Sub-Doppler cooling of sodium atoms in gray molasses

We report on the realization of sub-Doppler laser cooling of sodium atoms in gray molasses using the $D_1$ optical transition $\left(3s{}^{2}S_{1/2}\to 3p{}^{2}P_{1/2}\right)$ at 589.8 nm. The technique is applied to samples containing $3\times {10}^9$ atoms, previously cooled to $350\mu \mathrm{K}$ in a magneto-optical trap, and it leads to temperatures as low as $9\mu \mathrm{K}$ and phase-space densities in the range of ${10}^{-4}$. The capture efficiency of the gray molasses is larger than 2/3, and we observe no density-dependent heating for densities up to ${10}^{11}{\mathrm{cm}}^{-3}$.

Figure 1

(a) Sodium $D_1$ level structure. Cooler (blue, solid) and repumper (red, dashed) transitions are shown. (b) Sketch of the $D_1$ laser source.

Figure 2

(a) Capture efficiency (blue circles) and temperature (red squares) after a single GM pulse at constant intensity $I_C=80I_{\mathrm{sat}}$ and fixed detuning ${\mathrm{\Delta }}_{\mathrm{pulse}}=4\mathrm{\Gamma }$, as a function of the pulse duration $\tau$, (b) as a function of the detuning ${\mathrm{\Delta }}_{\mathrm{pulse}}$ with $\tau =0.5\mathrm{ms}$ and $I_C=80I_{\mathrm{sat}}$, and (c) as a function of the total cooler intensity $I_C$ with $\tau =0.5\mathrm{ms}$ and ${\mathrm{\Delta }}_{\mathrm{pulse}}=5\mathrm{\Gamma }$. Error bars correspond to one standard deviation. The repumper relative intensity and detuning are fixed to $I_{\mathrm{rep}}/I_{\mathrm{C}}\simeq 4\%$ and $\delta =0$, respectively.

Figure 3

(a) Cooler detuning $\mathrm{\Delta }/\mathrm{\Gamma }$ (blue line) and intensity $I_{\mathrm{C}}/I_{\mathrm{sat}}$ (red line) as a function of time for the single GM pulse and (b) for the GM sequence with ramp. Shown intensity values already take into account the reduction of AOMs diffraction efficiency for high detuning.

Figure 4

(a) Capture efficiency (blue circles) and temperature (red squares) after the complete ramp sequence characterized as a function of $I_{\mathrm{rep}}/I_{\mathrm{C}}$ for $\delta =0$ and ${\mathrm{\Delta }}_{\mathrm{ramp}}=10\mathrm{\Gamma }$, (b) as a function of the ramp detuning ${\mathrm{\Delta }}_{\mathrm{ramp}}$ for $\delta =-0.01\mathrm{\Gamma }$ and $I_{\mathrm{rep}}/I_{\mathrm{C}}=0.034$, and (c), (d) as a function of $\delta$ for ${\mathrm{\Delta }}_{\mathrm{ramp}}=12\mathrm{\Gamma }$ and $I_{\mathrm{rep}}=0.034I_{\mathrm{C}}$. Error bars correspond to one standard deviation. The initial pulse parameters are fixed $\tau =0.5\mathrm{ms},{\mathrm{\Delta }}_{\mathrm{pulse}}=5\mathrm{\Gamma }$, and $I_{\mathrm{C}}=80I_{\mathrm{sat}}$.