Quantum and classical localisation and the Manhattan lattice
Abstract
We consider a network model, embedded on the Manhattan lattice, of a quantum localisation problem belonging to symmetry class C. This arises in the context of quasiparticle dynamics in disordered spin-singlet superconductors which are invariant under spin rotations but not under time reversal. A mapping exists between problems belonging to this symmetry class and certain classical random walks which are self-avoiding and have attractive interactions; we exploit this equivalence, using a study of the classical random walks to gain information about the corresponding quantum problem. In a field-theoretic approach, we show that the interactions may flow to one of two possible strong coupling regimes separated by a transition: however, using Monte Carlo simulations we show that the walks are in fact always compact two-dimensional objects with a well-defined one-dimensional surface, indicating that the corresponding quantum system is localised.
Cite
@article{arxiv.cond-mat/0210359,
title = {Quantum and classical localisation and the Manhattan lattice},
author = {E. J. Beamond and A. L. Owczarek and John Cardy},
journal= {arXiv preprint arXiv:cond-mat/0210359},
year = {2007}
}
Comments
11 pages, 8 figures