We have set up a digital micromirror-device (DMD- DLP7000 from Texas Instruments) in a designated test set-up and created arbitrary light potentials using a designated feedback-optimisation algorithm. On our setup, the illuminated plane of the DMD is imaged in the focal plane of a test microscope identical to that of the quantum-gas microscope experiment, in which this configuration will be ultimately used. Light in this plane is then focused on a camera to monitor the created patterns. We have first developed Matlab-based software for the control of the DMD. This include the calibration protocols that map the dimensions of the DMD to those of the imaging area, as well as the error-diffusion and feedback algorithms necessary for the realisation of arbitrary light patterns with high fidelity. The algorithm is working well and we can produce tailored light-potentials, like flat-top square regions, which are being characterised and optimised at the moment to minimise inhomogeneities and maximise edge sharpness. Typical inhomogeneities are around a few percent in both the root-mean-square error between the desired and obtained pattern and the root-mean-square flatness of uniform light patterns, with the edge sharpness of the pattern being limited by the optical resolution of the setup. An example of the effect of the feedback algorithm is shown Figure 1.
The investigation of strongly correlated many-body states requires cooling the atoms to lower temperatures and entropies. In order to achieve this, we cool the atoms in a multi-stage process, starting with laser cooling in a magneto-optical trap and evaporative cooling in a dipole trap, before the atoms are transported to the science chamber. They are evaporatively cooled in a crossed dipole trap before we prepare a single layer of atoms in a vertical standing wave. We then shine in a so-called “dimple trap”, which consists of a single red-detuned laser beam. In this approach, which is also followed by other groups, the laser beam creates a tightly confining trap and increases the atomic density. The hotter atoms confined in the harmonic trap of the vertical lattice were removed by evaporation using a magnetic-field gradient, leaving a low-entropy sample. A large part during the time of the fellowship has been devoted to cool the atomic sample further to higher phase space densities. We had to add several laser beams (an elliptical “squeezing beam” and the aforementioned dimple trap) to our already complex setup. The Dimple beam allowed us to increase the number of atoms confined in a single layer of the vertical lattice by a factor of three. During the process of optimizing the cooling we had discovered and now fully understood a new gray molasses laser cooling technique and this result has recently been published [G. Bruce et al., JPhysB 50, 2017]. We have much better characterized and optimized the evaporative cooling procedure during the different stages to the experiment. As a result, we have achieved higher phase space densities and a sample of 1.2·10^4 atoms near quantum degeneracy at T/Tf =0.5. Part of the characterisations had the goal to determine whether the atoms populate the lowest band of the lattice and this was diagnosed with band mapping techniques and the evaporative cooling procedures and lattice loading was optimized accordingly.
