Curbing Leakage with the Honeycomb Floquet Code

Leakage, occuring when a qubit undetectably exits the computational subspace, is particularly challenging for quantum error correcting codes as it leads to correlated errors in space and time. These correlations, if left unchecked, cause a reduction in both the code threshold and effective code distance. In this work, we study the effect of leakage in the honeycomb Floquet code, which dynamically generates and protects logical qubits via a periodic schedule of anticommuting two-body measurements. We show that these weight-two measurements, conducted using new low-overhead leakage-reducing circuits, are effective at localizing leakage without sacrificing code distance. We place these results in the context of the more popular surface code, showing that the honeycomb Floquet code presents an advantage in terms of sub-threshold logical error scaling when leakage comprises a considerable portion of the error budget.