Dark-State Enhanced Loading of an Optical Tweezer Array

Open Access

Dark-State Enhanced Loading of an Optical Tweezer Array

Adam L. Shaw, Pascal Scholl, Ran Finklestein, Ivaylo S. Madjarov, Brandon Grinkemeyer, and Manuel Endres

Phys. Rev. Lett. 130, 193402 – Published 12 May 2023

Abstract

Neutral atoms and molecules trapped in optical tweezers have become a prevalent resource for quantum simulation, computation, and metrology. However, the maximum achievable system sizes of such arrays are often limited by the stochastic nature of loading into optical tweezers, with a typical loading probability of only 50%. Here we present a species-agnostic method for dark-state enhanced loading (DSEL) based on real-time feedback, long-lived shelving states, and iterated array reloading. We demonstrate this technique with a 95-tweezer array of ${}^{88}{Sr}$ atoms, achieving a maximum loading probability of 84.02(4)% and a maximum array size of 91 atoms in one dimension. Our protocol is complementary to, and compatible with, existing schemes for enhanced loading based on direct control over light-assisted collisions, and we predict it can enable close-to-unity filling for arrays of atoms or molecules.

Received 24 February 2023
Accepted 25 April 2023

DOI:https://doi.org/10.1103/PhysRevLett.130.193402

Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article’s title, journal citation, and DOI.

Published by the American Physical Society

Physics Subject Headings (PhySH)

Quantum information with atoms & light

Trapped atoms

Optical tweezers

Atomic, Molecular & OpticalQuantum Information, Science & Technology

Authors & Affiliations

Adam L. Shaw¹, Pascal Scholl ¹, Ran Finklestein¹, Ivaylo S. Madjarov¹, Brandon Grinkemeyer ², and Manuel Endres ^1,*

¹Division of Physics, Mathematics and Astronomy, California Institute of Technology, Pasadena, California 91125, USA
²Department of Physics, Harvard University, Cambridge, Massachusetts 02138, USA

^*mendres@caltech.edu

Article Text

Click to Expand

References

Click to Expand

Issue

Vol. 130, Iss. 19 — 12 May 2023

Reuse & Permissions

Author publication services for translation and copyediting assistance advertisement

Images

Figure 1
(a) A scheme for dark-state enhanced loading (DSEL) of an optical tweezer array. (1) Atoms (or molecules) are initially stochastically loaded into a tweezer array, and imaged to determine their locations. (2) Atoms are transferred to a long-lived shelving state which is dark to the loading dynamics, and (3) occupied traps are ramped down in depth to limit reloading. (4) The loading procedure is repeated, stochastically reloading the unoccupied deep traps while leaving the shelved dark-state atoms unaffected, and (5) all atoms are returned to the ground state for an effective increase in filling fraction. The procedure can be repeated for successively higher filling fraction. (b) DSEL is feasible to implement in arrays of alkaline-earth atoms (clock state shelving, e.g., Sr), alkali atoms (hyperfine state shelving, e.g., Rb), or molecules (rotational state shelving, e.g., CaF); the latter two both require tweezers be loaded from an ODT, rather than directly from a MOT. In this work we show an experimental demonstration with the alkaline-earth species ${}^{88}{Sr}$ .
Reuse & Permissions
Figure 2
(a) Low-lying state manifold for ${}^{88}{Sr}$ . Of particular importance is the ${}^{3}{P_{0}}$ state which is long lived, and dark to imaging, cooling, and MOT dynamics. (b) Consecutive, single-shot fluorescence images of (top) all atoms initially in the ground state; (middle) select atoms locally shelved via frequency-selective motional sideband driving to ${}^{3}{P_{0}}$ , and thus hidden from imaging; (bottom) all atoms brought back to the ground state, indicating the survival of the ${}^{3}{P_{0}}$ atoms. In all images 461, 689, and 707 nm light is illuminating the array. (c) ${}^{3}{P_{0}}$ state population (orange) at $8 μ K$ trap depth while the blue and red MOT light is on, showing a 23(4) s lifetime. In contrast, atoms in ${}^{1}{S_{0}}$ (gray) at this trap depth are quickly heated and expelled from the trap in less than $2 μ s$ with the MOT light on (inset).
Reuse & Permissions
Figure 3
(a) DSEL results in a 95-tweezer array. Shown in the main plot are histograms of the filling fraction of the array after each successive repeated loading cycle. The zeroth cycle is the standard, nonenhanced loading with a loading fraction of 45.88(5)%, while after five DSEL cycles a maximum of 84.02(4)% is achieved. Inset: Average and standard deviation of the filling fraction as a function of loading cycle (markers), compared against the prediction from Eq. (1) based on a per-cycle loading fidelity of 0.46, and a cycle-to-cycle survival probability of 0.93 (pink line, see text). (b) Exemplary single-shot fluorescence images from a single set of DSEL cycles, showing the largest-achieved 91 atom array. Sites above the atom-detection threshold are circled in black, while those below the threshold are circled in blue. Note that atoms are additionally rearranged to the middle of the array between each cycle, though this is not required. (c) Cycle-to-cycle survival (red markers) is limited by imperfect dark-state transfer fidelity (blue fill), dark-state lifetime (purple line), and vacuum-limited survival (open markers) during the reloading cycle, here 280 ms; $x$ -error bars indicate uncertainty in the exact ramp-down trap depth due to a fluctuating bias on a photodiode used for stabilization. The combination of these effects is shown as a pink line. An optimum in survival emerges at a ramp-down factor of $1 / 60$ th (a trap depth of $8 μ K$ ). At this trap depth, the probability of reloading an already occupied tweezer is $≲ 0.1 %$ (bottom). Data in (c) are taken in a 39-tweezer array with higher fidelity survival and dark-state transfer compared with the 95-tweezer array in (a) and (b).
Reuse & Permissions
Figure 4
(a) Predicted total filling fraction from Eq. (1) with varying number of DSEL cycles, $κ$ , and a fixed per-cycle filling fraction, $f_{0} = 0.5$ . (b) Total filling fraction after DSEL for various cycle numbers, per-cycle filling fractions, $f_{0}$ , and cycle-to-cycle survival probability, $p_{0}$ ; $f_{0}$ values greater than 0.5 are achievable by combining DSEL with existing methods for directly controlling the collisional processes during loading [22, 23, 24, 25, 26, 27].
Reuse & Permissions

Physical Review Letters