Figure - available from: New Journal of Physics
This content is subject to copyright. Terms and conditions apply.
![Schematic representation of image formation and information extraction. We retrieve the original atomic distribution $O(x,y)$ from the measured fluorescence image $S[{x}_{i},{y}_{j}]$ by parametric deconvolution with the point spread function $P(x,y)$ of the imaging system. Here we use the model assumption that atoms trapped in a 1D optical lattice are line-like radiation sources: their motion is tightly confined along the longitudinal direction (horizontal direction in the images) and optically not resolved (see also section 4), while it is only loosely confined along the transverse direction (vertical).](publication/288060130/figure/fig1/AS:1132484506980393@1647016755669/Schematic-representation-of-image-formation-and-information-extraction-We-retrieve-the.jpg)
Schematic representation of image formation and information extraction. We retrieve the original atomic distribution $O(x,y)$ from the measured fluorescence image $S[{x}_{i},{y}_{j}]$ by parametric deconvolution with the point spread function $P(x,y)$ of the imaging system. Here we use the model assumption that atoms trapped in a 1D optical lattice are line-like radiation sources: their motion is tightly confined along the longitudinal direction (horizontal direction in the images) and optically not resolved (see also section 4), while it is only loosely confined along the transverse direction (vertical).
Source publication
We report on image processing techniques and experimental procedures to
determine the lattice-site positions of single atoms in an optical lattice with
high reliability, even for limited acquisition time or optical resolution.
Determining the positions of atoms beyond the diffraction limit relies on
parametric deconvolution in close analogy to meth...
Similar publications
We report on experimentally observed addition, subtraction, and cancellation of orbital angular momentum (OAM) in the process of parametric four-wave mixing that results in frequency up- and down-converted emission in Rb vapor. Specific features of OAM transfer from resonant laser fields with different optical topological charges to the spatially a...
Citations
... Recent advancements have been made in the realization and manipulation of open quantum many-body systems in atomic, molecular, and optical (AMO) systems [8][9][10][11][12][13][14][15][16][17][18] , which offer precise control over individual quantum particles and enable the engineering of complex interactions and dissipation mechanisms. In addition, state-of-the-art measurement techniques, such as quantum state tomography and quantum non-demolition measurements, provide unprecedented opportunities to investigate the dynamics of these systems with high precision [19][20][21][22][23][24][25][26][27][28][29][30][31] . ...
The dynamics described by the non-Hermitian Hamiltonian typically capture the short-term behavior of open quantum systems before quantum jumps occur. In contrast, the long-term dynamics, characterized by the Lindblad master equation (LME), drive the system towards a non-equilibrium steady state (NESS), which is an eigenstate with zero energy of the Liouvillian superoperator, denoted as $\mathcal{L}$. Conventionally, these two types of evolutions exhibit distinct dynamical behaviors. However, in this study, we challenge this common belief and demonstrate that the effective non-Hermitian Hamiltonian can accurately represent the long-term dynamics of a critical two-level open quantum system. The criticality of the system arises from the exceptional point (EP) of the effective non-Hermitian Hamiltonian. Additionally, the NESS is identical to the coalescent state of the effective non-Hermitian Hamiltonian. We apply this finding to a series of critical open quantum systems and show that a local dissipation channel can induce collective alignment of all spins in the same direction. This direction can be well controlled by modulating the quantum jump operator. The corresponding NESS is a product state and maintains long-time coherence, facilitating quantum control in open many-body systems. This discovery paves the way for a better understanding of the long-term dynamics of critical open quantum systems.
... Which of these techniques is used varies throughout the community, and it is not a priory known which one achieves the highest atom detection fidelity in a specific experimental scenario. Recently, alternative approaches were introduced to reconstruct the atom distribution on a lattice using machine learning techniques [23] and a parametric deconvolution [29]. Other superresolution microscopy techniques have been invented to resolve ultracold atoms in 1D trap geometries beyond the diffraction limit, by probing them with a standing-wave, either in free space [30,31], or within a cavity [32]. ...
... As readout noise is not a dominant noise source, our simulations hold for both camera types. Other noise contributions such as clock-induced charges are not simulated as they only have a marginal impact on the overall noise level [29]. In the following we describe step by step how our simulated images are created. ...
... The LI method uses the underlying optical lattice grid for its optimisation protocol whereas the other methods use the lattice only to find the signal per lattice site. In a previous study [29], a deconvolution technique has been introduced that can be seen as a combination of the LI and Wiener method. In such a 'parametric deconvolution' , knowledge of the experiment is used, such as lattice structure and specific noise properties. ...
