Coulomb crystals---ordered structures of cold ions confined in ion traps---have applications in a variety of research fields. The number and temperature of the ions forming the Coulomb crystals are two key attributes of interest in many trapped-ion experiments. Here, we present a fast and accurate approach to determining these attributes from fluorescence images of the ions based on convolutional neural networks (CNNs). In this approach, we first generate a large number of images of Coulomb crystals with different numbers of ions at different temperatures using molecular-dynamics simulations and then train CNN models on these images to classify the desired attributes. The classification performance of several common pretrained CNN models is compared in example tasks. We find that for crystals with numbers of ions in the range 100--299 and secular temperatures of 5--15 mK, the best-performing model can discern variations of a single ion with an accuracy of 93% and temperature changes of 1 mK with an accuracy of 92%. Since the trained models can be directly integrated into experiments, in situ determination of these attributes can be realized in a noninvasive fashion. This has the potential to greatly enhance the analysis and control of trapped ions in real time.
Coulomb crystals---ordered structures of cold ions confined in ion traps---have applications in a variety of research fields. The number and temperature of the ions forming the Coulomb crystals are two key attributes of interest in many trapped-ion experiments. Here, we present a fast and accurate approach to determining these attributes from fluorescence images of the ions based on convolutional neural networks (CNNs). In this approach, we first generate a large number of images of Coulomb crystals with different numbers of ions at different temperatures using molecular-dynamics simulations and then train CNN models on these images to classify the desired attributes. The classification performance of several common pretrained CNN models is compared in example tasks. We find that for crystals with numbers of ions in the range 100--299 and secular temperatures of 5--15 mK, the best-performing model can discern variations of a single ion with an accuracy of 93% and temperature changes of 1 mK with an accuracy of 92%. Since the trained models can be directly integrated into experiments, in situ determination of these attributes can be realized in a noninvasive fashion. This has the potential to greatly enhance the analysis and control of trapped ions in real time.
Accurate determination of the temperature of trapped ion Coulomb crystals (ICCs) is critical for advancing quantum technologies. We present a deep learning approach to classify ICC temperature states from fluorescence images. A high-fidelity dataset was generated using the (py)LIon molecular dynamics toolkit, simulating ICCs containing 100--268 ions over a temperature range of 5--20 mK. In the initial 16-class problem (each class a 1-mK interval), a ConvNeXt convolutional network achieved a peak accuracy of 57%, demonstrating its ability to capture fine-grained structural features. To overcome low distinguishability among similar temperature classes, we redefined the task into four broader temperature intervals and applied a vision transformer (ViT). The ViT model achieved a test accuracy of 84%, indicating substantially improved generalization. These results highlight the complementarity of convolutional neural network and transformer architectures for noninvasive temperature diagnostics. We discuss limitations due to dataset diversity and outline future work on larger-scale simulations and physics-guided network designs to further improve ICC temperature estimation in quantum applications.
We describe a simple and fast method for simulating observed images of ion Coulomb crystals. In doing so, cold elastic collisions between Coulomb crystals and virtual very light atoms are implemented in a molecular dynamics (MD) simulation code. Such an approach reproduces the observed images of Coulomb crystals by obtaining density plots of the statistics of existence of each ion. The simple method has the advantage of short computing time in comparison with previous calculation methods. As a demonstration of the simulation, the formation of a planar Coulomb crystal with a small number of ions has been investigated in detail in a linear ion trap both experimentally and by simulation. However, also large Coulomb crystals including up to 1400 ions have been photographed and simulated to extract the secular temperature and the number of ions. For medium-sized crystals, a comparison between experiments and calculations has been performed. Moreover, an MD simulation of the sympathetic cooling of small molecular ions was performed in order to test the possibility of extracting the temperature and the number of refrigerated molecular ions from crystal images of laser-cooled ions. Such information is basic to studying ultracold ion-molecule reactions using ion Coulomb crystals including sympathetically cooled molecular ions.
Laser cooling and trapping techniques are foundational to modern atomic physics, with applications including atomic clocks and sensors as well as quantum simulators and computers. These applications require stable, well-calibrated laser cooling and accurate determination of atom number and temperature. We develop machine learning techniques for estimating physical properties of laser-cooled potassium-39 atoms in a magneto-optical trap using only the scattered light—i.e., fluorescence—that is intrinsic to the cooling process. In situ snapshot images of fluorescing atomic ensembles directly reveal the spatial structure of these millimeter-scale objects but contain no obvious information regarding internal properties such as the temperature. We first assembled and labeled a balanced dataset sampling 8 × 10 3 different experimental parameters that includes examples with large and dense atomic ensembles, a complete absence of atoms, and everything in between. We describe a range of models trained to predict atom number and temperature solely from fluorescence images. These run the gamut from a poorly performing linear regression model based only on integrated fluorescence to deep neural networks that give number and temperature with fractional uncertainties of 0.1 and 0.2, respectively. These results show that fluorescence images can provide fast, nondestructive diagnostics for cold-atom experiments and may enable real-time experimental feedback.
