Nuclear-electronic spin systems, magnetic resonance, and quantum information processing

TL;DR

This paper develops a predictive model for nuclear-electronic spin systems in silicon impurities, demonstrating its accuracy with experimental data and highlighting Si:Bi's advantages for quantum computing due to larger nuclear spin and hyperfine interaction.

Contribution

The paper introduces a model for predicting magnetic resonance properties of nuclear-electronic spin systems, validated with experimental data, and identifies unique advantages of Si:Bi for quantum information processing.

Findings

Model accurately predicts magnetic resonance properties of Si:Bi.

Si:Bi offers larger hyperfine strength and nuclear spin than Si:P.

Optimal working points exist in Si:Bi but not in Si:P.

Abstract

A promising platform for quantum information processing is that of silicon impurities, where the quantum states are manipulated by magnetic resonance. Such systems, in abstraction, can be considered as a nucleus of arbitrary spin coupled to an electron of spin one-half via an isotropic hyperfine interaction. We therefore refer to them as "nuclear-electronic spin systems". The traditional example, being subject to intensive experimental studies, is that of phosphorus doped silicon (Si:P) which couples a spin one-half electron to a nucleus of the same spin, with a hyperfine strength of 117.5 MHz. More recently, bismuth doped silicon (Si:Bi) has been suggested as an alternative instantiation of nuclear-electronic spin systems, differing from Si:P by its larger nuclear spin and hyperfine strength of 9/2 and 1.4754 GHz respectively. The aim of this thesis has been to develop a model that is capable of predicting the magnetic resonance properties of nuclear-electronic spin systems. The theoretical predictions of this model have been tested against experimental data collected on Si:Bi at 4.044 GHz, and have proven quite successful. Furthermore, the larger nuclear spin and hyperfine strength of Si:Bi, compared with that of Si:P, are predicted to offer advantages for quantum information processing. Most notable amongst these is that magnetic field-dependent two-dimensional decoherence free subspaces, called optimal working points, have been identified to exist in Si:Bi, but not Si:P.