Quantum-gas microscopes are used to study ultracold atoms in optical lattices at the single-particle level. In these system atoms are localised on lattice sites with separations close to or below the diffraction limit. To determine the lattice occupation with high fidelity, a deconvolution of the images is often required. We compare three different techniques, a local iterative deconvolution algorithm, Wiener deconvolution and the Lucy-Richardson algorithm, using simulated microscope images. We investigate how the reconstruction fidelity scales with varying signal-to-noise ratio, lattice filling fraction, varying fluorescence levels per atom, and imaging resolution. The results of this study identify the limits of singe-atom detection and provide quantitative fidelities which are applicable for different atomic species and quantum-gas microscope setups.
... Both methods, however, rapidly decrease in fidelity for β ≳ 2 24 . Moreover, unconstrained deconvolution methods can be improved by adding further information, such as the discrete lattice grid and a sophisticated noise model, as shown in ref. 30 for a one-dimensional (1D) system. ...
In quantum gas microscopy experiments, reconstructing the site-resolved lattice occupation with high fidelity is essential for the accurate extraction of physical observables. For short interatomic separations and limited signal-to-noise ratio, this task becomes increasingly challenging. Common methods rapidly decline in performance as the lattice spacing is decreased below half the imaging resolution. Here, we present an algorithm based on deep convolutional neural networks to reconstruct the site-resolved lattice occupation with high fidelity. The algorithm can be directly trained in an unsupervised fashion with experimental fluorescence images and allows for a fast reconstruction of large images containing several thousand lattice sites. We benchmark its performance using a quantum gas microscope with cesium atoms that utilizes short-spaced optical lattices with lattice constant 383.5 nm and a typical Rayleigh resolution of 850 nm. We obtain promising reconstruction fidelities ≳ 96% across all fillings based on a statistical analysis. We anticipate this algorithm to enable novel experiments with shorter lattice spacing, boost the readout fidelity and speed of lower-resolution imaging systems, and furthermore find application in related experiments such as trapped ions.
... crease in fidelity for β 2 [24]. Moreover, unconstrained deconvolution methods can be improved by adding further information, such as the discrete lattice grid and a sophisticated noise model, as shown in Ref. [30] for a one-dimensional (1D) system. ...
In quantum gas microscopy experiments, reconstructing the site-resolved lattice occupation with high fidelity is essential for the accurate extraction of physical observables. For short interatomic separations and limited signal-to-noise ratio, this task becomes increasingly challenging. Common methods rapidly decline in performance as the lattice spacing is decreased below half the imaging resolution. Here, we present a novel algorithm based on deep convolutional neural networks to reconstruct the site-resolved lattice occupation with high fidelity. The algorithm can be directly trained in an unsupervised fashion with experimental fluorescence images and allows for a fast reconstruction of large images containing several thousand lattice sites. We benchmark its performance using a quantum gas microscope with cesium atoms that utilizes short-spaced optical lattices with lattice constant $383.5\,$nm and a typical Rayleigh resolution of $850\,$nm. We obtain promising reconstruction fidelities~$\gtrsim 96\%$ across all fillings based on a statistical analysis. We anticipate this algorithm to enable novel experiments with shorter lattice spacing, boost the readout fidelity and speed of lower-resolution imaging systems, and furthermore find application in related experiments such as trapped ions.
... The final state is measured by recording a fluorescence image of the atoms [107,117]. Using a high-resolution objective lens, the positions of the individual atoms in the optical lattices can be reconstructed with a high fidelity, exceeding 99 %. Standard fluorescence-imaging techniques, however, only give information about the occupation of modes belonging to distinct lattice sites. ...
... We start with a handful of Cs atoms, which are sparsely loaded in a onedimensional polarization-synthesized optical lattice [87]. Using the atom sorting technique presented in Ref. [88], a pair of atoms is then selected and repositioned to a relative distance of twenty lattice sites with a success rate of about 99 %, mainly limited by an incorrect detection of the initial distance between the two atoms [117]. ...
... When the atoms are bunched in the same site, inelastic lightinduced collisions result in either one or zero atoms being detected. Super-resolution microscopy allows resolving the single lattice sites despite the diffraction-limited optical resolution of ≈ 4 sites[117]. ...