Coulomb crystal yields a wide range of applications including quantum computing and simulation, phase transitions, quantum logic spectroscopy, nonlinear dynamics and chaos, chemical reaction process, etc. The structure of the Coulomb crystal and the trajectory of each trapped ion are typically determined by the parameters of the trap and the ion species. However, dark ions are often inevitable in experiments, which introduces uncertainty to the desired crystal structures and ion trajectories. Few research has been conducted to investigate the configuration change of the crystal in the presence of dark ions and the influence of a dark ion on its surrounding ion trajectories in a multi-ion system. In this paper, we utilize the molecular dynamics simulation software LAMMPS and the (py)Lion package (modified to adapt the semi-classical theory of laser cooling) for the simulation of the 3D ion trajectories of Coulomb crystals. The formation process of 40Ca+ Coulomb crystals in a linear trap is simulated. With the micromotion and secular motion trajectories of each ion, we calculate the temperature of Coulomb crystal and the average velocity of specific ions. It’s observed that the crystal structure exhibits obvious layering phenomenon when the trapped ions yield a large difference in their charge-to-mass ratio (CMR), however, layering is not obvious with a small difference in the CMR. In addition, we simulate and compare the Coulomb crystal structure formed by pure 40Ca+ ions and that formed by 40Ca+ ions mixed with a small number of dark ions including isotopic ions (44Ca+) and impurity ions (CaH+). Three different cases are investigated, namely the 1D ion string, 2D plane structure and 3D helical structure. Results show that ions in the neighborhood of a dark ion exhibit around micron-order position changes compared to their positions before the dark ion is formed. Such a change can be measured in experiment through microscopic imaging, providing a way to identify dark ions in Column crystals with a large number of ions.
Coulomb crystals have applications in many areas such as quantum computing and simulation, quantum logic spectroscopy, nonlinear dynamics and chaos, phase transitions, and chemical reaction process. The structure of the Coulomb crystal and the trajectory of each trapped ion are typically determined by the parameters of the trap and the ion species. However, dark ions are often inevitable in experiment, which introduces uncertainty into the desired crystal structures and ion trajectories. Few researches have been conducted to investigate the configuration change of the crystal in the presence of dark ions and the influence of a dark ion on its surrounding ion trajectories in a multi-ion system. In this work, we utilize the molecular dynamics simulation software LAMMPS and the (py)LIon package (modified to adapt the semi-classical theory of laser cooling) for simulating the three-dimensional ion trajectories of Coulomb crystals. The formation process of 40Ca+ Coulomb crystal in a linear trap is simulated. With the micromotion and secular motion trajectories of each ion, we calculate the temperature of Coulomb crystal and the average velocity of specific ions. It is observed that the crystal structure exhibits obvious layering phenomenon when the trapped ions yield a large difference in their charge-to-mass ratio (CMR), however, layering is not obvious with a small difference in the CMR. In addition, we simulate and compare the Coulomb crystal structure formed by pure 40Ca+ ions with that formed by 40Ca+ ions mixed with a small number of dark ions including isotopic ions (44Ca+) and impurity ions (CaH+). Three different cases are investigated, namely the one-dimensional ion string, two-dimensional planar structure and three-dimensional helical structure. The results show that the ions in the neighborhood of a dark ion exhibit around micron-order position change compared with their positions before the dark ion is formed. Such a change can be measured in experiment through microscopic imaging, thereby providing a way to identify the formation of dark ions in Column crystals with a large ion number.
This thesis describes the study of a range of ion-molecule reactions at very low collision energies using a newly developed experimental technique which involves the reaction of velocity-selected beams of translationally cold neutral molecules with very low kinetic energy ion ensembles. These studies have been enabled by the construction of a new apparatus for trapping and laser-cooling gas phase atomic ions (40Ca⁺). The laser-cooling process results in the formation of ordered, low kinetic energy, lattice-like ion structures, also known as "Coulomb crystals". The properties of single and multicomponent Coulomb crystals (which may also involve molecular ions), and their manipulation via modulation of the applied fields, are explored experimentally and with the use of molecular dynamics simulations. Variations in the laser-cooling parameters are shown to result in different steady-state populations of the electronic states of 40Ca⁺ involved with the laser cooling cycle, and these are modelled within an appropriate theoretical framework. The imaging of 40Ca⁺ fluorescence as a function of time allows the study of various ion-molecule reactions at collision energies around 300 K, with single ion sensitivity. These reaction studies are extended to low-temperature (collision energies close to 1 K), by combination of the ion trap apparatus with a bent quadrupole guide velocity-selector. Ion-molecule collision energies are shown to be variable over a short range through a change in the quadrupole guide voltage, or the ion trapping parameters; the effect of these modulations on the rate constant is explored for Ca⁺ + CH₃F. Bimolecular rate constants for the reactions of 40Ca⁺ with CH₃F, CH₂F₂ and CH₃Cl have been determined for a range of 40Ca⁺ state populations, allowing resolution of the global rate contributions from the ground and combined excited states. These results are analysed in the context of capture theories and ab initio electronic structure calculations. In each case, suppression of the ground state rate constant is explained by the presence of either a submerged or real barrier on the ground state potential surface. Rates of reaction from the combined excited states are generally found to be in line with capture theories, and in some cases variation is found between the high and low collision energy regimes. Molecular product ions generated in these experiments have been shown to be sympathetically-cooled into the crystal structure, and subsequently identified through resonance-excitation mass spectrometry. Molecular ions were also produced by multiphoton laser ionisation of a thermal background gas of OCS molecules. An ion-molecule reaction involving a molecular ion, that of charge transfer between OCS⁺ and ND₃, has been studied at a collision energy near 1 K for the first time using sympathetically-cooled OCS⁺ and velocity-selected ND₃. These experiments illustrate the generality of the techniques described herein, and should lead to many possibilities for future studies.