Individual Qubit Manipulation Using Dual Acousto-Optic Deflectors in a Trapped Ion Quantum Processor

Abstract

Quantum processors offer computational advantages over classical computers by lever- aging quantum phenomena such as superposition and entanglement to efficiently solve certain classes of problems. Trapped-ion systems represent one of the leading platforms for quantum information processing, owing to their long coherence times, high-fidelity state preparation and measurement, high-fidelity quantum gates, and native all-to-all qubit con- nectivity. However, scaling these processors introduces significant challenges, including the need for increasingly complex control systems and improved stabilization against en- vironmental noise. In particular, scalable trapped-ion architectures require the ability to individually address and precisely control selected subsets of qubits while maintaining high-fidelity operations. Here I present my contributions toward the development and characterization of an individual ion addressing system for a scalable trapped-Ytterbium ion quantum processor. A common method of implementing coherent gates in 171Yb+ is via laser beams that are tuned in frequency far away from any atomic transition, relying on a stimulated Raman transition instead. Additionally, the light momentum is capable of coupling to the motional degrees of freedom of the ion, creating the ‘bus of entanglement’. By investigating the im- pact of experimental imperfections on single-qubit gate fidelities, I establish the tolerances that such an optical individual addressing system must satisfy to achieve high-fidelity gate operation. Based on these requirements, I develop an acousto-optic deflector (AOD)-based optical architecture that provides fast, programmable spatial control of addressing laser beams. The design is validated through optical simulations before being experimentally implemented and characterized in comparison to the requirements established for main- taing high-fidelity gates. Based on our trap parameters, I verify that the AOD optical relay is capable of providing an array of beams spaced apart by∼4 µm, though we have the controls required to match the beams to a non-uniformly spaced chain of ions. The global beam drift of this array of beams is intrinsically less than the proposed beam waist at the ion location, owing to the mechanical engineering practices employed in the optics design. Though the optical cross-talk at an intermediate image plane is greater than design re- quirements, there is room for improvement via custom-optics solutions or active aberration compensation. Finally, I model the decoherence of the qubit state due to heating up of ions along the chain, as these axial vibrational frequencies are lower and often hard to cool to their ground states – a necessity for high-fidelity quantum operations. In this thesis, I propose an optical solution that enables the real-time coupling of the laser momentum to the axial motional modes via a commerically available galvonometer mirror.

Overall, the optical scheme presented in this thesis should be able to individually ad- dress a linear chain of around 50 ions, spaced by approximately 4 µm. This is greater than the anticipated number of useable qubits in this processor, where the limiting factors are likely dominated by diminishing fidelity in coherent operations due to spectral crowding of the vibrational modes as the system size scales up.