Single Sr Atoms in Optical Tweezer Arrays for Quantum Simulation

TL;DR

This work presents a Sr-88 optical-tweezer platform that integrates blue and red MOT cooling, light-assisted collision-based single-atom loading, and high-fidelity fluorescence detection in programmable tweezer arrays. By employing a blue shielded MOT, broadband and single-frequency red MOT stages, and Sisyphus in-trap cooling, the system achieves ≈5×10^5 atoms at high densities, with single-atom occupancy realized via collisional blockade at ~50% probability. Single-atom readout reaches 99.986% fidelity with 97% survival, and release-and-recapture measurements yield a tweezer temperature around 12.9 μK, while vacuum lifetimes exceed 7 minutes, indicating excellent stability. The platform provides essential building blocks for scalable quantum simulation and information processing with alkaline-earth Sr atoms, including high-fidelity detection, long coherence potential, and pathways toward Rydberg-enabled many-body operations.

Abstract

We report on the realization of a platform for trapping and manipulating individual $^{88}$Sr atoms in optical tweezers. A first cooling stage based on a blue shielded magneto-optical trap (MOT) operating on the $^1S_0$ -> $^1P_1$ transition at 461 nm enables us to trap approximately $4\times 10^6$ atoms at a temperature of 6.8 mK. Further cooling is achieved in a narrow-line red MOT using the $^1S_0$ -> $^3P_1$ intercombination transition at 689 nm, bringing $4\times 10^5$ atoms down to 5 $μ$K and reaching a density of $\approx 10^{10}$ cm$^{-3}$. Atoms are then loaded into 813 nm tweezer arrays generated by crossed acousto-optic deflectors and tightly focused onto the atoms with a high-numerical-aperture objective. Through light-assisted collision processes we achieve the collisional blockade, which leads to single-atom occupancy with a probability of about $50\%$. The trapped atoms are detected via fluorescence imaging with a fidelity of $99.986(6)\%$, while maintaining a survival probability of $97(2)\%$. The release-and-recapture measurement provides a temperature of $12.92(5)$ $μ$K for the atoms in the tweezers, and the ultra-high-vacuum environment ensures a vacuum lifetime higher than 7 min. These results demonstrate a robust alkaline-earth tweezer platform that combines efficient loading, cooling, and high-fidelity detection, providing the essential building blocks for scalable quantum simulation and quantum information processing with Sr atoms.

Figures (6)

• Figure S1: Overview of the experimental setup. (a) Vacuum system composed of a commercial atomic source, connected through a differential pumping tube to a pumping chamber which holds an ion--NEG pump, a Bayard--Alpert gauge and an angle valve. The fused silica glass science cell, where trapping and detection occur, is surrounded by two sets of magnetic coils: a pair of high-field coils providing the MOT gradient and strong bias fields for clock-state manipulation, and six smaller compensation coils that enable full three-dimensional control and active cancellation of stray fields. (b) Optical layout for tweezer generation and imaging. A high-power $\qty{813}{\nano\meter}$ beam passes through a pair of crossed AODs to produce an array of deflected beams. A relay telescope and a high-numerical-aperture objective (NA = 0.55) couple the angular deflection to the focal plane of the objective to form an array of spatially-displaced tightly-focused optical tweezers. The same objective collects atomic fluorescence, which is separated by a dichroic mirror and imaged onto a low-noise qCMOS detection camera for single-atom readout.
• Figure S2: Temporal sequence of the atomic cooling, trapping in tweezers and imaging process. The diagram shows the different experimental stages (Blue MOT, Multifrequency Red MOT, Single Frequency Red MOT, LAC, Imaging and cooling) with their durations, along with the time evolution of the optical beam intensities, magnetic field gradient, and tweezer potential depth.
• Figure S3: Blue shielded MOT characterization. (a) Simplified electronic level structure of [88]Sr displaying the transitions of interest. (b) Loading curves of the shielded blue MOT for different repumping configurations. The presence of the $\qty{707}{\nano\meter}$ repumper (green) increases the number of trapped atoms by preventing optical pumping into the dark metastable $\ket{^3\mathrm{P}_2}$ state. When the $\qty{679}{\nano\meter}$ repumper is also simultaneously applied (red), the loss channel to the dark metastable $\ket{^3\mathrm{P}_0}$ state is also inhibited, which increases the atom number by more than one order of magnitude compared to the case without any repumper (blue). (c) Relative enhancement of the atom number by the red-shielding beam as a function of the frequency detuning from resonance, assuming the far from resonance values as baseline. Error bars in (b,c) represent standard deviations.
• Figure S4: Absorption images of laser-cooled [88]Sr clouds: (a) a blue shielded MOT with $\sim 1.25 \times 10^7$ atoms at $\qty{6.8}{\milli\kelvin}$ and (b) a single-frequency red MOT with more than $5\times10^{5}$ atoms at $\qty{5.4}{\micro\kelvin}$. The false-color optical density scale is the same in both absorption images to allow direct visual comparison.
• Figure S5: Characterization of single atoms in tweezers through fluorescence detection. (a) The histogram of photon counts collected in ROIs centered at the atom positions (averaged over the array) features a bimodal distribution corresponding to empty and occupied traps, that is fitted with the sum of two skewed gaussians (red line); the dashed line at $\sim 23$ photon counts indicates the threshold value that maximizes the fidelity of true positive single-atom detection. Insert: typical fluorescence image of a $3\times3$ array of single atoms, obtained by averaging 100 individual acquisitions. (b) Trapped atom survival probability and detection fidelity $\mathcal{F}$ (in red in the inset that focuses on the first points) as a function of the probe beam scattering rate. All data is taken with a fixed imaging duration of $\qty{150}{\milli\second}$, chosen to ensure that for every scattering rate the detected signal was sufficiently strong to allow accurate evaluation of both quantities. The shaded region indicates the optimal scattering-rate value. Error bars represent standard deviations calculated with the bootstrapping method Efron1993.