Mitigating Measurement Crosstalk via Pulse Shaping

Abstract

Quantum error correction protocols require rapid and repeated qubit measurements. While multiplexed readout in superconducting quantum systems improves efficiency, fast probe pulses introduce spectral broadening, leading to signal leakage into neighboring readout resonators. This crosstalk results in qubit dephasing and degraded readout fidelity. Here, we introduce a pulse shaping technique inspired by the derivative removal by adiabatic gate (DRAG) protocol to suppress measurement crosstalk during fast readout. By engineering a spectral notch at neighboring resonator frequencies, the method effectively mitigates spurious signal interference. Our approach integrates seamlessly with existing readout architectures, enabling fast, low-crosstalk multiplexed measurements without additional hardware overhead - a critical advancement for scalable quantum computing.

Figures (4)

• Figure 1: Removing probe signal leakage using DRAG. (a) Schematic of multiplexed readout: multiple resonators share a common feedline. The spectrum of the probe pulse overlaps with the resonance of untargeted readout resonators, leading to additional crosstalk errors and qubit dephasing. (b) Time-domain envelopes of readout pulses with cosine-shaped rise and fall edges and a 200-ns flat plateau, shown with and without DRAG. (c) Effect of DRAG in suppressing the 50-MHz spectral component for readout pulses with plateau durations of 200 ns and 2 µ s. For the 2 µ s case, the spectral component at 50 MHz is suppressed by over an order of magnitude when the DRAG pulse is applied. (d) Dependence of readout infidelity on the DRAG notch frequency. The readout pulse plateau is 2 µ s. The solid reference line indicates the readout infidelity without applying DRAG.
• Figure 2: Recovering phase coherence by DRAG. Ramsey measurement results with and without DRAG applied to the pseudo-readout pulse. The pulse is detuned by $+10$ MHz from the readout resonator, with a flat plateau of duration $\tau$ and 10-ns cosine-shaped rise and fall edges. The final readout pulse plateau is 2 µ s. Without DRAG, a clear frequency beating pattern appears (inset), whereas DRAG suppresses the beating and extends the qubit coherence time.
• Figure 3: Probe frequency and amplitude dependence of readout-induced qubit dephasing. (a) Pulse sequence used to measure readout induced dephasing. To avoid beating in the Ramsey experiment, the pseudo-readout pulse length is fixed. The pulse envelope has a cosine shape with a flat plateau, 10-ns total rise and fall times, and a 200-ns plateau. The phase of the second $\pi/2$ gate is scanned, and the resulting oscillation is fitted to extract the amplitude and phase. The final readout pulse plateau is 2 µ s. (b) Experimentally extracted parameters: $\kappa/2\pi = 2.2$ MHz, $2\chi/2\pi = 2.1$ MHz. (c-d) Experimental results of scanning the pseudo-readout pulse amplitude and detuning, without DRAG (c) and with DRAG (d). The red cross indicates the pseudo-readout pulse parameters used in Fig. \ref{['fig:ramsey']}. (e-f) Theoretical calculations of the same scans, without DRAG (e) and with DRAG (f).
• Figure 4: Dependence of readout-induced $T'_2$ on normalized probe detuning $\Delta_d/\kappa$ and resonator linewidth $\kappa$. Theoretical evaluation of readout-induced transverse relaxation time obtained by scanning the pseudo-readout detuning and resonator linewidth using Eq. \ref{['eq:dephasingratenew']}. (a) Without DRAG. (b) With DRAG. (c) Ratio of (b) to (a), illustrating the enhancement in $T'_2$ achieved by DRAG.