Exploring the $\theta$-vacuum structure in the functional renormalization group approach

TL;DR

This paper examines the $ heta$-vacuum structure and 't Hooft anomaly in a quantum mechanical system using the functional renormalization group, highlighting the method's limitations and the importance of nonlocal effects.

Contribution

It demonstrates the applicability and limitations of the local potential approximation in the fRG approach for capturing $ heta$-dependence and ground state degeneracy.

Findings

LPA reproduces ground state energy accurately at small $ heta$

LPA fails near energy level crossings and degeneracy points

Nonlocal effective actions are essential for capturing level crossing behavior

Abstract

We investigate the $\theta$-vacuum structure and the 't Hooft anomaly at $\theta=\pi$ in a simple quantum mechanical system on $S^1$ to scrutinize the applicability of the functional renormalization group (fRG) approach. Even though the fRG is an exact formulation, a naive application of the fRG equation would miss contributions from the $\theta$ term due to the differential nature of the formulation. We first review this quantum mechanical system on $S^1$ that is solvable with both the path integral and the canonical quantization. We discuss how to construct the quantum effective action including the $\theta$ dependence. Such an explicit calculation poses a subtle question of whether a Legendre transform is well defined or not for general systems with the sign problem. We then consider a deformed theory to relax the integral winding by introducing a wine-bottle potential with the finite depth $\propto g$, so that the original $S^1$ theory is recovered in the $g\to\infty$ limit. We numerically solve the energy spectrum in the deformed theory as a function of $g$ and $\theta$ in the canonical quantization. We test the efficacy of the simplest local potential approximation (LPA) in the fRG approach and find that the correct behavior of the ground state energy is well reproduced for small $\theta$. When the energy level crossing is approached, the LPA flow breaks down and fails in describing the ground state degeneracy expected from the 't Hooft anomaly. We finally turn back to the original theory and discuss an alternative formulation using the Villain lattice action. The analysis with the Villain lattice at $\theta=\pi$ indicates that the nonlocality of the effective action is crucial to capture the level crossing behavior of the ground states.