Noise Effects on the Wilczek-Zee Geometric Phase

TL;DR

This paper analyzes how different types of noise affect the non-abelian Wilczek-Zee geometric phase, proposing a measure to quantify noise impact and revealing unique effects at specific noise frequencies.

Contribution

It introduces a state-independent measure for noise effects on the Wilczek-Zee phase and analyzes the impact of sinusoidal noise, highlighting a special case at frequency m=2.

Findings

Noise of frequency m ≠ 2 behaves similarly to the abelian case.

Noise at frequency m=2 has a distinct and pronounced effect.

The proposed measure helps quantify noise impact on geometric phases.

Abstract

Non-abelian geometric phases have been proposed as an essential ingredient in logical gate implementation -- their geometric nature guarantees their invariance under reparametrizations of the associated cyclic path in parameter space. However, they are still dependent on deformations of that path, due to, e.g., noise. The first question that we tackle in this work is how to quantify in a meaningful way this effect of noise, focusing, for concreteness, on the nuclear quadrupole resonance hamiltonian -- other systems of this nature can clearly be treated analogously. We consider a precessing magnetic field that drives adiabatically a degenerate doublet, and is subjected to noise, the effects of which on the Wilczek-Zee holonomy are computed analytically. A critical review of previous related works reveals a series of assumptions, like sudden jumps in the field, or the presence of white noise, that might violate adiabaticity. We propose a state-independent measure of the effect, and then consider sinusoidal noise in the field, of random amplitude and phase. We find that all integer noise frequencies $m \neq 2$ behave similarly, in a manner reminiscent of the abelian case, but that noise of frequency $m = 2$ has a very different, and, at the same time, very pronounced effect, that might well affect robustness estimations.