Estudo de transições de fases em um antiferromagneto frustrado através do grupo de renormalização na representação de redes de tensores

Abstract

The tensor network renormalization group is a theoretical framework for the analysis of strong interacting physical systems. The technique is based on a representation of the system partition function by a tensor network, in which one associates to each of its site a translationally invariant tensor. The tensor encodes the states associated with the degrees of freedom of system. The indexes of the tensors, which are called “legs”, correspond to the links between the sites of the network. Thus, the calculation of the partition function is reduced to the contraction of a tensor network. That is, a sum over the legs common to each pair of tensors. Here, we apply the technique to the Ising model defined in a square lattice with antiferromagnetic interactions between the pairs of first neighbors and the pairs of second neighbors. In this case the square lattice is fully frustrated. In the process of renormalization of the tensor network, we introduced a scale factor that avoids unrestricted growth of the tensor norm. This scale factor captures any singularity presents in the free energy. Thus, the critical values of the parameters were obtained by locating the point at which the scale factor presents a singularity. In this way, we obtain the phase diagram of the model in the plane temperature ratio between the coupling constants between first and second neighbors.