Coherent quantum annealing in a programmable 2000-qubit Ising chain
Abstract
Quantum simulation has emerged as a valuable arena for demonstrating and understanding the capabilities of near-term quantum computers. Quantum annealing has been used successfully in simulating a range of open quantum systems, both at equilibrium and out of equilibrium. However, in all previous experiments, annealing has been too slow to simulate a closed quantum system coherently, due to the onset of thermal effects from the environment. Here we demonstrate coherent evolution through a quantum phase transition in the paradigmatic setting of the 1D transverse-field Ising chain, using up to 2000 superconducting flux qubits in a programmable quantum annealer. In large systems we observe the quantum Kibble-Zurek mechanism with theoretically predicted kink statistics, as well as characteristic positive kink-kink correlations, independent of system temperature. In small chains, excitation statistics validate the picture of a Landau-Zener transition at a minimum gap. In both cases, results are in quantitative agreement with analytical solutions to the closed-system quantum model. For slower anneals we observe anti-Kibble-Zurek scaling in a crossover to the open quantum regime. These experiments demonstrate that large-scale quantum annealers can be operated coherently, paving the way to exploiting coherent dynamics in quantum optimization, machine learning, and simulation tasks.
Cite
@article{arxiv.2202.05847,
title = {Coherent quantum annealing in a programmable 2000-qubit Ising chain},
author = {Andrew D. King and Sei Suzuki and Jack Raymond and Alex Zucca and Trevor Lanting and Fabio Altomare and Andrew J. Berkley and Sara Ejtemaee and Emile Hoskinson and Shuiyuan Huang and Eric Ladizinsky and Allison MacDonald and Gaelen Marsden and Travis Oh and Gabriel Poulin-Lamarre and Mauricio Reis and Chris Rich and Yuki Sato and Jed D. Whittaker and Jason Yao and Richard Harris and Daniel A. Lidar and Hidetoshi Nishimori and Mohammad H. Amin},
journal= {arXiv preprint arXiv:2202.05847},
year = {2022}
}