© 2026 Aqora Quantum S.A.S.

Terms Privacy Legal Notice

arXiv:2110.03931 · AVS Quantum Science

Machine learner optimization of optical nanofiber-based dipole traps

Ratnesh Kumar Gupta, Jesse L. Everett, Aaron D. Tranter, René Henke, Vandna Gokhroo, Ping Koy Lam, Síle Nic Chormaic

Published May 11, 2022• 22 citations

Abstract

We use a machine learning optimizer to increase the number of rubidium-87 atoms trapped in an optical nanofiber-based two-color evanescent dipole trap array. Collisional blockade limits the average number of atoms per trap to about 0.5, and a typical uncompensated rubidium trap has even lower occupancy due to challenges in simultaneously cooling atoms and loading them in the traps. Here, we report on the implementation of an in-loop stochastic artificial neural network machine learner to optimize this loading by optimizing the absorption of a near-resonant, nanofiber-guided, probe beam. By giving the neural network control of the laser cooling process, we observe an increase in peak optical depth of 66% from 3.2 ± 0.2 to 5.3 ± 0.3. We use a microscopic model of the atomic absorption to infer an increase in the number of dipole-trapped atoms from 300 ± 60 to 450 ± 90 and a small decrease in their average temperature from 150 to 140 μK. The machine learner is able to quickly and effectively explore the large parameter space of the laser cooling control process so as to find optimal parameters for loading the dipole traps. The increased number of atoms should facilitate studies of collective atom–light interactions mediated via the evanescent field.

DOI: 10.1116/5.0086507

Related papers

Machine learner optimization of atom loading in optical nanofiber evanescent dipole traps

Ratnesh Kumar Gupta, Jesse L. Everett, Aaron D. Tranter +4 more

We use an online machine learning algorithm to optimize cooling and loading of rubidium-87 atoms into an evanescent dipole trap array along an optical nanofiber, increasing the number of trapped atoms by 50%.

Quantum 2.0 Conference and Exhibition•Jan 1, 2022•0 citations

Machine-Learning Optimization and Characterization of a High-Optical-Depth Two-Color Nanofiber Trap

W. Crump, M. Sadeghi, M. D. Hoogerland

Optical nanofibers provide a way of coupling quantum information in cold atoms across large distances, however, this coupling requires atoms to reside close to the nanofiber surface. Atoms can be trapped close to the surface using a two-color dipole trap. Here we present our experimental realization of a two-color dipole trap. We optimize the number of trapped atoms using a machine learning algorithm and measure the optical density via the transmission. We estimate the number of atoms in the trap to be approximately 1400 and the lifetime of the atoms in the trap to be 28 ms. Machine-learning optimization improved the on-resonance optical depth from 0.5 in the initial optimization stage to optical depths exceeding 15.

quant-pharXiv:2606.06798•Jun 8, 2026•0 citations

Machine learning optimization and characterization of a high-optical depth two-color nanofiber trap

Wayne Crump, M. Sadeghi, Maarten D. Hoogerland

Optical nanofibers provide a way of coupling quantum information in cold atoms across large distances; however, this coupling requires atoms to reside close to the nanofiber surface. Atoms can be trapped close to the surface using a two-color dipole trap. Here, we present our experimental realization of a two-color dipole trap. We optimize the number of trapped atoms using a machine learning algorithm and measure the optical density via the transmission. We estimate the number of atoms in the trap to be approximately 1400 and the lifetime of the atoms in the trap to be 28 ms. Machine learning optimization improved the on-resonance optical depth from 0.5 in the initial optimization stage to optical depths exceeding 15.

AVS Quantum Science•Jun 1, 2026•1 citation

Using optical nanofibres to mediate cold atom interactions

Síle Nic Chormaic

We discuss several recent advances related to optical nanofibres in cold atom systems, including two-colour dipole trap optimisation using an in-loop stochastic artificial neural network machine learner, upper bound limitations on Rydberg atom excitation due to localised ion formation, spectral lineshapes arising from the high intensity evanescent fields, and two-photon processes mediated via the evanescent field of the optical nanofibre.

EPJ Web of Conferences•Jan 1, 2022•0 citations

$Λ$-enhanced gray-molasses loading and EIT cooling of neutral atoms in nanophotonic traps

Lucas Pache, Antoine Glicenstein, Philipp Schneeweiss +3 more

Nanophotonic traps for cold atoms typically have trap volumes that are orders of magnitude smaller than, e.g., free-space optical tweezers. This makes efficient loading of these traps challenging, thereby limiting the total number of atoms coupled to the nanophotonic waveguide. Here, we demonstrate that $Λ$-enhanced gray-molasses ($Λ$GM) can substantially increase the number of trapped atoms in a nanofiber-based cold-atom setup. Specifically, we observe a six-fold increase in the number of loaded atoms compared to conventional red-detuned polarization gradient cooling. Despite the unusually small depth of our optical trap of only 24 $μ$K, we load about 4000 individual Cesium atoms, achieving optical depths exceeding 140 and reaching the collisional blockade regime over a length of approximately 1 mm. After loading, we perform efficient EIT-assisted cooling that is found to increase the trap storage time to 400(9) ms. This is a 5-fold improvement over the passive storage time. Remarkably, EIT-cooling also works with two co-propagating nanofiber-guided light fields and requiries only about a few hundred picowatt of optical power. Our results provide an efficient method to boost both the number of loaded atoms and the storage time of nanophotonic atom traps.

