Preparation of ultracold atomic-ensemble arrays using time-multiplexed optical tweezers

We use optical tweezers based on time-multiplexed acousto-optic deflectors to trap ultracold cesium atoms in one-dimensional arrays of atomic ensembles. For temperatures between $2.5\mu \mathrm{K}$ and $50\mathrm{nK}$ we study the maximal time between optical tweezer pulses that retains the number of atoms in a single trap. This time provides an estimate of the maximal number of sites in an array of time-multiplexed optical tweezers. We demonstrate evaporative cooling of atoms in arrays of up to 25 optical tweezer traps and the preparation of atoms in a box potential. Additionally, we demonstrate three different protocols for the preparation of atomic-ensemble arrays by transfer from an expanding ultracold atomic cloud. These result in the preparation of arrays of up to 74 atomic ensembles consisting of $\sim 100$ atoms on average.

Figure 1

Experimental setup schematically showing the optical tweezers. The position of the optical tweezers' focal plane is controlled by the position of a lens before the AODs. The two-dimensional position of the optical tweezers is controlled by the AODs. The optical tweezers cross the horizontal dimple beam inside the experimental vacuum chamber. The guide beam (not shown) is perpendicular to the page.

Figure 2

Characterization of multiplexing. (a) Temperature of the atomic cloud as a function of the evaporation time in a single optical tweezer trap. The vertical dashed line marks the start of the magnetic-field-gradient ramp down. (b) An example of a $t_{\max }$ measurement ($t_{\mathrm{evap}}= 2700\mathrm{ms}$), showing the number of atoms retained in the trap for different $t_{\text{on-off}}$. $t_{\max }$ is determined as the time for which the number of retained atoms rapidly drops. (c) Maximal time $t_{\max }$ between optical tweezer pulses that maintains the number of atoms in the trap as a function of the temperature and the linear fit to the measured data.

Figure 3

Arrays loaded directly from a dipole trap. (a) An absorption image of atoms in an array of 20 tweezers of equal intensity. (b) Atoms in an array of 20 sites with a V-shaped profile of tweezer intensities. All other parameters are the same as in (a). The red bars illustrate the intensities of the optical tweezers. (c) Absorption image of a 25-site array of atomic ensembles in optical tweezers, loaded directly from the large dipole trap. The image is an average of 10 realizations. Below, we plot ${\sigma }_{N_a}/\sqrt{N_a}$ and $N_a$, where ${\sigma }_{N_a}$ is the variance of the number of atoms within each trap determined from 10 measurements, and $N_a$ is the average number of atoms in each trap. The dashed line marks ${\sigma }_{N_a}/\sqrt{N_a}=1$, which represents the Poissonian distribution of the number of atoms in each trap. (d) Splitting of 5 atomic ensembles into 10. Absorption images taken at 10-ms intervals of the 100-ms splitting. All images use the same color bar, except (c), where the same colors correspond to optical densities from 0 to 0.2. The optical tweezers are time multiplexed with a frequency of $100\mathrm{kHz}$. This means that each trap is active for $10\mu \mathrm{s}$, and its duty cycle is $1/N$, where $N$ is the number of traps forming the array.

Figure 4

Characterization of the number of traps forming the potential for preparation of cold atoms in a line box-shaped potential. (a) The number of atoms after $3000\mathrm{ms}$ of evaporation in a box-shaped potential as a function of the number of traps $N$ forming the box. (b) The number of atoms in a box potential after $2800\mathrm{ms}$ of evaporation in an array of five tweezers and an additional $400\mathrm{ms}$ of evaporation in a box potential formed by $N$ tweezer traps. (c) Absorption image (color bar in Fig. 3) and the linear density profile $p$ of atoms in a box-shaped potential, prepared with a transfer from 5 to 40 sites. Average of 10 measurements.

Figure 5

Arrays loaded from a dimple trap using one-dimensional time of flight or expansion in a harmonic potential. Protocol A: loading the tweezers from an expanding cloud. (a) The linear density of the atomic ensembles in an array of 10 optical tweezers as a function of time after the optical tweezers are turned on. A breathing-mode oscillation can be seen. Protocol B: expansion of the atomic cloud in a painted harmonic potential and switching to an array of optical tweezers after half the breathing oscillation. (b) The oscillation of the atomic cloud in the painted harmonic potential and (c) the resulting array (optical density image and the normalized variance ${\sigma }_{N_a}/\sqrt{N_a}$). Protocol C: ramping the optical tweezer array power and position with the expansion of the atomic cloud. (d) and (e) The results of protocol C with 60 and 80 traps (74 filled), respectively. (c)–(e) show an average of 10 realizations of the experiment, and the variance is determined from the same 10 measurements. The color bar is the same as in Fig. 3.