Scalable suppression of heating errors in large trapped-ion quantum processors
Trapped-ion processors are leading candidates for scalable quantum computation. However, motional heating remains a key obstacle to fault-tolerant operation, especially when system size increases. Heating error is particularly challenging to suppress due to is incoherence nature, and no general methods currently exist for mitigating their impact in large systems with multiple phonon modes. In this work, based on a careful analysis about the dependence of heating-induced infidelity on phase-space trajectories, we present a simple yet comprehensive framework for suppressing heating errors in large trapped-ion quantum processors. Our approach is flexible, allowing various control pulse bases, ion numbers, and noise levels. Our approach is also compatible with existing error-mitigation techniques, including those targeting laser phase and frequency noise. Crucially, it relies on an efficiently computable cost function that avoids the exponential overhead of full fidelity estimation. We perform numerical simulations for systems with up to 55 qubits, demonstrating up to a fivefold reduction in infidelities compared with the conventional method in typical regimes. We further show that our framework simultaneously reduces the sensitivity of the gate rotation angle to detuning errors, leading to an improvement in detuning-error tolerance. These results offer a practical route toward robust, large-scale quantum computation with trapped ions.