Integrability and conformal data of the dimer model

TL;DR

This paper provides evidence that the dimer model on a square lattice is best described by a conformal field theory with central charge c=-2, analyzing its integrability and conformal data through algebraic and scaling limit methods.

Contribution

The authors introduce a new solution of the dimer model using the Temperley-Lieb algebra and connect it to the critical dense polymers model, clarifying its conformal structure and integrability.

Findings

Supports c=-2 conformal description of the dimer model.

Constructs Virasoro modes as limits of Temperley-Lieb algebra elements.

Eigenvalues of lattice integrals match c=-2 conformal integrals.

Abstract

The central charge of the dimer model on the square lattice is still being debated in the literature. In this paper, we provide evidence supporting the consistency of a $c=-2$ description. Using Lieb's transfer matrix and its description in terms of the Temperley-Lieb algebra $TL_n$ at $\beta = 0$, we provide a new solution of the dimer model in terms of the model of critical dense polymers on a tilted lattice and offer an understanding of the lattice integrability of the dimer model. The dimer transfer matrix is analysed in the scaling limit and the result for $L_0-\frac c{24}$ is expressed in terms of fermions. Higher Virasoro modes are likewise constructed as limits of elements of $TL_n$ and are found to yield a $c=-2$ realisation of the Virasoro algebra, familiar from fermionic $bc$ ghost systems. In this realisation, the dimer Fock spaces are shown to decompose, as Virasoro modules, into direct sums of Feigin-Fuchs modules, themselves exhibiting reducible yet indecomposable structures. In the scaling limit, the eigenvalues of the lattice integrals of motion are found to agree exactly with those of the $c=-2$ conformal integrals of motion. Consistent with the expression for $L_0-\frac c{24}$ obtained from the transfer matrix, we also construct higher Virasoro modes with $c=1$ and find that the dimer Fock space is completely reducible under their action. However, the transfer matrix is found not to be a generating function for the $c=1$ integrals of motion. Although this indicates that Lieb's transfer matrix description is incompatible with the $c=1$ interpretation, it does not rule out the existence of an alternative, $c=1$ compatible, transfer matrix description of the dimer model.