We calculate high-temperature graph expansions for the Ising spin glass model with 4 symmetric random distribution functions for its nearest neighbor interaction constants J ij . Series for the Edwards-Anderson susceptibility χ EA are obtained to order 13 in the expansion variable (J/(k B T )) 2 for the general d-dimensional hyper-cubic lattice, where the parameter J determines the width of the distributions. We explain in detail how the expansions are calculated. The analysis, using the Dlog-Padé approximation and the techniques known as M1 and M2, leads to estimates for the critical threshold (J/(k B T c )) 2 and for the critical exponent γ in dimensions 4, 5, 7 and 8 for all the distribution functions. In each dimension the values for γ agree, within their uncertainty margins, with a common value for the different distributions, thus confirming universality.
We study the site and bond quantum percolation model on the two-dimensional square lattice using series expansion in the low concentration limit. We calculate series for the averages of P ij r k ij Tij(E), where Tij(E) is the transmission coefficient between sites i and j, for k = 0, 1, . . . , 5 and for several values of the energy E near the center of the band. In the bond case the series are of order p 14 in the concentration p (some of those have been formerly available to order p 10 ) and in the site case of order p 16 . The analysis, using the Dlog-Padé approximation and the techniques known as M1 and M2, shows clear evidence for a delocalization transition (from exponentially localized to extended or power-law-decaying states) at an energy-dependent threshold pq(E) in the range pc < pq(E) < 1, confirming previous results (e.g. pq(0.05) = 0.625 ± 0.025 and 0.740 ± 0.025 for bond and site percolation) but in contrast with the Anderson model. The divergence of the series for different k is characterized by a constant gap exponent, which is identified as the localization length exponent ν from a general scaling assumption. We obtain estimates of ν = 0.57 ± 0.10. These values violate the bound ν ≥ 2/d of Chayes et al.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.