Laser Cooling without Repumping: A Magneto-Optical Trap for Erbium Atoms

We report on a novel mechanism that allows for strong laser cooling of atoms that do not have a closed cycling transition. This mechanism is observed in a magneto-optical trap (MOT) for erbium, an atom with a very complex energy level structure with multiple pathways for optical-pumping losses. We observe surprisingly high trap populations of over atoms and densities of over , despite the many potential loss channels. A model based on recycling of metastable and ground state atoms held in the quadrupole magnetic field of the trap explains the high trap population, and agrees well with time-dependent measurements of MOT fluorescence. The demonstration of trapping of a rare-earth atom such as erbium opens a wide range of new possibilities for practical applications and fundamental studies with cold atoms.

Figure 1

(a) Energy levels of erbium vs $J$ value, showing the laser cooling transition and the many metastable levels. Odd parity is indicated by gray lines (red online). (b) Schematic of the recycling model described by Eq. (1). Atoms decay from the MOT with effective rate $f_{\mathrm{ex}}{R}_1$ to a manifold of metastable levels. Atoms are either lost from the metastable manifold with a rate $R_3$, or decay to the ground state with rate $R_2$ to form a reservoir of atoms held by the quadrupole magnetic field of the MOT. Reloading of the MOT occurs from the reservoir with a rate $R_{\mathrm{load}}$.

Figure 2

Typical time dependence of the MOT fluorescence, beginning when the MOT light is turned on after both the slower and MOT beams have been off for 0.2 s. A clear two-component exponential decay is observed. Experimental values are indicated by open circles, and a fit of the model of Eq. (1) is indicated by a solid line (red online).

Figure 3

Values of $f_{\mathrm{ex}}{R}_1$ extracted from the model fit vs MOT power for five different detunings. Symbols indicate experimental values, with uncertainties about the size of the plotting symbols. Solid lines indicate a fit of Eq. (2) to the data. Note that Eq. (2) was fit simultaneously to all detunings and intensities to yield a single set of parameters $R_1$, $a_1$, and $a_2$.