Direct evaporative cooling of ${}^{39}$K atoms to Bose-Einstein condensation

We report the realization of a Bose-Einstein condensate of ${}^{39}$K atoms without the aid of an additional atomic coolant. Our route to Bose-Einstein condensation comprises sub-Doppler laser cooling of large atomic clouds with more than ${10}^{10}$ atoms and evaporative cooling in an optical dipole trap where the collisional cross section can be increased using magnetic Feshbach resonances. Large condensates with almost ${10}^6$ atoms can be produced in less than 15 s. Our achievements eliminate the need for sympathetic cooling with Rb atoms, which was the usual route implemented until now due to the unfavorable collisional property of ${}^{39}$K. Our findings simplify the experimental setup for producing Bose-Einstein condensates of ${}^{39}$K atoms with tunable interactions, which have a wide variety of promising applications, including atom interferometry to studies on the interplay of disorder and interactions in quantum gases.

Figure 1

Schematic of the vacuum system. Note that the windows are directly glued to the 2D MOT and 3D MOT chambers, which are made of medical titanium.

Figure 2

Numerically calculated elastic collision rate $\sigma v$, where $\sigma$ is the elastic cross section and $v$ is the relative velocity of the atom pair as a function of the collision energy in the center-of-mass frame. The calculation is performed for atoms in the $|F=1,m_F=-1\rangle$ hyperfine state in a vanishing magnetic field, and it includes $s$ and $d$ partial waves. Collision model parameters are as in Ref. [29].

Figure 3

Atom number trapped from the quadrupole into the dipole trap as a function of the loading time. The measurement is performed for different values of the laser power.

Figure 4

Optimized variation of the dipole trap power during the evaporative cooling. Solid squares and solid circles represent the estimated trap depth and the measured temperature of the atoms, respectively.

Figure 5

Phase-space density vs atom number during the evaporative cooling for the realization of a condensate in the $m_F=1$ state close to the 402.4 G Feshbach resonance. The vertical dotted lines are placed where the $s$-wave scattering length is changed. The dashed line connecting the initial and final points corresponds to an overall evaporative cooling efficiency of 3.4. Note that throughout the lowering of the trapping beam power, we find more efficient evaporative cooling by tuning the scattering length to larger values.

Figure 6

Condensed fraction vs temperature around the BEC phase transition point. The theoretical curve is $1-{(T/T_c)}^3$ and includes finite size and interaction shifts of the critical temperature [35]. For condensed fractions above 80$\%$, the temperature was not measurable due to the low signal in the thermal tails of the distribution. In the insets, we show the absorption images and the atomic density profiles of the cloud above (a), around (b), and below (c) the condensation critical temperature.