Error attribution for singlet-triplet qubits through randomized benchmarking in the presence of temporally correlated noise

Temporally correlated noise is understood to be a significant source of error for spin qubits. Examples of such noise sources include magnetic field fluctuations due to spinful nuclei and charge noise that commonly exhibit a 1/f-like noise power spectral density. In this work, we model the influence of temporally correlated noise on a qubit in the singlet-triplet encoding, for which logic operations are generated by the exchange interaction and a magnetic field gradient between electrons in neighboring quantum dots. Singlet-triplet qubits are susceptible to charge and magnetic noise that may lead to variation of qubit performance in time. In our study, we simulate randomized benchmarking (RB) using compiled generators of the single-qubit Clifford group and a realistic noise model, calibrated with experimental results, that accounts for noise correlations at all time scales of a Clifford RB experiment. Based on this simulated data set, we investigate how error attribution may be performed on experimental data by comparing RB errors on a per-circuit basis.