Path integral approach for predicting the diffusive statistics of geometric phases in chaotic Hamiltonian systems

Abstract

From the integer quantum Hall effect, to swimming at low Reynolds number, geometric phases arise in the description of many different physical systems. In many of these systems the temporal evolution prescribed by the geometric phase can be directly measured by an external observer. By definition, geometric phases rely on the history of the system's internal dynamics, and so their measurement is directly related to temporal correlations in the system. They, thus, provide a sensitive tool for studying chaotic Hamiltonian systems. In this work we present a toy model consisting of an autonomous, low-dimensional, chaotic Hamiltonian system designed to have a simple planar internal state space, and a single geometric phase. The diffusive phase dynamics in the highly chaotic regime is thus governed by the loop statistics of planar random walks. We show that the na\"ive loop statistics result in ballistic behavior of the phase, and recover the diffusive behavior by considering a bounded shape space, or a quadratic confining potential.