Sampling from a quantum distribution can be exponentially hard for classical computers and yet could be performed efficiently by a noisy intermediate-scale quantum device. A prime example of a distribution that is hard to sample is given by the output states of a linear interferometer traversed by $N$ identical boson particles. Here, we propose a scheme to implement such a boson sampling machine with ultracold atoms in a polarization-synthesized optical lattice. We experimentally demonstrate the basic building block of such a machine by revealing the Hong-Ou-Mandel interference of two bosonic atoms in a four-mode interferometer. To estimate the sampling rate for large $N$, we develop a theoretical model based on a master equation. Our results show that a quantum advantage compared to today's best supercomputers can be reached with $N \gtrsim 40$.
... The amplification process of the EMCCD camera as well as the discretisation of the image due to the camara's pixel size introduce noise. These noise sources have been compared and quantified in a recent study [27] which we use as a basis for our simulation. In our case, a Poissonian filter is applied to simulate the signal amplification process of the EMCCD camera [28,29], and the effect of the pixel size is discussed at the end of Section IVc. ...
... The magnification, M , and pixelation, p (p = M a/d pix ), can be varied independently of this. In previous studies, the minimum number of pixels required to resolve atomic positions on a lattice without loss of information is reasoned to be p = 2.5 [27,37], based on linear-optics calculations (Shannon-Nyquist sampling). ...
Quantum-gas microscopes are used to study ultracold atoms in optical lattices at the single particle level. In these system atoms are localised on lattice sites with separations close to or below the diffraction limit. To determine the lattice occupation with high fidelity, a deconvolution of the images is often required. We compare three different techniques, a local iterative deconvolution algorithm, Wiener deconvolution and the Lucy-Richardson algorithm, using simulated microscope images. We investigate how the reconstruction fidelity scales with varying signal-to-noise ratio, lattice filling fraction, varying fluorescence levels per atom, and imaging resolution. The results of this study identify the limits of singe-atom detection and provide quantitative fidelities which are applicable for different atomic species and quantum-gas microscope setups.
... Most pertinently, light patterns used for trapping atoms tend to be scalable to very large numbers (thousands or more) [1], a built-in advantage over both ion trap quantum computing and superconducting quantum computing. Neutral atom quantum computing has met many necessary milestones such as trapping, detecting, controlling, and reading out the state of single atoms [2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17], single-qubit [18,19], two-qubit [20][21][22][23], and multi-qubit gates [24], the creation of entangled many-qubit states [25], and storage of many qubits in one- [26][27][28], two- [23,[29][30][31][32][33][34][35] and threedimensional structures [10,36]. Neutral atom transport also plays an important role in quantum information processing and computing. ...
Quantum computing is rapidly becoming a practical technology, driven by new and advanced quantum systems, each with its own advantages and disadvantages. For neutral atoms as a host system, the natural advantage of long lifetimes is limited by decoherence mechanisms like light scattering. In this work, we computationally investigate a blue-detuned one-dimensional optical dipole trap array to reduce light scattering by holding atoms in minimum light intensity regions. The array consists of two counter-propagating Gaussian beams of differing widths, such that light walls from the larger beam radially confine the standing wave generating the dipole traps. We find this system is viable for neutral atom quantum computing and able to produce similar trap properties to the traditional red-detuned case, but with lower laser power. It also features added versatility from the tunable beam waist ratio, and the ability to reduce light scattering, by twenty times in one example we present. Generally, there is a tradeoff between the scattering rate and the number of viable traps formed. This design is promising for transporting atoms, as a complimentary technique to two- and three-dimensional trap arrays for storing atoms.
... In the case of the FFPC setup shown in Fig. 1(a) this knowledge is extracted through a secondary imaging channel, independently of the primary fiber quantum channel. Fluorescence imaging, complemented by cooling measures to compensate scattering-induced heating and atom loss, has been established for long as the workhorse for positional detection of individual neutral atoms [37,38]. ...
... To maintain a sufficiently high scattering rate, this requires the repumper to be near resonant. An optimal imaging signal-to-noise ratio (SNR) [37] is found for a two-photon Raman Rabi frequency (on the carrier)˜ 0 /2π ≈ 340 kHz, a repumper blue detuning ≈ 3 MHz and a repumper intensity ≈ 0.057I sat (see Fig. 5). For such parameters we measure a scattering rate per atom of ≈ 2 × 10 4 photons/s which allows us to collect about 800 photons during a 1 s exposure time (atom survival probability >90 %). ...
... For such parameters we measure a scattering rate per atom of ≈ 2 × 10 4 photons/s which allows us to collect about 800 photons during a 1 s exposure time (atom survival probability >90 %). The fluorescence images shown in Fig. 1(b) have a SNR ≈ 13, which is sufficient to count individual atoms and determine their position with full site resolution [37]. ...
