Direct Measurement of the $5s5p\,{}^1P_1 \to 5s4d\,{}^1D_2$ Decay Rate in Strontium

Abstract

We investigate the decay process $5s5p\,{}^1P_1 \to 5s4d\,{}^1D_2 \to 5s5p\,{}^3P_2$ in a magneto-optical trap of Sr atoms operating on the $461\,\mathrm{nm}$ ($5s^2\,{}^1S_0 - 5s5p\,{}^1P_1$) transition by irradiating the trapped atoms with laser light resonant with the $448\,\mathrm{nm}$ ($5s4d\,{}^1D_2 - 5s8p\,{}^1P_1$) transition and observing the transient response of atom fluorescence. We measure, for the first time, the branching ratio of the $5s4d\,{}^1D_2 \to 5s5p\,{}^3P_2$ transition to be $0.177(4)$, which significantly deviates from the widely cited theoretical value of $0.322$ [C. W. Bauschlicher Jr. et al., J. Phys. B 18, 1523 (1985)]. Moreover, we determine the decay rate of the $5s5p\,{}^1P_1 \to 5s4d\,{}^1D_2$ transition to be $5.3(5)\times10^3\,\mathrm{s^{-1}}$, consistent within uncertainty with the widely cited experimental value [L. R. Hunter et al., Phys. Rev. Lett. 56, 823 (1986)], but substantially lower than the recent theoretical value of $9.25(40)\times10^3\,\mathrm{s^{-1}}$ [A. Cooper et al., Phys. Rev. X 8, 041055 (2018)]. These findings have significant implications for laser cooling of Sr and fluorescence detection of single atoms in optical tweezers. They also call for a reevaluation of the previous theoretical frameworks used to calculate the decay rates of the $5s5p\,{}^1P_1 \to 5s4d\,{}^1D_2$ and $5s4d\,{}^1D_2 \to 5s5p\,{}^3P_{1,2}$ transitions.

Figures (4)

• Figure 1: Energy level diagram of Sr relevant to this study.
• Figure 2: Change in the trapped atom number after switching off the $448\,\mathrm{nm}$ light, with the $461\,\mathrm{nm}$, $481\,\mathrm{nm}$, and $483\,\mathrm{nm}$ lights kept on. The data represent the average of 64 experimental traces. The inset shows the change in the trapped atom number after switching off the $481\,\mathrm{nm}$ light while the $461\,\mathrm{nm}$ and $483\,\mathrm{nm}$ lights remain on. The detuning of the MOT beams is set to $-36\,\mathrm{MHz}$.
• Figure 3: Extracted branching ratios for different trapping light detunings. The error bars represent the uncertainties derived from the fits to the decay curves as shown in Fig. \ref{['fig:448nmoff']}. The solid line indicates the weighted average of the data, and the gray band represents the overall $\pm 1\sigma$ statistical uncertainty. For reference, the dashed line shows the theoretical value ($0.322$) reported by Bauschlicher et al.C.W.Bauschlicher1985. The reduced $\chi^2$ value for the combined data is 0.96, corresponding to a $P$ value of 0.46, indicating that the branching ratio shows no statistically significant dependence on the detuning.
• Figure 4: Comparison of the literature values and the present measurement for the $5s5p\,{}^1P_1 - 5s4d\,{}^1D_2$ transition rate C.W.Bauschlicher1985L.R.Hunter1986J.Kwela1988H.G.C.Werij1992S.G.Porsev2001A.Cooper2018. Circular markers indicate experimental results, and triangular markers indicate theoretical values.