Density-dependent clock spectroscopy in arbitrary arrays of mesoscopic 88Sr ensembles

Abstract

For frequency metrology experiments, the fractional uncertainty is a key figure of merit: generating correlations between atoms allows this value to improve up to linearly with the number of atoms being interrogated. This has been demonstrated elsewhere using long-range Rydberg interactions in arrays of single strontium atoms and ensembles of caesium atoms. In our experiment, we are working towards demonstrating spin squeezing long-range Rydberg interactions to generate entanglement between strontium atoms in mesoscopic (N < 5) ensembles, trapped in clock-magic arbitrary tweezer arrays. We are interested in investigating how the fractional uncertainty of spin-squeezed clock measurements varies with atom number, in the case that tweezer sites are not restricted to containing a single atom at most. We hope to determine the optimum average number of atoms per site within the Rydberg blockade radius as a function of number of sites.

This thesis presents progress towards this goal, which has been made in two ways. First, we have developed a novel technique for loading vertically large (100 micron) tweezer arrays with uniform and linearly-varying distributions of atom numbers in as fast as 20 milliseconds. We are able to load individual sites with average populations of between one and ten atoms, and show variations in atom number of up to a factor of two between sites in the array. We outline a theoretical model which successfully predicts our experimental results.

Additionally, we show the first measurement of the 5s2 1S0 → 5s5p 3P0 clock transition in our experiment. We present experimental measurements of this transition, investigating how the linewidth and linecentre of the measured frequency spectrum varies with trap depth, atom temperature, probe beam intensity, and atomic density. We consider theoretical models of the clock transition based on the motional states of atoms in a tweezer and on collisional effects, and use these to study the effects of motional dephasing on the measured spectrum in our experiment. With these results in hand, we look ahead to the future of our experiment.