physics.atom-phCross-listedarXiv:2605.13387•May 14, 2026•0 citations

Experiments Towards the Realization of a Nanofiber Guided Dipole Trap

Di Chang

In this thesis I present a developing project to elaborate a scientific platform towards the realization of light-matter interaction based on tapered optical nanofiber (TNF) that enables optical fields to interface with trapped neutral atoms in their vicinity. In such a configuration, the guided light exhibits a large portion of evanescent field into the vacuum thus laser cooled atoms can couple to the fiber mode. By trapping and cooling atoms around the nanofiber I investigated the interaction of the atoms with the field. My work presented in this thesis includes: Calculating the evanescent modes propagating through our fiber; Implementing fiber pulling setup using oxyhydrogen flame and analyzing nanofiber’s thermal dynamics and degradation; Constructing a Rubidium 87 MOT and characterizing its critical parameters; Realizing some atom-light interface through an experiment on coupling spontaneously emitted photons from MOT into TNF; Calculating two-color dipole trap parameters for our nanofiber.

•Jan 1, 2015•0 citations

We use a machine learning optimizer to increase the number of rubidium-87 atoms trapped in an optical nanofiber-based two-color evanescent dipole trap array. Collisional blockade limits the average number of atoms per trap to about 0.5, and a typical uncompensated rubidium trap has even lower occupancy due to challenges in simultaneously cooling atoms and loading them in the traps. Here, we report on the implementation of an in-loop stochastic artificial neural network machine learner to optimize this loading by optimizing the absorption of a near-resonant, nanofiber-guided, probe beam. By giving the neural network control of the laser cooling process, we observe an increase in peak optical depth of 66% from 3.2 ± 0.2 to 5.3 ± 0.3. We use a microscopic model of the atomic absorption to infer an increase in the number of dipole-trapped atoms from 300 ± 60 to 450 ± 90 and a small decrease in their average temperature from 150 to 140 μK. The machine learner is able to quickly and effectively explore the large parameter space of the laser cooling control process so as to find optimal parameters for loading the dipole traps. The increased number of atoms should facilitate studies of collective atom–light interactions mediated via the evanescent field.

Optical nanofibers provide a way of coupling quantum information in cold atoms across large distances, however, this coupling requires atoms to reside close to the nanofiber surface. Atoms can be trapped close to the surface using a two-color dipole trap. Here we present our experimental realization of a two-color dipole trap. We optimize the number of trapped atoms using a machine learning algorithm and measure the optical density via the transmission. We estimate the number of atoms in the trap to be approximately 1400 and the lifetime of the atoms in the trap to be 28 ms. Machine-learning optimization improved the on-resonance optical depth from 0.5 in the initial optimization stage to optical depths exceeding 15.

Optical nanofibers provide a way of coupling quantum information in cold atoms across large distances; however, this coupling requires atoms to reside close to the nanofiber surface. Atoms can be trapped close to the surface using a two-color dipole trap. Here, we present our experimental realization of a two-color dipole trap. We optimize the number of trapped atoms using a machine learning algorithm and measure the optical density via the transmission. We estimate the number of atoms in the trap to be approximately 1400 and the lifetime of the atoms in the trap to be 28 ms. Machine learning optimization improved the on-resonance optical depth from 0.5 in the initial optimization stage to optical depths exceeding 15.

Λ

-enhanced gray-molasses loading and EIT cooling of neutral atoms in nanophotonic traps

Nanophotonic traps for cold atoms typically have trap volumes that are orders of magnitude smaller than, e.g., free-space optical tweezers. This makes efficient loading of these traps challenging, thereby limiting the total number of atoms coupled to the nanophotonic waveguide. Here, we demonstrate that

Λ

-enhanced gray-molasses (

Λ

GM) can substantially increase the number of trapped atoms in a nanofiber-based cold-atom setup. Specifically, we observe a six-fold increase in the number of loaded atoms compared to conventional red-detuned polarization gradient cooling. Despite the unusually small depth of our optical trap of only 24

μ

K, we load about 4000 individual Cesium atoms, achieving optical depths exceeding 140 and reaching the collisional blockade regime over a length of approximately 1 mm. After loading, we perform efficient EIT-assisted cooling that is found to increase the trap storage time to 400(9) ms. This is a 5-fold improvement over the passive storage time. Remarkably, EIT-cooling also works with two co-propagating nanofiber-guided light fields and requiries only about a few hundred picowatt of optical power. Our results provide an efficient method to boost both the number of loaded atoms and the storage time of nanophotonic atom traps.

In this thesis I present a developing project to elaborate a scientific platform towards the realization of light-matter interaction based on tapered optical nanofiber (TNF) that enables optical fields to interface with trapped neutral atoms in their vicinity. In such a configuration, the guided light exhibits a large portion of evanescent field into the vacuum thus laser cooled atoms can couple to the fiber mode. By trapping and cooling atoms around the nanofiber I investigated the interaction of the atoms with the field. My work presented in this thesis includes: Calculating the evanescent modes propagating through our fiber; Implementing fiber pulling setup using oxyhydrogen flame and analyzing nanofiber’s thermal dynamics and degradation; Constructing a Rubidium 87 MOT and characterizing its critical parameters; Realizing some atom-light interface through an experiment on coupling spontaneously emitted photons from MOT into TNF; Calculating two-color dipole trap parameters for our nanofiber.