Apparatus and method for continuously loading a neutral atom register

Abstract

The present disclosure relates to trapping and manipulating neutral atoms for quantum computing, quantum simulation, and metrology. In one aspect, a method includes: loading a plurality of neutral atoms into a loading trap array arranged adjacent to or at least partially co-located with a storage trap array; selectively removing neutral atoms in a first electronic state |1> from the storage trap array; determining trap occupancy of neutral atoms in a second electronic state |2> in the storage trap array and trap occupancy of neutral atoms in the loading trap array; and moving neutral atoms from the loading trap array to one or more unoccupied trapping sites of the storage trap array based on the determined trap occupancy of neutral atoms in the storage trap array and trap occupancy of neutral atoms in the loading trap array. The method can further include shelving the neutral atoms in a shelving state.

Description

Technical Field

[0001] This disclosure relates to apparatus, methods, and computer programs for capturing and manipulating neutral atoms, and in particular to apparatus, methods, and computer programs for continuously loading neutral atom registers, which can be used as quantum registers for quantum technologies such as quantum computing, atomic clocks, and quantum simulations.

[0002] introduction

[0003] Ensembles of neutral atoms trapped in optical traps are a promising approach for realizing quantum computers and other quantum technologies. Quantum computing or simulating quantum simulations using neutral atoms requires initializing the atomic ensembles using known quantum gas microscopy techniques: loading laser-cooled atoms into multiple optical trapping potentials formed by arrays of optical tweezers or optical lattices, or combinations thereof. Initializing the atomic ensembles to suitable registers, for example, for quantum computing, requires precisely fabricating one neutral atom at each trapping site. This has been demonstrated by actively moving atoms within their trapping potentials using moving optical tweezers. Fabricating defect-free atomic ensembles in configurable two-dimensional optical traps provides the initialization of the quantum registers required for quantum computing.

[0004] Various quantum technologies (e.g., those mentioned above) benefit from the largest possible system size, i.e., the largest number of neutral atoms in the register, and the fastest possible system initialization. The preparation of defect-free ensembles typically requires detection of initial random trapping site occupancy via established imaging techniques (e.g., fluorescence imaging using microscope objectives) before rearrangement, for example, via a movable optical trap or similar device.

[0005] Against this backdrop, M. Norcia et al. disclosed an atomic array assembly based on a combination of optical tweezers and cavity-enhanced optical lattices, using incremental filling of the target array from a repeatedly filled library, in their paper "Iterative assembly of 171Yb atom arrays incavity-enhanced optical lattices" (arXiv:2401.16177v1) published on January 29, 2024. In this scheme, optical tweezers are provided for the trapping and micro-rearrangement of atoms, while the cavity-enhanced lattice enables rapid, low-loss imaging of atoms.

[0006] In addition, the applicant’s publication, High-fidelity detection of large-scale atom arrays in an optical lattice, R. Tao, J. Zeiher, I. Bloch et al., arXiv:2309.04717v2, September 12, 2023 (hereinafter: Reference [1]), discloses high-fidelity and high-survival imaging of strontium atoms using repulsive Sisyphus cooling. The optical lattice is used as the pinning potential of atoms in a large-scale optical tweezers array with up to 399 optical tweezers, and enables repeatable high-fidelity lattice-optical tweezers-lattice transfers. The scalability of the platform is demonstrated by the direct loading of more than 10,000 atoms in a single plane of the optical lattice, which can be used as a locally addressable and sortable library of atoms for the continuous refilling of optical tweezers arrays and similar trapping arrays. For the avoidance of redundancy, Reference [1] is incorporated herein by reference in its entirety. Summary of the Invention

[0007] Extending the methods mentioned above to large system sizes requires fabricating atoms and loading them into arrays of optical traps. However, fabrication time increases with system size and constitutes a significant bottleneck in assembling large ordered arrays from typically randomly loaded optical trap arrays. Specifically, a typical experimental sequence for controlling neutral atoms begins with the fabrication of an ensemble, followed by a sequence of quantum simulations, quantum computations, or metrologies. Then, destructive measurements of the system's states are typically performed, making it impossible to recycle atoms from one experimental cycle to the next. Fabricating new atomic ensembles typically requires significantly more time than the actual experimental sequence, resulting in dead times that become significantly longer for large arrays. Therefore, there is always a need for rapid and large-scale fabrication and / or loading of neutral atom quantum registers while overcoming at least some of the limitations discussed.

