Effects of Systematic Error on Quantum-Enhanced Atom Interferometry

We develop a framework for describing the effects of systematic state preparation error in quantum-enhanced atom interferometry on sensing performance. We do this in the context of both spin-squeezed and non-Gaussian states for the two-axis-twisting (TAT), one-axis-twisting (OAT), and twist-and-turn (TNT) state preparation schemes, and derive general conditions for robustness and susceptibility of quantum states to state preparation error. In the spin-squeezing regime, we find that OAT is more susceptible to state preparation error than TAT due to its parameter-dependent phase space rotation. In the non-Gaussian regime, we find that OAT is robust to state preparation errors, which can be explained by a small ratio of off-diagonal to diagonal elements in its Fisher-covariance matrix. In contrast, TNT does not exhibit this robustness. We find that the single parameter unbiased estimators that are habitually used in quantum-enhanced atom interferometry are not always optimal, and that there may be occasions where biased estimators, or two-parameter unbiased estimators, lead to lower net error.