Extracting critical exponents for sequences of numerical data via series extrapolation techniques.

Abstract

We describe a generic scheme to extract critical exponents of quantum lattice models from sequences of numerical data, which is, for example, relevant for nonperturbative linked-cluster expansions or nonperturbative variants of continuous unitary transformations. The fundamental idea behind our approach is a reformulation of the numerical data sequences as a series expansion in a pseudoparameter. This allows us to utilize standard series expansion extrapolation techniques to extract critical properties such as critical points and critical exponents. The approach is illustrated for the deconfinement transition of the antiferromagnetic spin-1/2 Heisenberg chain.