Curator's Take
This article marks a significant step forward in open‑system quantum simulation by scaling collision‑model experiments to seven system qubits and forty time steps across both trapped‑ion and superconducting platforms, far beyond the one‑ or two‑qubit demonstrations that have dominated the field. By tailoring ancilla handling to each hardware’s strengths—reset overhead on superconductors versus connectivity on ions—the work shows how practical implementation choices can dramatically extend what noisy intermediate‑scale devices can achieve for nonunitary dynamics. The results not only broaden the toolbox for studying dissipative quantum phenomena such as thermalization and error mitigation, but also underscore that future algorithmic advances will need to be co‑designed with hardware characteristics.
— Mark Eatherly
Summary
Hamiltonian dynamics have been widely implemented on noisy intermediate-scale quantum devices in recent years. In contrast, experimental demonstrations of Markovian quantum dynamics remain limited, because implementing nonunitary evolution on quantum computers is challenging. Quantum collision models provide a natural approach to this problem by coupling the system to ancillas to realize dissipation. However, previous implementations of quantum collision models on quantum computers have typically been restricted to one or two system qubits and fewer than 12 time steps, owing to noise, circuit depth, the overhead of ancilla reset, and limited qubit resources. In this work, we experimentally simulate Markovian quantum processes with local and nonlocal dissipation on both trapped-ion and superconducting quantum computers. By employing hardware-specific ancilla strategies, we realize simulations with up to seven system qubits, corresponding to 13 qubits in total, and 40 time steps. Our results demonstrate that, even for the same physical model, the optimal implementation strategy depends strongly on the hardware characteristics of the quantum computer.