High-bandwidth, fiber-based optical cavities are a promising building block for future quantum networks. They are used to resonantly couple stationary qubits such as single or multiple atoms with photons routing quantum information into a fiber network at high rates. In high-bandwidth cavities, standard fluorescence imaging on the atom-cavity resonance line for controlling atom positions is impaired since the Purcell effect strongly suppresses all-directional fluorescence. Here, we restore imaging of 87 Rb atoms strongly coupled to such a fiber Fabry-Pérot cavity by detecting the repumper fluorescence which is generated by continuous and three-dimensional Raman sideband cooling. We have carried out a detailed spectroscopic investigation of the repumper-induced differential light shifts affecting the Raman resonance, dependent on intensity and detuning. Our analysis identifies a compromise regime between imaging signal-to-noise ratio and survival rate, where physical insight into the role of dipole-force fluctuations in the heating dynamics of trapped atoms is gained.
... The estimation of the absolute value of the light shift with our method is not very precise, because (i) most of the traces do not have sufficient data points to fit, (ii) the errors for the fluorescence step are for us on the order of 10% (see 6.6b), mainly given by the photon-shot noise and conversion processes in the camera (see e.g. Ref. [Alberti et al., 2016]). Furthermore, the precision could be improved with a calibration of the camera, to be able to extract the number of photons (see e.g. ...
Tweezer atom arrays are a promising platform for the quantum simulation of spin models. As for most quantum simulation platforms, scaling up the number of individually controlled quantum objects is a major challenge. In this thesis, I present our work on lifting principal limitations to achieving large defect-free atom arrays with high fidelities. These limitations of the preparation fidelities include the vacuum-limited lifetime of a single atom in the tweezer, and the time needed to prepare large arrays atom-by-atom with a moveable optical tweezer. We first improved the assembly of large defect-free atom arrays by developing a new algorithmic framework. Using the new framework, we increased the number of atoms from around forty to two hundred on our room-temperature setup. We then built a novel cryogenic atom tweezer platform in which the single-atom lifetime is over 6000 seconds, an approximately 300-fold improvement over our room-temperature experiment. We describe the design and construction of the new cryogenic setup and evaluate its performance in a series of tests. Finally, we demonstrate the trapping of single atoms in tweezer arrays at cryogenic temperatures and analyze different loss mechanisms present during the lifetime measurement. These results open the way to large-scale quantum simulations on our platform.
... In the case of the FFPC setup shown in Fig. 1(a) this knowledge is extracted through a secondary imaging channel, independently of the primary fiber quantum channel. Fluorescence imaging, complemented by cooling measures to compensate scatteringinduced heating and atom loss, has been established for long as the workhorse for positional detection of individual neutral atoms [34,35]. ...
... To maintain a sufficiently high scattering rate, this requires the repumper to be near resonant. An optimal imaging signal-to-noise ratio (SNR) [34] is found for a two-photon Raman Rabi frequency (on the carrier)Ω 0 /2π ≈ 400 kHz, a repumper blue detuning ∼ 3 MHz and a repumper intensity ∼ 0.057I sat (see Fig. 5). For such parameters we measure a scattering rate per atom of ∼ 2 × 10 4 photons/s which allows to collect about 800 photons during a 1 s exposure time (atom survival probability > 90 %). ...
... For such parameters we measure a scattering rate per atom of ∼ 2 × 10 4 photons/s which allows to collect about 800 photons during a 1 s exposure time (atom survival probability > 90 %). The fluorescence images shown in Fig. 1(b) have a SNR of ∼ 13, which is sufficient to count individual atoms and determine their position with full site resolution [34]. ...
High-bandwidth, fiber-based optical cavities are a promising building block for future quantum networks. They are used to resonantly couple stationary qubits such as single or multiple atoms with photons routing quantum information into a fiber network at high rates. In high-bandwidth cavities, standard fluorescence imaging on the atom-cavity resonance line for controlling atom positions is impaired since the Purcell effect strongly suppresses all-directional fluorescence. Here, we restore imaging of $^{87}$Rb atoms strongly coupled to such a fiber Fabry-P\'erot cavity by detecting the repumper fluorescence which is generated by continuous and three-dimensional Raman sideband cooling. We have carried out a detailed spectroscopic investigation of the repumper-induced differential light shifts affecting the Raman resonance, dependent on intensity and detuning. Our analysis identifies a compromise regime between imaging signal-to-noise ratio and survival rate, where physical insight into the role of dipole-force fluctuations in the heating dynamics of trapped atoms is gained.