[0008] The current limitation to further proportionally increasing the number of atoms in the atomic array is the absence of a continuous operation mode where atoms are stored sequentially. In this case, continuous operation means that atoms are stored losslessly in the memory register, possibly in a coherent superposition of two internal states, and lost atoms can be continuously replenished from the reserve register. Problems arise in such a scheme when reloading atoms from the memory register to the reserve register without accidentally loading them into the memory register, or equivalently, selectively removing newly loaded atoms from the memory register without removing already stored atoms.

[0009] To address such and similar problems, in a first aspect, this disclosure provides a method for capturing and manipulating neutral atoms, the method comprising: loading a plurality of neutral atoms into a loading well array (e.g., an array of optical tweezers, a portion of an optical lattice, or a combination thereof), the loading well array being arranged adjacent to or at least partially co-located with a storage well array (e.g., an array of optical tweezers, a portion of an optical lattice, or a combination thereof). The method further comprises: selectively removing atoms in a first electronic state from the storage well array. (For example, alkaline earth atoms, for example) 88 Sr or 171 Neutral atoms in the electronic ground state of Yb; determine the atoms in the storage trap array in the second electronic state. (For example, electronically excited states, e.g.) 88 Sr's long lifespan 3 The neutral atoms in the storage well array (P0) and the well occupancy of neutral atoms in the loaded well array; and based on the determined well occupancy of neutral atoms in the storage well array and the loaded well array, moving neutral atoms from the loaded well array to one or more unoccupied trap sites in the storage well array (see discussion below for details). Figures 5 to 9 ).

[0010] As discussed below, for example, a loading trap array can be formed by an array of optical tweezers co-located with a subset of the trapping sites of an optical lattice, and a storage trap array can be formed by a separate array of optical tweezers, a separate subset of the trapping sites of an optical lattice, or a different optical lattice operating at a different wavelength than the optical lattice used for the loading trap array (see discussion below for details). Figures 1 to 8 For example, a storage trap array can be formed from multiple trapping sites of a so-called magic light lattice (i.e., a light lattice that provides substantially the same polarization for the |0> and |1> states). Further details of possible implementations of load trap arrays and storage trap arrays are discussed in references [1] and [2].

[0011] In a typical implementation, the first electronic state |1> of a neutral atom can be a state with a lower energy compared to the second electronic state |2>. For example, the first electronic state |1> could be... 88 Sr 1 The S0 state, while the second electronic state |2> can be 3 The P0 state, due to its orientation towards... 1 The S0 state exhibits a double-forbidden transition and thus a natural lifetime on the order of 100 s. Clearly, other states can be used, for example... 3 The P0 state can be used as the |1> state, and 1 The S0 state can be used as the |2> state.

[0012] In some implementations, the load trap array and the storage trap array can be overlapped by interleaving the load sites and storage sites (see below). Figure 7 Using such a configuration is advantageous because it increases the available space of the storage well array and reduces the average required movement distance from the load well array to the storage well array, which reduces heating and speeds up cycle time. Typically, the load well array can be loaded from the source region of laser-cooled neutral atoms, for example, from a magneto-optical trap (MOT) of neutral atoms, as known in the art, that can operate in the same spatial region (e.g., in an ultra-high vacuum UHV environment). In some embodiments, the source region of laser-cooled atoms can be spatially separated from the load well array. In such a configuration, the laser-cooled atoms first need to be transferred to a region of the load well array, for example, by means of a transfer dipole trap, a transfer lattice, or similar devices known in the art.

