Fast, Nondestructive Quantum-state Readout of a Single, Trapped, Neutral Atom

Experimental systems that trap single, neutral atoms have recently emerged as a promising platform for experiments in a range of disciplines such as quantum information science, quantum simulation and fundamental light-atom interaction. In this thesis, I build such a system and use it to trap and study a single, neutral atom of 87Rb. I confront and overcome several experimental challenges while designing and building the system. For example, I develop a MOT of unusual geometry with which to load the single-atom trap and also a detection scheme that robustly detects the trapped atom nondestructively, that is, without pushing it out of the trap. The result of this design and construction process is a system that stably traps a single atom in an optical dipole trap. I achieve trap lifetimes of over 1 minute in the absence of near-resonant laser light.

In addition to the experimental apparatus, I develop a thorough rate-equation model to predict the population dynamics of the trapped atom's internal quantum state when probed by near-resonant light. This model gives unique insight into the influence of the atom's internal dynamics on the detected scattering rate. I use this model to predict several important experimental parameters and compare it to the experimental data. This allows me to characterize the parameters that govern how the atom interacts with near-resonant laser light and how that interaction affects the experimental data. For example, I perform an absolute calibration of the collection efficiency of the experimental system, a first for a single, neutral-atom trap.

Using these experimental and modeling tools, I investigate the scattering rate of an atom in the presence of near-resonant linearly-polarized laser light. This is of great interest to the field because it is used to measure the atom's internal quantum state, in a process known as quantum-state readout. Fast and accurate quantum-state readout is crucial to the success of many protocols in quantum information science and quantum simulation. Using the tools described here, I achieve quantum-state readout with an average fidelity of 97.6±0.2% using a linearly-polarized probe beam. The readout requires a measurement time of 160±20 μs, and the atom remains in the trap after the readout in 97.1±0.1% of the trials. I use linearly-polarized light instead of circularly-polarized light because it makes the readout less sensitive to the atom's occupation of a specific magnetic sublevel, and hence does not require sublevel-specific state preparation. It also allows for a more flexible experimental geometry. This is the fastest and highest-fidelity nondestructive readout of a single neutral atom performed with a linearly-polarized probe beam reported to date.

In addition, I identify a decay in the atom's scattering rate over the course of the readout time that limits the quantum-state readout fidelity. I investigate possible sources of this decay using the rate-equation model and a model of the readout protocol, and I conclude that it is likely caused by a combination of Raman transitions and heating. The heating is related to the near-resonant probe light and also to the optical dipole trap that holds the atom. I discuss ways that this decay can be avoided, but point out that these possible solutions result in longer readout times. This investigation has applications across a wide variety of experiments that require fast quantum-state readout.