Adaptive quantum measurements with spins in diamond

In the last decade, outstanding experimental progress has led to sophisticated techniques to control and measure individual quantum systems. The combination of quantum measurements with real-time feedback control has opened new opportunities for quantum information and sensing.

In this seminar, I will discuss some of my recent experiments at TU Delft (NL) on adaptive quantum measurements using single spins associated to nitrogen-vacancy (NV) centres in diamond. Over the last few years, spins in NV centres have become one of the leading candidates for quantum information and quantum sensing, due to long coherence time and high-fidelity manipulation by optical and radiofrequency fields, up at room temperature. After an introduction to the field of spin physics in diamond, I will address two questions regarding adaptive quantum measurements:

1) Given a quantum system, what is the best measurement protocol to extract the most information about unknown parameters? This is very important in sensing applications, because individual quantum systems can be used as small localized probes of physical quantities, at the ultimate limit of spatial resolution. For example, single spins in NV centres are routinely used to map magnetic fields, temperature and strain with nanometric resolution. Given a sequence of measurements on a single quantum bit, outcomes obtained by earlier measurements within the sequence could be used to optimize the settings for later measurements (adaptive estimation). What kind of advantage can this provide? Is adaptive estimation advantageous also in the presence of noise, decoherence and imperfect measurements? By numerical simulations and experiments, we demonstrate an optimized adaptive sensing protocol (based on machine learning techniques, such as swarm optimization) that performs better than the best known non-adaptive protocol, leading to a record sensitivity (Nature Nano 11, 247 - 2016).

2) Can quantum measurement back-action be used as a state preparation tool? In other words, can we steer an initial quantum state towards a different per-determined target state by simply peeking at it? Since measurement outcomes are probabilistic, a sequence of measurements on a quantum system results in an uncontrolled final state. However, we show that, by adapting in real-time the measurement parameters based on the previous measurement outcomes, deterministic preparation of a quantum state by measurement-only is possible (Nature Physics 10, 189 - 2014).