Mobility edges in pseudo-unitary quasiperiodic quantum walks

TL;DR

This paper introduces a Floquet quasicrystal model with non-reciprocal hopping, revealing a novel mobility edge and unique phase transitions related to pseudo-unitarity and spectral properties.

Contribution

It presents the first observation of a mobility edge in a pseudo-unitary quasiperiodic quantum walk and identifies a new discrete-time specific transition.

Findings

Discovered a sharp mobility edge dividing metallic and insulating phases.

Identified a unique transition in the discrete-time setting.

Related phase transitions to spectral properties and symmetry breaking.

Abstract

We introduce a Floquet quasicrystal that simulates the motion of Bloch electrons in a homogeneous magnetic field in discrete time steps. We admit the hopping to be non-reciprocal which, via a generalized Aubry duality, leads us to push the phase that parametrizes the synthetic dimension off of the real axis. This breaks unitarity, but we show that the model is still ``pseudo-unitary''. We unveil a novel mobility edge between a metallic and an insulating phase that sharply divides the parameter space. Moreover, for the first time, we observe a second transition that appears to be unique to the discrete-time setting. We quantify both phase transitions and relate them to properties of the spectrum. If the hopping is reciprocal either in the lattice direction or the synthetic dimension, the model is $\mathcal{PT}$-symmetric, and the spectrum is confined to the unit circle up to a critical point. At this critical point, $\mathcal{PT}$-symmetry is spontaneously broken and the spectrum leaves the unit circle. This transition is topological and measured by a spectral winding number.