${}^{39}\mathrm{K}$ Bose-Einstein Condensate with Tunable Interactions

We produce a Bose-Einstein condensate of ${}^{39}\mathrm{K}$ atoms. Condensation of this species with a naturally small and negative scattering length is achieved by a combination of sympathetic cooling with ${}^{87}\mathrm{Rb}$ and direct evaporation, exploiting the magnetic tuning of both inter- and intraspecies interactions at Feshbach resonances. We explore the tunability of the self-interactions by studying the expansion and the stability of the condensate. We find that a ${}^{39}\mathrm{K}$ condensate is interesting for future experiments requiring a weakly-interacting Bose gas.

Figure 1

Phase transition to a ${}^{39}\mathrm{K}$ BEC. The three profiles are taken at different times during the final stage of forced evaporation in the optical trap (see Fig. 2). The profiles are obtained by vertically integrating the column density measured after release from the trap and 15 ms of ballistic expansion.

Figure 2

Sketch of the relevant parameters of the system during the evaporation in the optical potentials. The trap depth for K (solid line) is always larger than the one for Rb (dashed line). The two exponential ramps for K have time constants of 0.9 and 0.45 s, respectively.

Figure 3

Sample images of a ${}^{39}\mathrm{K}$ condensate at various magnetic fields in the vicinity of the 52 G-wide Feshbach resonance centered at 402.4 G, taken after 31.5 ms of expansion. The size shrinks as the scattering length $a_{\mathrm{K}}$ is decreased, and the condensate eventually collapses for negative $a_{\mathrm{K}}$. The field of view is $300\times 300\mu \mathrm{m}$.

Figure 4

Tuning the interaction in a ${}^{39}\mathrm{K}$ condensate. (a) atom number; (b) width of the cloud after 31.5 ms of ballistic expansion; (c) theory prediction for the scattering length. The two dashed lines indicate the expected position of the zero-crossing and resonance center. Condensates are either fitted with a Thomas-Fermi profile (circles) in the region of large interactions or a Gaussian profile (dots) in the region of weak interactions. Atom number and width of uncondensed clouds are directly extracted from the raw images (triangles). Each data point is the average of at least three measurements.

Figure 5

(a) Scattering-length dependence of measured (dots) and calculated (lines) mean width $\sigma$. The two dashed lines indicate the range of variation of $\sigma$ due to a 30% systematic uncertainty on atom numbers. (b) Zoom into the zero-crossing region. Both condensed samples (dots) and collapsed samples (triangles) are shown. The horizontal error bars are determined by the uncertainty on $a_{\mathrm{K}}(B)$. The hatched region indicates the critical $a_{\mathrm{K}}$ for collapse as predicted by the numerical calculation, including its variation due to the uncertainty in the atom number.