Loading a fountain clock with an enhanced low-velocity intense source of atoms

We present experimental work for improved atom loading in the optical molasses of a cesium fountain clock, employing a low-velocity intense source of atoms [Lu et al., Phys. Rev. Lett 77, 3331 (1996)], which we modify by adding a dark-state pump laser. With this modification the atom source has a mean flux of $4\times {10}^8$ atoms/s at a mean atom velocity of $8.6$ m/s. Compared to fountain operation using background gas loading, we achieve a significant increase of the loaded and detected atom number by a factor of 40. Operating the fountain clock with a total number of detected atoms $N_{\mathrm{at}}=2.9\times {10}^6$ in the quantum projection noise-limited regime, a frequency instability ${\sigma }_y\left(1s\right)=2.7\times {10}^{-14}$ is demonstrated.

Figure 1

Cesium $D_2$ line and the frequency splitting between the hyperfine components of the $6{}^2{S}_{1/2}$ ground and $6{}^2{P}_{3/2}$ excited states (not to scale). The transitions driven by the various laser fields, present in our LVIS trap, are shown as well.

Figure 2

Schematic drawing of the enhanced low-velocity intense source system. The magnetic field gradient coils are not shown.

Figure 3

Snapshot of the LVIS MOT region made with a CCD camera. The bright central spot is the fluorescence from the trapped cesium atoms and the tail emerging from it is due to scattered photons from atoms that leave the trap and form the cold-atom beam. The conditions at which the picture is taken are 18-mW cooling laser optical power on each beam and 0.79-mT/cm axial and 0.39-mT/cm radial magnetic field gradients.

Figure 4

Detected atom number ${\mathit{N}}_{\mathrm{at}}$, resulting from fountain OM loading by means of the LVIS system, as a function of the probe laser frequency detuning from (a) the $|F=3\rangle$ $\to$ $|F^{\prime }\rangle$ and (b) the $|F=4\rangle$ $\to$ $|F^{\prime }\rangle$ transition. Results, with and without pump laser present, are shown by a solid green (top) and a dashed red (bottom) trace, respectively. The pump laser is operated on the $|F=4\rangle$ $\to$ $|F^{\prime }=3\rangle$ transition with an optical power of $0.01$ mW. At exact resonance frequencies, quantum numbers $F^{\prime }$ are indicated, while the crossover resonances in between are not denoted.

Figure 5

Plot of the detected atom number $N_{\mathrm{at}}$ as a function of the pump laser frequency at four different laser powers. For comparison, the dashed magenta trace represents $N_{\mathrm{at}}$ obtained in the standard mode of fountain operation (background gas OM loading).

Figure 6

Plot of the detected atom number $N_{\mathrm{at}}$ vs the frequency of the pump laser for different LVIS repump laser configurations. Each configuration is characterized by the operating frequency of the repump laser (driving either the $|F=3\rangle$ $\to$ $|F^{\prime }=3\rangle$ transition or the $|F=3\rangle$ $\to$ $|F^{\prime }=4\rangle$ transition) and by the particular spatial orientation of the repumping light in the LVIS trap [either the repumping light is available along the $z$ axis (atomic beam axis) or it is applied only transversely to the $z$ axis]. For all four cases the LVIS repump laser optical power was kept the same and the delivered power by the pump laser was $0.45$ mW.

Figure 7

Allan deviation of the CSF2 frequency measured against the DRO referenced to an optical cavity (symbols). The fountain frequency instability of ${\sigma }_y\left(\tau \right)=2.7\times {10}^{-14}{\tau }^{-1/2}$ is shown with a dashed line.