[0013] For example, selectively removing devices in the first electronic state from the storage trap array. The presence of neutral atoms ensures that only atoms in the second electronic state |2> are present when well occupancy is determined, thus allowing for the reliable assembly of large neutral atom arrays in storage well arrays. As discussed below, the selective removal of neutral atoms in the |1> state from the storage well array can be achieved via state- and well-selective heating laser pulses that essentially leave the neutral atoms in the loaded well array and those in the |2> state unaffected (see below). Figure 12 ).

[0014] Another aspect of this disclosure relates to a method for capturing and manipulating neutral atoms. The method includes: transferring a plurality of neutral atoms captured in a storage well array from a first electronic state |1> to a second electronic state |2>; loading a plurality of laser-cooled neutral atoms into a loaded well array arranged adjacent to or at least partially co-located with the storage well array; determining the well occupancy of neutral atoms in the second electronic state |2> in the storage well array and the neutral atoms in the loaded well array; and, based on the determined well occupancy, moving neutral atoms from the loaded well array to one or more unoccupied sites in the storage well array.

[0015] See below for reference. Figure 10 The proposed transition of a neutral atom from its first electronic state |1> to its second electronic state |2> can be achieved coherently, for example, via interaction with an alkaline earth atom (e.g., 88 This can be achieved by generating π pulses from an ultranarrow laser system with clock transition resonance of Sr, or incoherently, for example, via optical pumping (see [link to article]). Figure 12 For example, the |2> state can be an electronically excited state, for example... 88 Sr 3The P0 clock state does not have a dipole-allowed transition to the electronic ground state. In the following text, the transfer of the first electronic state |1> to the second electronic state |2> can also be designated as shelving, and the second electronic state |2> can also be designated as a shelving state.

[0016] For example, placing atoms in storage wells into a rested state (e.g., the |2> state) can have the advantage of allowing the identification and / or selective removal of other atoms randomly loaded into the storage well array (e.g., atoms randomly loaded into the storage well array during loading of the loading well array), as described in detail above and below (see [link to documentation]). Figure 12 In this way, complete control over the well occupancy of neutral atoms in the storage well array can be achieved, which allows for reliable initialization of neutral atom registers used for quantum simulation, metrology, and quantum computing.

[0017] Another aspect of this disclosure relates to a method for capturing and manipulating neutral atoms, the method comprising: loading a plurality of neutral atoms into a loaded well array adjacent to or co-located with a storage well array, wherein the loaded well array may be formed by an optical lattice and includes at least 1,000 capture sites, preferably at least 5,000 capture sites, and more preferably at least 10,000 capture sites overlapping with source regions of laser-cooled neutral atoms; determining the well occupancy of neutral atoms in the storage well array and the well occupancy of neutral atoms in the loaded well array; and moving neutral atoms from the loaded well array to one or more unoccupied capture sites in the storage well array based on the determined well occupancy of neutral atoms in the storage well array and the well occupancy of neutral atoms in the loaded well array.

[0018] Such a large N-loaded trap array can be realized, for example, by multiple trapping sites of a folded optical lattice, as described in detail in the applicant’s own international patent application PCT / EP2023 / 074768 entitled APPARATUS AND METHOD FOR TRAPPING AND MANIPULATING LARGE NUMBERS OF INDIVIDUAL NEUTRAL ATOMS (hereinafter referred to as [2]), which is incorporated herein by reference in its entirety and claims priority thereto.

[0019] As discussed in detail in reference [2], a large N-loaded trap array can be formed by multiple trapping sites of an optical lattice, wherein the trap depth of the optical lattice is greater than 0.1 mK, preferably greater than 1.0 mK, and / or wherein the trap frequency in the z-direction substantially orthogonal to the xy-plane comprising the optical lattice is greater than 2π × 1 kHz, preferably greater than 2π × 2 kHz, more preferably greater than 2π × 4 kHz, and even more preferably greater than 2π × 8 kHz. For example, such an optical lattice, specified in reference [2] as an optical trapping lattice, can be realized by a device for trapping and manipulating neutral atoms, the device comprising: a trapping laser system and optics for generating an optical trapping lattice at a trapping volume within a vacuum cavity, wherein the optics for generating the optical trapping lattice are configured to generate a single elliptical trapping laser beam based on the output of the trapping laser system, the single elliptical trapping laser beam being back-reflected and focused onto the trapping volume using a bowtie configuration to generate the optical lattice (see below for details). Figure 1 , Figure 2 and Figure 3 (and [Reference 2]).

