , we denote by @xmath495 the transformed pre - limit process living on the grid @xmath496\subset{\mathbb r}\ ] ] with mesh @xmath497 . as @xmath498 and @xmath499 get large , the mesh of the grid tends to zero and its right end @xmath500 shifts to @xmath501 , so in the limit the grid fills the whole real line . the result can be stated for the two variants , @xmath146 and @xmath147 , simultaneously . we assume that @xmath136 tends to @xmath501 or @xmath502 , respectively , and we set @xmath503 we present a heuristic argument showing that in this limit regime , @xmath504 because of the symmetry @xmath505 , it suffices to examine the case of @xmath146 . let us denote by @xmath506 the difference operator on @xmath23 that is defined by the jacobi matrix associated with @xmath146 . the action of @xmath506 on a test function @xmath87 is given by @xmath507 recall we have the freedom of multiplying the pre - limit operator by a positive constant depending on our large parameter @xmath136 . [ prop7.a ] fix an arbitrary smooth function @xmath508 on @xmath3 and assign to it a function @xmath297 on @xmath23 by setting @xmath509 let @xmath498 and @xmath499 depend on @xmath136 as in , and set @xmath510 as @xmath511 , we have @xmath512 we expand @xmath513 and substitute this in . then we obtain @xmath514 next , we substitute @xmath515 and use the expansion @xmath516 this gives us : @xmath294 the coefficient of @xmath508 is equal to @xmath517 because , by the very definition of @xmath498 and @xmath499 , @xmath518 @xmath294 the coefficient of @xmath519 is @xmath520 @xmath294 the coefficient of @xmath521 is equal to @xmath522 hence , after multiplication by @xmath523 we obtain in the limit @xmath524 as desired . we state the result for the plus and minus versions of the discrete laguerre simultaneously . the initial large parameters are @xmath422 and @xmath170 . we assume that @xmath136 and @xmath401 have the same order of growth , but @xmath525 in the case of @xmath526 , and @xmath527 in the case of @xmath399 . more precisely , there exists @xmath528 such that @xmath529 keeping this in mind , the limit regime is specified by setting @xmath530 this entails , in particular , that @xmath531 and @xmath532 . we are going to show that in this limit regime , @xmath533 the argument is heuristic , the exact assertion is proposition [ prop7.b ] below . let us denote by @xmath534 the difference operator on @xmath23 defined by the jacobi matrix associated with @xmath176 . the action of @xmath534 on a test function @xmath87 on @xmath23 is given by @xmath535 [ prop7.b][pr:7.2 ] fix an arbitrary smooth function @xmath508 on @xmath3 and assign to it a function @xmath297 on @xmath23 by setting @xmath509 let @xmath498 and @xmath499 depend on @xmath136 as in , and set @xmath536 as @xmath511 , we have @xmath537 we need a lemma . the parameters @xmath498 and @xmath499 as defined above satisfy the system of equations @xmath538 here and below the upper sign is taken for @xmath526 and the lower sign is taken for @xmath399 . the origin of this system will be clear from the proof of proposition [ prop7.b ] . let us check . we have @xmath539 whence @xmath540 it follows that @xmath541 as desired . let us check . this equation is equivalent to @xmath542 while the definition of @xmath499 says that @xmath543 using it is readily seen that the both expressions are the same . we argue as in the proof of proposition [ prop7.a ] . keeping the first three terms of the taylor expansion of @xmath544 we write the right - hand side of in the form @xmath545 now we substitute @xmath515 and compute the coefficients of @xmath508 , @xmath519 , and @xmath521 up to negligible terms . @xmath294 the coefficient of @xmath508 is equal to @xmath546 the expression in the curly brackets is equal to @xmath547 hence , the whole expression can be written as @xmath548-\tau\left(\frac{2{\sigma}+{\beta}\pm2\sqrt{{\sigma}({\sigma}+{\beta})}}{\sqrt{{\sigma}({\sigma}+{\beta})}}\right)v+\ldots\,.\ ] ] by virtue of , the expression in the square brackets vanishes . we conclude that the expression for the coefficient of @xmath508 has the form @xmath549 @xmath294 as in the proof of proposition [ prop7.a ] , we verify that the coefficient of @xmath519 is negligable . @xmath294 the coefficient of @xmath521 has the form @xmath550 finally , by virtue of and the definition of @xmath388 , we see that dividing by @xmath388 and passing to the limit we obtain the airy differential operator . our exposition in this section largely follows @xcite , see also @xcite . we only consider a homogeneous version of the model , for the fully general inhomogeneous version see @xcite . consider an ensemble @xmath551 of infinite oriented up - right paths drawn in the first quadrant @xmath552 of the square lattice , with all the paths starting from a left - to - right arrow entering each of the points @xmath553 on the left boundary ( no path enters through the bottom boundary ) . assume that no two paths share any horizontal piece ( but common vertices and vertical pieces are allowed ) . see figure [ fig : intro ] . [ scale=.7,thick ] in 1, ... ,6 ( -1,5.5)++(0,-5 ) ; at ( -1,.5 ) @xmath554 ; in 1,2,3,4,5 ( 0,)++(5.5,0 ) ; at ( -1 , ) @xmath554 ; ( -1,)++(.5,0 ) ; ( -1,5)++(1 - 3,0)++(,)++(0,1- ) ; ( -1,4)++(1 - 2,0)++(,)++(0,1 - 2)++(,2)++(0,1- ) ; ( -1,3)++(1-,0)++(,)++(0,1 - 2)++(,2)++(0,1 - 2)++(,2)++(0,1- ) ; ( -1,2)++(1,0)++(0,1-)++(,)++(1-,0 ) ++(0,1)++(2-,0)++(,)++(0,2- ) ; ( -1,1)++(3,0)++(0,2)++(1,0)++(0,1-)++(,)++(1-,0 ) ++(0,2 ) ; ( -.5,4.5)++(4,-4 ) node[above , anchor = west , yshift=16,xshift=-9 ] @xmath555 ; define a probability measure on the set of such path ensembles in the following markovian way . for any @xmath556 , assume that we already have a probability distribution on the intersection @xmath557 of @xmath551 with the triangle @xmath558 . we are going to increase @xmath151 by 1 . for each point @xmath559 on the upper boundary of @xmath560 , i.e. , for @xmath561 , every @xmath557 supplies us with two inputs : ( 1 ) the number of paths that enter @xmath559 from the bottom denote it by @xmath562 ; ( 2 ) the number of paths that enter @xmath559 from the left denote it @xmath563 . now choose , independently for all @xmath559 on the upper boundary of @xmath560 , the number of paths @xmath564 that leave @xmath559 in the upward direction , and the number of paths @xmath565 that leave @xmath559 in the rightward direction , using the probability distribution with weights of the transitions @xmath566 given by @xmath567 assuming that all above expressions are nonnegative , this procedure defines a probability measure on the set of all @xmath551 s because we always have @xmath568 , and @xmath569 vanishes unless @xmath570 . in what follows we will use the following assumptions on the parameters : @xmath571 if @xmath572 then @xmath573 , which means that no two paths can share the same vertical segment ; this is the case of the stochastic six vertex model introduced in @xcite and recently studied in @xcite , see also @xcite . in this case we impose a stronger condition @xmath574 . it is easy to see that our assumptions on the parameters guarantee the positivity of the weights . each path ensemble @xmath551 can be encoded by a _ height function _ @xmath575 , that assigns to each vertex @xmath576 the number @xmath577 of paths in @xmath551 that pass through or to the right of this vertex . the value @xmath578 clearly depends only on the behavior of the paths in the @xmath579 rectangle formed by @xmath580 first columns and @xmath91 first rows . our notation for partitions , symmetric functions , etc . below is mostly the standard one used in @xcite . let @xmath581 be the set of all partitions and @xmath582 be the algebra of symmetric functions in indeterminates @xmath583 . a particularly nice linear basis of @xmath582 is formed by the schur symmetric functions @xmath584 indexed by @xmath585 . the schur symmetric polynomials are defined as the restriction of @xmath586 s to a finite number of variables @xmath587 and written as @xmath588 . if @xmath589 then @xmath590 . for any two sets of indeterminates @xmath591 and @xmath592 define @xmath593 where @xmath594 stand for a partition whose young diagram is conjugate to that of @xmath595 . the well - known cauchy and dual cauchy identities for the schur functions read @xmath596 the branching rule for the schur symmetric functions ( * ? ? ? * section i.5 ) immediately implies that @xmath597 for any @xmath598 with @xmath599 . the following definition first appeared in @xcite . [ macprocessdef ] for any two sets of nonnegative variables @xmath600 and @xmath601 such that @xmath602 ( or @xmath603 ) define the corresponding schur measure as the probability measure on @xmath581 that assigns to a partition @xmath604 the weight @xmath605 let us now establish a correspondence between certain instances of the schur measures @xmath606 , @xmath607 and classical orthogonal polynomial ensembles . for any @xmath608 , denote by @xmath609 the set of partitions with no more than @xmath227 nonzero parts : @xmath610 and by @xmath611 the subset of @xmath609 of partitions with the largest part @xmath612 : @xmath613 the importance of these sets to us is explained by the fact that @xmath614 is supported by @xmath609 if the set of @xmath202-variables , or the set of @xmath615-variables , contains no more than @xmath227 nonzero elements . similarly , @xmath616 is supported by @xmath611 if there are no more than @xmath227 nonzero @xmath202 s and no more than @xmath151 nonzero @xmath615 s . this is simply because the schur function @xmath584 vanishes if @xmath617 is greater than the number of nonzero @xmath618 s . [ pr : schur - meixner ] fix @xmath619 and @xmath620 with @xmath621 . consider the schur measure @xmath622 ; it is supported by @xmath623 . the pushforward of this measure under @xmath624 coincides with the @xmath625-point meixner ensemble @xmath626 . a straightforward computation based on weyl s dimension formula @xmath627 [ pr : schur - krawtchouk ] fix @xmath619 and @xmath620 . the the pushforward of the schur measure @xmath628 ( which is supported by @xmath629 ) under the map @xmath630 coincides with the @xmath631-point krawtchouk ensemble @xmath632 . another straightforward computation with . in @xcite it was shown that averages of certain observables on the higher spin six vertex model are equal to other averages over the schur ( more generally , macdonald ) measures . let us restate here the results we will need . [ pr : match ] ( * ? ? ? * example 4.3 , corollary 4.4 ) take any @xmath633 , @xmath634 , and @xmath635 . then for @xmath572 and any @xmath636 we have @xmath637 where in the right - hand side we assume that @xmath638 for @xmath639 , and the schur measure expectation is over @xmath640 . on the other hand , for @xmath641 we have @xmath642 where the expectation in the right - hand side is over @xmath643 . let us now use propositions [ pr : schur - meixner ] and [ pr : schur - krawtchouk ] to rewrite the right - hand sides of the above identities in terms of the orthogonal polynomial ensembles . for @xmath644 , @xmath645 , and a point configuration @xmath646 , define @xmath647 note that this is a bounded continuous function on @xmath648 . if @xmath52 is distributed according to a random point process @xmath649 on @xmath650 , we define @xmath651 [ cr:6v - poly ] in the notation of proposition [ pr : match ] , in the case @xmath572 , assuming @xmath652 we have @xmath653 and in the case @xmath641 we have @xmath654 where , as before , @xmath655 stands for the particle / hole involution of @xmath649 viewed as a point process on @xmath23 . this is a direct substitution of propositions [ pr : schur - meixner ] and [ pr : schur - krawtchouk ] into proposition [ pr : match ] , where one uses the obvious fact that for a partition @xmath656 , @xmath657 . [ rm : meixner - shift ] the relation can be easily extended to the case @xmath658 as well . the extra caveat is that now @xmath659 , the corresponding schur measure is supported by @xmath660 , and @xmath661 thus , we obtain @xmath662 where we used the notation @xmath663 to denote the random point process obtained from @xmath649 by the deterministic shift of all points of the random point configuration @xmath52 by @xmath664 : @xmath665 . in the case @xmath666 the two relations and coincide , as it should be . the need for the deterministic shift in can also be explained by the fact that @xmath667 , and thus we can not possibly have the equality of the form unless all the particles of the point process in the right - hand side are located strictly to the right of @xmath668 . it is not hard to show , see @xcite , that @xmath669 thus , can be rewritten as @xmath670 in this section we collect a few simple probabilistic statements that we will need later on . we will only deal with random variables that take values in @xmath671 . for any such random variable @xmath387 and any @xmath633 we define @xmath672 this is obviously a meromorphic function of @xmath673 with possible poles in @xmath674 . using the @xmath675-binomial theorem and the fact that @xmath676 , for @xmath677 with @xmath678 , we have @xmath679 as @xmath680 , it is natural to view @xmath681 as a @xmath675-analog of the laplace transform . the inverse transform is provided by the following statement : ( * ? ? ? * proposition 3.1.1 ) one may recover the probability distribution of a @xmath671-valued random variable @xmath387 from @xmath682 as follows : @xmath683 where @xmath684 is any positively oriented contour which encircles the poles @xmath685 for @xmath686 . note that one can also recover the moments of @xmath687 directly from : @xmath688 let @xmath689 be a sequence of @xmath671-valued random variables , and assume that there is a random variable variable @xmath387 such that @xmath690 in distribution . then @xmath691 as @xmath692 , uniformly in @xmath673 varying over any compact subset of the open unit disc . the convergence in distribution implies the convergence of the @xmath675-moments @xmath693 for any @xmath694 ( because @xmath695 for @xmath696 and @xmath697 ) , and this implies the convergence of the right - hand sides of uniformly over compact sets in the unit disc , where the series are term - wise majorated by @xmath698 a converse statement is also true . [ lm : ql - limit ] let @xmath699 be a sequence of @xmath671-valued random variables , and assume that the sequence of the @xmath675-laplace transforms @xmath700 converges to a function @xmath701 uniformly in @xmath673 varying over any compact subset of the open unit disc . then there exists a random variable @xmath387 such that @xmath702 as @xmath692 in distribution , and @xmath703 . the uniform convergence of holomorphic functions in a domain implies the convergence of all their derivatives at every point of the domain . on the other hand , the convergence of the derivatives at a point implies , by taylor expansion , the uniform convergence of the functions in compact subsets of any open disc centered at this point , if all the functions are holomorphic in this disc . iterating these statements implies the uniform convergence of @xmath704 in any compact subset of @xmath705 , which implies the convergence of with @xmath706 instead of @xmath387 and @xmath692 . set @xmath707 , @xmath697 . clearly , @xmath708 and @xmath709 . we thus have for any @xmath710 @xmath711 where we used for the second relation . substituting @xmath712 yields @xmath713 , which means that we can define a random variable @xmath387 by @xmath714 . finally , the above limiting relations imply , via , that @xmath715 . the asymmetric simple exclusion process , or asep for short , is a well - known interacting particle system on ( an interval of ) the one - dimensional lattice that has been extensively studied since its introduction in @xcite and @xcite . the system consists of particles occupying vertices of @xmath34 , no more than one per site , that randomly move in continuous time . the evolution is markovian , and can be informally described as follows : each particle has two exponential clocks of rates @xmath716 and @xmath717 , call them left and right . when the left clock rings , the particle moves to the left by one if the corresponding target site is unoccupied , and nothing happens if that site is occupied ( the jump is suppressed ) . similarly , if the right clock rings , the particle moves to the right by one if the corresponding site is empty . all the clocks are independent . a more detailed description , as well as a proper definition of this dynamics with infinitely many particles , can be found in @xcite . we next observe that the stochastic six vertex model as described in section [ sc:6v ] can be used to approximate the asep with a particular _ packed _ or _ step _ initial condition , when at time 0 the particles occupy all negative integers . this is explained in ( * ? ? ? * section 6.5 ) in detail and proved in @xcite , but the idea is fairly simple ; let us describe it . one reads the path ensemble @xmath718 row by row , thinking of the places , where the paths intersect a given horizontal section , as of particle locations . moving the horizontal section upward will correspond to increasing the time . further , one tunes the parameters in such a way that the probability for a path not to turn is infinitesimally small , i.e. @xmath719 and @xmath720 are both small . this leads to paths becoming staircase - like , with all the steps having height and width one . if we now measure the particle positions using a moving frame with the @xmath34-origin in each row corresponding to the points @xmath721 in the quadrant , we will observe particles accumulating at the negative integer locations as our horizontal sections move higher . finally , taking @xmath722 , @xmath723 , and scaling the height of the horizontal section as @xmath724 , @xmath725 , we see the convergence to the asep , as the appearance of vertices of the type @xmath726 means that the corresponding particle jumps to the right , and , similarly , vertices of the type @xmath727 give jumps to the left . let us now make a precise statement . for asep particle configurations , we introduce a height function @xmath728 that counts the number of particles weakly to the right of a given location . since the particle configurations are random , this is a random function . we will only consider the situation when asep particles do not accumulate at @xmath501 , so that the height function is always finite . [ pr:6v - asep ] consider the stochastic six vertex model in the quadrant with @xmath729 also consider the asep on @xmath34 with particles occupying all negative integers at time 0 , and with the jump rates @xmath730 , @xmath37 . then for any @xmath731 , we have @xmath732+x+1,\quad n=[\epsilon^{-1}t],\ ] ] where the convergence is uniform on compact sets in the open unit disc , and on the left @xmath733 stands for the height function of the six vertex model , while on the right @xmath733 stands for the height function of the asep at time @xmath31 . the @xmath675-moments of the height function of the stochastic six vertex model in the quadrant were computed in ( * ? ? ? * theorem 4.12 ) , while the @xmath675-moments of the asep height function with the step initial condition were computed in ( * ? ? ? * theorem 4.20 ) . the fact that one formula converges to the other in the limit regime of the proposition is essentially obvious , the only needed limiting relation can be found in ( * ? ? ? * corollary 10.2 ) . on the other hand , convergence of the @xmath675-moments implies the convergence of the @xmath675-laplace transforms @xmath681 via . we can now establish a connection between the asep and discrete laguerre ensembles . [ th : asepdl ] consider the asep on @xmath34 with particles occupying all negative integers at time 0 , and with the jump rates @xmath36 , @xmath37 , and let @xmath733 denote its height function , as above . then at any time moment @xmath40 , and for any @xmath734 , we have @xmath735 where @xmath736 stands for the discrete laguerre ensemble that is deterministically shifted ( to the right ) by @xmath737 . this is a combination of proposition [ pr:6v - asep ] , corollary [ cr:6v - poly ] , and theorem [ thm4.a ] . it suffices to consider the case @xmath41 ; the case @xmath738 is then obtained by the particle / hole involution that maps the asep into itself reflected around the origin ( alternatively , one can use remark [ rm : meixner - shift ] ) . first , @xmath739 is a limit of the six vertex @xmath740 , see . next , use to write @xmath740 in terms of a meixner ensemble . the weak limit of the meixner ensemble is afforded by theorem [ sc:6.2 ] ; the limiting point process is @xmath741 as in the notation of proposition [ pr:6v - asep ] we have @xmath742 the particle / hole involution obviously does not impact the weak convergence of the point processes on @xmath671 , thus we have the weak convergence of the @xmath743 ensemble from to @xmath744 . the latter process is @xmath745 , cf . section [ sect3.1 ] . finally for @xmath746 and @xmath42 , @xmath747 is a bounded continuous function in @xmath748 , thus the weak convergence of the point processes implies the convergence of its corresponding averages . theorem [ th : asepdl ] admits a limit as @xmath749 . this turns the asep into the tasep ( ` t ' is for ` totally ' ) , and the result itself turns into ( * ? ? ? * proposition 1.4 ) that was a crucial step for johansson s celebrated proof of the tracy - widom asymptotics for the tasep ( theorem 1.6 _ ibid . the limiting version of theorem [ th : asepdl ] looks as follows : ( * ? ? ? * proposition 1.4 ) consider the tasep with the unit jump rate on @xmath34 , with particles occupying all negative integers at time 0 . let @xmath733 denote its height function . then for any position @xmath750 and any time moment @xmath40 , @xmath751 equals the probability of the event that the right - most particle in the @xmath91-particle laguerre orthogonal polynomial ensemble @xmath752 is to the left of @xmath31 . for @xmath753 , the same relation holds with @xmath38 replaced by @xmath754 and @xmath752 replaced by @xmath755 . as in theorem [ th : asepdl ] , it suffices to consider @xmath41 . observe that for @xmath756 , @xmath757 , @xmath758 , we have @xmath759 and all the expressions remain bounded throughout the limit transition . taking the expectations of these limit relations with @xmath760 and @xmath761 being distributed according to @xmath745 , we see that theorem [ th : asepdl ] tells us that @xmath38 for the tasep is distributed exactly as the left - most particle of @xmath762 . employing the connection between the discrete and continuous laguerre ensembles from section [ sc : duality ] , see theorem [ thm3.a ] , finishes the proof . the goal of this section is to use theorem [ th : asepdl ] to analyze the behavior of the asep with step initial data ( particles occupy all negative integers at time 0 ) at large times . we will prove three results , one of which corresponds to the degeneration of the discrete laguerre ensemble to the discrete hermite ensemble ( as in section [ sc:6.7 ] ) , while the other two correspond to the degeneration to the airy ensemble ( as in section [ sc:7.3 ] ) . let us start with the one that corresponds to the discrete hermite ensemble ; it is simpler as it does not require any scaling of the state space @xmath671 of the determinantal point process . [ pr : asep - hermite ] consider the asep on @xmath34 with particles occupying all negative integers at time 0 , and with the jump rates @xmath36 , @xmath763 , and let @xmath733 denote its height function at time @xmath40 . then as @xmath764 and for any @xmath765 , @xmath766 converges in distribution to a @xmath671-valued random variable @xmath767 characterized by @xmath768 since the limiting values of the height function are in @xmath769 , proposition [ pr : asep - hermite ] describes the behavior of the finitely many first ( that is , right - most ) asep particles . an alternative description of this limiting regime was obtained by tracy - widom in ( * ? ? ? * theorem 2 ) ( conjectured earlier in @xcite ) in terms of fredholm determinants . matching our result to tracy - widom s one is an interesting problem , but we do not pursue it in this work . the argument is very similar to the proof of theorem [ th : asepdl ] above . first , by theorem [ thm6.lh ] , in the described limit regime the discrete laguerre ensembles of theorem [ th : asepdl ] weakly converges to the discrete hermite ensemble @xmath770 . hence , the averages of the bounded continuous function @xmath771 converge too . this shows the convergence of the right - hand side of to @xmath772 . it is straightforward to strengthen this convergence to the uniform one in @xmath673 varying over compact sets in the open unit disc . ( for example , one can first show that the particles outside a large enough subset @xmath773 can only affect the values of @xmath774 by a uniformly close to 1 multiplicative factor , and then note that there are only finitely many possible particle configurations in @xmath436 . hence , the collection of their probabilities with respect to the discrete laguerre ensembles converges uniformly to the corresponding probabilities of the discrete hermite ensemble . ) finally , lemma [ lm : ql - limit ] implies the claim . now we proceed to the convergence of the discrete laguerre ensembles to the airy ensemble . [ th : asep - tw ] consider the asep on @xmath34 with particles occupying all negative integers at time 0 , and with the jump rates @xmath36 , @xmath763 , and let @xmath733 denote its height function at time @xmath40 . assume that @xmath775 at the same rate , and @xmath776 . then the random variable @xmath777 with @xmath778 weakly converges to the gue tracy - widom distribution . the above result is equivalent to the celebrated theorem of tracy - widom ( * ? ? ? * theorem 3 ) that says that @xmath779 which in its turn generalized johansson s ( * ? ? ? * theorem 1.6 ) for @xmath780 . in the above relation , @xmath781 is the position of the @xmath227th right - most asep particle at time @xmath30 ( we assume the same step initial data ) , and @xmath782 is the gue tracy - widom distribution . the equivalence is established by noting that the event that at time @xmath31 the particle number @xmath227 is at position @xmath783 is exactly the same as the event that at time @xmath31 and we have @xmath784 , and by matching the notations ( for @xmath41 , the case @xmath738 is similar ) @xmath785 as @xmath786 , up to terms of order smaller than @xmath787 . a different proof of theorem [ th : asep - tw ] for @xmath451 was given in ( * ? ? ? * appendix d ) . the argument is similar to the proof of ( * ? ? ? * theorems 6.1 , 6.3 ) . first , we need to recall the definition of asymptotic equivalence from ( * ? ? ? * definitions 5.1 and 5.2 ) . with those definitions we can proceed to the proof . we will only consider the case @xmath41 , the case @xmath738 is analogous . using ( * ? ? ? * proposition 5.3 , example 5.5 ) we see that the family of functions @xmath788 is asymptotically equivalent to @xmath789 as @xmath764 . on the other hand , ( * ? * corollary 5.7 ) implies that the family of functions @xmath790 is asymptotically equivalent to @xmath791 , which is the negative location of the left - most particle in @xmath745 , as @xmath764 . since @xmath792 by theorem [ th : asepdl ] , we conclude that @xmath38 is asymptotically equivalent to the position of the left - most particle in @xmath745 . the latter converges in distribution , under the scaling of section [ sc:7.3 ] and proposition [ pr:7.2 ] , to the distribution of the right - most particle of the airy ensemble , which is the gue tracy - widom distribution @xcite . ( proposition [ pr:7.2 ] does not actually proof the needed convergence . however , by virtue of theorem [ thm3.a ] , this convergence is equivalent to the corresponding convergence of the right - most particle of the laguerre orthogonal polynomial ensembles to the right - most particle of the airy ensemble , which is a well - known fact that goes back to @xcite . ) matching the scaling of theorem [ th : asep - tw ] and propositon [ pr:7.2 ] completes the proof . before moving into a full description of the third asymptotic result , let us first explain how it can be foreseen . the proof of theorem [ th : asep - tw ] above essentially consists of two ingredients : ( 1 ) the convergence of the discrete laguerre ensembles to the airy ensemble ; ( 2 ) the approximation of the observable @xmath793 , @xmath794 , by a characteristic function of @xmath795 . the latter claim follows from the fact that each factor of the form @xmath796 in the infinite product that defines @xmath797 , converges to either 0 or 1 . the new limit transition will now follow from varying the parameter @xmath675 that so far remained fixed in @xmath798 . namely , we want to send @xmath675 to 1 so that , after choosing an appropriate @xmath673 , each of the factors @xmath796 would converge to a nontrivial limit in @xmath798 for any particle of the limiting airy ensemble , with different limits for different particles . since the pre - airy particles in the discrete laguerre ensembles have inter - particle distances of order @xmath799 , this means that we have to choose @xmath800 . on the other hand , the @xmath675-laplace transform @xmath801 in the limit @xmath802 should approximate the usual laplace transform of the scaled height function . all this leads to the following statement . we will only consider the case @xmath41 , the case @xmath738 is completely analogous . [ th : asep - kpz ] consider the asep on @xmath34 with particles occupying all negative integers at time 0 , and with the jump rates @xmath36 , @xmath763 , and let @xmath733 denote its height function at time @xmath40 . let @xmath803 be a small parameter , and assume that @xmath804 then the random variables @xmath805 have a weak limit as @xmath806 ; denote it by @xmath387 . the laplace transform of @xmath807 is given by @xmath808 = \mathbb{e}_{z\in airy } \prod_{z\in z } \frac{1}{1+\widehat\zeta \exp(\widehat \tau z)}\,,\qquad\qquad \widehat\zeta>0 , \quad \widehat \tau=\dfrac{(\widehat t^2-\widehat x^2)^{2/3}}{2^{4/3}\widehat t}.\ ] ] [ rm : asep - kpz ] this result ( in a slightly different form ) goes back to @xcite and @xcite . according to @xcite , the limiting random variable @xmath387 has the same distribution as @xmath809 , where @xmath810 , and @xmath811 is the hopf - cole solution of the kpz ( kardar - parisi - zhang ) stochastic partial differential equation with the so - called narrow wedge initial data . we refer to @xcite for a description of this result that is similar to the notations of the present paper . a general result that the _ weak asymmetry _ limit of the asep ( @xmath802 ) is related to solutions of the kpz equation for certain regular initial conditions goes back to @xcite . a concrete realization for the somewhat singular narrow wedge initial data , that is equivalent to theorem [ th : asep - kpz ] , is due to @xcite . a discussion on the levels of mathematical rigor of the above developments can be found in the survey @xcite . we need to take limits of the two sides of the identity @xmath812 let us first explain how the limit works and then point out the steps that would turn this explanation into an actual proof . for the right - hand side , we use , cf . section [ sc:7.3 ] and proposition [ pr:7.2 ] , with @xmath813 as in theorem [ th : asep - tw ] , @xmath814 choosing @xmath673 so that @xmath815 and observing that @xmath816 for finite @xmath817 , we conclude that the right - hand side of should converge to that of . on the other hand , if we denote @xmath818 , then @xmath819 and the last expression obviously converges to @xmath820 as @xmath821 ( we used the q - binomial theorem to turn the product into the sum ) . to convert this computation into a proof , we follow the following steps . _ we start by restricting the infinite products in the right hand - sides to include only the particles that lie ( directly or after the discrete laguerre to airy scaling ) in a subset of the form @xmath822 of the airy ensemble s state space @xmath823 . we claim that such a modification changes the observables by a multiplicative constant that is uniformly close to 1 with high probability , and that remains bounded almost surely , when we choose @xmath259 close enough to @xmath502 . in the airy case , this amounts to investigating @xmath824 first , this is obviously bounded with probability 1 . next , the closeness of this product to 1 can be controlled by the smallness of the additive statistic @xmath825 . we have @xmath826 both these quantities can be made as small as one wishes by a suitable choice of @xmath259 . hence , by chebyshev s inequality one can make @xmath827 arbitrarily close to 0 with arbitrarily high probability . the argument for the discrete laguerre ensemble is very similar . _ in the restricted range of @xmath828 , we have the trace - class convergence of the discrete laguerre kernel to the airy one . this implies the convergence of the probabilities of observing a fixed number of particles in the corresponding ranges of the two ensembles , as well as the weak convergence of the distributions of the positions of those particles . hence , the right - hand side of converges to that of . the q - laplace transform - like observables in the left - hand side of uniformly approximate the exponential observable @xmath829 as @xmath806 , with @xmath830 varying over @xmath831 . by the weak compactness of the space of positive measures , the distributions of @xmath832 must have limiting points in the space of positive measures on @xmath831 of total mass @xmath833 ; let @xmath834 be one such limiting point . then the left - hand sides of converge to the laplace transforms @xmath835 of @xmath834 . since the laplace transforms determine such measures uniquely , @xmath834 is actually unique . furthermore , since we already know the limit of the right - hand side of , and it has the property of approaching 1 as @xmath836 , we conclude that the total integral of @xmath834 is exactly 1 , i.e. @xmath834 is a _ bona fide _ probability measure on @xmath831 . denoting by @xmath387 a random variable with distribution @xmath834 yields the desired claim . in this section we study asymptotic regimes of the stochastic six vertex model in a quadrant as defined in section [ sc:6v ] , that are parallel to the asymptotic behavior of the asep from the previous section . we start with the simplest regime that is analogous to proposition [ pr : asep - hermite ] . as in section [ sect11 ] , let us denote by @xmath837 a @xmath671-valued random variables determined by ( cf . ) @xmath838 the argument is exactly the same as in the proof of proposition [ pr:6v - dhermite ] , with the use of corollary [ cr:6v - poly ] and the convergences of the meixner and krawtchouk ensembles to the discrete hermite ensemble , see sections [ sc:6.3]-[sc:6.4 ] above . then change of the sign in front of @xmath851 , as compared to and , is explained by the fact that there is the particle - hole involution @xmath852 in , , , and @xmath853 . proposition [ pr:6v - dhermite ] describes the behavior of the stochastic six - vertex model near the linear boundaries that separate ` liquid ' and ` frozen ' zones . this can be seen from theorems 6.1 and 6.3 of @xcite that describe the asymptotic behavior in the liquid zone ( theorem 6.1 there goes back to @xcite ) . in fact , those two theorems are exact analogs of theorem [ th : asep - tw ] above . their proofs relied on convergences of the meixner and krawtchouk ensembles to the airy ensemble ( in @xcite the exposition is in the language of the schur measures , the match to the meixner and krawtchouk ensembles is made via propositions [ pr : schur - meixner ] and [ pr : schur - krawtchouk ] above ) . very similarly to what was done with the asep in the previous section , one can modify the limiting argument by taking @xmath802 in such a way that all points of the airy ensemble yield nontrivial factors in the averaged observable . this leads to the following statement . [ th:6v - kpz ] consider the ( higher spin stochastic ) six vertex model in a quadrant , as described in section [ sc:6v ] . let @xmath803 be a small parameter , and assume that @xmath854 with @xmath576 in the liquid zone , i.e. @xmath855 for @xmath572 , and @xmath856 for @xmath641 . then for the following choice of the normalizing constants @xmath857 the random variables @xmath858 have a weak limit as @xmath806 ; denoting it by @xmath387 , the laplace transform of @xmath807 is given by . the proof of this statement is very similar to the proof of theorem [ th : asep - kpz ] , and we omit it . it rides on the heels of theorems 6.1 and 6.3 of @xcite in the same way as the proof of theorem [ th : asep - kpz ] does with theorem [ th : asep - tw ] instead . in particular , the normalizing constants @xmath859 and @xmath860 come from @xcite . g. amir , i. corwin , j. quastel , probability distribution of the free energy of the continuum directed random polymer in 1 + 1 dimensions , communications on pure and applied mathematics 64 ( 2011 ) , 466537 . arxiv:1003.0443 . a. borodin and v. gorin , lectures on integrable probability . in : probability and statistical physics in st . proceedings of symposia in pure mathematics vol . 91 , pp . 162221 . soc . , 2016 . arxiv:1212.3351 . j. breuer and m. duits , central limit theorems for biorthogonal ensembles and asymptotics of recurrence coefficients . , to appear ( published electronically on january 28 , 2016 ) ; arxiv:1309.6224 . p. deift , integrable operators . in : differential operators and spectral theory : m. sh . 70th anniversary collection ( v. buslaev , m. solomyak , d. yafaev , eds . ) . american mathematical society translations , ser . 2 , vol . 189 . providence , r.i . : ams , 1999 . r. koekoek and r. f. swarttouw , the askey - scheme of hypergeometric orthogonal polynomials and its @xmath675-analogue , report no . 98 - 17 ( 1998 ) , delft university of technology , http://homepage.tudelft.nl/11r49/askey/. t. tao , the dyson and airy kernels of gue via semiclassical analysis ( tao s blog , october 23 , 2010 ) , https://terrytao.wordpress.com/2010/10/23/the-dyson-and-airy-kernels-of-gue-via-semiclassical-analysis/ .	we introduce a family of discrete determinantal point processes related to orthogonal polynomials on the real line , with correlation kernels defined via spectral projections for the associated jacobi matrices . for classical weights , we show how such ensembles arise as limits of various hypergeometric orthogonal polynomials ensembles . we then prove that the q - laplace transform of the height function of the asep with step initial condition is equal to the expectation of a simple multiplicative functional on a discrete laguerre ensemble a member of the new family . this allows us to obtain the large time asymptotics of the asep in three limit regimes : ( a ) for finitely many rightmost particles ; ( b ) gue tracy - widom asymptotics of the height function ; ( c ) kpz asymptotics of the height function for the asep with weak asymmetry . we also give similar results for two instances of the stochastic six vertex model in a quadrant . the proofs are based on limit transitions for the corresponding determinantal point processes .
