Two-species five-beam magneto-optical trap for erbium and dysprosium

We report on the first realization of a two-species magneto-optical trap (MOT) for the highly magnetic erbium and dysprosium atoms. The MOT operates on an intercombination line for the respective species. Owing to the narrow-line character of such a cooling transition and the action of gravity, we demonstrate a trap geometry employing only five beams in the orthogonal configuration. We observe that the mixture is cooled and trapped very efficiently, with up to $5\times {10}^8$ Er atoms and ${10}^9$ Dy atoms at temperatures of about 10 $\mu \mathrm{K}$. Our results offer an ideal starting condition for the creation of a dipolar quantum mixture of highly magnetic atoms.

Figure 1

Illustration of the vacuum apparatus, including the optical setup for the 5B MOT, and atomic properties of Er and Dy. (a) Er-Dy vacuum apparatus, including the high-temperature oven, transversal cooling chamber, Zeeman slower (ZS), and main chamber. The atomic beam propagates from right to left. The ZS beam is reflected by a metallic mirror in vacuum. (b) Sketch of the working principle of the open-top MOT. Arrows depict MOT beams, and the blue region indicates the ZS beam. For clarity, we depict the atomic clouds displaced from each other. (c, d) Energy level diagrams for Er and Dy up to 26 000 ${\mathrm{cm}}^{-1}$ at different total electronic angular momentum quantum numbers $J$. States with odd (even) parity are indicated by black (orange) horizontal lines. Arrows show the broad and narrow laser-cooling transitions. (e) Table listing the atomic properties of Er and Dy.

Figure 2

Loading curve of the two-species MOT for the 6B and 5B configurations. Filled and open squares (circles) show data for the 6B and 5B Er (Dy) MOTs, respectively. The corresponding lines are fits to the data, as detailed in the text. Fit parameters are listed in Table 1. Inset: Loading rate of the two species dependent on the effusion cell temperature $T_{\text{EC}}$. Note that the hot lip is always kept at $T_{\text{HL}}=T_{\text{EC}}+100{}^{\circ }\mathrm{C}$.

Figure 3

Dependence of atom numbers in the cMOT on the initial MOT detuning, in units of the respective linewidth of the narrow transition of ${\mathrm{\Gamma }}_{583}/2\pi =190\mathrm{kHz}({\mathrm{\Gamma }}_{626}/2\pi =136\mathrm{kHz}$) for Er (Dy). Both species show broad ranges of detuning for efficient MOT loading in the 6B configuration, while 5B shows a narrower range. The optimal detunings are nearly equal for the 5B and 6B configurations. Data were taken for $5\mathrm{s}$ of MOT loading.

Figure 4

Lifetime of the cMOT for both Er (red) and Dy (blue) in the 5B (open symbols) and 6B (filled symbols) configurations. We adjust the MOT loading time to compare samples with equal atom numbers. Solid lines show exponential fits to the respective data. From the fits we extract lifetimes of 515(65) and 475(5) ms [374(44) and 345(23) ms] for Er [Dy] in the 5B and 6B cMOTs, respectively.