[0020] As explained below, the number N atoms that can be loaded into the load trap array during each iterative loading cycle is... L It is an important parameter, which in particular determines the number N of saturated atoms that can be continuously maintained in a storage trap array for substantially arbitrarily long holding times. S For example, the number N saturated atoms in a storage trap array S It can be modeled as:

[0021] (Equation 1)

[0022] in, It is the normalized cycle loss, and This is the rearrangement nonfidelity associated with moving neutral atoms from the loaded trap array to unoccupied sites in the storage trap array. For example, it has been shown that shelving efficiency can reach 99.7% [e.g., see W. Tian et al., Parallel assembly of arbitrary defect-free atom arrays with a multi-tweezer algorithm, Physical Review Applied 19, 034048 (2023)]. Furthermore, for second-state |2> lifetimes typically achievable in storage trap arrays with optimized trapping parameters on the order of ~100 s [S. Dörscher et al., Lattice-induced photon scattering in an optical latticeclock, Physical Review A 97,063,419 (2018)] and typical MOT-level durations of 100 ms, the loss of shelved atoms can be as low as 0.4% or even lower. Considering a typical vacuum lifetime of approximately 250 seconds (e.g., in a room-temperature UHV system), a total cycle loss of α ≈ 0.8% can be achieved using current operating techniques. Furthermore, atomic losses due to rearrangement shifts can be as low as α. r ≈2% (e.g., see AWYoung et al., An atomic boson sampler, arXiv: 2307.06936). Based on this, the achievable amplification factor β = (1-α) r ) / α c (See Equation 1) Values ​​β > 100 can be achieved. With this amplification, approximately 10,000 atoms can be continuously held in a single array in a storage trap array, with only 100 atoms loaded into the load trap array during each loading cycle. Thus, the use of large N load trap arrays as disclosed in references [1] and [2] allows for further increases in this number, paving the way for near-term quantum advantage in neutral atom-based quantum technologies (see, for example, Bluvstein, D. et al., Logical quantum processor based on reconfigurable atom arrays, Nature 626, 58-65 (2024)).

[0023] This disclosure also relates to a method for the continuous operation of quantum computing, simulation, and / or metrology devices, the method comprising: repeating the steps of the method discussed above to iteratively assemble and maintain an ensemble of neutral atoms in a storage trap array; and using a subset of the neutral atom ensemble to execute quantum computing sequences, quantum simulation sequences, and / or quantum metrology sequences. Such a method may further include: receiving instructions from a remote computing device (e.g., from an interface, via a network to a cloud computing system, or a supercomputing facility, etc.) for executing the quantum computing sequences, quantum simulation sequences, and / or quantum metrology sequences; executing the quantum computing sequences, quantum simulation sequences, and / or quantum metrology sequences based on the received instructions; and (e.g., via a network) transmitting the executed quantum computing sequences, executed quantum simulation sequences, and / or executed quantum metrology sequences to the remote computing device. Such a method enables the provision of modern quantum computing sequences, quantum simulation sequences, and / or quantum metrology sequences to a variety of remote computing devices that may be spatially separated from the actual quantum hardware, even by long distances.

[0024] This disclosure also relates to quantum computing, simulation, and / or metrology devices (see examples). Figures 1 to 3 (Refs. [1] and [2]), the quantum computing, simulation and / or metrology device includes any of the methods for performing the methods disclosed herein (e.g., referred to above or below). Figures 8 to 10 The present disclosure also relates to an apparatus for (the methods discussed herein). This disclosure also relates to a computer program comprising instructions for causing a control unit to control quantum computing, simulation, and / or metrology devices to perform any of the methods disclosed herein.

