A random-walk benchmark for single-electron circuits

Mesoscopic integrated circuits achieving high-fidelity control of elementary quantum systems require new methodology for benchmarking. We offer circuit-level statistical description of rare-error accumulation in terms of a universal random-walk model for on-demand electron transfer. For a high-fidelity single-electron circuit, realized in the experiment as a chain of quantum dots in a GaAs/AlGaAs heterostructure, the error of the transfer operation is probed by charge counting. Error rates for extra ($P_+$) or missing ($P_-$) electrons of the electron shuttle are measured to $P_{-}=(6.92 \pm 0.14) \times 10^{-5}$ and $P_{+}=(2.13 \pm 0.08)\times 10^{-5}$ with uncertainty due to correlated noise in the environment. Furthermore, precise control over the timing of the random walk allows to explore the role of memory as the clock frequency is increased.