Entanglement in permutation symmetric states, fractal dimensions, and geometric quantum mechanics

TL;DR

This paper investigates the entanglement entropy scaling in symmetric many-body quantum states, revealing a connection between fractal dimensions of measures on complex projective spaces and entropy divergence, extending previous spin-1/2 results.

Contribution

It generalizes earlier findings to arbitrary spin-s systems, linking entanglement entropy to fractal dimensions and geometric entropy in quantum mechanics.

Findings

Entanglement entropy diverges as (d/2)log(m) for fractal support.

Support on submanifolds yields corrections related to geometric entropy.

Recovers and generalizes the spin-1/2 scaling law.

Abstract

We study the von Neumann and R\'enyi bipartite entanglement entropies in the thermodynamic limit of many-body quantum states with spin-s sites, that possess full symmetry under exchange of sites. It turns out that there is essentially a one-to-one correspondence between such thermodynamic states and probability measures on CP^{2s}. Let a measure be supported on a set of possibly fractal real dimension d with respect to the Study-Fubini metric of CP^{2s}. Let m be the number of sites in a subsystem of the bipartition. We give evidence that in the limit where m goes to infinity, the entanglement entropy diverges like (d/2)log(m). Further, if the measure is supported on a submanifold of CP^{2s} and can be described by a density f with respect to the metric induced by the Study-Fubini metric, we give evidence that the correction term is simply related to the entropy associated to f: the geometric entropy of geometric quantum mechanics. This extends results obtained by the authors in a recent letter where the spin-1/2 case was considered. Here we provide more examples as well as detailed accounts of the ideas and computations leading to these general results. For special choices of the state in the spin-s situation, we recover the scaling behaviour previously observed by Popkov et al., showing that their result is but a special case of a more general scaling law.