[0025] Essentially, this disclosure therefore relates to a novel method for replenishing atomic registers from a storage device without affecting the stored atoms. The method disclosed herein relies on a combination of several aspects and techniques, including, for example, efficiently shelving atoms in a metastable state, in which the atoms are protected from fluorescence light associated with the magneto-optical trap used as a storage device; and combining two atomic registers: a load register and a storage register, which in some implementations can operate at different wavelengths. This mode of operation allows for the selective removal of atoms loaded into the storage register while retaining the atoms in the load register before rearrangement. This method can be used to assemble and maintain large-scale neutral atom quantum registers. Therefore, this disclosure aims to assemble and store densely packed arrays of atoms in storage trap arrays such as optical lattices or optical tweezers arrays. In particular, aspects of this disclosure allow for the assembly of even larger arrays of atoms that can be stored indefinitely and can be used as registers and / or storage devices for neutral atoms for quantum computing, simulation, and metrology. Reference is made below. Figure 9 , Figure 10 and Figure 11Let's discuss other aspects of the methods disclosed in this paper.

[0026] Further details of the above-described apparatus and methods, along with related computer programs, are discussed below with reference to exemplary implementations shown in the accompanying drawings. The foregoing provides a broad overview of the features and technical advantages of examples according to this disclosure in order to better understand the detailed description that follows. Additional features and advantages will be described below. The disclosed concepts and specific examples can be readily used as the basis for modifications or the design of other structures for achieving the same purpose as this disclosure. The characteristics of the concepts disclosed herein, their organization and operation, and their associated advantages will be better understood from the following description when considered in conjunction with the accompanying drawings. Each drawing is provided for illustrative and descriptive purposes and not as a limitation of the claims. Attached Figure Description

[0027] Figure 1 Exemplary apparatus for capturing and manipulating neutral atoms is shown that may be used in and / or be part of possible implementations of this disclosure;

[0028] Figure 2 Typical setups for generating load trap arrays and storage trap arrays are shown that can be used in and / or are part of possible implementations of this disclosure.

[0029] Figure 3 A side view is shown of a part of an apparatus for capturing and manipulating neutral atoms that may be used in and / or be part of a possible implementation of the present disclosure;

[0030] Figure 4 Exemplary experimental sequences for assembling and sequentially operating an array of neutral atoms in a storage trap array are shown, which may be used in and / or are part of possible implementations of this disclosure.

[0031] Figure 5 Possible implementations of a load trap array arranged adjacent to a storage trap array according to various aspects of this disclosure are shown;

[0032] Figure 6 The transfer of atoms from a loading trap array to an adjacent storage trap array with minimal heating is illustrated according to various aspects of this disclosure.

[0033] Figure 7 Possible implementations of a load trap array according to various aspects of this disclosure are shown, the load trap array being arranged to be at least partially co-located with the capture sites of a storage trap array;

[0034] Figure 8 The illustration shows an iterative assembly of an array of N > 1000 neutral atoms in an exemplary storage trap array, which can be used in and / or is part of a possible implementation of the present disclosure and runs continuously for more than 30 minutes.

[0035] Figure 9 Methods for capturing and manipulating neutral atoms, for example, using devices as described herein, in reference [1] and / or in reference [2], are shown according to various aspects of this disclosure.

[0036] Figure 10 Methods for capturing and manipulating neutral atoms, for example, using devices as described herein, in reference [1] and / or in reference [2], are shown according to various aspects of this disclosure.

[0037] Figure 11 Methods for capturing and manipulating neutral atoms, for example, using devices as described herein, in reference [1] and / or in reference [2], are shown according to various aspects of this disclosure.

