Beating the density limit by continuously loading a dipole trap from millikelvin-hot magnesium atoms

We load ${10}^5$ magnesium atoms in a dipole trap from a millikelvin-hot magneto-optical trap (MOT) using a continuous-loading scheme. Light-assisted two-body processes limit the maximum achievable density in a MOT, resulting in a reduced transfer efficiency into a dipole trap when using the conventional sequential scheme. It is overcome in a continuous-loading scheme where a loss channel is opened in the MOT. This allows the accumulation of atoms in the dipole trap over the trap lifetime, determined by collisions with the background gas. This results in a significantly higher number of trapped atoms even at a lower steady-state peak density in the MOT.

Figure 1

Partial energy diagram of ${}^{24}$Mg with the transitions relevant for the operation of a dual MOT in our experiment. (Dashed) arrows indicate transitions driven by (repump) lasers. The S-MOT operates on the singlet (${}^1{S}_0\to {}^1{P}_1$) transition while the T-MOT operates on the triplet (${}^3{P}_2\to {}^3{D}_3$) transition. The (${}^1{S}_0\to {}^3{P}_1$) transition is used to transfer atoms to the triplet manifold.

Figure 2

Transfer time ${\tau }_{AT}$ for atoms in the S-MOT to be transferred to the triplet system via the ${}^1{S}_0\to {}^3{P}_1$ intercombination transition in a sequential manner. The loss of atoms from the S-MOT is used to measure the transferred population. The transfer laser intensity is 3.9 kW/m${}^2$. The absolute frequency is measured with a high-precision wave meter.

Figure 3

Evolution of the number of atoms in the T-MOT. The fast decay for short times is attributed to binary collisions, whereas the exponential decay for longer times (with a time constant of 0.9 s) is due to insufficient repumping of the ${}^3{P}_1$ state. In comparison, the background-gas collision-limited lifetime (observed for the dipole trap) is 5 s. The (red) curve is a theoretical fit of a two-body loss equation to the experimental data using the model in [13]. The fluorescence of the T-MOT is detected by a CCD camera. Inset: T-MOT decay with the repumper for the ${}^3{P}_0$ state switched off. The decay has a time constant of 30 ms.

Figure 4

Loading the dipole trap using stepwise (red line with squares) and continuous (blue line with circles) schemes. The loading time for the stepwise scheme is the S-MOT loading time followed by transfer to the metastable state and T-MOT cooling, which takes a total of 50 ms. The inset shows the loading curve of the dipole trap using the continuous scheme for an increased incoming atomic flux and higher T-MOT cooling light intensity.

Figure 5

Temperature (black squares) and number (red circles) of atoms captured in the dipole trap for different trap depths. The dipole trap waist was 72 $\mu$m. Temperatures as low as 5 $\mu$K were observed.