Efficient readout of superconducting qubits

Abstract

Quantum computers have the potential to revolutionize various fields, from cryptography to material science, by employing the principles of quantum mechanics. At the core of these quantum computers are quantum processors, where the ability to perform fast and reliable manipulation and readout of qubit states is crucial for their overall performance. This thesis presents a systematic approach to optimize the readout of superconducting transmon qutrits—quantum systems with three states that offer a more expansive computational space compared to traditional qubits. The study involves simulating a transmon-resonator system and exploring two key approaches: the optimization of integration weights and the use of frequency-modulated readout pulses.

The first method, integration weights optimization, yielded significant improvements in the readout process, with an enhancement of up to 60%. This technique is most successful when a proper balance between pulse duration and coupling strength is maintained. However, its effectiveness diminishes as coupling between the resonator and the environment increases, with improvements becoming more limited to a narrower parameter space.

In contrast, the second strategy, frequency modulation of the readout pulse, only resulted in a slight improvement under certain parameter ranges, with the best performance is in the region with high coupling and short duration. While modulated pulses can improve state distinguishability in naturally low distinguishability regimes, they require careful optimization of the modulation depth and readout frequency offset, introducing additional complexity and potential time costs in implementation.