The Effect of Device Drift on Elzerman Readout Fidelity

High-fidelity, single-shot readout of electron spins is needed for fault-tolerant quantum computation. High-fidelity has been achieved using the common method known as Elzerman readout. However, Elzerman readout has the requirement of very high signal-to-noise and often requires device retuning. We perform simulations of a charge sensor with an experimentally motivated noise model, which includes charge sensor drift. We introduce a method for calculating the charge sensor operating voltage as a function of elapsed time from a simple noise measurement, which allows us to deduce the reduction in readout fidelity in the absence of retuning. These calculations provide a simple framework for comparing device performance between different architectures and estimating experimental retuning times.