one success of network science has been to identify that some complex systems can be simplified by considering just the topology of the pairwise interactions between their parts . abstracting a complex system as a graph can bring physical insights and predictive power . yet these graphs can still be very complicated . network geometry is an approach which further abstracts the system by modelling the nodes of the network as points in a geometric space . most existing approaches use riemannian spaces , the simplest example of which is euclidean space . random geometric graphs ( rgg ) are graphs embedded in euclidean space @xcite . recently there has been much interest in geometric approaches to the study of networks in non - euclidean spaces @xcite . embedding in hyperbolic spaces can yield scale free , clustered networks with community structure illustrating the remarkable power that geometric approaches have to recover complex network properties . a well established geometric approach to data analysis is _ multidimensional scaling _ ( mds ) , a technique to give data expressed as distances or similarities a spatial representation @xcite . in most mds analysis , the space used for that spatial representation has been euclidean , and the technique , as usually described , requires a riemannian manifold , where the triangle inequality is maintained . mds has been used in network science , to fit models of rggs to networks from real data , for example , from protein interactions @xcite . normally , the mds algorithm takes as an argument pairwise distances between objects , so when applying it to simple networks , where only binary pairwise relations exist , these distances have to be inferred from the network structure . in the simplest euclidean case ( as in @xcite ) , the shortest path on the network is used as an estimate for the distances between vertices , from which mds is used to calculate coordinates . once these coordinates have been calculated , a new rgg can be built from them , and if it is similar to the original graph , the initial geometric assumption was a good one . in this paper we will consider networks where each node is associated with a particular time and directed edges between nodes represent causal relations . such a network forms a * directed acyclic graph * ( dag)@xcite . instead of embedding a network in geometric space alone , the causal ordering of the nodes in a dag suggests that an embedding in space and time is needed . the causal structure of such a network has the same constraints as the causal structure of spacetime as used in special and general relativity @xcite.this suggests that the geometries used in relativity , which are pseudo - riemannian are the appropriate ones to use because of the special properties of a time dimension . in particular , we will consider _ lorentzian spacetimes _ , a special case of pseudo - riemannian manifolds in which there is one time dimension with some number of spatial dimensions . euclidean space , being flat and isotropic , is the simplest riemannian manifold . analogously , flat isotropic spacetime is minkowski spacetime , the simplest lorentzian manifold . in this paper , we first generalise classical mds to allow converting distances into coordinates for pseudo - riemannian manifolds . we then show how this allows embedding of dags in minkowski spacetime ^{1/\lambda}$ ] . this is _ not _ what we mean by a minkowski metric in this paper , rather we mean the minkowski spacetime of special relativity@xcite . ] , and that the coordinates of geometric graphs in minkowski spacetime can be successfully recovered . we then illustrate this technique by finding coordinates for some citation networks , which naturally form dags and suggest applications such as paper recommendation . we will begin by briefly summarising the details of standard mds in euclidean space . suppose we have @xmath0 objects , and are given the squared euclidean distance , @xmath1 between each pair . we wish to find the co - ordinates of the objects , which will be @xmath2 dimensional vectors , @xmath3 for each object @xmath4 , such that they fit the constraint that @xmath5 . the classical mds algorithm solves this problem by using this @xmath6 matrix of square distances , @xmath7 , and then constructing the double centred matrix @xmath8 where @xmath9 . it can then be shown ( details are available in @xcite ) that @xmath10 where @xmath11 is an @xmath12 matrix of co - ordinate vectors @xmath13 which satisfy the constraint of recovering the original distances , and with the centre of mass of the coordinates at the origin . @xmath14 is guaranteed to be semi - positive definite ( i.e. it has no negative eigenvalues ) . so we can then find ( up to a distance - preserving symmetry transformation ) the coordinates in @xmath11 by decomposing @xmath14 into @xmath15 where @xmath16 is a diagonal matrix of the eigenvalues of @xmath14 , and @xmath17 a matrix of its eigenvectors . a solution is given by @xmath18 this process yields coordinates in @xmath0 dimensions , but only @xmath2 of the eigenvalues will be non - zero . it is possible retrieve coordinates in fewer dimensions , by using only the largest @xmath19 eigenvalues and their corresponding eigenvectors . the larger eigenvalues correspond to principle components , meaning that using them as the coordinates minimises the square difference between the original distances we started with , and the ones calculated from these inferred coordinates . these coordinates are in this sense the most accurate @xmath19 dimensional representation of the original data . minkowski spacetime is a combination of a @xmath20-dimensional euclidean space , and one time dimension forming a @xmath21-dimensional manifold . a point @xmath4 in this space , has coordinates @xmath3 consisting of a time coordinate , @xmath22 , and spatial coordinates @xmath23 , with @xmath24 . the minkowski separation between two such spacetime points @xmath4 and @xmath25 is given by @xmath26 pairs of points can then be classified into three types : for a positive separation the pair is spacelike separated , for a negative separation the pair is timelike separated , while exactly zero separation means the pair is lightlike separated . we can now ask the same question that classical euclidean mds poses : given pairwise separations @xmath27 , for points in this space , can we recover coordinates which respect these separations ? proceeding with the classical euclidean algorithm we can construct the double centred matrix @xmath14 as before using @xmath28 . however we now encounter a problem when decomposing @xmath14 . previously the eigenvalues were guaranteed to be non - negative , but now we find one negative eigenvalue corresponding to the time dimension s negative sign in equation [ m_metric ] . since we need to take the square root of these eigenvalues , and we want real coordinates this is a problem . it turns out that the changes required to the classical mds algorithm are remarkably simple ( details are given in appendix a ) . instead of looking for a matrix of coordinates @xmath11 such that @xmath29 , we now search solutions to @xmath30 where @xmath31 is matrix representing the metric of the embedding space . for traditional mds with its euclidean space @xmath31 is just the identity matrix so this factor drops out from the analysis . however for dags where nodes are also associated with a time coordinate , we choose @xmath31 to represent the minkowski metric . in our conventions , this is a diagonal matrix with @xmath32 in the first column and @xmath33 in the others . we again decompose @xmath14 and now need solutions to @xmath34 the difference to traditional mds is that the @xmath32 present in @xmath31 changes the sign of the one negative eigenvalue in @xmath35 . this allows us to take the square root of @xmath16 as we do in classical euclidean mds . to use lorentzian mds on networks , we require a method of estimating the separations between every pair of nodes from the network structure . we will do this using ideas from causal set theory . a causal set is a locally finite partially ordered set . in the causal set approach to quantum gravity , the underlying structure of spacetime is postulated to be a causal set . spacetime is seen as discrete at the planck scale , and the continuous spacetime we perceive emerges at larger length scales.@xcite . the only structure present are elements of the set and the relation , @xmath36 between pairs of elements . in the correspondence between the discrete causal set , and the continuous spacetime that emerges , the relation @xmath36 corresponds to timelike separation where @xmath37 corresponds to @xmath38 being in the past of @xmath39 , and spacelike separated pairs are not related . causal sets give us a natural way of discretising spacetime with related elements , and we can use them in our approach because they have the same structure as a dag . in physics a timelike separated pair are allowed to be causally related , as information could pass from one to the other , from past to future . however a spacelike separated pair can never be causally related as light could not reach one event from the other . this is why we will associate timelike separation in the continuous spacetime with a link representing a causal relation in the network . we can now construct the equivalent of an rgg in minkowski space . we begin by assigning coordinates in @xmath40 to @xmath0 points , by sampling uniformly at random from @xmath41^d$ ] . we will denote the coordinates of point @xmath4 by @xmath42 , @xmath43 , where @xmath44 is the time coordinate . we then construct the causal set graph , @xmath45 by placing an edge from @xmath4 to @xmath25 if @xmath46 and directing that edge from the past to the future . the fact that edges in the graph now represent causal relations illustrates why the graph is necessarily a dag , as closed causal loops are forbidden . figure [ fig : eigenval ] shows the distribution of the eigenvalues @xmath16 for causal sets and random dags . , @xmath47 , and @xmath48 dimensions , and for a random dag . in all cases , one large negative eigenvalue is seen , corresponding to the one timelike dimension . for the causal sets , this timelike dimension is the one time dimension in the minkowski spacetime they are embedded in . for the random dag , this corresponds to the time ordering which could be created as a consequence of the acyclic property of this graph . we observe @xmath20 large positive eigenvalues corresponding to the spatial dimensions in each causal set , illustrating that the coarse graining of the causal set does not prevent the mds algorithm from successfully identifying the principle components of the space . for a given number of points , higher dimensionality will mean fewer relations between the causal set s elements and since these relations are the information used by the mds algorithm its ability to cleanly pick out @xmath20 dimensions diminishes as @xmath20 increases . however for the random dag there is no clear separation of large eigenvalues suggesting there is no natural dimension to an embedding minkowski spacetime.,scaledwidth=100.0% ] to use mds on a network we must estimate the separation matrix @xmath7 using the network s structure . for euclidean mds , the separation is always a non - negative number and the shortest path between nodes in the graph is a natural and effective estimator for the distance . however in minkowski spacetime , the separation of points is not always positive , so the the number of steps along some path is not going to be measure of all lorenztian separations . the solution is to estimate spacelike and timelike separations separately when studying a dag . suppose we have two connected nodes in the graph , meaning they are timelike separated . it was conjectured in @xcite and later shown in @xcite that for timelike separated points @xmath4 and @xmath25 in @xmath45 the length of _ longest path _ @xmath49 between two connected nodes , say @xmath4 and @xmath25 , is proportional their timelike separation ( in the limit of long separations , and where the longest path must respect the edge s direction ) . so in this case we set @xmath50 . finding the distance between spacelike pairs is more challenging and to our knowledge there is no solution as easily calculated as the longest path is for timelike pairs . good approximations are known however , and we will use a very simple one , described in @xcite as ` naive spatial distance ' . suppose we have two disconnected points @xmath4 and @xmath25 in the @xmath51 meaning they are spacelike separated . we then look for a pair of nodes , @xmath52 and @xmath53 , where @xmath52 is in the future of both nodes @xmath4 and @xmath25 while @xmath53 is in the past of both @xmath4 and @xmath25 and @xmath25 equal to some maximal distance which is a parameter of the algorithm . in the examples shown here , we used the length of the longest path in the graph as this parameter . ] . we then chose nodes @xmath52 and @xmath53 so as to minimise the length of the longest path between @xmath52 and @xmath53 ( which are necessarily connected , via paths through @xmath38 and @xmath39 ) . the timelike separation between @xmath52 and @xmath53 is then used as an estimate for the spacelike separation between @xmath4 and @xmath25 . figure [ fig : causal_set ] shows an example of how we estimate timelike and spacelike separations . this estimate is simple and at first appealing , but fails in more than two dimensions for large graphs ( hence the ` naive ' in the name ) . nonetheless we find it is sufficient for our purposes nodes too many cases had no two - links and so the spacelike separation could nt be calculated ] . this is partly because it is inaccurate only for large causal sets but also because in the mds algorithm each point s coordinates is fixed by many distances , both timelike and spacelike which limits the effect of noise some poor estimations of spacelike separations . we can then set @xmath54 for the chosen @xmath52 and @xmath53 . these timelike and spacelike distances define our separation matrix @xmath7 ( where timelike separation has the @xmath55 sign in our conventions ) . finally we use the algorithm described in the previous section to assign coordinates in some @xmath19-dimensional minkowski spacetime @xmath56 to each vertex in the dag . given a causal set graph , @xmath45 it is possible , in principle and for large enough @xmath57 , to recover all properties of the spacetime ( up to a factor of the density of the sprinkling)@xcite . in minkowski spacetime there is only one parameter to recover , @xmath2 , and this can be estimated for the process described above and for dags in general @xcite . our task here is to recover not properties of the manifold in which the nodes are embedded , but to find the full details of that embedding . if the graph was originally made sprinkling points into @xmath40 then we know that an exact embedding is possible ( since the original sprinkled coordinates must be a solution ) , and so the embedding algorithm should approximately recover the original coordinates ( up to distance preserving factors ) . if the graph was not , then we may only be able to find an approximate embedding . as is the case for classical mds , our lorentzian mds is guaranteed to recover the coordinates of points when the exact distances are used as the algorithms input . however , when distances are estimated using graph topology , the pairwise separations will be noisy as they are coarse grained by the discrete graph ( see figure [ fig : m_mds ] ) . to assess the reliability of this algorithm we will test it first on the causal set model described above . we will take the coordinates produced by the lorentzian mds algorithm use them to rebuild the graph by again placing edges only between timelike separated pairs . if the overlap between the edges between nodes in the recreated graph , and in the original graph is high , the embedding is an accurate one , and similarly if the overlap is low the embedding is poor . as in @xcite we will measure this using the sensitivity ( the fraction of the correct edges which were predicted ) and specificity ( the fraction of correct non - edges which were predicted ) . embedding of the top @xmath58 most cited papers in the ` hep - th ` citation network . node size is proportional to the number of citations , and lines correspond to citations amongst these papers . a large group of papers is visible in the middle of the plot , forming a long chain of citations , as well as some more isolated papers on either side . a small number of spacelike citations are visible ( those edges more than @xmath59 from vertical ) because this two - dimensional embedding is not perfect , but only the optimal set of coordinates found by the mds algorithm.,scaledwidth=100.0% ] our mds algorithm is able to find accurate embeddings for randomly sprinkled causal sets . we can now attempt to embed networks formed from real social systems , and here we will use citation networks from the arxiv ( 2003 kdd cup datasets ) and the us supreme court@xcite , as well as random dags for comparison . recall that the dimensionality of the embedding is something we can choose , by selecting the @xmath19 largest eigenvalues in the mds algorithm . to measure the effectiveness of the embedding we will compare the original network , with a new network generated from the coordinates determined by the mds algorithm . we want to measure the effectiveness of a classifier which predicts edges in the network from the mds coordinates . to do this we will use the established method of the area under the receiver - operator curve , auc . varying a continuous parameter , the sensitivity and specificity of the embedding is measured , and plotted , as in figure [ fig : auc_example ] and the area under this curve describes the quality of the classifier . the continuous parameter we will vary is the speed of light ( or the speed information can be transferred ) in the minkowski space , @xmath60 . previously , we have set @xmath61 , but varying this speed will change which nodes are connected in new network generated from the mds coordinates . now , nodes @xmath4 and @xmath25 are connected if their coordinates satisfy @xmath62 for small values of @xmath60 , very few nodes are connected and so the specificity is high ( few false positives ) but the sensitivity is low ( many false negatives ) . for large values of @xmath60 , many nodes are connected and so the reverse is true . values for @xmath63 dimensional embeddings of @xmath64 networks . where the curve reaches the top - left of the plot we are in the @xmath65 regime ( @xmath61 denoted by the large black point on each curve ) where the trade - off between false positives and false negatives is balanced . the shape of these roc curves measures the effectiveness of the embedding and it is clear that the the causal set performs best , the random dag worst and the citation networks in between.,scaledwidth=100.0% ] we will measure the ease of embedding a network by taking the mean of the area under this curve for networks of size @xmath66-@xmath67 . in the case of the citation networks this was done by randomly sampling intervals of this size from the citation network . in the cases of the causal set graph , and the random dags , many instances of the required sizes are stochastically generated . -dimensional spacetime , then embedded back into that same dimensional spacetime . their auc values therefore represent the ideal case , where we know an embedding is possible . the random dags show the worst embeddings . their auc values represent the success of an embedding for a graph which has no structure . they are noticeably more embeddable in higher dimensions which is because there are more degrees of freedom in which to assign coordinates while maintaining the randomly placed links . the bars show means and standard deviations of the auc for random dags of size @xmath68 to @xmath69 . citation networks from the arxiv , and the us supreme court fall between these extremes , illustrating the presence of some structure which makes them easier to embed in spacetime . we sampled @xmath70 intervals with between @xmath66 and @xmath67 nodes randomly from each citation network , and the bars show means and standard deviations for the auc in each case.,scaledwidth=100.0% ] finding a good geometric embedding of a network provides a powerful tool for the analysis of that network as it allows standard geometric techniques and intuition to be used . calculations of network properties can be made more efficient , for example , when finding optimal routes from one node to another , the node coordinates provide local information which can improve routing algorithms@xcite . coordinates resulting from geometric embedding also provide a natural visualisation for a network . such visualisations are used in bibliometrics to help identify distinct fields or assist literature reviews@xcite . in the cases of citation analysis , where we conjecture that the spatial dimensions that result from a geometric approach correspond to similarity in the topic of a paper , our approach yields spatial similarities between papers while accounting for the time difference in their publication . once embedding coordinates are known , the idea that nodes may be ` similar ' can be expressed as nodes being close by using an appropriate metric , which need not be the lorentzian one we used in the construction of the coordinates . two nodes might be spacelike separated , so have no direct links , yet be close in a euclidean sense because mds calculates coordinates globally using information from all vertices and edges . papers could be recommended if spatially close to others a reader has interest in even if there are no local connections between them , potentially bringing work or authors to the attention of readers who are not aware of them . spatial similarity can also be used to define clusters . the idea of ` centrality ' or importance of a node has a natural representation in terms of the density of points in the geometric neighbourhood of the point linked with the chosen node . another use of this approach is where edges in a network are placed primarily according to some geometric rule but their connections are also governed by some smaller second order effect . it may only be possible to measure the smaller effect once we have accounted for the primary geometric one by assigning coordinates using lorentzian mds or a similar method we can see this effect clearly when the geometric embedding is one in real geographic space , such as in @xcite where accounting for geographic distance in phone - call data allows more accurate prediction of the second order effect of shared language . other approaches to geometric embedding exist in the literature . mds is characterised by maintaining global separations between pairs . others , such as isomap@xcite maintain shortest paths between pairs linked by local interactions . that approach may not be as appropriate in pseudo - riemannian geometries . firstly , the shortest path locally may not correspond to the geodesic distance like it does in euclidean space ; as discussed above in graphs embedded in minkowski space it is the longest path which corresponds to the geodesic . secondly , the idea of local neighbours is less clearly defined if there are different types of separation , or if , as is the case for minkowski space , the number of nearest neighbours diverges . another class of embedding approaches are probabalistic , such as stochastic neighbour embedding @xcite . although it is beyond the scope of this work , we do not see why such approaches could not be adapted to pseudo - riemannian manifolds , and the ability to use a mixture of separated images for the same object may prove very useful . finally we note that inserting a metric signature into the equations for classical mds allows it to be used on any metric signature , even though we have focused only on the lorentzian signature here . to our knowledge this pseudo - riemannian output is a new development , although some kind of manifold learning techniques exist which can take pseudo - riemannian manifolds as their input @xcite . we note that when performing the lorentzian network mds algorithm we often find multiple negative eigenvalues , suggesting that embeddings in spaces with more than one timelike dimension is also possible , as are potential embeddings into lorentzian manifolds other than minkowski space , incorporating curvature or preferred directions . 