Fully autonomous tuning of a spin qubit

TL;DR

Fully autonomous tuning of a spin qubit tackles the needle-in-a-haystack problem in semiconductor qubits by combining deep learning, Bayesian optimization, and computer vision to autonomously configure a Ge/Si core/shell nanowire device. The authors implement a modular, stage-based framework that defines a double quantum dot, tunes tunnel barriers, detects Pauli Spin Blockade, and identifies readout resonances, while introducing an efficient measurement algorithm to accelerate stability-diagram data collection. Ten autonomous runs demonstrate qubit operation and enable characterization of how the Rabi frequency $f_{\text{Rabi}}$ and the g-factor $g$ depend on barrier gate voltages, illustrating the method's ability to navigate device variability. This approach enables scalable, statistically driven studies of qubit quality metrics and could catalyze the operation of large semiconductor quantum circuits.

Abstract

Spanning over two decades, the study of qubits in semiconductors for quantum computing has yielded significant breakthroughs. However, the development of large-scale semiconductor quantum circuits is still limited by challenges in efficiently tuning and operating these circuits. Identifying optimal operating conditions for these qubits is complex, involving the exploration of vast parameter spaces. This presents a real 'needle in the haystack' problem, which, until now, has resisted complete automation due to device variability and fabrication imperfections. In this study, we present the first fully autonomous tuning of a semiconductor qubit, from a grounded device to Rabi oscillations, a clear indication of successful qubit operation. We demonstrate this automation, achieved without human intervention, in a Ge/Si core/shell nanowire device. Our approach integrates deep learning, Bayesian optimization, and computer vision techniques. We expect this automation algorithm to apply to a wide range of semiconductor qubit devices, allowing for statistical studies of qubit quality metrics. As a demonstration of the potential of full automation, we characterise how the Rabi frequency and g-factor depend on barrier gate voltages for one of the qubits found by the algorithm. Twenty years after the initial demonstrations of spin qubit operation, this significant advancement is poised to finally catalyze the operation of large, previously unexplored quantum circuits.

Figures (3)

• Figure S1: Qubit measurements of all successful runs. The panels a,-j, are ordered by the total run time of the algorithm for each qubit respectively. Each panel includes four current measurements: the pair of bias triangles (upper left), spectroscopy measurement, varying magnetic field and driving frequency (upper right), Rabi chevron pattern, varying magnetic field and burst duration (lower left), and averaged Rabi oscillations (lower right) taken at the dashed lines in the Rabi chevron measurement. All measurements were performed autonomously. The Rabi chevron measurement does not have a dedicated re-centering stage, accounting for the off-centered measurements. The spectroscopy measurements were purely taken for documentation and always with the same ranges; these measurements did not inform any other part of the algorithm. Some measurements for panels d, e, and f were taken again using automated measurements after the initial runs finished because a setting of the lock-in amplifier led to slight measurement artifacts.
• Figure S2: Examples of measurements taken with the efficient measurement algorithm. Areas where no measurement have been taken have been filled with the threshold value.
• Figure S3: Examples of search trees from full runs. The fastest run only has two branches and then successfully found a qubit. The slowest run explored much more, with several branches reaching all the way to qubit measurements. However, only the last branch shows conclusive qubit signatures. We rejected the first tries as noise.