The Signal
Bendersky, Gomez and Romero Correct Strong-Drive Qubit Gates
A paper posted to arXiv on October 29 presents frequency and timing corrections for single-qubit gates driven beyond the rotating-wave approximation. Numerical simulations show at least a thousandfold reduction in Hadamard-gate infidelity a
Quantum.world Team · Written with AI · fact-checked ·
A paper posted to arXiv on October 29 sets out a way to reduce errors in strongly driven single-qubit gates by adjusting both the drive frequency and the gate duration. Alejandro D. Bendersky, Sergio S. Gomez and Rodolfo H. Romero use a second-order effective Hamiltonian to calculate corrections beyond the rotating-wave approximation, the simplification commonly used to derive gate parameters for weak drives.
“Numerical solutions of the Schrödinger equation show that these corrections lower the gate infidelity by several orders of magnitude, extending the range of usable amplitudes”
Alejandro D. Bendersky, Sergio S. Gomez and Rodolfo H. Romero, arXiv:2610.40595
Why stronger drives miss their target
Single-qubit gates can be produced by driving a two-level system with a harmonic field for a chosen time. The rotating-wave approximation (RWA) makes that calculation tractable by dropping the drive’s counter-rotating component. At larger amplitudes, however, the neglected term affects the dynamics and shifts the resonance, so a gate designed with RWA parameters can rotate the qubit by the wrong amount or around the wrong axis.
The authors examine Z, X and Hadamard gates using an effective Hamiltonian derived through second order in the inverse drive frequency. Their prescription changes the frequency according to the drive amplitude and rescales the operation time. In the paper’s notation, the relative drive strength is λ = A/ω, where A is the drive amplitude and ω its frequency; the calculations explore values up to roughly one.
A detuning paired with a shorter pulse
For the fixed drive phase and timing protocol studied, the authors derive a gate detuning ΔBS = −(3/4)λ²ω. They also shorten the RWA gate time by a factor of approximately 1 − λ²/16. The time adjustment alone is marginal, the paper says; the larger practical benefit in the model is lower infidelity, which could allow stronger drives and therefore faster gates.
The authors compare corrected parameters with RWA settings by numerically solving the time-dependent Schrödinger equation. For a Hadamard gate at A/ω = 1/14, the resonant RWA setting produces a broad infidelity minimum of about 10−3. Applying both corrections lowers the minimum by at least three orders of magnitude. Across the simulated amplitude range, the corrected Z and Hadamard gates reach lower infidelity than their RWA-designed counterparts.
Not the usual spectral Bloch–Siegert shift
The paper’s qualification matters: ΔBS is an optimal gate detuning for a drive with a fixed initial phase and gates evaluated at specific points in the drive cycle. It is not the Bloch–Siegert shift of the system’s spectral resonance, which the authors distinguish as a separate, phase-independent property. The correction is tied to the control protocol, not a claim that the underlying spectral resonance itself moves by that amount.
Hardware performance is still an open test
The results come from calculations for an idealized driven two-level system, not an experiment on a physical qubit. The authors say residual errors at high drive strength arise from higher-order terms omitted from the effective Hamiltonian, as well as finite precision in the amplitude and transition-frequency values. They also leave resistance to calibration errors and noise for future analysis.
The study therefore establishes a mathematical correction and tests it in numerical gate simulations; it does not show that a device has achieved the simulated error reduction. Its most concrete benchmark is the modeled Hadamard gate at A/ω = 1/14, where the corrected frequency and timing reduce the simulated infidelity by at least a factor of 1,000.