[0038] Figure 12 Possible implementations of trap and state selective removal of neutral atoms from a storage trap array according to various aspects of this disclosure are shown. Detailed Implementation

[0039] In the following description, various aspects of this disclosure are described in more detail with reference to the accompanying drawings. However, this disclosure may be implemented in many different forms and should not be construed as limited to any particular structure or function presented herein. Rather, these aspects are provided so that this disclosure will be thorough and complete and will fully convey the scope of this disclosure to those skilled in the art. Based on the teachings herein, those skilled in the art will understand that the scope of this disclosure is intended to cover any aspect of this disclosure disclosed herein, whether implemented independently of or in combination with any other aspect of this disclosure. For example, an apparatus, device, or system may be implemented using any number of aspects set forth herein, or a method may be practiced. Furthermore, the scope of this disclosure is intended to cover such apparatus, devices, systems, or methods practiced using structures, functions, or structures and functions other than those set forth herein or different from those set forth herein. Any aspect of this disclosure disclosed herein may be implemented by one or more elements of the claims. While specific combinations of features are described below with respect to certain aspects of this disclosure, it should be understood that not all features of the examples discussed must be present to achieve the technical advantages of the devices, apparatuses, systems, methods, and computer programs disclosed herein. A disclosed aspect can be modified by combining certain features of one aspect with one or more features of other aspects. Those skilled in the art will understand that features, steps, components, and / or functional elements of one aspect can be combined with compatible features, steps, components, and / or functional elements of any other aspect of this disclosure.

[0040] The capture and manipulation of neutral atoms (e.g., imaging, gate manipulation, spectroscopy, heating, shelving, rearrangement, etc.) will now be represented with reference to various devices, apparatuses, systems, and methods described in the following detailed description and illustrated in the accompanying drawings by various boxes, modules, components, circuits, steps, processes, algorithms, etc. (collectively, “elements”). These elements can be implemented using hardware, software, or a combination thereof. Whether such elements are implemented as hardware or software depends on the design constraints imposed on the overall system and the specific application. Furthermore, the apparatuses and methods disclosed herein can be part of complex quantum technology systems such as neutral atom quantum computers. Those skilled in the art will understand that several components of such systems, such as laser sources, experimental control, and timing units, are not explicitly described below. Further details regarding such quantum technology elements are provided in the applicant’s own earlier European patent application 23181879.0 entitled HARDWARE-EFFICIENT NEUTRAL ATOM QUANTUM COMPUTING METHODAND DEVICE, the entire contents of which are incorporated herein by reference.

[0041] While aspects of this disclosure are presented below with regard to specific laser wavelengths for the boson strontium isotope 88Sr and / or for optically trapping lattices in the range of 1010 nm to 1100 nm, it should be understood that any other kind of neutral atom (e.g., other alkaline earth atoms (e.g., 171Y)) and any other suitable trapping laser system with a suitable internal energy level structure may also be used in other implementations of this disclosure. Furthermore, while the internal states 1S0 and 3P0 are discussed below as examples, other state pairs, such as 1S0 and 3P2, may also be used.

[0042] Figure 1 Aspects of an exemplary apparatus 100 for capturing and manipulating neutral atoms, as provided in this disclosure, are illustrated. The illustrated apparatus 100 may include a capturing laser system (not shown) and optics 110 (e.g., lenses, mirrors, waveplates, fiber collimators, etc.) for generating an optical lattice 120 at a capturing volume 140 within a vacuum chamber 150. The optics 110 for generating the optical lattice 120 may be configured to generate a single elliptical capturing laser beam 130 based on the output of the capturing laser system. This single elliptical capturing laser beam 130 is reflected back and focused onto the capturing volume 140 within the vacuum chamber 150 using a bow-tie configuration with a crossing angle θ, to generate the optical lattice 120, for example, via a quadruple interference of the crossed and reflected beams of the bow-tie configuration. The interference plane of the bow-tie configured beam defines the xy plane of the xyz coordinate system, where the z-axis is orthogonal to the xy plane.

[0043] In some aspects, the aspect ratio of the elliptical trap laser beam 130 can be configured / selected such that the well depth of the optical lattice is greater than 0.1 mK, preferably greater than 1 mK, and / or such that the well frequency in the z-direction, which is substantially orthogonal to the plane comprising the optical trap lattice 120, is greater than 2π × 1 kHz, preferably greater than 2π × 2 kHz, more preferably greater than 2π × 4 kHz, and even more preferably greater than 2π × 8 kHz.

