Measurement of inelastic losses in a sample of ultracold ${}^{85}\mathrm{Rb}$

We report on measurements of inelastic loss processes in ultracold ${}^{85}\mathrm{Rb}$ $|F=2\rangle$ atoms. Our apparatus creates ultracold ${}^{85}\mathrm{Rb}$ clouds by sympathetic cooling with a ${}^{87}\mathrm{Rb}$ reservoir in a quadrupole-Ioffe magnetic trap and subsequently in a weak, large-volume optical dipole trap. We demonstrate strong sympathetic cooling of ${}^{85}\mathrm{Rb}$ in the magnetic trap, increasing its phase-space density by three orders of magnitude with no detectable loss in number. Ultracold samples created in this way are used to observe the variation of inelastic loss in ${}^{85}\mathrm{Rb}$ $|F=2,\hspace{0.5em}{m}_F=-2\rangle$ clouds as a function of magnetic field near the $155$-G Feshbach resonance and to measure the decay due to inelastic losses in all five Zeeman sublevels of the $F=2$ manifold, finding a particularly high three-body recombination rate in the lowest energy state. We have also observed and characterized a previously unobserved loss feature at $219.9(1)$ G with a width of $0.28(6)$ G, which we associate with a narrow Feshbach resonance predicted by theory.

Figure 1

Atom number as a function of temperature during rf-induced evaporative cooling in the magnetic trap. The evaporation trajectories of ${}^{85}\mathrm{Rb}$ (circles) and ${}^{87}\mathrm{Rb}$ (squares) are shown during both single-species (open) and sympathetic cooling (solid).

Figure 2

(a) Measured inelastic loss of ${}^{85}\mathrm{Rb}$ from the optical dipole trap as a function of magnetic field, with the elastic scattering length from Ref. [17] overlaid (solid line). The minimum inelastic loss occurs at 168 G (dashed line). (b) Zoomed-in section showing a sharp increase in inelastic loss associated with a Feshbach resonance at 220 G. The dashed line is a Gaussian fit to the experimental data.

Figure 3

Decay curves for ${}^{85}\mathrm{Rb}$ samples prepared in each of the five Zeeman substates of the $F=2$ manifold. The error bars shown are statistical uncertainties calculated from five runs of the experiment. The solid lines are fits to the data of the form $e^{-t/\tau }$, from which a loss rate $\tau$ can be extracted.