Towards Autonomous Tuning of Double Quantum Dots

Defining quantum dots in semiconductor or semiconductor/superconductor heterostructures is an essential step in initializing solid-state qubits. To date, the majority of the repetitive procedure of finding suitable gate voltages defining a quantum dot is done manually. However, as the complexity of devices as well as the number of devices per chip grows, this approach becomes unfeasible.
Current approaches either require significant, manually measured input, or address specific parts of the tune-up only. In both cases, only one well known device has been addressed.
In this talk we present an alternative autonomous automated solution for the tuning of gate-defined quantum dots which requires as input only the device layout. Our approach is able to characterize unknown devices and determine suitable gate voltages defining either a single or double quantum dot. We also discuss the extraction of the full capacitance matrix and performance of different classifiers when applied to both theoretical and experimental data.