High performance imaging of $^{171}$Yb atom in shallow clock-magic tweezer by alternating dual-tone narrowline cooling

Abstract

We demonstrate imaging $^{171}$Yb single atoms in clock-magic tweezers of 759.4 nm wavelength, with above 99.9% fidelity and survival. We use alternating dual-tone narrowline imaging for more efficient three-dimensional cooling in tweezers, allowing several-millisecond imaging in 200 $μ$K trap depth, which is half of typical depth used for imaging in clock-magic tweezers. Accordingly, even without repumping, imaging survival is still close to 99.9% with the high fidelity, which can enable high performance nondestructive qubit measurements based on metastable shelving. Moreover, our simulation predicts that more optimal configuration could further reduce the trap depth, as improving the imaging performance. This imaging capability in shallow traps opens high performance imaging for more general trap wavelength, and lays the foundation for large scale systems over 1,000 qubits, and highly repeatable tweezer clocks.

Figures (7)

• Figure 1: (a) Schematic of alternating cooling in a tweezer trap. (b) Dual-tone driving of ${^1S_0},F=1/2-{^3P_1},F=3/2$ transitions of $^{171}$Yb atom is shown with relative strength. With a moderate B field (15 G for our case) and the natural linewidth $\Gamma=2\pi\times182$ kHz, the ground state splitting is negligible, while the excited state can be selectively driven by each tone. (c,d) $3\times3$$^{171}$Yb array image (c) of a single shot and (d) averaged over 20,000 shots by 5.4 ms alternating dual-tone narrowline imaging. (e) Correlation plot between detected photons of two sequential shots in the $5\times5$ pixels region of interest (ROI) in (c) and (d). Dashed lines indicate optimized thresholds. Inset exhibits log-scale histogram for the total detected photons in the ROI.
• Figure 2: (a,b) Narrowline driving of ${^1S_0},F=1/2-{^3P_1},F=3/2,\left|m_F\right|=1/2$ transition. (a) Single-tone driving results in a lambda system, which is MS-transformed to a two-state system and a dark state. (b) Dual-tone driving results in a double lambda system, which is MS-transformed to 2 separate two-state systems. (b,c) Dependence of detected photon rate for dual-tone driving on polarization angle $\theta$ with phase difference (b) $\phi=\pi/2$ and (c) $\phi=0$, respectively, of which electric field is $\vec{E}=e^{i\phi}E\cos\theta\hat{e}_\perp+E\sin\theta\hat{e}_\parallel$. Solid (dashed) lines show theoretical values based on Eq. \ref{['eq_scattering']} with (without) considering transmission difference between P and S polarizations at glass cell. For these measurements, fixed input imaging beam intensity ($s=2$ for $\theta=0^\circ$) and detuning ($\Delta=-0.92\Gamma$) are used. Error bar indicates standard error of the mean.
• Figure 3: (a) MC simulation of Doppler cooling in a tweezer trap as a function of polar angle $\theta_k$ between k-vector direction of imaging beam and tweezer axis, with intensity $s=2$ and detuning $\Delta=-\Gamma$. 1-axis cooling in radially symmetric trap, 1-axis and 2-axis cooling in asymmetric trap (10% anisotropy) are compared. (b) Release and recapture measurement for atom temperature after 2-axis alternating cooling depending on alternating frequency, with $s=2$, $\Delta=-\Gamma$, and release time of 20$~\rm\mu s$. Colored shades indicate the measurement results without alternating in various configurations. Purple solid line is MC simulation for 2-axis cooling.
• Figure 4: (a) Temperature measurement by 20-$\rm\mu s$ release and recapture for alternating dual-tone imaging. Colored dots are measured data with specified intensity saturation parameters, and solid lines are corresponding MC simulation results. (b) Atom loss per detected photon during the imaging. Colored dots are experimental data with $s=1,2,3$ and $\Delta=-0.77\Gamma,-\Gamma,-1.1\Gamma$, respectively. Red, blue, and green dashed lines indicate corresponding MC simulations, and black dotted line is calculated trap Raman scattering to metastable states. Colored solid lines are summation of the losses from the MC simulation and the Raman scattering. For (a) and (b), alternating frequency is 2.5 kHz.
• Figure 5: High fidelity and survival imaging of $^{171}$Yb array of 0.2 mK trap depth with intensity $s=1$, detuning $\Delta=-0.77\Gamma$, imaging time $t_{\rm imag}=5.4~{\rm ms}$, and alternating frequency of 5 kHz. (a) Calculation of loss per emitted photon as a function of relative differential AC Stark shift of the imaging excited state. Dashed and dotted lines indicate the Stark shifts of ${^3P_1},F=3/2,m_F=\pm1/2$ states, respectively, when the angle $\theta_B$ between B field (15 G) and trap E field (depth 0.2 mK) is about $16.5^\circ$. Dash-dotted line corresponds to the Stark shift at $\theta_B=0^\circ$. (b) Atom retention after repeated imagings, with or without repumping beams. (c) Histogram of detected photons for each site fitted by Gaussian plus exponential distributions (black solid lines). (d) Imaging lifetime measurement with or without repumping beams. For comparison, data with only trap beam is also plotted.