arXiv:2610.30056
High-fidelity and robust controlled-Z gates for superconducting transmon qubits. I. Intrinsic errors and decoherence
Chia-Hsien Huang, Hsi-Sheng Goan
Abstract
Fault-tolerant quantum computing requires quantum gates that combine short duration, high fidelity, and robustness to noise. As quantum processors scale, robustness becomes increasingly important because exhaustive characterization and frequent recalibration of individual gates become impractical. We present a comprehensive optimal-control framework for controlled-Z (CZ) gates between directly coupled, frequency-tunable transmons that systematically accounts for intrinsic Hamiltonian errors and realistic decoherence mechanisms, including energy relaxation and dephasing. Using smooth, single-shot flux pulses, we obtain 12-ns CZ gates---close to the theoretical speed limit of 11.6 ns---with intrinsic errors suppressed to the $2\times10^{-5}$ level. We construct dephasing noise models, calibrated against published experimental measurements, that capture fast dephasing from room-temperature control electronics and slow dephasing from $1/f$ flux noise. Using these models, we identify dephasing sweet spots that suppress sensitivity to dephasing noise. We further exploit transmon asymmetry as a design parameter to access distinct noise-resilient operating regimes. These results establish an experimentally applicable framework for balancing gate speed, fidelity, and noise robustness while reducing characterization and recalibration overhead.
AI Summary
This paper studies how to balance speed, intrinsic error and decoherence in CZ gates between directly coupled, flux-tunable transmons, including how junction asymmetry changes noise sensitivity. It uses numerical optimal control rather than new gate experiments. The control waveform combines smooth ramps to an interaction point with a sinusoidal fine-tuning component, optimized using Nelder–Mead methods. Intrinsic infidelity is decomposed into phase, swap and leakage contributions, with excitation-sector truncations used to reduce optimization cost and larger Hilbert spaces used for verification. White-noise and $1/f$-noise models are fitted to published Ramsey measurements and gate-error budgets, then incorporated into gate optimization and stochastic simulations. Optimized $12\,\mathrm{ns}$ gates approach the estimated resonant-operation timescale of $11.6\,\mathrm{ns}$ and attain intrinsic infidelity around $2\times10^{-5}$. At junction-energy ratio $E_{J1}/E_{J2}=1.31$, the predicted total infidelity is $4.66\times10^{-4}$ for $T_1=30\,\mu\mathrm{s}$, including $4\times10^{-4}$ from relaxation. Substituting $T_1=360\,\mu\mathrm{s}$ gives a projection of $9.9\times10^{-5}$. Asymmetry-dependent operating points trade sensitivity to fast and slow dephasing; selected designs remain below $10^{-3}$ infidelity with noise standard deviations increased by more than $10\times$ and approximately $90\times$, respectively. Filter-aware reoptimization restores a representative bandwidth-distorted gate from $1.14\times10^{-2}$ to $2.05\times10^{-5}$ intrinsic infidelity. The analysis is restricted to direct coupling and calibrated noise models. Extensions to other coupling architectures are proposed, not demonstrated. Robustness to systematic parameter uncertainty and drift is deferred to Part II; the higher-coherence fidelities are projections.
Central claim may not hold: The quoted 11.6-ns theoretical speed limit is derived as an approximate resonant two-state return time, without establishing a bound for the full allowed multilevel control problem. The reported flux-noise tolerance margins omit nonlinear flux sensitivity because the noise Hamiltonian retains only the first derivative of frequency with respect to flux, which vanishes at the exploited sweet spots. The predicted low-frequency fast-dephasing advantage extrapolates a phenomenological model fitted at 4.15–5.09 GHz on a symmetric transmon to substantially lower frequencies and other junction asymmetries without independent validation. The printed leakage diagnostic gives −3 rather than zero for a leakage-free unitary, although the separate trace-overlap optimization objective is correctly normalized.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 37 pages, 11 figures