arXiv Analytics

Sign in

arXiv:2307.07634 [math.PR]AbstractReferencesReviewsResources

Features of a spin glass in the random field Ising model

Sourav Chatterjee

Published 2023-07-14Version 1

A longstanding open question in the theory of disordered systems is whether short-range models, such as the random field Ising model or the Edwards-Anderson model, can indeed have the famous properties that characterize mean-field spin glasses at nonzero temperature. This article shows that this is at least partially possible in the case of the random field Ising model. Consider the Ising model on a discrete $d$-dimensional cube under free boundary condition, subjected to a very weak i.i.d. random external field, where the field strength is inversely proportional to the square-root of the number of sites. It turns out that in $d \ge 2$ and at sufficiently low temperatures, this model has some of the key features of a mean-field spin glass. Namely, (a) the site overlap exhibits one step of replica symmetry breaking, (b) the quenched distribution of the overlap is non-self-averaging, and (c) the overlap has the Parisi ultrametric property. Furthermore, it is shown that for Gaussian disorder, replica symmetry does not break if the field strength is taken to be stronger than the one prescribed above, and non-self-averaging fails if it is weaker, showing that the above order of field strength is the only one that allows all three properties to hold. However, the model does not have two other features of mean-field models. Namely, (a) it does not satisfy the Ghirlanda-Guerra identities, and (b) it has only two pure states instead of many.

Related articles: Most relevant | Search more
arXiv:2212.14571 [math.PR] (Published 2022-12-30)
Hypergraph Counting and Mixed $p$-Spin Glass Models under Replica Symmetry
arXiv:0908.1920 [math.PR] (Published 2009-08-13, updated 2009-12-01)
Replica Symmetry and Combinatorial Optimization
arXiv:2211.09879 [math.PR] (Published 2022-11-17)
Existence of the free energy for heavy-tailed spin glasses