Towards heat-bath algorithmic cooling in a superconducting circuit

Abstract

Despite being a fundamental requirement for reliable and efficient quantum computing, qubit initialization remains a major challenge, even with current state-of-the-art qubit reset techniques [1]. To address this challenge, heat-bath algorithmic cooling provides a theoretical framework for re-distributing entropy within a qubit system to initialize a subset of these into their ground states [2]. Here, we present a proofof- concept for the experimental implementation of extended heat-bath algorithmic cooling on a superconducting two-qubit chip connected by a tunable coupler [3]. The cooling is realized through a β-SWAP, which transfers entropy from the target qubit to the coupler. To achieve optimal cooling over repeated cycles, each β-SWAP is followed by a resonator-assisted coupler re-thermalization by dissipating the extracted entropy to the environment. We experimentally demonstrate the β-SWAP and re-thermalization, achieving over 99 % population transfer for the β-SWAP and a coupler re-thermalization reaching a 4 % residual excited state population in 3 μs. Finally, we also test the full cooling protocol. From the results, we cannot observe any definite cooling effect, and the experimental data does not follow the theoretical predictions. By optimizing the device parameters and employing more advanced calibration techniques for the β-SWAP and re-thermalization, we expect to achieve the cooling effect, which would serve as an important step towards optimal qubit initialization in superconducting circuits.