10 j. dall and m. christensen . _ random geometric graphs_. phys . e , 66(1):016121 , july 2002 . m. penrose . _ random geometric graphs_. oxford university press , 2003 . z. xie and t. rogers . _ scale - invariant random geometric graphs_. arxiv:1505.01332 , 2015 . d. krioukov , f. papadopoulos , m. kitsak , a. vahdat , and m. boguna . _ hyperbolic geometry of complex networks_. phys . e , 82(3):036106 , 2010 . d. krioukov , m. kitsak , r. sinkovits , d. rideout , d. meyer , and m. boguna . _ network cosmology . _ scientific reports , 2:793 , 2012 . j. r. clough and t. s. evans . _ what is the dimension of citation space ? _ physica a , 448:235247 , 2016 . zhihao wu , giulia menichetti , christoph rahmede , and ginestra bianconi . _ emergent complex network geometry_. scientific reports , 5:10073 , 2015 . trevor f. cox and m.a.a . _ multidimensional scaling , second edition_. crc press , 2000 . d. j. higham , m. rasajski , and n. przulj . _ fitting a geometric graph to a protein - protein interaction network_. bioinformatics , 24(8):10931099 , 2008 . j. r. clough , j. gollings , t. v. loach , and t. s. evans . _ transitive reduction of citation networks_. complex networks , 3(2):189:203 , 2015 . s. w. hawking and g. f. r. ellis . _ the large scale structure of space - time_. cambridge university press , 1973 . g. brightwell and m. luczak . _ the mathematics of causal sets_. arxiv:1510.05612 , 2015 . roger n. shepard . _ attention and the metric structure of the stimulus space_. journal of mathematical psychology , 1(1):5487 , 1964 . h. minkowski . _ die grundgleichungen fur die elektromagnetischen vorgange in bewegten korpern _ , 1910 . l. bombelli , j. lee , d. meyer , and r. d sorkin . _ space - time as a causal set_. phys . lett . 59(5):521524 , 1987 . l. bombelli and d. meyer . _ the origin of lorentzian geometry_. physics letters a , 141(5 - 6):226228 , 1989 . f. dowker . _ causal sets as discrete spacetime_. contemporary physics , 47(1):19 , 2006 . j. myrheim . _ statistical geometry_. cern2538 , 1978 . b. bollobs and g. brightwell . _ box - spaces and random partial orders_. transactions of the ams , 324(i):5972 , 1991 . g. brightwell and r. gregory . _ structure of random discrete spacetime_. phys . 66(3 ) , 1991 . d. rideout and p. wallden . _ spacelike distance from discrete causal order_. classical and quantum gravity , 155013:32 , 2009 . f. dowker , j. henson , and r. d. sorkin . _ quantum gravity phenomenology , lorentz invariance and discreteness_. modern physics letters a , 19(24):18291840 , 2004 . d. d. reid . _ manifold dimension of a causal set : tests in conformally flat spacetimes_. phys . d , ( october 2002):18 , 2003 . j. h. fowler and s. jeon . _ the authority of supreme court precedent_. social networks , 30(1):1630 , 2008 . d. krioukov , f. papadopoulos , a. vahdat , and m. boguna . _ curvature and temperature of complex networks_. phys . e , 80(3):035101 , 2009 . n. j. van eck and l. waltman . _ citnetexplorer : a new software tool for analyzing and visualizing citation networks_. journal of informetrics , 8(4):802 - 823 , 2014 . p. expert , t. s. evans , v. d. blondel , and r. lambiotte . _ uncovering space - independent communities in spatial networks . _ pnas , 108(19):76638 , 2011 . j. b. tenenbaum , v. de silva , and j. c. langford . _ a global geometric framework for nonlinear dimensionality reduction . _ science , 290(5500):231923 , 2000 . s. t. roweis and l. k. saul . _ nonlinear dimensionality reduction by locally linear embedding . _ science , 290(5500):23236 , 2000 . g. e. hinton and s. t. roweis . _ stochastic neighbor embedding_. advances in neural information processing systems , 833840 , 2002 . d. liu , s. trajanovski , and p. van mieghem . _ reverse line graph construction : the matrix relabeling algorithm marinlinga versus roussopoulos s algorithm_. arxiv:1005.0943 , 2010 . r. liu , z. lin , z. su , and k. tang . _ feature extraction by learning lorentzian metric tensor and its extensions_. pattern recognition , 43(10):32983306 , 2010 . following closely the derivation for euclidean mds in @xcite we begin with the timelike ( negative ) and spacelike ( positive ) square distances between each pair of points , and wish to derive the inner product matrix @xmath14 , where @xmath71 . we will denote the minkowski separation between @xmath4 and @xmath25 as @xmath72 . as usual we first fix the coordinates centre of mass , placing it at the origin , such that @xmath73 for each @xmath74 in a @xmath75-dimensional space . remembering that @xmath76 with @xmath31 a diagonal @xmath77 matrix with @xmath32 in the first row and @xmath78 in every other row ( the minkowski metric ) we have : @xmath79 the last term on the right vanishing if the order of the sums is reversed due to the condition we stipulated above . doing the same summing over @xmath80 as well gives : @xmath81 we can now write down each element in the matrix @xmath82 we are trying to calculate . @xmath83 \\ b_{ij } & = -\frac{1}{2 } \left [ s_{ij } - \frac{1}{n } \sum_{i=1}^{n } s_{ij } + \frac{1}{n } \sum_{i=1}^n \mathbf{x}_i^{\mathrm{t}}{{{\mathbf{\textsf{g}}}}}\mathbf{x}_i - \frac{1}{n } \sum_{j=1}^{n } s_{ij } + \frac{1}{n } \sum_{j=1}^n \mathbf{x}_j^{\mathrm{t}}{{{\mathbf{\textsf{g}}}}}\mathbf{x}_j \right ] \\ b_{ij } & = -\frac{1}{2 } \left [ s_{ij } - \frac{1}{n } \sum_{i=1}^{n } s_{ij } - \frac{1}{n } \sum_{j=1}^{n } s_{ij } + \frac{1}{n^2 } \sum_{i=1}^{n } \sum_{j=1}^{n } s_{ij } \right]\end{aligned}\ ] ] which gives us the the matrix @xmath14 from the distances . in the euclidean case all of the eigenvalues of @xmath14 are positive ( or zero ) . in general , the signs of eigenvalues of the matrix @xmath14 will follow the signature of the metric @xmath31 . for any metric , @xmath14 is symmetric and so can can be decomposed into orthogonal eigenvectors , @xmath17 and eigenvalues @xmath16 . @xmath84 we aim to find a solution to the equation @xmath85 . trying @xmath86 where @xmath87 is some real diagonal matrix gives @xmath88 assuming that the metric @xmath31 is diagonal we then have that @xmath89 and since @xmath87 is real , the signs of the elements of @xmath16 must equal those of @xmath31 .	geometric approaches to network analysis combine simply defined models with great descriptive power . in this work we provide a method for embedding directed acyclic graphs into minkowski spacetime using multidimensional scaling ( mds).first we generalise the classical mds algorithm , defined only for metrics with a euclidean signature , to manifolds of any metric signature . we then use this general method to develop an algorithm to be used on networks which have causal structure allowing them to be embedded in lorentzian manifolds . the method is demonstrated by calculating embeddings for both causal sets and citation networks in minkowski spacetime . we finally suggest a number of applications in citation analysis such as paper recommendation , identifying missing citations and fitting citation models to data using this geometric approach .
during the last decades a variety of metal surfaces , such as be(0001 ) and the ( 111 ) surfaces of the noble metals cu , ag , and au , have become a testing ground for many experimental and theoretical investigations.@xcite these surfaces are known to support a partially occupied band of shockley surface states with energies near the fermi level . since these states are strongly localized near the surface and disperse with momentum parallel to the surface , they can be considered to form a quasi two - dimensional ( 2d ) surface - state band with a 2d fermi energy @xmath1 equal to the surface - state binding energy at the @xmath2 point . in the absence of the three - dimensional ( 3d ) substrate , partially occupied shockley surface states would support a 2d collective oscillation , the energy of this plasmon being given by ( unless stated otherwise , atomic units are used , i.e. , @xmath3)@xcite @xmath4 where @xmath5 represents the density of occupied surface states : @xmath6 , @xmath7 represents the magnitude of a 2d wave vector , and @xmath8 is a 2d effective mass . ( [ one ] ) shows that at very long wavelengths plasmons in a 2d electron gas have low energies ; however , they do not affect electron - hole ( e - h ) and phonon dynamics near the fermi level , due to their square - root dependence on the wave vector . much more effective than ordinary 2d plasmons in mediating , e.g. , superconductivity would be the so - called acoustic plasmons with sound - like long - wavelength dispersion.@xcite recently , it has been demonstrated that in the presence of the 3d substrate the dynamical screening at the surface provides a mechanism for the existence of a _ new _ acoustic collective mode , whose energy exhibits a linear dependence on the 2d wave vector.@xcite we refer to this mode as _ acoustic surface plasmon _ ( asp ) , to distinguish it from the conventional surface plasmon predicted by ritchie.@xcite the energy of this latter plasmon is known to be @xmath9 , where @xmath10 is the plasmon energy of a homogeneous electron gas of density @xmath11 : @xmath12 . . all other states of the semi - infinite metal are assumed to comprise a plane - bounded 3d electron gas at @xmath13 . the metal surface is located at @xmath14.,title="fig : " ] + in a simplified model in which surface - state electrons comprise a 2d electron gas at @xmath15 ( see fig . [ fig1 ] ) , while all other states of the semi - infinite metal comprise a 3d substrate at @xmath16 , one finds that both e - h and collective excitations occurring within the 2d gas can be described with the use of an effective 2d dielectric function , which in the random - phase approximation ( rpa ) takes the form@xcite @xmath17 @xmath18 being the 2d fourier transform of the so - called screened interaction in the presence of the 3d substrate alone,@xcite and @xmath19 being the noninteracting density - response function of a 2d electron gas.@xcite in the absence of the 3d substrate , @xmath18 yields the 2d fourier transform of the bare coulomb interaction and @xmath20 coincides , therefore , with the rpa dielectric function of a 2d electron gas , which in the long - wavelength ( @xmath21 ) limit has one single zero corresponding to collective oscillations at @xmath22 . in the presence of a 3d substrate , the long - wavelength limit of @xmath20 has two zeros.@xcite one zero corresponds to a high - frequency oscillation of energy @xmath23 in which 2d and 3d electrons oscillate in phase with one another . the other zero corresponds to a low - frequency _ acoustic _ oscillation in which 2d and 3d electrons oscillate out of phase . the energy of this low - frequency mode is found to be of the form@xcite @xmath24 where @xmath25 represents the 2d fermi velocity @xmath26 and @xmath27 ^ 2\over\pi\left[\pi+2\,i(z_d)\right]}},\ ] ] with @xmath28 the coefficient @xmath29 ( whose value depends on the electron density of the 3d substrate and increases with @xmath30 ) ranges from a constant value ( on the order of @xmath31 for metallic densities ) for a 2d sheet far inside the 3d susbtrate to the asymptotic value @xmath32 ( see ref . ) for a 2d sheet far outside the metal surface . in this paper , we extend the self - consistent calculations of the dynamical response of be(001 ) reported in ref . to the case of the ( 111 ) surface of the noble metals cu , ag , and au . our results indicate that the partially occupied surface - state band in these materials yields the existence of acoustic surface plasmons whose energy is of the form of eq . ( [ two ] ) , but with an @xmath29 coefficient that is much closer to unity than expected from the simplified model described above . furthermore , we demonstrate that the sound velocity ( @xmath33 ) of this low - energy collective excitation is dictated not only by the fermi velocity of the 2d surface - state band but also by the nature of the decay and penetration of the surface - state orbitals into the solid . we also investigate the width of the corresponding plasmon peak , which dictates the lifetime of this collective excitation . in order to achieve a full description of the dynamical response of real metal surfaces , we first consider the one - dimensional potential of ref . . this allows us to assume translational invariance in the plane of the surface , which yields , within linear - respone theory , the following expression for the electron density induced by an external perturbation @xmath34 : @xmath35 @xmath36 representing the 2d fourier transform of the density - response function of our _ interacting _ many - electron system . the collective oscillations created by an external potential of the form@xcite @xmath37 can be traced to the peaks of the imaginary part of the so - called surface response function @xmath38:@xcite @xmath39 which at @xmath40 exhibits a pole at the conventional surface plasmon @xmath41.@xcite the starting point of our calculations is a set of single - particle states @xmath42 and energies @xmath43 of the form @xmath44 and @xmath45 where @xmath46 , @xmath47 is a normalization area , and @xmath48 and @xmath49 are the eigenfunctions and eigenvalues of a one - dimensional schrdinger equation of the form @xmath50\phi_n(z)=\varepsilon_n \phi_n(z ) , \label{schrodinger}\ ] ] @xmath51 being the model potential described in ref . . this potential reproduces the key features of the surface band structure , which in the case of the ( 111 ) surface of the noble metals are the presence of a band gap at the center of the 2d brillouin zone ( 2dbz ) and the existence of shockley and image states in it . alternatively , for a description of the screened interaction @xmath18 entering eqs . ( [ eff2 ] ) and ( [ limit2 ] ) ( which accounts for the presence of 3d bulk states alone ) , the wave functions @xmath48 and @xmath49 can be taken to be the eigenfunctions and eigenvalues of a jellium kohn - sham hamiltonian of density - functional theory ( dft),@xcite which we evaluate in the local - density approximation ( lda ) with the parametrization of perdew and zunger.@xcite in order to solve either eq . ( [ schrodinger ] ) or the jellium kohn - sham equation of dft , we consider a thick slab with a given number of atomic layers and assume that the electron density vanishes at a distance @xmath52 from either crystal edge.@xcite the one - dimensional wave functions @xmath48 are then expanded in a fourier series of the form@xcite @xmath53 , \label{phi}\end{aligned}\ ] ] where the distance @xmath54 is given by the equation @xmath55 @xmath56 and @xmath57 being the number of atomic layers and the interlayer spacing , respectively ; in the case of the ( 111 ) surfaces of the noble metals , we take @xmath58 , @xmath59 , and @xmath60 corresponding to an energy of @xmath61 . due to the symmetry of the model potential entering eq . ( [ schrodinger ] ) , the eigenfunctions @xmath48 are easily found to be either even ( @xmath62 ) or odd ( @xmath63 ) . of the shockley surface state versus the 2d momentum @xmath64 , as obtained with the measured value of the effective mass @xmath8 . the grey area ( with the upper limit at @xmath65 ) represents the projected bulk band structure . the values of @xmath1 , @xmath8 , @xmath66 , and @xmath67 are reported in table [ table ] . solid lines correspond to the bulk states that we have obtained by employing a slab with 81 atomic layers to simulate the semi - infinite solid . we have used the following parameters in the description of the model potential @xmath51 entering eq . ( [ schrodinger ] ) : @xmath68 , @xmath69 , @xmath70 , and @xmath71 ( for a description of the model potential see ref . ) . due to the presence of the band gap , for optical ( @xmath40 ) transitions to occur from an occupied 3d bulk state to an unoccupied 2d surface state the minimum energy @xmath72 is required , which decreases as the momentum transfer @xmath7 increases . for @xmath7 larger than @xmath73 , transitions from occupied ( unoccupied ) 3d bulk states to unoccupied(occupied ) surface states can occur at arbitrary values of the energy transfer @xmath74 . ] figure [ fig2 ] shows the energies @xmath43 that we have obtained from eq . ( [ emu ] ) cu(111 ) by solving eq . ( [ schrodinger ] ) as described above and using the experimental values of the effective masses @xmath75 . electronic structures for ag(111 ) and au(111 ) are similar . the corresponding parameters are reported in table [ table ] . once we have an accurate description of the single - particle orbitals @xmath48 and energies @xmath49 , we evaluate the 2d fourier transform @xmath76 of noninteracting electrons moving in either the model potential @xmath77 or the jellium effective kohn - sham potential of dft : @xmath78 here , the sum over @xmath79 and @xmath80 includes both occupied and unoccupied states , @xmath81 is a positive infinitesimal , and @xmath82 are fermi factors , which at zero temperature are simply given by the heaviside step function @xmath83 @xmath84 being the fermi energy of the solid . in particular , the noninteracting density - response function of the quasi-2d band of occupied shockley states in the absence of the 3d substrate can be obtained from eq . ( [ chi0zz1 ] ) by omitting all bulk states in the sum over @xmath79 and @xmath80 . introducing the one - dimensional wave functions of eq . ( [ phi ] ) into eq . ( [ chi0zz1 ] ) , one finds the following fourier representation of the noninteracting density - response function ( see appendix a):@xcite @xmath85 in the framework of the rpa,@xcite the 2d fourier transform @xmath36 of the density - response function of an interacting many - electron system is obtained by solving the following integral equation @xmath86 the ingredients of this equation being the 2d fourier transforms @xmath76 and @xmath87 of the noninteracting density - response function and the bare coulomb interaction , respectively . all quantities entering eq . ( [ chichi0 ] ) can be represented in the form of eq . ( [ chi0gg ] ) , which yields the following matrix equation for the coefficients @xmath88 : @xmath89 @xmath90 being the corresponding coefficients of the bare coulomb interaction @xmath87 . [ cols="^,>,^,^,<,^,^,^,^,^",options="header " , ] of the ( 111 ) surfaces of the noble metals cu , ag , and au , shown by solid , dashed , and dashed dotted lines , respectively , versus the excitation energy @xmath74 , as obtained from eq . ( [ g ] ) for @xmath91 and @xmath92 mev . the vertical solid lines are located at the energies @xmath93 , which would correspond to eq . ( [ two ] ) with @xmath94 . ] figure [ fig3 ] shows the energy - loss function @xmath95 of the ( 111 ) surfaces of the noble metals cu , ag , and au , as obtained from eq . ( [ g ] ) for @xmath91 . this figure shows the presence of a low - energy collective excitation , whose energy is of the form of eq . ( [ two ] ) but with an @xmath29 coefficient that is close to unity . furthermore , we have carried out calculations of @xmath96 for several low values of @xmath7 and we have found that this low - energy collective excitation is indeed an acoustic surface plasmon with linear dispersion , as shown in fig . 