Simple framework for systematic high-fidelity gate operations

Semiconductor spin qubits demonstrated single-qubit gates with fidelities up to $99.9\%$ benchmarked in the single-qubit subspace. However, tomographic characterizations reveals non-negligible crosstalk errors in a larger space. Additionally, it was long thought that the two-qubit gate performance is limited by charge noise which couples to the qubits via the exchange interaction. Here, we show that coherent error sources such as a limited bandwidth of the control signals, diabaticity errors, microwave crosstalk, and non-linear transfer functions can equally limit the fidelity. We report a simple theoretical framework for pulse optimization that relates erroneous dynamics to spectral concentration problems and allows for the reuse of existing signal shaping methods on a larger set of gate operations. We apply this framework to common gate operations for spin qubits and show that simple pulse shaping techniques can significantly improve the performance of these gate operations in the presence of such coherent error sources. The methods presented in the paper were used to demonstrate two-qubit gate fidelities with $F>99.5\%$ in Ref.~[Xue et al, Nature 601, 343]. We also find that single and two-qubit gates can be optimized using the same pulse shape. We use analytic derivations and numerical simulations to arrive at predicted gate fidelities greater than $99.9\%$ with duration less than $4/(\Delta f)$ where $\Delta f$ is the difference in qubit frequencies.