[0044] For example, in one possible implementation, an optical lattice 120 is created using a single 1040 nm laser beam 130 back-reflected in a 4f configuration in the xy plane. The beam crossing angle θ is chosen to be ~26.56°, resulting in a lattice spacing of ax = 1186.5 nm along the x-axis and ay = 578.5 nm along the y-axis. The focal points of all four beams overlap at the center of the field of view of objective 160, with beam waists of ~20 µm (~100 µm) along the z-axis (in the xy plane), respectively. With the polarization of all beams aligned along the z-axis, the trap frequencies generated by multipath interference at 12 W optical power can reach 2π × 150 kHz (2π × 300 kHz) along the x-axis (y-axis), respectively, with a trap depth of 1.4 mK. On the z-axis, the trapping frequency of the lattice is greater than 2π × 4 kHz. Increasing the laser power and / or beam aspect ratio of an elliptical / anisotropic beam can produce up to 2π × 10 kHz within technically feasible parameters (e.g., using an aspect ratio of 1 / 10). Such trapping parameters facilitate the efficient loading of an optical lattice 120 from a MOT operating in a trapping volume 140, which provides a source of laser-cooled neutral atoms to be trapped and manipulated in the optical lattice 120, and in some implementations, after transferring the neutral atoms to a physical well array 170 or a physical trapping lattice that can operate at different wavelengths and / or lower trap depths, for example, by reducing non-resonant scattering events, increasing the lifetime of the neutral atoms in the physical well array 170 or the physical trapping lattice.

[0045] In some implementations, the captured laser system may include a ytterbium-doped fiber amplifier operating at wavelengths λ in the range of 1010 nm to 1100 nm, within which output power greater than 10 W is readily available. This contrasts with wavelength ranges that may be required due to, for example, magic capture conditions, or the lack of commercially available laser sources with high output power. Figure 1 In the example, the physical trap array can be implemented as a physical trap array 170 (e.g., an array of optical tweezers), generated, for example, by a spatial light modulator or similar device and imaged to the capture volume 140 by a high-resolution objective lens 160.

[0046] exist Figure 1 In the example, the optical tweezers spacing on the two axes can be chosen as 6 ay and 3ax, respectively (see Figure 1(Illustration in the image). Here, using 650 mW of optical power and a computer-generated hologram, a 2D array of 21 × 19 = 399 optical tweezers 170 can be generated at a wavelength of ~520 nm, each with a well depth of ~140 μK. Through independent calibration of the well depth and well frequency, the beam waist of the optical tweezers can be estimated to be approximately 472 (3) nm. The well frequency at a depth of 140 μK is approximately 2π × 77 kHz in the radial direction and approximately 2π × 19 kHz in the axial direction (i.e., in the z-direction). 88 Using Sr atoms as a reference, an initial optical tweezers pattern is pre-corrected for diffraction efficiency. 1 S0- 3 Minimizing the variation in differential optical displacement measured on the P1 transition allows the well depth of the optical tweezers 170 to be equalized to a root-mean-square intensity variation of ~2.2%. This is discussed in reference [2]. Figure 1 Further details of the device.

[0047] Figure 2 Typical optical setups for generating optical and physical capture lattices are shown, using a bowtie configuration at capture volume 140 within vacuum chamber 150 and, in some examples, in the field of view of high-resolution imaging objective 160. To achieve such a bowtie configuration, the optics may include a set of three mirrors 220a to 220c and a set of focusing lenses 210a arranged in the xy plane. To achieve the optical capture lattice as described herein, a single elliptical capture laser beam (see...) Figure 1 The laser beam can be entered into a bow configuration via a first focusing lens 210a, which focuses the optically captured beam to a focal point at the capture volume 140. Then, a second lens 210b collimates the captured laser beam, which is then reflected by turning mirrors 220a and 220b to be focused a second time to the focal point at the capture volume 140 via lens 210c. Lens 210d again collimates the laser beam to the retroreflector 220c and focuses it a third time to the focal point at the capture volume 140. The reflected laser beam is then collimated by lens 210c, turned by mirrors 220b and 220c, and focused a fourth time to the focal point at the capture volume 140, producing a quadruple interference, thereby generating the image as described above. Figure 1 The optical trapping lattice described above. As discussed above, the cross angle θ of the bow configuration formed by mirror 220 and lens 210 can be adjusted to tune the lattice constants in the x and y directions. Also, as discussed above, an elliptical / anisotropic laser beam is used as a target for... Figure 2 The bow configuration shown provides strong constraints in the z-direction, even though the optical trapping lattice consists of only a single 2D lattice plane.