4 for the case of cu(111 ) . . the dashed line represents the energy of an acoustic surface plasmon whose linewidth starts to be considerable due to the presence of interband transitions . the dashed dotted line is the plasmon dispersion of a 2d electron gas in the absence of the 3d system . the grey area indicates the region of the @xmath97 plane ( with the upper limit at @xmath98 ) where e - h pairs can be created within the 2d shockley band of cu(111 ) . the area below the thick solid line corresponds to the region of momentum space where transitions between 3d and 2d states can not occur . the quantities @xmath72 and @xmath99 are determined from the surface band structure of fig . [ fig2 ] . ] in fig . [ fig4 ] , we show the energy of the acoustic surface plasmon of cu(111 ) versus @xmath7 ( solid line ) , as derived from the maxima of our calculated @xmath96 , together with the well - defined plasmon energies that we obtain when only the surface - state band is considered in the evaluation of the noninteracting density - response function of eq . ( [ chi0zz1 ] ) ( dashed dotted line ) . while the plasmon energies of electrons in the isolated surface - state band nicely reproduce in the long - wavelength ( @xmath21 ) limit the conventional 2d plasmon dispersion @xmath100 of eq . ( [ one ] ) , the combination of this surface - state band with the underlying 3d system yields a _ new _ distinct mode whose energy lies just above the upper edge @xmath98 of the 2d e - h pair continuum , as occurs in the case of be.@xcite furthermore , fig . [ fig4 ] shows that in the long - wavelength ( @xmath21 ) limit the energy of the acoustic surface plasmon in cu(111 ) is of the form of eq . ( [ two ] ) but with an @xmath29 coefficient that is considerably closer to unity than expected from eq . ( [ alpha ] ) ( see table [ table ] ) . this discrepancy can be originated in ( i ) the absence in the simplified model leading to eqs . ( [ two ] ) and ( [ alpha ] ) of transitions between 2d and 3d states , and ( ii ) the nature of the decay and penetration of the surface - state orbitals into the solid . plane ( see fig . [ fig1 ] ) with a decay constant @xmath101 , acoustic surface plasmons are found to exist whose energy is of the form of eq . ( [ two ] ) with the coefficient @xmath29 of eq . ( [ alpha ] ) replaced by the @xmath101-dependent @xmath29 coefficient that we have presented in this figure in the case of cu(111 ) for various values of @xmath30 . the horizontal solid line corresponds to the coefficient @xmath29 of eq . ( [ alpha ] ) with @xmath102 . the horizontal dotted line represents the coefficient @xmath29 derived from our full self - consistent calculation that treats bulk and surface states on the same footing . as @xmath30 is shifted from the interior of the solid towards the vacuum , the coefficient @xmath29 increases , in agreement with ref . . ] of the occupied shockley surface state of cu(111 ) at the @xmath2 point , as obtained by solving eq . ( [ schrodinger ] ) . the dashed and dotted lines show wave functions of the form @xmath103 with @xmath104 a.u . and two different values of the decay constant : @xmath105 ( dashed line ) and @xmath106 ( dotted line ) . @xmath107 is the bohr radius : @xmath108 . ] in order to investigate the origin of the small differences between the plasmon energies obtained here and those expected from eqs . ( [ two ] ) and ( [ alpha ] ) , we have carried out calculations of @xmath96 along the lines of the simplified model leading to eqs . ( [ two ] ) and ( [ alpha ] ) ( see fig . [ fig1 ] ) but with the 2d electron gas of fig . [ fig1 ] replaced by a more realistic quasi-2d gas of electrons described by a wave function that decays exponentially away from the @xmath15 plane with a decay constant @xmath101 . we have found that an acoustic surface plasmon is present whose energy is indeed of the form of eq . ( [ two ] ) but with an @xmath29 coefficient that strongly depends on the decay constant @xmath101 , as shown in fig . this figure demonstrates that while in the limit as @xmath109 ( where the quasi-2d electron gas is indeed an ideal 2d sheet ) the coefficient @xmath29 approaches the value expected from eq . ( [ alpha ] ) ( horizontal solid line ) , as @xmath101 decreases the dispersion of the acoustic surface plasmon approaches ( for all negative values of @xmath30 ) the more realistic situation where @xmath29 is close to unity ( horizontal dotted line ) . a comparison between the model wave functions that we have used in this calculation and the actual surface - state wave functions that are involved in the full calculation of figs . [ fig3 ] and [ fig4 ] is presented in fig . although the _ model _ wave functions do not reproduce the actual shape of the surface - state wave function , a finite penetration of the model surface - state wave functions into the solid allows the formation of an acoustic surface plasmon whose sound velocity is very close to the fermi velocity of the 2d surface - state band ( @xmath110 ) , as predicted by our more realistic calculation . indeed , the finite penetration of the surface - state wave function into the solid provides a more complete screening of the quasi-2d collective excitations by the surrounding 3d substrate , which brings the acoustic surface plasmon closer to the upper edge of the 2d e - h pair continuum ( @xmath111).@xcite of eq . ( [ g ] ) with ( solid symbols ) and without ( open symbols ) inclusion of transitions between 2d and 3d states in the evaluations of the noninteracting density - response function of eq . ( [ chi0zz1 ] ) and for various values of the parameter @xmath81 : 0.1 , 1 , and 5 mev . the solid and dashed lines represent fits from the solid and open symbols , respectively . for this surface , the threshold for interband transitions between 2d and 3d states occurs at @xmath112 mev ( see fig . [ fig4 ] ) . ] finally , we have carried out lifetime calculations of the acoustic surface plasmon , as derived from the width of the imaginary part of the full surface - response function of eq . ( [ g ] ) . [ fig7 ] shows the results we have obtained for cu(111 ) with ( solid symbols and lines ) and without ( open symbols and dashed lines ) inclusion of transitions between 2d and 3d states in the evaluation of the noninteracting density - response function of eq . ( [ chi0zz1 ] ) and for three different values of the parameter @xmath81 . plasmon decay can occur by exciting e - h pairs either through transitions between 2d and 3d states , which would not be present in the model leading to eqs . ( [ two ] ) and ( [ alpha ] ) , or through transitions within the 3d continuum of bulk states.@xcite at small energies below the threshold of @xmath113 ( see fig . [ fig4 ] ) , where acoustic surface plasmons can only decay by exciting e - h pairs within the 3d continuum of bulk states ( the solid and dashed lines of fig . [ fig7 ] coincide ) , the linewidth is entirely determined by the choice of the parameter @xmath81 , showing that at these low energies the impact of intraband transitions between 3d bulk states is negligibly small . as the plasmon energy increases , there is a small contribution to the plasmon linewidth from transitions between 2d and 3d states ( the difference between solid and dashed lines)@xcite and an increasing contribution from intraband 3d transitions yielding a finite linewidth which for @xmath114 is not sensitive to the precise value of @xmath81 employed , but which still allows the formation of a well - defined acoustic - surface collective excitation for plasmon energies at least up to @xmath115 . a similar behavior is observed in the case of ag and au . we have carried out self - consistent calculations of the surface - loss function of the ( 111 ) surfaces of the noble metals cu , ag , and au , by considering a one - dimensional potential that describes the main features of the surface band structure . we have found that the partially occupied surface - state band in these materials yields the existence of acoustic surface plasmons , as it had already been demonstrated to occur in the case of be(0001).@xcite the energy of these collective excitations has been shown to exhibit linear dispersion at small wave vectors , with the sound velocity being very close to the fermi velocity of the 2d surface - state band and considerably smaller than expected from a simplified model in which surface - state electrons comprise a 2d electron gas while all other states of the semi - infinite metal comprise a 3d substrate . the origin of the differences between the plasmon energies obtained here and those expected from simplified models has been investigated by performing simplified calculations with the ideal 2d sheet of shockley electrons replaced by a more realistic quasi-2d gas of electrons whose wave functions decay exponentially away from the surface . these calculations have been found to yield an acoustic surface plasmon whose energy is linear in the magnitude of the wave vector , the sound velocity being dictated not only by the fermi velocity of the 2d surface - state band but also by the nature of the decay and penetration of the surface - state orbitals into the solid . with an appropriate choice of the exponential decay of surface - state wave functions these simplified calculations ( which do not account for transitions between 2d and 3d states ) accurately account for the energy dispersion of acoustic surface plasmons , which indicates that the impact of interband transitions between 2d and 3d states on the asp s energy dispersion is negligible . we have also carried out self - consistent calculations of the linewidth of acoustic surface plasmons in the ( 111 ) surfaces of the noble metals . we have found that while the impact of interband transitions between 2d and 3d states is small , intraband transitions between 3d bulk states contribute considerably to the finite linewidth of acoustic surface plasmons , which are found to represent a well - defined acoustic collective excitation for plasmon energies at least up to @xmath116 . finally , we note that as in the case of conventional surface plasmons , acoustic surface plasmons should also be expected to be excited by light , as discussed recently.@xcite we gratefully acknowledge partial support by the university of the basque country , the basque unibertsitate eta ikerketa saila , the spanish ministerio de ciencia y tecnologa , and the ec 6th framework network of excellence nanoquanta ( nmp4-ct-2004 - 500198 ) . here we give explicit expressions for the coefficients @xmath117 entering the expansion of the noninteracting density - response function of eq . ( [ chi0gg ] ) , as obtained by introducing the one - dimensional wave functions of eq . ( [ phi ] ) into eq . ( [ chi0zz1 ] ) . replacing the sum over * k * in eq . ( [ chi0zz1 ] ) by an integral , we find if one characterizes the 3d substrate by a local dielectric function , the characteristic collective oscillations of the 2d electron gas would be completely screened by the sorrounding 3d substrate and no low - energy acoustic mode would exist for a 2d sheet inside the 3d substrate , i.e. , for @xmath127 . this result has suggested over the years that acoustic plasmons should only exist in the case of _ spatially separated _ plasmas , as pointed out by das sarma and madhukar [ s. das sarma and a. madhukar , phys . b * 23 * , 805 ( 1981 ) ] . this is precisely the result first obtained by chaplik [ a. v. chaplik , zh . . fiz . * 62 * , 746 ( 1972 ) ( sov . jetp , * 35 * , 395 ( 1972 ) ) ] in his study of charge - carrier crystallization in low - density inversions layers , by assuming that both the 3d and 2d subsystems can be characterized by local dielectric functions of the drude form . a similar double - cosine fourier representation was introduced by eguiluz [ a. g. eguiluz , phys . lett . * 51 * , 1907 ( 1983 ) ] to evaluate the density - response function of free electrons in a jellium surface . since the 3d fermi velocity @xmath128 of the noble metals is larger than the fermi velocity @xmath25 of the corresponding 2d surface - state band , acoustic surface plasmons can always decay by exciting e - h pairs within the 3d continuum of bulk states .	we have performed self - consistent calculations of the dynamical response of the ( 111 ) surface of the noble metals cu , ag , and au . our results indicate that the partially occupied surface - state band in these materials yields the existence of acoustic surface plasmons with linear dispersion at small wave vectors . here we demonstrate that the sound velocity of these low - energy collective excitations , which had already been predicted to exist in the case of be(0001 ) , is dictated not only by the fermi velocity of the two - dimensional surface - state band but also by the nature of the decay and penetration of the surface - state orbitals into the solid . our linewidth calculations indicate that acoustic surface plasmons should be well defined in the energy range from zero to @xmath0 mev .
fetal distress is generally used to describe the lack of the oxygen of the fetus , which may result in damage or death if not reversed or the fetus is delivered immediately . thus the emergent action by discriminating the fetal distress is an important issue in the obstetrics . in particular , it is important to discriminate the normal fetal heart rate(fhr ) and two types of fetal distress groups(the presumed distress and the acidotic distress ) and reduce the wrong diagnosis rate in which about 65% of the presumed distress fetus(not serious but in need of careful monitoring ) are diagnosed as the acidotic distress fetus(serious , needing an emergent action ) , experiencing a useless surgical operation . in this paper , we try to discriminate the normal and pathologic fetal heart rate group with a robust and reliable method . + the cardiovascular system of the fetus is a complex system . the complex signal , the heart rate , from the complex system contains enormous information about the various functions . the incoming information , on which various heart activities are projected , can not be linearly separated into each function . but we may be able to estimate the signal from a pathologic fetal heart which contains information on the weakness or a complete loss of a particular function . based on this concept , multiple time scale characteristic of heart dynamics have received much attention . as a quantitative method , the multi - scale entropy , which quantifies multi - time scale complexity in the heart rate , was introduced and widely applied to the classification between normal and pathologic groups and also different age groups . it was found that in a specific time scale region , the normal and the pathologic adult heart rate groups and the different age effects on the heart rhythms are significantly distinguished @xcite . + the multiple - scale characteristic in the fetal heart rate is due to the autorhythmicity of its different oscillatory tissues , its interaction with other neural controller and the maternal circulatory system and other neural or hormonal activities , which have a wide range of time scales from secondly to yearly@xcite . + in our previous work on the multiple - scale analysis of heart dynamics , we extended the analysis with the time scale to both the event and time scales@xcite . we found that the event scale in heart dynamics plays more important role than the time scale in classifying the normal and pathologic adult groups . in this paper , therefore , we will investigate characteristic event or time scales of heart dynamics of the normal and the fetal distress group in order to determine the criteria for classifying the normal and the fetal distress groups . + previous works on the fetal heart rate were based on the various nonlinear measures such as the multi - scale entropy , approximate entropy , power spectral density and detrended fluctuation analysis @xcite . they were able to significantly differentiate the mean of normal and pathologic fhr groups but their classification performance was poor , which prevent practical applications of these methods . in this paper , we investigate the typical scale structure of each fhr group by scanning both the event and time scale regions in order to find an appropriate scale region for classifying the normal and the pathologic fetal groups in an optimal way . + we introduce a new analysis method , called the unit time block entropy(utbe ) , which scans both the event and time scale regions of the heart rate based on symbolic dynamics to find the characteristic scale region of the normal and the pathologic heart rate groups@xcite . this method matches the unevenly sampled rr interval data length and the measurement time of the heart rate data set simultaneously , where the rr interval means the time duration between consecutive r waves of the electrocardiogram(ecg ) . in most previous studies the number of rr interval sequences of all data set are fixed in spite of a large variability in the measurement time@xcite . using the utbe method , we can directly compare the entropy of data sets without any ambiguity caused by the nonstationarity and noise effect that inevitably appears in the data set with different data length or measurement time . we find that the normal and two pathologic fhr groups are reliably discriminated . + in section 2 , we will introduce the normal and pathologic fhr data set and their linear properties . in section 3 , the new analysis method , called unit time block entropy(utbe ) , is introduced and applied to the fetal heart rate data set . in section 4 , we show that the normal , the presumed distress and the acidotic distress fhr groups can be discriminated through the utbe method in some characteristic scale regions . finally , we end with the conclusion . the rr interval time of three fetuses . ( a ) a normal fetus . ( b ) a presumed fetus . ( c ) a distress fetus . ( d ) the mean and standard deviation of normal , presumed and distress rr interval sequences in units of seconds . ( e ) the log - log plot of rr interval acceleration . the log - log distribution of rr acceleration shows a power - law distribution.,title="fig : " ] + the rr interval time of three fetuses . ( a ) a normal fetus . ( b ) a presumed fetus . ( c ) a distress fetus . ( d ) the mean and standard deviation of normal , presumed and distress rr interval sequences in units of seconds . ( e ) the log - log plot of rr interval acceleration . the log - log distribution of rr acceleration shows a power - law distribution.,title="fig : " ] the rr interval time of three fetuses . ( a ) a normal fetus . ( b ) a presumed fetus . ( c ) a distress fetus . ( d ) the mean and standard deviation of normal , presumed and distress rr interval sequences in units of seconds . ( e ) the log - log plot of rr interval acceleration . the log - log distribution of rr acceleration shows a power - law distribution.,title="fig : " ] + the fetal heart rate is acquired from 77 pregnant women who were placed under computerized electronic fetal monitoring during the ante partum and intra partum periods . the fetal heart rate is digitized with the data received with the corometrics 150 model(corometrics , connecticut , usa ) through the catholic computer - assisted obstetric diagnosis system(ccaod ; dobe tech , seoul , korea ) . the computerized electronic fetal heart rate monitoring was done on the fetal heart rate data two hours before the delivery , from each pregnant woman without any missing data . first , 77 pregnant women are divided into two groups ; 36 women into the normal fetus group , who showed the normal heart rate tracing , and 41 women with the abnormal fetal heart rate tracing(severe variable , late deceleration , bradycardia , or tachycardia , and delivery by caesarean section ) . the heart rate tracing has been done using as a standard criteria to determine normal or pathologic states of a fetus . after an immediate delibery by a caesarean section , the umbilical artery @xmath0 of the fetus is examined . then , 41 women were further divided into the presumed distress group with 26 women whose umbilical artery @xmath0 was higher than 7.15 , and the acidotic distress group with 15 women whose umbilical artery @xmath0 was lower than 7.15 and the base excess was lower than @xmath1 . by this umbilical artery @xmath0 test we can retrospectively know whether a fetus was in dangerous situation or not . in this work , the wrong diagnosis rate is @xmath2 , 26 women undertaken a useless surgical operation . the improvement of this wrong diagnosis rate through the presurgical classification of different groups based on the fhr data is the main goal of this work . + in order to treat errors in the measuring equipment and ectopic beats we selected the interval that did not have any missing data , and manually treated the ectopic beats using a linear interpolation , which is less than @xmath3 of all heart beats selected . we replaced the ectopic beats with resampled data around the vicinity of @xmath4 using the average sampling frequency of each heart rate sequence . the resampling , which is usually used for even sampling from the unevenly measured heart beat , is restrictively applied for replacing ectopic beats . however , this method inevitably produces a number of unexpected artificial effects on the original rr sequences . for example , it can cause the distortion of the short term correlation of heart rate data@xcite . thus a new method is developed and applied to overcome these problems as in the following section . + before applying the nonlinear method , we investigate conventional linear properties such as the mean and the standard deviation and its student t - test between healthy and pathologic groups(normal , presumed distress and acidotic distress ) . in fig.1(a)-(c ) , the rr interval time of three fetuses are presented . the normal fetus shows a typical characteristic of the small , fast variations , while the pathologic fetus shows relatively slower , larger variations . in fig.1(d ) and ( e ) , linear properties such as the mean , the standard deviation and the log - log distribution of rr acceleration from the normal , the presumed distress and the acidotic distress groups are presented . the result of the t - test on the mean and the standard deviation distributions shows that three fhr groups are distinguishable by these two linear properties , resulting in significant p - values for the mean and the standard deviation[p - value(for the mean / the standard deviation ) : the normal and the presumed distress(@xmath5 ) , the normal and the acidotic distress(@xmath6 ) , the presumed distress and the acidotic distress(0.28/0.903 ) ] . although their significant difference in the mean and the standard deviation , their classification performance based on the sensitivity and specificity is poor . in the next section , we will compare the linear properties with the nonlinear ones obtained from the utbe and the multi - scale entropy method . in fig.1(e ) , the rr interval acceleration shows a power - law distribution . this suggests that the rr interval acceleration rather than rr interval has a nonlinear characteristics which can be best utilized in classifying the different fhr groups . the investigation of linear properties such as the mean and the standard deviation of the rr sequence shows that the heart dynamics of the normal fetus group is faster than that of the pathologic fetus group , leading to a significant difference in the mean . however , the total measurement time shows a large variability for the 77 subjects studied . in fig.