English

Engineering random spin models with atoms in a high-finesse cavity

Quantum Gases 2023-08-21 v1 Disordered Systems and Neural Networks Atomic Physics Quantum Physics

Abstract

All-to-all interacting, disordered quantum many-body models have a wide range of applications across disciplines, from spin glasses in condensed-matter physics, over holographic duality in high-energy physics, to annealing algorithms in quantum computing. Typically, these models are abstractions that do not find unambiguous physical realisations in nature. Here, we realise an all-to-all interacting, disordered spin system by subjecting an atomic cloud in a cavity to a controllable light shift. Adjusting the detuning between atom resonance and cavity mode, we can tune between disordered versions of a central-mode model and a Lipkin-Meshkov-Glick model. By spectroscopically probing the low-energy excitations of the system, we explore the competition of interactions with disorder across a broad parameter range. We show how disorder in the central-mode model breaks the strong collective coupling, making the dark state manifold cross over to a random distribution of weakly-mixed light-matter, "grey", states. In the Lipkin-Meshkov-Glick model the ferromagnetic finite-size ground state evolves towards a paramagnet as disorder is increased. In that regime, semi-localised eigenstates emerge, as we observe by extracting bounds on the participation ratio. These results present significant steps towards freely programmable cavity-mediated interactions for the design of arbitrary spin Hamiltonians.

Keywords

Cite

@article{arxiv.2208.09421,
  title  = {Engineering random spin models with atoms in a high-finesse cavity},
  author = {Nick Sauerwein and Francesca Orsi and Philipp Uhrich and Soumik Bandyopadhyay and Francesco Mattiotti and Tigrane Cantat-Moltrecht and Guido Pupillo and Philipp Hauke and Jean-Philippe Brantut},
  journal= {arXiv preprint arXiv:2208.09421},
  year   = {2023}
}

Comments

8 pages, 4 figures, methods, supplementary material

R2 v1 2026-06-25T01:49:34.486Z