Distributed quantum codes as the quantum sensor for chip-level catastrophic errors

The widely accepted framework for quantum error correction rests on a foundational assumption of local, uncorrelated errors. However, this assumption is fundamentally inconsistent with the physical reality of catastrophic errors (CE) —millisecond‐scale, chip‐wide bursts of relaxation, dephasing, and correlated faults driven by cosmic rays. This anomaly in our standard noise models threatens the viability of both near‑term and fault‑tolerant processors. To address this issue, we propose a distributively encoded quantum sensor scheme that uses low‑weight graph codes of the form [[n,0,d]] spread across multiple chips to continuously monitor for CE via non‑demolition stabilizer readout. We cast CE monitoring as binary channel discrimination and use Jensen–Shannon divergence and Fisher information to quantify detection power. With small codes, DEQS (i) detects the occurrence and location of a CE on a specific chip, (ii) discriminates dominant error types (amplitude vs. phase damping), and (iii) identifies non‑local, chip‑level correlations. We implement numerical simulations with realistic gate/measurement noise to demonstrate the effectiveness of DEQS. These results establish distributed, low‑overhead QEC primitives as a valuable and scalable health monitor for quantum processors.