Single Cs atoms as collisional probes in a large Rb magneto-optical trap

We study cold interspecies collisions of cesium and rubidium in a strongly imbalanced system with single and few Cs atoms. Observation of the single-atom fluorescence dynamics yields insight into light-induced loss mechanisms, while both subsystems can remain in steady state. This significantly simplifies the analysis of the dynamics, as Cs-Cs collisions are effectively absent and the majority component remains unaffected, allowing us to extract a precise value of the Rb-Cs collision parameter. Extending our results to ground-state collisions would allow to use single neutral atoms as coherent probes for larger quantum systems.

Figure 1

Schematic of the experimental setup to operate and detect a strongly imbalanced two-species MOT. Both MOTs of Rb and Cs are produced at the same position using the same quadrupole coils. The Rb MOT is loaded by an atomic beam (orange) provided by a second vapor pressure MOT, whereas Cs is loaded from the background of the ultrahigh vacuum. The Rb and Cs atom numbers are determined by fluorescence detection setups located at the side and at the bottom of the glass cell, respectively. The Cs fluorescence light collected by a high numerical aperture objective is detected by an APD-SPCM with a quantum efficiency of $42\%$. To suppress stray light as well as Rb fluorescence and laser light a 300-$\mu$m pinhole and two interference filters serve as spatial and spectral filters, respectively. The Rb fluorescence is collected by a lens, spectrally filtered and detected by an APD.

Figure 2

(a)–(c) Typical fluorescence traces of the single-Cs-atom fluorescence signal for a selection of three different Rb atom numbers: The detected pulses are binned in time intervals of 20 ms each. Each panel shows two typical fluorescence traces of single Cs atoms in the presence of about 550, 2300, and 3000 Rb atoms, respectively. Within the gray region the cooling laser light and the magnetic field is switched off. For the last 200 ms (yellow shading) the laser light is switched on again to determine the background light level. The Cs-Cs collision event shown in (a) at 2.5 s is one of the very few observed ones. (d)–(f) Corresponding histograms of the Cs fluorescence signal for the same Rb atom number as in (a)–(c). Each peak corresponds to 0,1,2, … , atoms, respectively. The red lines are Gaussian fits to the peaks. We find that, not only in steady state, but for all data sets the probability distribution can be well described by a Poissonian function. The insets compare the measured probability distribution (bars) with a fitted Poissonian distribution (red dots).

Figure 3

Dynamics of the Cs MOT depending on the number of Rb atoms trapped: (a) Loading rate. The red line is a linear fit to the data yielding $R_0=(1.48\pm 0.06)$ s${}^{-1}$ and $\alpha =(2.3\pm 0.3)·{10}^{-4}$ s${}^{-1}$. (b) Time-averaged loss counts, and (c) ratio of loading rates and time-averaged loss counts. The vertical dashed line divides the graphs into two sections. Within section 1 the loading rate dominates the Cs MOT dynamics. For larger Rb atom number (section 2) the ratio between loading rates and time-averaged loss counts tends to unity (horizontal dashed line), which corresponds to Cs being in steady state.

Figure 4

Mean of the Poissonian distribution of Cs atom numbers as a function of the Rb atom number and peak density. The red solid line is a fit of Eq. (2) to the data within the steady-state regime. The red dashed line is the extrapolation to the non-steady-state regime of the Cs MOT.