arXiv:2610.14778
Avoiding Defects within Heavy-Hexagon Architecture
Hansol Kim, Younghun Kwon
Abstract
When implementing scalable fault-tolerant quantum computing using superconducting quantum processors, one of the critical challenges to overcome is the presence of defective qubits arising from fabrication and calibration imperfections. A defect on a specific qubit severely restricts the connectivity between adjacent qubits, disrupts stabilizer measurements, induces frequent detection events, and consequently inflates the logical error rate. To mitigate this issue, various strategies have been proposed, such as syndrome extraction circuits based on stabilizer folding and the construction of super-stabilizers via gauge operators. However, these methods have primarily been developed for architectures with relatively high connectivity of 3 to 4. Furthermore, experimental demonstrations showing an actual reduction in logical error rates in the presence of physical hardware defects have been lacking. In this work, we propose a defect-avoidance scheme tailored for the heavy-hexagon architecture, which features a lower connectivity of 2 to 3. While previous approaches required four measurement cycles to measure all gauge operators comprising a super-stabilizer, our proposed scheme completes all gauge measurements within only two cycles. We experimentally validate this approach on IBM quantum hardware by identifying physical defects and implementing the defect-avoidance circuits. As a result, avoiding the defect yields an approximately 10\% reduction in the logical error rate per cycle for both logical $X$ and $Z$ errors. These findings demonstrate that defects in superconducting hardware can be successfully handled even under the constrained connectivity of heavy-hexagon lattices, and that defect avoidance substantially reduces logical errors on real quantum hardware.
AI Summary
The paper addresses how to bypass defective qubits and couplers when implementing surface-code syndrome extraction on degree-$2$–$3$ heavy-hexagon hardware. It adapts stabilizer folding and gauge-based super-stabilizers, repurposing syndrome and bridge qubits so that all constituent gauge measurements fit into two measurement cycles rather than the four required by the compared folding-based schemes. Temporally staggered parity comparisons reconstruct the super-stabilizer outcomes. The accompanying rotated-surface-code layout requires $N=(5d^2+2d-5)/2$ physical qubits: 23, 65, and 127 for nominal distances 3, 5, and 7. Experiments on the 156-qubit IBM Boston processor compare circuits including a persistently poor data qubit against circuits bypassing it. Using calibration-derived noise models and minimum-weight perfect matching, distance-5 logical error rates per cycle decrease from 0.0578 to 0.0508 for $Z$ errors and from 0.1188 to 0.1069 for $X$ errors, reductions of 12.1% and 10.0%. At distance 7, the corresponding reductions are 8.1% and 6.7%. Nearby detector-event probabilities decrease from approximately 0.5 to 0.4. The layout has hook-error propagation along one logical direction, reducing its effective circuit distance to $\lceil d/2\rceil$. Circuit-level depolarizing-noise simulations show threshold behavior, but the hardware experiments show no logical-error suppression with increasing distance. Defect avoidance adds gauge-measurement overhead and can worsen performance when calibration improves. Multiple nonoverlapping defects are considered; overlapping super-stabilizers are outside scope. Experimental validation focuses on a naturally occurring data-qubit defect rather than all proposed defect types.
Error found: The reported per-cycle improvements use a stationary logical-flip inversion without testing its assumptions or separating preparation/readout offsets, so they establish improvements in the chosen estimator rather than independently isolated cycle-error probabilities. The paper provides no uncertainty analysis for its logical-error reductions despite a nonlinear estimator that strongly amplifies sampling uncertainty when cumulative logical errors approach one half.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.