[fig : fig1 ] , the mean and the standard deviation of the measurement time of all rr interval sequence is @xmath7 and the time difference between the shortest one and the longest one is @xmath8 . in treating the rr interval sequence , we typically face this contradictory situation . if we try to match the number of rr interval sequence in advance , the measurement time for all rr interval sequences becomes different . if we try to match the measurement time in advance , the number of rr interval sequence becomes different . this naturally occurs for unevenly sampled data such as the rr interval sequence . in order to solve this problem , we used the unit time block entropy(utbe ) , which simultaneously matches both the measurement time and the number of rr interval sequence for all subjects . the variation of measurement time after fixing the number of rr interval sequence in healthy and pathologic fhr groups . the mean and standard deviation of @xmath9 subjects is @xmath10(min ) , it shows a large variation . ] the utbe estimates the entropy of the symbolic sequence composed at the specific event and time scale of a heart beat sequence . since it matches both the measurement time and the number of words of the alphabet from the rr sequence , we can make a direct comparison of all involved fhr groups reliably . + first , we define a word sequence from the rr interval sequence . to construct a word , we use a unit time windowing method , in which each word is constructed in a given unit time window , so the number of symbols involved in each word can be different between windows . + the rr interval sequence is given by @xmath11 and the rr interval acceleration is by @xmath12 . a word is defined as follows @xmath13 a word @xmath14 composed by a unit time windowing method consists of n(i ) symbols . the number of symbols , n(i ) , is different at each unit time scale @xmath15 , satisfying @xmath16 . a symbol @xmath17 , which is used to construct a word , is 1 for @xmath18 larger than @xmath19 and 0 , otherwise . here , @xmath15 is a unit time for constructing a word and @xmath19 is a rr interval threshold or a rr acceleration threshold for binary symbolization . thus a word @xmath14 contains both information about the time scale and the event scale of the heart rate . therefore , it defines a state of specific scale events of the cardiac system during a unit time . the event scale varies with 20 steps from 5% to 95% of the cumulative rank for all rr acceleration values from normal and pathologic hr data set . the time scale varies with 20 steps from 1 second to 10 seconds , which 1 second is sufficiently short to avoid an empty word . + fig.[fig : fig2 ] briefly illustrates how to compose a word sequence from a rr interval sequence . in the block windowing case , a word is defined regularly with the block size n=2 and the block is shifted to the one step next . in the unit time windowing case , a word is defined with a unit time @xmath15 and the unit time windows shifted by 0.5 second to define the next word , allowing the window overlap . the window shifting time , 0.5 second , is appropriately chosen to accumulate an enough number of sampled words . in this way , the unit time widowing method can match the number of words and the total measurement time of different rr intervals sets , while the block windowing method can not . through this procedure we can investigate the complexity of the symbolic rr sequence composed at each time and event scale combination . + with the above word constructing method , we can compose a word sequence from a rr interval sequence . let @xmath20 be a finite , nonempty set which we refer to as an alphabet . a string or word over w is simply a finite sequence of elements of w. an alphabet w has the following relation , @xmath21 where @xmath22 is the set of all words over w , including the empty word , @xmath23 , and @xmath24 the set of all words over w of length n or less and @xmath25 the set of all words over w of length n. the number of possible words of @xmath24 and @xmath25 are @xmath26 and @xmath27 , respectively . and s is a s - ary number , which in this paper is binary(s=2 ) . and the number of possible symbols in a word , @xmath28 , varies with the unit time @xmath15 , which is used to construct a word . @xmath29 is the smallest rr interval time and @xmath30 is an integer number . since @xmath30 depends on @xmath29 , which is sensitive on the ectopic beat or short term noisy beats , noise reduction should be done carefully in preprocessing . the @xmath14 word sequence composed by the unit time windowing method is a natural generalization of one based on the block windowing method . + unit time windowing and block windowing method : two different ways to compose a word sequence from a rr interval sequence . the upper case presents how to compose a word sequence from a rr interval sequence with the unit time windowing method , on the other hand , the lower case presents how to compose a word sequence with block windowing method . the former focus on unit time to define a word but the latter focus on the number of symbols , in this example , the block size is n=2 , to define a word . ] + based on this unit time windowing method , we calculate the symbolic entropy , called the utbe , which quantifies complexity of a word sequence at specific event and time scale region . the entropy , @xmath31 , is a function of the time scale @xmath15 and the event scale @xmath19 . @xmath32 where @xmath33 is the estimation of the frequency of word @xmath14 and @xmath34 is the occurrence number of a word @xmath14 and @xmath35 is the total number of sampled words . to calculate the exact probability of a word @xmath14 , infinite words are considered with @xmath36 then for s=2 , utbe varies between a lower bound and a upper bound at each time scale . the lower bound occurs for the completely regular case and the upper bound occurs for the completely random case where all words are equally probable . + in this paper , to construct an appropriate word ensemble from a rr sequence , the unit time scale is chosen up to 5 steps(about 2.8 seconds ) for the utbe . at this time scale , the largest number of symbols contained in a word from our data set is 8 , which contribute to 2.9% of the population . the number of possible words is @xmath37 and the number of sampled words from our rr interval sequence is 3,313 at this unit time scale . in the followings , we show that among two scales , the event scale is more significant for classification of healthy and two pathologic fhr groups than the time scale . + + + in fig.4 , we applied the utbe method to the three fhr groups and show the result of the search for all event and time scale regions . in order to compare utbe distributions between the normal and two pathologic groups , all p - values of the student t - test in the @xmath38 parameter plane are presented in fig . 4(a - f ) . since the scale characteristics of three groups are not known a priori , we scanned all event and time scales for optimization of the classification . for the presumed distress and the acidotic distress groups in the fig.[fig : fig4](a ) and fig.[fig : fig4](d ) , the event sequences composed by only the 50 - 80% cumulative rank region in the rr interval acceleration are significantly distinguished over all time scales(@xmath39 ) . this suggests that the presumed distress and the acidotic distress groups have relatively different characteristics scale complexity on this specific rr interval acceleration region . these two groups could not be discerned by their linear properties[the p - value of ( mean / standard deviation):(0.28/0.903 ) ] . for the normal and the acidotic distress groups in fig.[fig : fig4](b ) and fig.[fig : fig4](e ) , the event sequences composed by most rr interval acceleration scales except the 50 - 75% cumulative rank region are distinguished in their complexity(@xmath40 ) , while in event sequences composed by only narrow rr interval scale regions , with 5% and 95% cumulative rank ones , show the difference . it suggests that the rr interval acceleration of the normal and the acidotic distress group contains significant information on different internal dynamics of the cardiac system , while the rr interval does not . for the normal and the presumed distress groups in fig.[fig : fig4](c ) and fig.[fig : fig4](f ) , most rr interval scales could not discriminate these two groups except narrow regions , with 5 - 35% ranks and 90 - 95% cumulative ranks , while a wide rr acceleration scale region below the 50% cumulative rank significantly discriminates these two groups(@xmath41 ) . + the above results suggest that the normal , the presumed distress and the acidotic distress groups exhibit characteristic scales with typical dynamics in those scale regions . the information on the characteristic event and time scale regions of three different groups can be used to determine the optimal parameters for the classification of different group . between two types of scales , the event scale is more effective than the time scale in classifying these groups . when two groups are discriminated at an event scale region , these groups are also well discriminated over most time scales . these results suggest that the rr interval acceleration contains typical information about cardiac dynamics of three groups , whereas the rr interval does not . since slow or fast acceleration of the cardiac system is associated with fetal vagal activity and the motherly - fetal respiratory exchange system , it may provide some clues to which functional difference of cardiac systems causes such difference between healthy and pathological groups . + in fig.[fig : fig4](g)-(i ) , the best sensitivity and specificity in the @xmath42 parameter plane is presented for the four measures of statistics and complexity for comparison , respectively . + the sensitivity determines that when a utbe value is given for classifying two groups(a , b ) as a threshold , what percentage of the subjects involved in the group a is correctly classified by the given threshold utbe value . the specificity determines that when a utbe value is given for classifying two groups as a threshold , what percentage of the subjects involved in the other group b is correctly excluded from the group a. so if the sensitivity and specificity are all @xmath43 , the two groups are completely classified by the given threshold . here , the best sensitivity and specificity are achieved after calculating them at all points of the ( @xmath44 ) parameter space , varying the threshold from the minimum utbe to the maximum utbe value of the calculated utbe set@xcite . here , the best sensitivity and specificity is determined as the highest values along the diagonal in the plane of the sensitivity and specificity . as a result , utbe using the rr interval acceleration as a threshold provides the best performance in classification of the presumed distress and the acidotic distress , the normal and the acidotic distress , and the normal and the presumed distress groups . surprisingly , for the normal group and two types of distress groups , utbe using the rr interval acceleration as a threshold completely discriminates these groups with sensitivity 100% , and the specificity 100% , which could not be archived with two linear properties and the utbe using the rr interval threshold . for the presumed distress and the acidotic distress groups , both utbes lead to the same result(sensitivity=71.4% , specificity=72% ) . ( a ) the apen with error bar for different fhr groups . the normal(bold solid line ) has the highest complexity and the presumed(dotted line ) and the acidotic fhr(bold dotted line ) have the lower apen values at each time scale . ( b ) the best sensitivity and specificity of all fhr groups . the best sensitivity and specificity is ( 94% , 95% ) for the normal group and the presumed distress group(solid line ) , ( 85%,86% ) for the normal group and the acidotic distress group(bold dotted line ) and ( 48%,48% ) for the presumed group and acidotic distress group(dotted line).,title="fig:",scaledwidth=45.0% ] ( a ) the apen with error bar for different fhr groups . the normal(bold solid line ) has the highest complexity and the presumed(dotted line ) and the acidotic fhr(bold dotted line ) have the lower apen values at each time scale . ( b ) the best sensitivity and specificity of all fhr groups . the best sensitivity and specificity is ( 94% , 95% ) for the normal group and the presumed distress group(solid line ) , ( 85%,86% ) for the normal group and the acidotic distress group(bold dotted line ) and ( 48%,48% ) for the presumed group and acidotic distress group(dotted line).,title="fig:",scaledwidth=45.0% ] in this section , we compare the performances of the utbe and the multiscale entropy(mse ) , which has been widely used in the hear rate analysis@xcite . the multi - scale entropy calculates the approximate entropy(apen ) or the sample entropy at different scales with heart rate data , which measures the regularity of a given data . since the same data length(n=4061 ) is used in order to remove the dependency on the data length , the approximate entropy is chosen instead of the sample entropy . first , in order to calculate the apen , the rr sequence of length n(=4061 ) is divided into segments of length n and the mean value is calculated for each segment . with the coarse - grained sequence at each scale n , the apen is computed with the following , the parameters ; the embedding dimension m=2 and the delay @xmath45@xcite . in the calculation of the multi - scale entropy , we use two types of r values ; one is determined from the original data at n=1[@xmath46 and the other is variable to be determined at all n scales[@xmath47 . by using the variable r(n ) we can remove the effect of variation due to the coarse - graining process , where sd(n ) denotes the standard deviation of the coarse - grained sequence at a scale n@xcite . since the multi - scale entropy for two cases lead to the similar results , we present here one for the first case . in fig.5(a ) and ( b ) , we present the performance of multi - scale entropy in the classification of three fetal heart rate groups . in fig 5.(a ) , the best p - values for each pair of groups in the student t - test are @xmath48 for the normal vs the acidotic distress , @xmath49 for the normal vs the presumed distress and p=0.143(n=1 ) for the presumed distress vs the acidotic distress . this result shows that the mean multi - scale entropy values of each group are significantly different between the normal and two pathologic groups , but the mean multi - scale entropy values of two pathologic groups are not distinguishable . in order to check the possibility of classification , we investigate the sensitivity and specificity as in the utbe . fig.5(b ) presents the best sensitivities and specificities selected from the calculation in all scales . the best sensitivity and specificity is @xmath50 at the scale(n=4 ) of the normal and the presumed distress case , @xmath51 for the normal and the acidotic distress case and @xmath52 for the presumed distress and the acidotic distress case . the classification performance of multi - scale entropy is not better than that of the utbe in all classification cases from the three groups . this is because utbe searches all the event and time scales to find the optimal classification of different characteristics of healthy and two pathologic data , while multi - scale entropy searches only the time scale . ( a ) the normal and the pathologic fhr groups are distinguished at a specific time and event scale(3sec,5% rank ) . here , the two groups are divided by the threshold(utbe=7.8bit ) . ( b ) the presumed fhr and the acidotic fhr groups are divided by the threshold(utbe=1.4 ) . the marker `` e '' indicates the three fetuses in each group , whose umbilical artery phs are closest to 7.15.,title="fig:",scaledwidth=45.0% ] ( a ) the normal and the pathologic fhr groups are distinguished at a specific time and event scale(3sec,5% rank ) . here , the two groups are divided by the threshold(utbe=7.8bit ) . ( b ) the presumed fhr and the acidotic fhr groups are divided by the threshold(utbe=1.4 ) . the marker `` e '' indicates the three fetuses in each group , whose umbilical artery phs are closest to 7.15.,title="fig:",scaledwidth=45.0% ] in this section , using the difference in the characteristic event scale between the normal and two pathologic groups , we distinguish the normal , the presumed and acidotic groups systematically . in fig.4(e ) and ( f ) , the normal group is significantly distinguished from the presumed distress group and the acidotic distress group in the lower event scale region . in fig.4(d ) , the presumed distress group and the acidotic distress group are distinguished in the relatively higher event scale region . thus , with these characteristics , we can make a strategy to systematically differentiate these groups based on the utbe . first , as in fig.6(a ) , we separate the normal and pathologic groups by a threshold of 7.8 in the utbe , which is determined as the smallest utbe value of the normal group , at the specific time and event scale(3 seconds,@xmath53 rank),respectively . from this procedure , we can differentiate the normal and the pathologic groups with a 100% accuracy . then , for the fetuses deviating from the normal group , usually having the lower utbe , we try to separate the presumed and the acidotic distress groups . in this case , we determine the threshold utbe as 1.4 bits in the time and event scale , which are the first step(1 sec ) and the tenth step(about 50% rank ) , respectively . these time and event scales are selected from the scan of the parameter space as in fig.4(d ) , in which these two groups are distinguished well on the scale region above the tenth event scale . with this threshold , 9 acidotic fetuses out of 14 are distinguished . however , some ambiguity still remains . the clinical determination of the presumed distress fetus and the acidotic distress fetus was carried out by the umbilical artery phs . in fig.6(b ) , we mark the three fetuses with e in each group , who have the umbilical artery phs closest to the threshold value of phs , 7.15 . the range of umbilical artery phs measured is from 6.863 to 7.38 . the phs values of the six fetuses are 7.229 , 7.24 and 7.226 in the presumed distress group and 7.15 , 7.16 and 7.144 in the acidotic distress group . since all the pathologic fetuses were delivered by caesarean surgery after several signs of distress(severe variable , later deceleration , bradycardia , or tachycardia ) , it is not certain if all the acidotic distress fetuses would go to the emergency state or all the presumed distress would be in the safety state . therefore , if the fetuses marked with e are excluded in this analysis , the characteristic of two groups can be clarified more clearly . as a result , most acidotic distress fetuses are less complex than the most presumed distress fetuses at the time and event scale regions . in the comparison of three groups , two pathologic groups are less complex than the normal group , while in the comparison of two pathologic groups the acidotic distress group is less complex than the presumed group . we find that in order to distinguish the normal and pathologic fetuses a small event scale(5% rank ) and a relatively large time scale(3 sec ) of fetal heart dynamics is useful . on the other hand , in order to distinguish the presumed distress and the acidotic distress fetuses a large event scale(about 50 % rank ) and a relatively small time scale(1 sec ) of fetal heart dynamics is more appropriate . with these scale regions , we were able to reduce the wrong diagnosis rate from @xmath54 