Remote entanglement need not be the bottleneck for modular trapped-ion quantum computing

Architecture for Scalable Modular Computing

As trapped-ion quantum processors grow in size, they face physical limits related to fabrication yield, heat dissipation, and wiring complexity. Modular architectures, which link multiple smaller quantum processing units (QPUs), offer a path to scale. However, the current bottleneck is the production of shared entanglement between modules, which historically suffers from low rates, poor density, and fidelity gaps compared to local gate operations.

Overcoming the Entanglement Bottleneck

This paper proposes a comprehensive architecture that synthesizes four key techniques: single-photon heralding, coherent recoil correction, projective distillation, and trap-integrated photonics. By shifting from traditional two-photon coincidence schemes to a single-photon heralding protocol, the system changes the scaling of the entanglement success probability from quadratic to linear with respect to detection efficiency. This allows compact, integrated photonic components to replace bulky collection optics, significantly increasing the density of entanglement channels.

Fidelity and Fault Tolerance

To address the inherent errors of the single-photon scheme, the authors implement a multi-stage error mitigation strategy. Coherent recoil correction addresses spin-motion entanglement at the source, while a single round of projective distillation removes double-excitation errors and phase drifts. This approach decouples fidelity from the entanglement rate, allowing the system to achieve a projected Bell-pair fidelity of 99.9% at rates and densities compatible with fault-tolerant quantum computing requirements. The authors conclude that with this architecture, the performance limit of modular trapped-ion systems shifts from the photonic link to the local gate operations, which are already highly optimized.