Systems and Control Protocols for Neutral-Atom-Array Quantum Processors

Abstract

Neutral atom arrays are a leading platform for programmable quantum processors, offering individual qubit addressability, long-lived hyperfine ground states, and strong Rydberg interactions. Recent progress has demonstrated coherent control over thousands of atoms. However, achieving scalable control requires precise mitigation of environmental and hardware imperfections that degrade gate performance.

This thesis presents an integrated neutral-atom array platform built from the ground up that incorporates quantum sensing directly into the processor. Each atom functions both as a qubit and a local magnetometer. We design, build, and characterize from first principles three subsystems: 1) a microwave control system for driving hyperfine transitions in ground-state rubidium atoms; 2) a Raman laser system for site-selective single-qubit gates; and 3) a Rydberg laser system with quantum optimal control for robust two-qubit gates. This work provides a universal gate set and quantifies which error sources limit performance.

First, we develop an in-situ magnetic field imaging technique using the atom array as a quantum sensor. Through site-resolved Ramsey spectroscopy, we image magnetic fields across a 260 μm × 160 μm region with 3 μm spatial resolution. We then apply computed corrections that compensate for the bias magnetic fields, producing uniform global microwave single-qubit rotations.

Second, we introduce a hardware-aware simulation framework to evaluate Raman laser systems for hyperfine qubit manipulation. Simulations predict a single-qubit gate infidelity of 4.4 × 10⁻⁴ using BB1 composite pulses to mitigate thermal motion errors. We validate the Raman laser system by building and characterizing its phase noise.

Third, we develop a Rydberg laser system for high-fidelity entangling gates. We apply linear response theory to map laser phase noise to single-atom Rydberg excitation fidelity. We then demonstrate fast phase-noise engineering by optimizing laser servo parameters. We employ hardware-aware quantum optimal control to design both Rydberg excitation and two-qubit gate pulses with built-in robustness against physical and control parameter fluctuations, outperforming analytical benchmarks.

This integrated platform demonstrates high-fidelity universal control of neutral-atom registers with hundreds of qubits. By systematically addressing environmental inhomogeneities through integrated sensing and hardware-aware control design, this work provides a validated path for scaling quantum processors while maintaining gate fidelity.