to @xmath55 . in this paper , we investigated the event and time scale structure of the normal and pathologic groups . we also introduced and calculated the utbe method in the appropriate event and time scale region to distinguish the three groups . to extract meaningful information from the data set , the scale structure of the rr interval acceleration is found to be more helpful than that of the rr interval . in the comparison of the utbe over all event and time scale regions , we found that the normal , the presumed distress and the acidotic distress groups have relatively different event scale structures in the rr interval acceleration . in particular , for the normal and two pathologic groups the utbe from the rr acceleration threshold completely classifies these groups in a chosen scale regions , although both linear properties and utbe using the rr interval threshold performs worse . in the case of the presumed distress and the acidotic distress groups , it also provides better classification performance than other measures . the comparison with the multi - scale entropy also shows that the utbe method performs better . it is due to the fact that the utbe approach searches the event and time scale region , while the multi - scale entropy method searches only the time scale . + based on the difference in the scale structure , we are able to systematically distinguish three fhr groups . the normal and the pathologic groups are separated in a small event scale(5% rank ) and a large time scale(3 sec ) . then , the fetuses deviating from the normal group are separated into the presumed fetuses and the acidotic fetuses at a relatively large event scale(50% rank ) and a small time scale(1sec ) . from these scale regions , we reduce the wrong diagnosis rate significantly . + the results suggest that the utbe approach is useful for finding the characteristic scale difference between healthy and pathologic groups . in addition , we can make a more reliable comparison between all fetuses by simultaneous matching of the measurement time and the number of words . this approach can be applied to the other unevenly sampled data taken from complex systems such as biomedical , meteorological or financial tick data . in this study , we also reconfirm that the more pathological a fetus is the less complex its dynamics , following the pathological order , from the acidotic distress , the presumed distress and to the normal fetus . this indicates that in spite of the peculiarity of the fetal cardiac system , the generic notion of the complexity loss can be applied to the fetal cardiac system . + but these results come from a retrospective test under the well elaborated condition . in a practical view point , the prospective test is necessary to confirm the selected scale regions and its classification performance . we will further test these results with more subjects for the purpose of practical application of this analysis method . @xmath56 we thank dr . sim and dr.chung in dep . obstetrics and gynecology , medical college , catholic univ . for their assistance on clinical data and helpful comments . this work has been supported by the ministry of education and human resources through the brain korea 21 project and the national core research center program . 99 j. altimiras , comp . physiol . a * 124,447 ( 1999 ) s. havlin , s. v. buldyrev , a. bunde _ et al . _ , physica a 273 * , 46 ( 1999 ) m. costa , a. goldberger and c .- k . peng , phys . rev . lett . * 89 * , 068102 ( 2002 ) d. r. chialvo , nature * 419 * , 263 ( 2002 ) m. costa , a. goldberger and c - k . peng , comput . * 29 * , 137 ( 2002 ) m. costa and j. a. healey , comput . cardiol . * 30 * , 705 ( 2003 ) m. costa and a. goldberger and c .- k . peng , phys . e * 71 * , 021906 ( 2005 ) c. e. wood and h. tong , am . j. physiol . * 277 * , r1541 ( 1999 ) u. c. lee , s. h. kim and s. h. yi , phys . rev . e*71 * , 061917 ( 2005 ) g. magenes , m. g. signorini and d. arduini , proc . the second joint embs / bmes conference , october 23 - 26 , 2002 g. magenes , m. g. signorini and m. ferrario _ et al . _ , the 25th annual international conference of the ieee / embs , sep . 17 - 21 , 2003 m. g. signorini , g. magenes and s. cerutti__et al . _ _ , ieee trans . biomed . 50 * , 365 ( 2003 ) m. e. d. gomes , h. n. guimaraes , a. l. p. ribeiro and l. a. aguirre , comput . 32 * , 481 ( 2002 ) the sensitivity is a true positive rate defined as @xmath57 , and the specificity is a true negative rate defined as @xmath58 . ( a : true positive , b : false positive , c : false negative , d : true negative ) . more details are available at http://bmj.bmjjournals.com/ or see t. greenhalgh , bmj * 315 * , 540 ( 1997 ) v. v. nikulin and t. brismar , phys . * 92 * , 089803 ( 2004 )	recently , multiple time scale characteristics of heart dynamics have received much attention for distinguishing healthy and pathologic cardiac systems . despite structural peculiarities of the fetal cardiovascular system , the fetal heart rate(fhr ) displays multiple time scale characteristics similar to the adult heart rate due to the autorhythmicity of its different oscillatory tissues and its interaction with other neural controllers . in this paper , we investigate the event and time scale characteristics of the normal and two pathologic fetal heart rate groups with the help of the new measure , called the unit time block entropy(utbe ) , which approximates the entropy at each event and time scale based on symbolic dynamics . this method enables us to match the measurement time and the number of words between fetal heart rate data sets simultaneously . we find that in the small event scale and the large time scale , the normal fetus and the two pathologic fetus are completely distinguished . we also find that in the large event scale and the small time scale , the presumed distress fetus and the acidotic distress fetus are significantly distinguished . event scale , time scale , symbolization , fetal heart rate , dynamics , complexity 87.19.hh , 87.10.+e , 89.75.-k
one of the promising mechanisms of primordial black holes and wormholes formation is a collapse of vacuum bubbles during phase transitions in the early universe @xcite . the supermassive black holes in the centers of galaxies may be also originated by this mechanism . we will describe the possible types of dynamical evolution of vacuum bubbles in the schwarzschild - desitter metric by using the thin shell approximation for the boundary between the true and false vacuum . the formalism of thin shells in general relativity was first developed by w. israel @xcite . later this formalism was elaborated and adjusted for the case of cosmological vacuum phase transitions @xcite . the boundary of vacuum bubble divides the schwarzschild - desitter space - time into internal and external regions . in the following these regions are designated by indexes `` @xmath0 '' and `` @xmath1 '' respectively . our aim is a full classification of possible evolution scenarios of vacuum bubbles versus of parameters of the schwarzschild - desitter space - time and initial conditions for a thin shell . it appears that a useful classification quantity for this problem is a mass parameter @xmath2 which will be defined below . depending on this mass parameters the bubble is either expanding to infinity or contracting with a final formation of black hole or wormhole @xcite . this work is a generalization of the earlier analysis in @xcite , where only some particular scenarios for this problem were investigated . in previous works @xcite were considered only the particular cases of solutions corresponding to the zero value of our inner mass parameter @xmath3 . we describe a more general case when @xmath4 , and then the new types of solutions appear . the bubbles are originated in the phase transition in the early universe and may contain in principle smaller bubbles inside ( see e. g. @xcite ) . to model this in a formal way we include an interior mass parameter @xmath5 . this parameter may be considered as a seed black hole . in general , solutions with @xmath4 are quite the similar ones to considered in @xcite . at the same time the addition of inner mass parameter @xmath5 results in a complication of the classification scheme for possible solutions . in section [ sec2 ] the equation of motion for a thin shell in the schwarzschild - desitter metric is analyzed . basing on this equation in section [ sec3 ] we develop a classification scheme for possible evolution scenarios and construct also the carter - penrose diagrams for corresponding global geometries . the concluding remarks are shortly summarized in section [ sec4 ] . the schwarzschild - desitter metric can be written in the form @xmath6 where @xmath7 is the schwarzschild mass , @xmath8 is a vacuum energy density and @xmath9 this metric has the following specific properties . the positive roots of equation @xmath10 define the radii of event horizons in this metric . the number of positive roots ( and so the number of event horizons ) depends on the ratio of @xmath7 and @xmath11 . there are no event horizons , if @xmath12 . there is only one event horizon , @xmath13 , if @xmath14 . in the case @xmath15 there are two distinctive event horizons : @xmath16 ; \\ r_{\rm h3 } & = & 2\sqrt{\frac{p}{3}}\cos\left[\frac{\pi}{3}- \frac{1}{3}\arctan\sqrt{\frac{4p^3}{27q^2}-1}\right],\end{aligned}\ ] ] where @xmath17 and @xmath18 . it can be shown that @xmath19 . see in the fig . [ figure1 ] the corresponding carter - penrose diagrams for global geometry of the schwarzschild - desitter space - time @xcite . the carter - penrose diagrams for the schwarzschild - de sitter metric . in the case ( a ) , when @xmath15 , there are event horizons . there is only one event horizon in the case ( b ) , when @xmath14 . in the case ( c ) , when @xmath20 , the event horizon is absent . diagrams ( b ) and ( c ) are shown for the case of expansion . the regions @xmath21 and @xmath22 and also zero and infinity lines are swapped in the corresponding ( nonshown ) diagrams for the case of contraction . the oppositely directed null lines in these diagrams and in the following ones are not necessarily perpendicular to each other due to arbitrary deformation of the coordinate systems.,scaledwidth=80.0% ] the basic equation of motion for a thin vacuum shell in the schwarzschild - de sitter space - time , resulting from the matching of the inner and outer metrics on the shell , can be written in the following form @xcite : @xmath23 in this equation @xmath24 is a shell radius , @xmath25 is a proper time measured by an observer at rest with respect to this shell , @xmath26 is derivative with respect to a proper time , @xmath27 is a surface energy density on the shell ( in the discussed vacuum case @xmath27=const ) , @xmath28 and @xmath29 is the schwarzschild mass of the inner and outer region respectively , @xmath30 is an energy density of the inner region , @xmath31 is an energy density of the outer region and symbols @xmath32 . these symbols equal to @xmath33 , if radius of a two - dimensional sphere is growing in the direction of an outgoing normal , and equal to @xmath34 in the opposite case . for a further analysis we rewrite the equation of motion ( [ eqmot ] ) in the `` energy conservation '' form @xmath35 , with an effective potential @xmath36 which is shown in the fig . [ figure2 ] . values of @xmath37 and @xmath38 in the equation of motion ( [ eqmot ] ) in the `` energy conservation '' form are defined by relations @xmath39 ; \\ \label{sigmain}\ ] ] @xmath40 . \label{sigmaout}\ ] ] we will consider the general case when a surface energy density of the shell @xmath27 may be as positive and negative . it must be noted that for the positive value of @xmath27 there is the exceptional case , when @xmath41 and @xmath42 . in this case the equation of motion ( [ eqmot ] ) has no solution . it would be also seen in the following carter - penrose diagrams . it is useful to define the following two quantities , one with a dimension of mass , @xmath43 , and one dimensionless , @xmath44 . with this definitions it is seen from ( [ sigmain ] ) that @xmath45 changes its sign at @xmath46 , where @xmath47 provided that ( i ) @xmath48 and @xmath49 or ( ii ) @xmath50 and @xmath51 . respectively , @xmath52 changes its sign when @xmath53 , where @xmath54 provided , that ( iii ) @xmath48 and @xmath55 or ( iv ) @xmath50 and @xmath56 . the graphs of potential function @xmath57 from ( [ potential ] ) at @xmath58 ( left ) and @xmath59 ( right).,scaledwidth=80.0% ] now we consider a behavior of the effective potential ( [ potential ] ) . the zeros of this potential , @xmath60 , define the bounce points of solution when @xmath61 . the maximum of potential ( [ potential ] ) is at the point @xmath62 , where @xmath63}}\bigg\ } \nonumber\\ & \big/ & \big[(\varepsilon_{\rm in}+\varepsilon_{\rm out } + 6\pi\sigma^2)^2 - 4\varepsilon_{\rm in}\varepsilon_{\rm out}\big].\end{aligned}\ ] ] this maximum corresponds to the zero of potential , @xmath64 , at @xmath65 , where . @xmath66 \nonumber \\ & & + \mu\frac{\varepsilon_{\rm out}- \varepsilon_{\rm in}}{6\pi\sigma^2}+ \frac{\mu^2}{16\pi^2\sigma^2y_{\max}}\bigg\}^{3/2}. \label{m0}\end{aligned}\ ] ] the potential @xmath67 at @xmath68 and vise versa . a second derivative of the potential ( [ potential ] ) is @xmath69 it is negative everywhere , and so there is no of a stable equilibrium point for the equation of motion ( [ eqmot ] ) . at the next step let us define the values of parameter @xmath2 when radii @xmath70 and @xmath71 ( where @xmath45 and @xmath52 change the sign ) coincide with the bounce points of the equation of motion ( [ eqmot ] ) . the corresponding solution of equation @xmath72 is @xmath73 , where @xmath74 respectively , the corresponding solution of equation @xmath75 is @xmath76 , where @xmath77 by using ( [ m0 ] ) , ( [ m1 ] ) and ( [ m3 ] ) it can be shown that both @xmath78 and @xmath79 . there are degenerate cases when the inner and outer regions with respect to the shell have only one event horizon . the inner schwarzschild - de sitter metric has only one event horizon when @xmath80 , where @xmath81 respectively , the outer schwarzschild - de sitter metric has only one event horizon when @xmath82 , where @xmath83 it can be verified that both @xmath84 and @xmath85 . for the following analysis of the dynamical evolution of the shell it is important to know the values of potential ( [ potential ] ) at the event horizons of both the inner and outer metrics , @xmath86 and @xmath87 , respectively . by using equation ( [ horizon ] ) for the inner event horizon radius , @xmath88 , and equation ( [ potential ] ) for potential , after some algebraic manipulation we obtain @xmath89\ ! + \!\rho_{\rm h\,in}\left(1\!+ \!\frac{\varepsilon_{\rm out}- \varepsilon_{\rm in}}{6\pi\sigma^2}\right)\right\}^2 } { \left[8\rho_{\rm h\,in}(\rho_{\rm h\,in}-2m_{\rm in})\right]}\leq0 . \label{u - grin}\ ] ] analogously , for the outer event horizon @xmath90 we obtain @xmath91 + \!\rho_{\rm h\,out}\left(1\!+\!\frac{\varepsilon_{\rm in}- \varepsilon_{\rm out}}{6\pi\sigma^2}\right)\right\}^2 } { \left[8\rho_{\rm h\,out}(\rho_{\rm h\,out}-2m_{\rm out})\right]}\leq0 . \label{u - grout}\ ] ] the equality in ( [ u - grin ] ) and ( [ u - grout ] ) is achieved only at @xmath73 ( inner metric ) , where @xmath92 , and , respectively , at @xmath76 ( outer metric ) , where @xmath93 . in other words , the points of event horizons of both the inner and outer metrics , @xmath94 and @xmath95 can not be below the potential in the fig . [ figure2 ] . the gravitational radius @xmath96 for both ` in ' and ` out ' metrics is changed in the range @xmath97 , and , therefore , it is placed at the left of the potential curve . in a similar way , the gravitational radius @xmath98 is changed in the range @xmath99 , and , therefore , it is placed at the right of the potential curve . now we have all necessary ingredients for the investigation of possible motions of the thin vacuum shell in the schwarzschild - de sitter metric . an effective potential ( [ potential ] ) and the corresponding carter - penrose diagrams for the case @xmath48 , @xmath101 and values of the rolling mass parameter @xmath2 in the range @xmath102 ( graphs a and b ) and @xmath80 ( graphs c and d ) . see section 3.1 for details.,scaledwidth=80.0% ] first of all we consider a simple case , when @xmath48 and @xmath101 ( another similar case when @xmath50 and @xmath49 may be analyzed in a similar way ) . in this case @xmath103 , as it follows from ( [ sigmain ] ) and ( [ sigmaout ] ) . therefore , there are no radii @xmath70 and @xmath71 . it can be shown that in this case @xmath104 , @xmath105 and @xmath106 . the rolling parameter of our classification scheme is a mass parameter @xmath2 and we start from the large value of this parameter . if @xmath102 , then an event horizon is absent and initial expansion ( or contraction ) of the shell is unbounded ( there is no bounce point ) . the corresponding potential @xmath57 and the carter - penrose diagram is shown in the figs . [ figure3]a and [ figure3]b respectively . here and further below the carter - penrose diagrams are shown only for an initially expanding envelope . the corresponding diagrams for contracting envelope are easily reproduced from the expanding ones by symmetry reflection with respect to the median horizontal line ( except for some special cases ) . at @xmath80 , the first event horizon appears in the inner metric , @xmath107 . see figs . [ figure3]c and [ figure3]d for the corresponding potential and diagram . an effective potential ( [ potential ] ) and the corresponding carter - penrose diagrams for the case @xmath48 , @xmath101 and values of the rolling mass parameter @xmath2 in the range @xmath108 ( graph a ) , @xmath82 ( graphs b and c ) and @xmath109 ( graph d).,scaledwidth=80.0% ] if @xmath108 , then there are two event horizons , @xmath110 and @xmath111 . see fig . [ figure4]a for the carter - penrose diagram , while the potential has a similar form as in the previous two cases ( figs . [ figure4]a and [ figure3]c ) . at @xmath82 , the first event horizon appears in the outer metric , @xmath112 . we prove that @xmath113 . indeed , this inequality may be written in the form @xmath114 . for @xmath48 , if we will prove the inequality @xmath115 , then we prove our statement . the last inequality is evident by remembering that @xmath101 . so for @xmath82 the event horizon @xmath116 is located between @xmath110 and @xmath111 . see figs . [ figure4]b and [ figure4]c for the corresponding potential and the carter - penrose diagram of an expanding shell . if @xmath109 , where @xmath117 is defined from equation @xmath118 , there are two event horizons , @xmath119 and @xmath120 , in the outer metric . these event horizons are located between the event horizons of the inner metric @xmath110 and @xmath111 . see fig . [ figure4]d for a corresponding diagram . if @xmath121 and @xmath122 , then the arrangement of event horizons is @xmath123 . the carter - penrose diagram for an expanding shell is shown in the fig . [ figure8]d . an effective potential ( [ potential ] ) and the corresponding carter - penrose diagrams for the case @xmath48 , @xmath101 and values of the rolling mass parameter @xmath2 in the range @xmath68 ( graphs a , b and c ) , @xmath124 ( graphs b and d ) and @xmath125 ( graphs b and c).,scaledwidth=80.0% ] if @xmath68 , the region appears , where potential @xmath57 is positive ( see fig . [ figure5]a ) . now solutions for @xmath126 have the bounce points . the arrangement of event horizons is @xmath127 . the carter - penrose diagrams for an expanding and contracting shell are shown in the fig . [ figure5]b and fig . [ figure5]c respectively . if @xmath128 , then in the case @xmath124 the possible carter - penrose diagrams for an expanding and contracting shell are shown in the figs . [ figure5]b and [ figure5]d . in the last case @xmath129 the arrangement of event horizons is @xmath127 . see in the figs . [ figure5]b and [ figure5]c the carter - penrose diagrams for an expanding and contracting shell respectively . in particular , these diagrams illustrate the formation of a black hole or wormhole after a final contraction of the shell . an effective potential ( [ potential ] ) and the corresponding carter - penrose diagrams for the case @xmath48 and @xmath131 and values of the rolling mass parameter @xmath2 in the range @xmath82 ( graphs a ) , @xmath132 ( graphs b and c ) and @xmath122 ( graph d).,scaledwidth=80.0% ] from this point we start classification of a more tangled case , when @xmath48 and @xmath133 ( the case when @xmath50 and @xmath134 may be analyzed in a similar way ) . now both @xmath45 and @xmath52 can change the signs at radii @xmath70 and @xmath71 . in the considered case the following inequalities are valid : @xmath135 and @xmath136 . it is useful to define the additional 5 dimensionless parameters : @xmath137 , @xmath138 , @xmath139 , @xmath140 and @xmath141 . a first parameter is a solution @xmath142 of the equation @xmath143 , where @xmath144 ( @xmath145 at @xmath146 and vice versa ) . a second parameter is a solution @xmath147 of the equation @xmath148 at @xmath82 , where @xmath149 this parameter exists only if @xmath150 . under this condition @xmath151 at @xmath152 and vice versa . in the opposite case , when @xmath153 , it is always @xmath154 . a third parameter is a solution @xmath155 of the equation @xmath156 at @xmath80 , where @xmath157 ( @xmath158 at @xmath159 and vice versa ) . a fourth parameter is a solution @xmath160 of the equation @xmath161 at @xmath82 , where @xmath162 this parameter exists only if @xmath163 . under this condition @xmath164 at @xmath165 and vise versa . in the opposite case , when @xmath166 , it