An Economic Case for Erasure-Biased OAM Front-Ends in Programmable Photonic Quantum Buffers

Abstract

Programmable photonic quantum buffers—systems that synchronize, store, and route photonic qubits using active quantum memory—are an active area of both academic and commercial development, with several published memory protocols now demonstrated at useful efficiencies. Two companion notes in this series developed an orbital-angular-momentum (OAM) cat-qubit encoding whose channel noise is structurally erasure-biased: local mode-mixing converts almost all population leakage into a flagged, detectable erasure rather than a silent, unlocated error. This note asks what that structural property is worth if presented as a front-end noise model to an existing active photonic memory, independent of which specific platform or vendor implements the memory. Using established, published quantum error correction results on erasure-biased noise (surface-code thresholds reported 4-6x higher for erasure than for generic Pauli noise, translating to a documented 2-5x reduction in physical-qubit overhead per logical qubit), we compute an illustrative 3.5x overhead reduction at a representative fault-tolerance operating point. We compare three general architectural options—a passive fiber-recirculation buffer, a cryogenic solid-state color-center register, and a room-temperature warm-atomic-vapor memory—on the axes relevant to this value proposition, and outline the licensing-oriented (rather than vertically-integrated hardware) path this suggests. We are open to discussing this result with groups working on photonic quantum memory or buffer architectures.