For over forty years, studies of chaos in isolated quantum systems have relied on universal signatures such as level repulsion and eigenstate thermalization. Open systems interacting with their environment have long remained a theoretical domain: predictions about dissipative quantum chaos awaited experimental verification. Alongside this work, the archive contains studies of quantum chaos in closed systems, time crystals, and the effect of disorder on measurements, but none of them go beyond unitary dynamics.
Two circuits were implemented on a five-qubit superconducting processor chain: an integrable (free-fermionic) one and a chaotic one. Four qubits served as the system, and one as an auxiliary environment, which was traced out after evolution. Gradient tomography was used to reconstruct the quantum map, accounting for state preparation and measurement errors. The map's eigenvalues were diagonalized, and complex ratios of distances to nearest neighbors were computed. Pauli twirling was applied to suppress coherent errors.
For the chaotic circuit, the distribution of complex ratios of distances has a "bitten bagel" shape: density is suppressed near the center and the positive real axis, indicating level repulsion. For the integrable circuit, a sharp peak at zero is observed, indicating no correlations. As the depth of the integrable circuit increases from 5 to 50 layers, the distribution smoothly transitions from a peak to a "bagel," demonstrating a crossover from integrability to chaos. Quantitatively, the mean values ⟨r⟩ and ⟨−cos θ⟩ shift from free-fermionic statistics toward the random matrix class AI.
This work makes dissipative quantum chaos experimentally observable for the first time. Unlike neighboring studies focused on closed systems, here noise is not an obstacle but a tool: it generates chaotic dynamics and allows control over them. This opens the way to using quantum processors as testbeds for open many-body phenomena.
The method can be extended to study dissipative many-body localization, phase transitions in open systems, and various symmetry classes of quantum channels. It may also enable an operational metric for benchmarking quantum platforms based on their susceptibility to dissipative chaos.
The results will influence the development of error correction methods and quantum simulation strategies, where chaos inevitably affects computational outcomes.
Test the universality of the observed distributions on a larger number of qubits and for other architectures. Investigate the role of symmetries and types of dissipation in the transition to chaos.
Connection to unsolved problems in quantum physics: how universal spectral correlations emerge in open many-body systems and how noise changes the universality class. This work provides an experimental basis for testing random matrix predictions in the non-Hermitian case.
This section is addressed to the author of the paper and their colleagues. Neighbouring works were found by meaning-based search across our archive, not by word overlap. We do not review: we show the field around the work.
What the work does
In a study on a five-qubit superconducting processor, dissipative quantum chaos has been experimentally observed for the first time. Using gradient tomography, the quantum map of the open system was reconstructed, and complex ratio distances were computed for its eigenvalues. For chaotic circuits, the distribution takes a 'doughnut' shape; for integrable circuits, a peak at zero appears, and as the depth of the integrable circuit increases, a crossover to chaos is observed due to internal noise.
Where its strength is
This work makes dissipative quantum chaos experimentally observable and uses noise as a tool to generate and control chaotic dynamics, paving the way for using quantum processors as testbeds for open many-body phenomena.
Where it could go next
Worth trying
Why this may matter
This experimental platform for dissipative quantum chaos could serve as a foundation for testing theoretical predictions in open many-body systems, including the connection between chaos, decoherence, and thermalization.
Nearby works in our archive
The archive is sparse around this work: the closest review is only 0.56, against a usual 0.64. It looks like few have drilled nearby.
The processor's intrinsic noise, usually considered an enemy of precision, here played the role of a chaos source: the deeper the circuit, the more strongly noise destroyed integrability and led to chaotic statistics.