Maximized information on the environment by dynamically controlled qubit probes

ANALIA ZWICK

Cambridge

Seminario; Seminar in Massachusetts Institute of Technology (MIT), invited by Dr Paola Cappellaro; 2015

Massachusetts Institute of Technology (MIT)

Talk: Controlled quantum spins are sensitive probes of the environment and a powerful tool for characterizing highly complex quantum systems at a molecular or atomic scale. Novel quantum technologies requiring high sensitivity at the nanoscale re based on quantum spin probes serving as magnetometers, thermometers, sensors for imaging or monitoring biological process. We study the use of a single spin as a quantum probe to characterize unknown parameters of an environment with which it interacts. By resorting to quantum estimation theory tools, we analytically calculate the lower bound for the estimation precision and optimize the control procedure to attain the best precision. The procedure employs spin-echo sequences for the quantum control. We find the optimal initial probe state, the optimal observable to monitor the probe state and the optimal time to apply the time reversal of the spin-echo sequences. Specifically, we estimate the correlation time of the environmental fluctuations, which are given by a Lorentzian spectrum typical of the omnipresent Ornstein-Uhlenbeck processes. We compare the best performance obtained by free evolution of the quantum probe with different dynamical control schemes, such us Hahn and CPMG sequences or continuous-wave driving, for the estimation of the environmental correlation time. We demonstrate that dynamically controlled probes can improve the estimation, which is found to be optimal in either the Markovian or the strongly non-Markovian regimes. We reveal on a critical behavior of the precision bound at the transition between these two regimes, allowing a sharp characterization of what is usually a smooth and imprecise transition. This analysis is compared to a method for finding the optimal time to monitor the probe via a real-time estimation protocol based on a Bayesian estimator and an online experimental learning design that maximizes the information gain.