Optimal detuning and gate time for strongly driven single-qubit gates beyond the rotating-wave approximation

Technical Summary: Optimal Detuning and Gate Time for Strongly Driven Single-Qubit Gates Beyond the Rotating-Wave Approximation

Problem Statement
Single-qubit quantum gates are typically implemented by driving a two-level system with a harmonic field. For weak coupling, the Rotating-Wave Approximation (RWA) provides closed-form analytical expressions for gate parameters (amplitude, frequency, and duration). However, as driving amplitudes increase to achieve faster gate speeds and outpace decoherence, the RWA becomes inaccurate. This is due to the neglect of the counter-rotating term in the Hamiltonian, which induces the Bloch-Siegert (BS) shift (a frequency shift of the spectral resonance) and modifies Rabi oscillations. These neglected terms introduce coherent errors, causing gate infidelity to grow as a power law with the ratio of driving amplitude to frequency ($A/\omega$). Current fault-tolerant protocols require gate errors below $\approx 10^{-3}$, a threshold that RWA-designed gates fail to meet at high amplitudes, forcing the use of slower, weaker driving.

Methodology
The authors analyze the dynamics of a harmonically driven two-level system governed by the Bloch-Siegert Hamiltonian:
$$H_{BS}(t) = \frac{\omega_0}{2}\sigma_z + A \sin(\omega t)\sigma_x$$
where $\omega_0$ is the energy gap and $A$ is the driving amplitude.

To address the limitations of the RWA, the paper employs a second-order Magnus-Taylor effective Hamiltonian framework. This perturbative method expands the evolution operator in powers of $1/\omega$ (inverse driving frequency) and $A$, ensuring unitarity is preserved. The authors derive an effective time-independent Hamiltonian $\tilde{H}$ up to second order in $1/\omega$. By parameterizing $\tilde{H}$ in terms of effective field components ($h_x, h_y, h_z$), they analyze the conditions required to reproduce specific unitary gates (Z, X, and Hadamard) with high fidelity.

The study involves:

1. Deriving analytical corrections to the driving frequency (detuning) and operation time based on the effective Hamiltonian.
2. Analytically estimating the infidelity of standard RWA gates to determine the power-law dependence on $A/\omega$.
3. Numerically solving the Schrödinger equation for the full time-dependent Hamiltonian $H_{BS}(t)$ using the derived corrected parameters to verify the analytical predictions.

Key Contributions and Results

1. Analytical Correction of Gate Parameters:
   The paper derives closed-form prescriptions for corrected gate parameters that remain accurate for large driving amplitudes ($A/\omega \lesssim 1$). The corrections involve:

   • Amplitude-Dependent Detuning: A frequency shift of the Bloch-Siegert type, $\Delta_{BS} = -\frac{3}{4}A\lambda = -\frac{3}{4}\lambda^2\omega$ (where $\lambda = A/\omega$), to compensate for the effective detuning induced by the counter-rotating term.
   • Time Rescaling: A rescaling of the operation time $t_g$ by a factor of $(1 + \lambda^2/16)^{-1} \approx (1 - \lambda^2/16)$.

   With these corrections, the RWA gate conditions are recovered, but with the driving frequency shifted to $\omega = \omega_0 - \Delta_{BS}$ and the amplitude effectively renormalized.

2. Analytical Explanation of RWA Infidelity:
   The authors provide an analytical derivation showing that the infidelity ($I = 1 - F$) of RWA-designed gates follows distinct power laws depending on the gate type:

   • Z-gate: Infidelity scales as $I_Z \propto (A/\omega)^4$. This gate is less sensitive because it is insensitive to the tilt of the effective field axis; the dominant error arises from the rotation angle deviation.
   • Hadamard (and X) gates: Infidelity scales as $I_H \propto (A/\omega)^2$. These gates are more sensitive because they are affected by both the rotation angle error and the tilt of the effective field axis (caused by the mismatch between resonant driving and the optimal $\Delta_{BS}$).

   This analysis explains the "dip" structure observed in numerical infidelity plots, where the minima follow these power laws.

3. Numerical Validation:
   Numerical solutions of the Schrödinger equation confirm that applying the derived corrections (optimal detuning and rescaled time) reduces gate infidelity by several orders of magnitude compared to standard RWA parameters.

   • The corrected gates maintain high fidelity up to $A/\omega \approx 1$.
   • For the Hadamard gate at $A/\omega = 1/14$, the corrected parameters produce a sharp infidelity minimum at least three orders of magnitude lower than the broad minimum observed with resonant RWA driving.
   • The numerically optimized frequency shift follows $\delta \approx -0.7(A/\omega)^{2.0}$, closely matching the analytical prediction of $-0.75(A/\omega)^2$.

Significance and Claims
The paper claims that by utilizing a unitarity-preserving effective Hamiltonian framework, it is possible to recover the simplicity of RWA gate design while extending the usable range of driving amplitudes significantly. The primary contribution is a systematic, analytical method to correct for non-RWA errors (specifically the Bloch-Siegert shift) without requiring complex pulse shaping or dedicated experimental schemes.

The authors emphasize that while the approach is perturbative, the numerical agreement holds well within the explored range ($A/\omega \lesssim 1$). They note that the derived detuning $\Delta_{BS}$ is the optimal gate detuning for a drive with a fixed initial phase and gates evaluated at stroboscopic times, rather than a general shift of the spectral resonance. The work provides a practical route to implementing faster, high-fidelity gates in systems such as superconducting circuits, spin qubits, and quantum dots, where strong driving is desirable to mitigate decoherence. The paper concludes that robustness against calibration errors and noise remains an open area for future analysis.