Impact of control signal phase noise on qubit fidelity

Abstract

As qubit decoherence times are increased and readout technologies are improved, nonidealities in the drive signals, such as phase noise, are going to represent a growing limitation to the fidelity achievable at the end of complex control pulse sequencies. Here we study the impact on fidelity of phase noise affecting reference oscillators with the help of numerical simulations, which allow us to directly take into account the interaction between the phase fluctuations in the control signals and the evolution of the qubit state. Our method is based on the generation of phase noise realizations consistent with a given power spectral density, that are then applied to the pulse carrier in simulations, with Qiskit-Dynamics, of the qubit temporal evolution. By comparing the final state obtained at the end of a noisy pulse sequence with that in the ideal case and averaging over multiple noise realizations, we estimate the resulting degradation in fidelity, and exploiting an approximate analytical representation of a carrier affected by phase fluctuations, we discuss the contributions of the different spectral components of phase noise.

Figures (14)

• Figure 1: Fidelity as a function of the frequency offset and of the number of 50 ns $\pi_x$-pulses applied with constant amplitudes of -85 dBc/Hz. The qubit is initially prepared in $|0\rangle$
• Figure 2: Fidelity as a function of the number of applied $\pi_x$-pulses, with constant amplitudes of -85 dBc/Hz, and of the central frequency of the Gaussian filter used to model phase noise. In both panels the state of the qubit is initially prepared in $|0\rangle$. The top panel (a) shows the results for a sequence of 25 ns pulses separated by 10 ns, while the bottom panel (b) refers to 150 ns pulses with an inter-pulse interval of 60 ns.
• Figure 3: Fidelity as a function of the number of applied $\pi_x$-pulses, with constant amplitudes of -85 dBc/Hz, and of the central frequency of the Gaussian filter used to model phase noise. In both panels the qubit is initially prepared on the equatorial plane of the Bloch sphere, along the $y$-axis. The top panel (a) shows the results for a sequence of 25 ns pulses separated by 10 ns, while the bottom panel (b) refers to 150 ns pulses with an inter-pulse interval of 60 ns.
• Figure 4: Fidelity as a function of frequency offset and of the number of 25 ns $\pi_x$-pulses applied with a constant amplitudes of -95 dBc/Hz. The qubit is initially prepared in $|0\rangle$
• Figure 5: Fidelity as a function of the frequency offset and of the number of 25 ns $\pi_x$-pulses applied with a constant amplitude of -100 dBc/Hz and a bandwidth of the Gaussian FIR filter of 3 MHz. The qubit is initially prepared in $|0\rangle$