Thermodynamics and phase transition of the O(N) model from the two-loop Phi-derivable approximation

TL;DR

This paper investigates the thermodynamics and phase transition nature of the O(N) model using a two-loop Phi-derivable approximation, revealing a second-order transition in the chiral limit and ensuring Goldstone's theorem is satisfied.

Contribution

It introduces a two-loop Phi-derivable approximation for the O(N) model, demonstrating a second-order phase transition and improved renormalization over previous methods.

Findings

Transition is second order in the chiral limit

Goldstone's theorem is obeyed in the broken phase

Sigma mass around 450 MeV in the N=4 case

Abstract

We discuss the thermodynamics of the O(N) model across the corresponding phase transition using the two-loop Phi-derivable approximation of the effective potential and compare our results to those obtained in the literature within the Hartree-Fock approximation. In particular, we find that in the chiral limit the transition is of the second order, whereas it was found to be of the first order in the Hartree-Fock case. These features are manifest at the level of the thermodynamical observables. We also compute the thermal sigma and pion masses from the curvature of the effective potential. In the chiral limit, this guarantees that the Goldstone's theorem is obeyed in the broken phase. A realistic parametrization of the model in the N=4 case, based on the vacuum values of the curvature masses, shows that a sigma mass of around 450 MeV can be obtained. The equations are renormalized after extending our previous results for the N=1 case by means of the general procedure described in [J. Berges et al., Annals Phys. 320, 344-398 (2005)]. When restricted to the Hartree-Fock approximation, our approach reveals that certain problems raised in the literature concerning the renormalization are completely lifted. Finally, we introduce a new type of Phi-derivable approximation in which the gap equation is not solved at the same level of accuracy as the accuracy at which the potential is computed. We discuss the consistency and applicability of these types of "hybrid" approximations and illustrate them in the two-loop case by showing that the corresponding effective potential is renormalizable and that the transition remains of the second order.