An integrable modification of the critical Chalker-Coddington network model

TL;DR

This paper introduces an exactly solvable, integrable modification of the Chalker-Coddington network model for the Integer Quantum Hall Effect, revealing new critical behaviors and universality classes through analytical and numerical methods.

Contribution

It develops a truncated, integrable version of the Chalker-Coddington model, connecting it to known algebraic structures and identifying its distinct universality class.

Findings

Identifies four integrable branches related to braid-monoid algebra

Two branches correspond to decoupled Coulomb-Gas theories

The truncated model's universality class differs from the original model

Abstract

We consider the Chalker-Coddington network model for the Integer Quantum Hall Effect, and examine the possibility of solving it exactly. In the supersymmetric path integral framework, we introduce a truncation procedure, leading to a series of well-defined two-dimensional loop models, with two loop flavours. In the phase diagram of the first-order truncated model, we identify four integrable branches related to the dilute Birman-Wenzl-Murakami braid-monoid algebra, and parameterised by the loop fugacity $n$. In the continuum limit, two of these branches (1,2) are described by a pair of decoupled copies of a Coulomb-Gas theory, whereas the other two branches (3,4) couple the two loop flavours, and relate to an $SU(2)_r \times SU(2)_r / SU(2)_{2r}$ Wess-Zumino-Witten (WZW) coset model for the particular values $n= -2\cos[\pi/(r+2)]$ where $r$ is a positive integer. The truncated Chalker-Coddington model is the $n=0$ point of branch 4. By numerical diagonalisation, we find that its universality class is neither an analytic continuation of the WZW coset, nor the universality class of the original Chalker-Coddington model. It constitutes rather an integrable, critical approximation to the latter.