Vector magnetometry using cavity-enhanced microwave-optical double resonance in an atomic vapour

Microwave-optical double resonance is a two-photon absorption process that makes use of two electromagnetic fields with vastly separated energies. In these fields, the two polarization degrees of freedom adds an interesting opportunity to extract additional informaiton from the absorption process, in addition to the detunings one might first consider. In these experiments, we use an ambient-temperature vapour enclosed in a three-dimensional copper cavity to enhance the microwave coupling to ground-state rubidium atoms. Additionally, we probe the atoms across the cavity's width with a resonant optical beam, driving the two-photon double resonance. By measuring the absorption as we scan the microwave frequency across the comb of microwave resonances, the relative absorptions of from the different m_F levels provide information about the background, quantizing magnetic field. To this end, we apply these principles to create a vector magnetometer, which can determine both the pointing direction and the magnitude of the magnetic field. To analyse this multi-level system that is complicated by Doppler broadening and multiple optical pumping processes, we apply machine-learning techniques to extract the magnetic field information from our data.