Analyzing the optical pumping on the $5s4d\,{}^1D_2-5s8p\,{}^1P_1$ transition in a magneto-optical trap of Sr atoms

TL;DR

This work addresses the loss of Sr MOT atoms caused by decay from the cooling cycle to metastable states and demonstrates that optical pumping on the $5s4d^1D_2-5s8p^1P_1$ transition at $448$ nm can markedly boost MOT population in a 3D trap. By saturating the $448$ nm transition, the authors achieve an atom-number enhancement of $12.0(6)$, about six times larger than the previously explored $717$ nm repumper, and attribute the remaining loss to bypass channels $5s5p^1P_1 \to 5s4d^3D_{1,2}$. They measure the decay rates $A_1 = 66(6)$ s$^{-1}$ and $A_2 = 2.4(2)\times10^2$ s$^{-1}$ for the critical decays and show that upper-state decay to $5s5p^3P_J$ is negligible, while escape losses at small trap-beam diameters are suppressed by the 448 nm light. These results improve laser cooling, fluorescence imaging, and state-detection fidelity in Sr platforms, including optical tweezer arrays for quantum information processing.

Abstract

We explore the efficacy of optical pumping on the $5s4d\,{}^1D_2 - 5s8p\,{}^1P_1$ ($448\,\mathrm{nm}$) transition in a magneto-optical trap (MOT) of Sr atoms. The number of trapped atoms is enhanced by a factor of $12.0(6)$ relative to the case without repumping light, which is six times as large as that obtained using the pumping transition $5s4d\,{}^1D_2 - 5s6p\,{}^1P_1$ ($717\,\mathrm{nm}$). This enhancement is limited by decay pathways that bypass the $5s4d\,{}^1D_2$ state, namely $5s5p\,{}^1P_1 \to 5s4d\,{}^3D_1 \to 5s5p\,{}^3P_0$ and $5s5p\,{}^1P_1 \to 5s4d\,{}^3D_2 \to 5s5p\,{}^3P_2$, which account for 8% of the total loss of the trapped atoms. We determine the decay rates for the $5s5p\,{}^1P_1 \to 5s4d\,{}^3D_1$ and $5s5p\,{}^1P_1 \to 5s4d\,{}^3D_2$ transitions to be $66(6)\,\mathrm{s^{-1}}$ and $2.4(2)\times10^2\,\mathrm{s^{-1}}$, respectively. Furthermore, we experimentally demonstrate for the first time that, when the trap beam diameter is small, escape of atoms in the $5s4d\,{}^1D_2$ state, which has a relatively long lifetime of $400\,\mathrm{μs}$, becomes a dominant loss mechanism, and that the $448\,\mathrm{nm}$ pumping light effectively suppresses this escape. Our findings will contribute to improved laser cooling and fluorescence imaging in cold strontium atom platforms, such as quantum computers based on optical tweezer arrays.

Figures (6)

• Figure 1: Energy level diagram of strontium relevant to this study. Dashed arrows indicate decay processes, among which those highlighted in red represent decay pathways that lead to MOT loss.
• Figure 2: (a) Loading curves of the MOT under different repumping schemes. The detuning of the MOT beams is set to $-36\,\mathrm{MHz}$. The inset is an enlarged view of the loading curve for no repumping scheme (461 nm only) with a fit to the data. (b) An enlarged view of all the loading curves.
• Figure 3: Decay rates and branching ratios relevant to MOT loss.
• Figure 4: Dependence of the MOT loss rate on the excitation fraction of the $5s5p\,{}^1P_1$ state. The inset shows the MOT decay at a detuning of $-26\,\mathrm{MHz}$ ($f=0.14$).
• Figure 5: Evolution of the trapped atom number for MOT beam diameters of $11\,\mathrm{mm}$ and $6\,\mathrm{mm}$. Initially, the MOT is loaded with the ${}^3P_2$ repumping light at $481\,\mathrm{nm}$, the ${}^3P_0$ repumping light at $483\,\mathrm{nm}$, and the ${}^1D_2$ optical pumping light at $448\,\mathrm{nm}$. At $t=10\,\mathrm{s}$, the $448\,\mathrm{nm}$ light is switched off. After the number of atoms stabilizes, the $448\,\mathrm{nm}$ light is turned on again. The detuning of the MOT beams is set to $-36\,\mathrm{MHz}$.