is always @xmath167 . finally , the fifth parameter is a solution @xmath168 of the equation @xmath169 at @xmath80 , where @xmath170 ( @xmath171 at @xmath172 and vice versa ) . it is easy to verify that @xmath173 at @xmath82 and @xmath174 at @xmath80 . a mutual arrangement of these 5 parameters is @xmath175 . as a first step in classification of the situation , when @xmath48 and @xmath176 , we consider the case , when @xmath146 ( the case @xmath177 is quite a similar ) . again we begin from the large value of the rolling parameter @xmath2 . if @xmath102 , the event horizons are absent and there are only two radii @xmath70 and @xmath71 , where the signs of @xmath178 are changed . the potential @xmath57 is similar to one shown in the fig . [ figure3]a ( where @xmath70 and @xmath71 are not shown ) . a diagram for an expanding shell is analogous to one shown in the fig . [ figure3]b . the next case is @xmath80 . now a first event horizon @xmath179 appears , which is at left to @xmath70 ( i. e. @xmath180 ) . now potential is similar to one in the fig . [ figure3]c . the crucial point is that now a shell intersects the radius @xmath107 , when @xmath181 . the carter - penrose diagram is similar to one shown in the fig . [ figure3]d . if @xmath182 , there are two event horizons @xmath183 and @xmath111 , but qualitatively this case is similar to the preceding one . the carter - penrose diagram for an expanding shell is shown in the fig . [ figure4]a . the case @xmath82 is divided into subcases . we begin from the subcase @xmath184 , when inequalities @xmath167 and @xmath185 are fulfilled . the resulting diagram for an expanding shell is shown in the fig . [ figure6]a . the other subcase will be considered later . if @xmath132 , where @xmath117 is a solution of the equation @xmath186 . now instead of one event horizon @xmath116 there are two event horizons , @xmath119 and @xmath120 . the corresponding potential and the carter - penrose diagram are shown in the fig . [ figure6]b and [ figure6]c respectively . it must be noted that a moving shell intersects the event horizons @xmath95 when @xmath187 . therefore , a shell intersects these event horizons in the region @xmath188 . on the contrary , a shell intersects the event horizons @xmath189 in the region @xmath190 when @xmath191 . the only distinguishing feature of the case @xmath122 from the preceding one is swapping round the event horizons @xmath119 and @xmath111 . the arrangement of radii is ( @xmath192 . a corresponding diagram is shown in the fig . [ figure6]d . an effective potential ( [ potential ] ) and the corresponding carter - penrose diagrams for the case @xmath48 , @xmath101 and values of the rolling mass parameter @xmath2 in the range @xmath193 ( graphs a and b ) , @xmath194 ( graphs c ) and @xmath195 ( graph d).,scaledwidth=80.0% ] if @xmath193 , the potential intersects the axis @xmath196 and a shell will bounce from the potential . the arrangements of radii is the same as in the previous case . the diagrams for expanding and contracting shell are shown in the fig . [ figure7]a and [ figure7]b respectively . if @xmath194 , the radii @xmath119 and @xmath71 are swapped round . the potential is shown in the fig . [ figure7]c ( the event horizons @xmath120 and @xmath110 are not shown ) . a diagram for an expanding shell is the same as in the fig . [ figure7]a . a contracting shell intersects the event horizon @xmath119 in the region @xmath190 ( when @xmath197 ) . a diagram for a contracting shell is shown in the fig . [ figure5]d . if @xmath195 , the radii @xmath111 and @xmath70 are swapped round . two radii @xmath198 and @xmath199 ( where signs of @xmath178 are changed ) is now under the potential graph . the diagram for a contracting shell is the same as in the previous case ( see fig . [ figure5]d ) . a corresponding diagram for an expanding shell is shown in the fig . [ figure7]d . an effective potential ( [ potential ] ) and the corresponding carter - penrose diagrams for the case @xmath48 , @xmath101 and values of the rolling mass parameter @xmath2 in the range @xmath165 and @xmath165 ( graphs a ) , @xmath109 ( graphs b ) and @xmath193 ( graphs c).,scaledwidth=80.0% ] now we return to another subcase of the case @xmath82 , when @xmath200 , the inequality @xmath201 is fulfilled and parameter @xmath202 exists . if @xmath165 , then @xmath112 . the arrangement of radii is ( @xmath203 . a diagram for an expanding shell is shown in the fig . [ figure8]a . if @xmath109 , where @xmath117 defined from equation @xmath204 , instead of the one event horizon @xmath116 there are two event horizons @xmath205 . the arrangement of radii is @xmath206 . a corresponding diagram for an expanding shell is shown in the fig . [ figure8]b . if @xmath122 , the arrangement of radii is @xmath207 ) . the only difference with preceding subcase is swapping round the event horizons @xmath208 and @xmath120 . a diagram for an expanding shell shell shown in the fig . [ figure6]d . if @xmath193 , the potential intersects the axis @xmath196 and the bounce point appears . the arrangement radii is a similar to the previous case , but a shell now can bounce from the potential . the diagrams for an expanding and contracting shells are shown in the figs . [ figure7]a and the [ figure7]b respectively . at @xmath128 and at @xmath209 there will be the following arrangement of radii : @xmath210 . the diagrams for an expanding and contracting shells are shown in the figs . [ figure8]c and [ figure7]b respectively . the next two subcases depend on the relation @xmath211 and are described in a similar way . the remaining cases for @xmath212 are similar to ones for @xmath184 . if @xmath213 in the case @xmath214 , then @xmath167 . as a result this case is reduced to the case , when @xmath215 . if @xmath216 for @xmath82 , then the inequality @xmath217 is fulfilled and besides the parameter @xmath202 there exists also @xmath138 . at @xmath152 , the inequalities @xmath218 and @xmath151 are valid . as a result the arrangement of radii is @xmath219 a corresponding diagram for an expanding shell is shown in the fig . [ figure4]c . if @xmath220 , instead of the one event horizon @xmath221 there are two event horizons @xmath119 and @xmath120 and the arrangement of radii is @xmath222 . a corresponding diagram for an expanding shell is shown in the fig . [ figure4]d . if @xmath193 , the potential intersects the axis @xmath196 and a shell can bounce from the potential . the corresponding diagrams for an expanding and contracting shells are shown in the figs . [ figure5]b and [ figure5]d respectively . if @xmath223 , where @xmath117 is a solution of the equation @xmath224 , the event horizon @xmath120 changes the place with @xmath71 . the coincidence of these radii occurs at the radius , where @xmath60 . the radius @xmath71 now is under the potential curve . a shell will intersect @xmath120 with @xmath187 . a corresponding diagram for an expanding shell is shown in the fig . [ figure8]c . the case @xmath225 is similar to one for @xmath215 . the case @xmath226 is similar to one for @xmath214 . finally , the case @xmath213 is similar to one for @xmath184 . analogously is considered the case , when @xmath227 , and the case , when exists only one of the two radii , @xmath70 or @xmath71 . this case completes the classification . we classified all possible evolution scenarios of a thin vacuum shell in the schwarzschild - desitter metric and constructed the carter - penrose diagrams for corresponding global geometries . these geometries illustrates the possibilities for final formation of black holes and wormholes or eventual expansion of bubbles . we acknowledge viktor berezin and yury erochenko for helpful discussions . this work was supported in part by the russian foundation for basic research grants 06 - 02 - 16342-a and the russian ministry of science grants lss 5573.2006.2 .	we construct the classification scheme for all possible evolution scenarios and find the corresponding global geometries for dynamics of a thin spherical vacuum shell in the schwarzschild - desitter metric . this configuration is suitable for the modelling of vacuum bubbles arising during cosmological phase transitions in the early universe . the distinctive final types of evolution from the local point of view of a rather distant observer are either the unlimited expansion of the shell or its contraction with a formation of black hole ( with a central singularity ) or wormhole ( with a baby universe in interior ) .
the gluonium states of qcd have remained intriguing for almost four decades@xcite . their `` exotic '' nature reflects itself not least in several long - standing problems which the qcd sum rule approach faces in the spin-0 glueball channels@xcite . in the scalar ( @xmath0 ) glueball correlator , in particular , the departure from asymptotic freedom sets in at unusually small distances@xcite and the perturbative wilson coefficients of the standard operator product expansion ( ope ) proved inadequate to establish consistency both among the @xmath0 glueball sum rules and with an underlying low - energy theorem@xcite . although nonperturbative contributions due to direct ( i.e. small ) instantons@xcite were early candidates for the missing short - distance physics@xcite , insufficient knowledge of the instanton size distribution@xcite prevented their quantitative implementation at the time . only recently , the derivation of the exact instanton contributions ( to leading order in @xmath1 ) , their duality continua and the corresponding borel sum - rule analysis@xcite showed that direct instantons indeed solve the mentioned key problems in the scalar glueball channel . below we will outline a more thorough and systematic treatment@xcite which eliminates artefacts of earlier approximations , significantly modifies the sum - rule results in both spin-0 channels and improves their reliability . moreover , we implement for the first time topological charge screening contributions to the ope in the @xmath2 channel@xcite and show how those resolve earlier sum - rule inconsistencies@xcite . finally , we review several new predictions for spin-0 glueball properties . the spin-0 glueball sum rules are based on the scalar @xmath3 and pseudoscalar @xmath4 correlations functions @xmath5 where @xmath6 with @xmath7 are the gluonic interpolating fields ( of lowest mass dimension ) @xmath8 and @xmath9 with @xmath10 ( the definition in ref . contains a typo ) . the zero - momentum limit of the correlator ( [ corr ] ) is governed by the low - energy theorems ( lets ) @xmath11 in the scalar@xcite and ( for three light flavors and @xmath12)@xmath13 in the pseudoscalar channel@xcite . ( note that eq . ( [ pslet ] ) vanishes in the chiral limit . ) consistency with the low - energy theorems places stringent constraints on the sum rules which can not be satisfied without nonperturbative short - distance physics@xcite . contact with the hadronic information in the glueball correlators is established by means of the dispersive representation@xmath14 where the necessary number of subtractions is implied but not written explicitly . the standard sum - rule description of the spectral functions @xmath15 contains one or two resonance poles in zero - width approximation , @xmath16 , and the local - duality continuum @xmath17 from the iope discontinuities in the duality range which starts at an effective threshold @xmath18 . in order to write down qcd sum rules , the borel - transformed iope and dispersive representations of the correlators - weighted by @xmath19 powers of @xmath20 - are matched in the fiducial @xmath21-region ( @xmath21 is the borel variable ) and rearranged in terms of the continuum - subtracted borel moments @xmath22 as @xmath23 the pole contributions of interest ( and for @xmath24 the crucial subtraction terms ) are then isolated on the rhs , and the hadronic parameters @xmath25 and @xmath18 can be determined numerically . our theoretical framework , the instanton - improved operator product expansion ( iope ) , factorizes the correlators at large , spacelike momenta @xmath26 into contributions from `` hard '' field modes ( with momenta @xmath27 ) in the wilson coefficients @xmath28 and `` soft '' field modes ( with @xmath29 ) in the `` condensates '' @xmath30 of operators @xmath31 with increasing dimension @xmath32 . previous glueball sum rules based on the ope with purely perturbative wilson coefficients were plagued by notorious inconsistencies between the predictions of different moment sum rules and by massive let violations . moreover , the soft nonperturbative condensate contributions were exceptionally small . we have therefore analyzed _ hard _ nonperturbative contributions to the wilson coefficients . they are strongly channel dependent and due to direct instantons and topological charge screening . the instanton contributions to the spin-0 coefficients@xcite , @xmath33 are large ( while those in the @xmath34 tensor channel vanish ) and add to the unit - operator coefficients @xmath35 . the imginary part of their fourier transform at timelike momenta generates the borel moments@xcite@xmath36 which are similar or larger in size than the perturbative ones . the evaluation of these moments requires as the sole input the ( anti- ) instanton distribution @xmath37 which is implemented by means of a lattice - based gaussian - tail parametrization with the correct small-@xmath38 behavior , @xmath39 ( for @xmath40 ) , which was introduced in ref . and shown to prevent several artefacts of the oversimplified `` spike '' approximation @xmath41 on which all previous direct - instanton calculations had relied . as an additional benefit , the realistic size distribution allows for a gauge - invariant ir renormalization which excludes large instantons with size @xmath42 from the wilson coefficients , @xmath43 ( @xmath44 is a soft step function ) . the instanton - induced @xmath45 dependence turns out to be relatively weak for @xmath46 , as necessary to compensate its perturbative counterpart . neglect of this renormalization , although common practice in perturbative wilson coefficients , would significantly contaminate the results , e.g. by missing the reduction of the direct - instanton density@xmath47 another important renormalization effect is the reduction of the instanton contributions to the pseudoscalar relative to the scalar sum rules . the instanton s self - duality causes a strongly repulsive contribution to the @xmath2 channel , with seemingly detrimental impact on the sum rules@xcite : the glueball signal disappears and both unitarity and the let ( [ pslet ] ) are badly violated . the origin of these problems can be traced to the neglect of topological charge screening@xcite . due to their high channel selectivity and a small screening length @xmath48 fm , the model - independent screening correlations@xcite @xmath49 ( @xmath50 is the overall topological charge density ( @xmath51 for instantons ) and @xmath52 the flavor - singlet part of the @xmath53 ) affect almost exclusively the @xmath2 wilson coefficients . they arise from the axial - anomaly induced attractive ( repulsive ) interaction between topological charge lumps of opposite ( equal ) sign due to @xmath52 exchange@xcite and , after correcting for @xmath54 mixing , add the terms @xmath55 to the pseudoscalar iope moments . they strongly reduce the direct - instanton induced repulsion and resolve the disastrous problems mentioned above@xcite : positivity of the spectral function is restored , the four @xmath2 borel sum rules ( [ csubtrbmoms ] ) are stable and contain consistent pseudoscalar glueball information . ( all previous analyses had discarded the @xmath24 sum rule and thereby missed valuable first - principle information and let consistency checks . ) we have implemented direct instanton and topological charge screening contributions into the ope coefficients of the spin-0 glueball correlators , evaluated their duality continuua and demonstrated how these contributions resolve the problems of previous qcd sum rule analyses@xcite . a lattice - based instanton size distribution and the gauge - invariant ir renormalization of the nonperturbative wilson coefficients were also introduced . quark admixtures , and thereby quarkonium mixing effects , enter through quark loops , the instanton size distribution and the condensates . in the scalar channel , the sizeable direct instanton contributions are indispensable for mutually and let consistent sum rules . their improved treatment reduces our earlier ( spike - distribution based ) result for the @xmath0 glueball mass to @xmath56 gev . ( the mass stays well beyond 1 gev , however , in contrast to obsolete predictions based on purely perturbative coefficients . ) this value is somewhat smaller than the quenched lattice results@xcite ( which will probably be reduced by light - quark effects ) and consistent with the broad glueball state found in a comprehensive @xmath57-matrix analysis@xcite . the systematics among our different borel moments likewise indicates a rather large width of the scalar glueball , @xmath58 gev . our prediction for the scalar glueball decay constant , @xmath59 gev , is several times larger than the value obtained when ignoring the nonperturbative wilson coefficients . this implies an exceptionally small @xmath0 glueball size , in agreement with several lattice results@xcite . another stringent , ope- and sum - rule - independent consistency check of the instanton contributions and their @xmath60 enhancement provide numerical simulations in an instanton ensemble@xcite ( ilm ) . for @xmath61 fm the scalar ilm and iope correlators are indeed very similar , and the ilm prediction @xmath62 gev is similarly large . this indicates a robust instanton effect and rules out that the large @xmath60 may signal some eventual internal inconsistencies in the treatment of the instanton contributions@xcite . a subsequent model - independent confirmation of the @xmath60 enhancement was supplied by the first direct ( quenched ) lattice result@xcite @xmath63 gev which is consistent both with our prediction and the ilm value ( the latter is practically independent of quark quenching@xcite ) . our prediction for @xmath60 implies enhanced partial widths for radiative @xmath64 and @xmath65 decays into scalar glueballs and is therefore relevant for experimental glueball searches , e.g. in the cleo and bes data on @xmath66 and other decay branches . since the exceptionally small size and large decay constant are particular to the _ scalar _ glueball and since only part of the scalar decay constant contributes to the radiative production rates@xcite , however , the above results are not ruled out by experimental data on decays into @xmath2 glueballs ( cf . ref . ) . in the pseudoscalar ( @xmath2 ) glueball correlator we have identified and implemented a new type of nonperturbative contributions to the wilson coefficients , due to topological charge screening . roughly speaking , the screening effects `` unquench '' the direct instanton contributions , thereby restoring unitarity , the axial ward identity and the resonance signals . consistency among all moment sum rules and with the underlying let is also achieved , and the resulting mass prediction @xmath67 gev lies inside the range of quenched and unquenched lattice data . the coupling @xmath68 gev is somewhat enhanced by the topological short - distance physics , affecting radiative production rates , the @xmath69 cross section at high momentum transfers and other glueball signatures . the crucial impact of the nonperturbative wilson coefficients on both spin-0 glueball correlators is particularly evident in the interplay between their subtraction constants . indeed , the notorious consistency problems which plagued previous @xmath0 glueball sum rules were primarily caused by the large let - induced subtraction constant @xmath70 gev@xmath71 ( cf . ( [ slet ] ) ) : it can not be matched by perturbative wilson coefficients ( since @xmath72 ) and requires the direct instanton contribution @xmath73 gev@xmath71 . at first sight this seems to imply a conflict with the much smaller let subtraction constant @xmath74 gev@xmath71 ( cf . ( [ pslet ] ) ) in the @xmath75 channel , however , since the instanton contributions to both spin-0 correlators are equal ( up to a sign ) . here the topological screening contributions @xmath76 gev@xmath71 from eq . ( [ rscr ] ) prove indispensable : they restore consistency by canceling most ( and in the chiral limit , where @xmath77 , all ) of the instanton contributions : @xmath78 . to summarize : contrary to naive expectation , the nonperturbative contributions to the ope of the spin-0 glueball correlators reside primarily in the wilson coefficients ( i.e not in the condensates ) and are closely related to the topological vacuum structure . the nonperturbative short - distance physics resolves long - standing consistency problems of the associated qcd sum rules and generates a rather diverse set of new glueball predictions . a large part of the @xmath0 glueball mass and binding originates from direct instantons , for example , while their net effects in the @xmath75 channel are smaller and more subtle , due to cancellations between instanton and topological charge screening contributions .	several key problems of qcd sum rules in the spin-0 glueball channels are resolved by implementing nonperturbative short - distance physics from direct instantons and topological charge screening . a lattice - based instanton size distribution and the ir renormalization of the nonperturbative wilson coefficients are also introduced . results of a comprehensive quantitative sum rule analysis are reviewed and their implications discussed .

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.