Robust iSWAP gates for semiconductor spin qubits with local driving

TL;DR

The paper tackles the challenge of implementing high-fidelity two-qubit gates in semiconductor spin qubits amid decoherence and control imperfections. It introduces a robust iSWAP gate driven by local microwaves that continuously decouples the qubits from low-frequency noise, avoiding the difficulties of ac exchange modulation. Three progressively robust schemes are presented: A direct implementation, B a composite sequence, and C a dynamically corrected gate, with Scheme C providing first-order immunity to amplitude errors. Numerical simulations under realistic noise and decoherence show fidelities exceeding fault-tolerance thresholds across practical parameter ranges, establishing this approach as a viable building block for scalable quantum processors and adaptable to other exchange-based platforms.

Abstract

Scalable quantum computation demands high-fidelity two-qubit gates. However, decoherence and control errors are inevitable, which can decrease the quality of implemented quantum operations. We propose a robust iSWAP gate protocol for semiconductor spin qubits, which is a promising platform for scalable quantum computing. Our scheme uses only local microwave drives on conventional exchange-coupled spin qubits. This approach simultaneously addresses two critical challenges on semiconductor quantum computing: it suppresses low-frequency noise via continuous dynamical decoupling, and it circumvents the control difficulties associated with the ac modulation of the exchange interaction. We further develop a composite pulse sequence to remove drive-strength constraints and a dynamically corrected method to provide first-order immunity to microwave amplitude errors.Numerical simulations confirm that our scheme can achieve fidelity above the fault-tolerance threshold under current experimental conditions, offering a building block for practical quantum processors.

Figures (8)

• Figure 1: Illustration of the microwave-driven iSWAP gate scheme. The two qubits $Q_1$ and $Q_2$ are coupled via an exchange interaction $J$. A uniform external magnetic field ($B_z$) provides the Zeeman splitting, while a local micromagnet creates an inhomogeneous field. The resulting longitudinal field gradient ($\partial B_z/\partial x$) ensures qubit addressability, while the transverse gradient ($\partial B_y/\partial z$) facilitates EDSR. Local EDSR microwave-drives $MV_1$ and $MV_2$ (shown as smooth blue lines) are used to drive the qubits. These drive fields simultaneously provide continuous dynamical decoupling from low-frequency noise $\delta_1$ and $\delta_2$ (shown as jagged grey lines).
• Figure 2: Performance for the conventional resonant iSWAP gate scheme under control errors and decoherence. Gate infidelity ($1-F$, right y-axis) for the conventional scheme as a function of the relative error in the control voltage ($\epsilon_v$). The left y-axis shows the corresponding maximum relative error in the exchange interaction ($|\epsilon_J|_{\textrm{max}}$). Numerical simulations include high-frequency noise ($T_2^{\textrm{echo}} = 20\,\mu\textrm{s}$) and low-frequency noise ($\sigma/2\pi = 0.1\,\textrm{MHz}$).
• Figure 3: Gate performance under decoherence. The average gate fidelity $F$ as a function of low-frequency noise, modeled as quasi-static Zeeman shifts $\delta_{1,2}$. All simulations also account for the effects of high-frequency noise through a dephasing superoperator in the Lindblad master equation, with a rate corresponding to $T_2^{\text{echo}} = 20 \mu\text{s}$. The subplots compare the performance of (a) the conventional resonant iSWAP gate with the proposed (b) Scheme A, (c) Scheme B, and (d) Scheme C. The results demonstrate that our schemes are more robust, maintaining high fidelity across a wide parameter range, coin contrast to the rapid fidelity degradation seen in the conventional scheme.
• Figure 4: Gate fidelity F as a function of the low-frequency noise with standard deviation $\sigma$ and a symmetric amplitude error $\epsilon$, with $T_2^{\textrm{echo}} = 20\,\mu\textrm{s}$. The white dotted line at $\sigma/2\pi = 0.1\,$MHz indicates an experimentally relevant noise level. (a) Scheme A is sensitive to amplitude errors. Plots (b) and (c) correspond to Scheme B and Scheme C respectively, and shows their robustness against both noise sources simultaneously.
• Figure 5: Gate performance under control errors and decoherence. All simulations include high-frequency noise ($T_2^{\textrm{echo}} = 20\,\mu\textrm{s}$) and low-frequency noise ($\sigma/2\pi = 0.1\,\textrm{MHz}$). Plots (a), (b), and (c) show the gate fidelity ($F$) for the proposed Schemes A, B, and C, respectively, as a function of relative errors in the microwave drive amplitudes ($\epsilon_1, \epsilon_2$). The results clearly show the superior robustness of the dynamically corrected gate (Scheme C) against such errors, maintaining high fidelity across a wide parameter range.