What is it about?

Quantum dots are nanoscale experiments enabling the precise control of single electrons at the quantum scale. This allows them to be used as components for a variety of novel quantum technologies, ranging from ultra-precise sensors to quantum computers. Nonetheless, their practical application is far from trivial and relies on an accurate calibration in a time-consuming interplay between device measurement and fine-tuning of experimental parameters. Our work presents an algorithm for an automated characterization of single-electron pumps, which enable a robust transfer of a fixed number of electrons per time interval through a quantum dot. Our algorithm leverages the power of machine learning, allowing us to reduce the time required to extract the relevant measurement information by a factor of eight over conventional approaches.

Featured Image

Why is it important?

Due to their ability to transfer well-defined electron numbers through the device, single-electron pumps are ideal candidates for the realization of highly precise reference currents according to the fundamental definition of the ampere. Our work paves the way for the scalable application of such technologies, requiring a stable co-operation of many single devices to produce currents on relevant scales. This is an important step to establish more accurate calibration capabilities for high-precision measurement devices, and probe the relationship between fundamental constants defining our measurement system with higher accuracies. Our algorithm demonstrates how machine learning can be a useful aid to overcome bottlenecks in the practical realization of quantum technologies.

Perspectives

This work has been a fruitful collaboration at the intersection of computational algorithm development and the practical reality of quantum experiments. I believe this article provides a really novel perspective on the question: "What are the minimal measurements needed to characterize the performance of a quantum device?". I hope our algorithm will prove to be a useful tool of practical significance, streamlining laborious calibration procedures and reducing overall time spend in the fine-tuning of experiments.

Yannic Rath
National Physical Laboratory

Read the Original

This page is a summary of: Fast characterization of multiplexed single-electron pumps with machine learning, Frontiers in Human Neuroscience, September 2024, American Institute of Physics,
DOI: 10.1063/5.0221387.
You can read the full text:

Read

Contributors

The following have contributed to this page