[0048] A similar bow configuration (e.g., using θ=45°) can also be used to generate a physical trapping lattice at trapping volume 140. By adjusting the cross angle between the optical and physical trapping lattices and the wavelength of the corresponding laser beam, it is possible to ensure that the two lattices are substantially commensurable (see reference [2]). Figure 9 Furthermore, the bowtie configuration of one or two lattices may also include adjustable phase-shifting elements (e.g., EOM, rotatable glass plates, etc.) that allow tuning of the relative phase between the optical lattice and the physical lattice.

[0049] Figure 3 The following configuration is shown. Figure 1 A side view of a portion of the device 100, in which an optical trapping laser beam 130 generates an optical trapping lattice at a trapping volume 140, and a physical trapping beam 310 generates a physical trapping lattice at the trapping volume 140, for example using a second bow configuration with a different crossover angle (e.g., 45°). At the trapping volume 140, the MOT can be operated to provide a reservoir for laser-cooled atoms, which can be loaded into the optical trapping lattice and / or the physical trapping lattice. Furthermore, a high-resolution objective lens 160 can be used to image the physical trap array and the movable optical tweezers 310b onto the trapping volume 140. Using the same objective lens 160, fluorescence from neutral atoms inside the optical trapping lattice and / or the physical trap can be collected and imaged onto an imaging sensor 340, for example using a dichroic mirror 320 and an additional focusing lens 320.

[0050] The exemplary apparatus discussed above can be used in particular to provide load trap arrays and adjacent or partially co-located storage trap arrays as disclosed herein.

[0051] Figure 4 An exemplary experimental sequence is shown that can be used to iteratively assemble and maintain large arrays of neutral atoms in a storage well array. Therefore, implementations of this disclosure can involve a combination of storage registers (also designated herein as a storage well array) and load registers (also designated herein as a load well array). The storage registers are used to hold atoms for quantum operations, while the load registers are repeatedly loaded from a storage device (e.g., a magneto-optical trap or an optical dipole trap). The exemplary sequence of steps allows atoms to be continuously reloaded into the storage registers. Typically, such a sequence, as discussed herein, can include:

[0052] 1. Place the atoms located in the storage register in the metastable state of alkaline earth or alkaline earth-like atoms.

[0053] 2. Construct a magneto-optical trap to load atoms in their internal ground state into a load register, which is maintained at a wavelength optimized for loading and cooling the atoms.

[0054] 3. Execute parity projection pulses that affect only the atoms in state two in the load register, and remove double occupancy of the sites in the storage register without affecting the atoms in state one.

[0055] 4. Using a specifically tuned laser, ground-state atoms are removed from the storage register via transitions that selectively remove ground-state atoms from the storage register, while leaving ground-state atoms in the loaded register and atoms in the storage register that have been placed in metastable excited states unaffected. For this critical step, the wavelengths of the storage and loaded registers should typically be carefully selected.

[0056] 5. Move the atom from the load register to the storage register.

[0057] 6. Release the atoms from the storage register, that is, transfer them from the suspended state to the ground state.

[0058] 7. Image all atoms and, based on the obtained image, rearrange the atoms to new positions in the storage register to free the load register.

[0059] 8. Continue with step 1 to end the loop.

[0060] For example, for a load register, an array of optical traps generated by, for example, a spatial light modulator or an acousto-optic deflector can be used. For a storage register, an array of optical traps formed by, for example, an optical lattice or a spatial light modulator can be used. For array-specific heating pulses, a repulsive Sisyphus cooling technique can be used, which acts on transitions to heat atoms in internal state two in the storage register while keeping atoms in state one in the storage register and atoms in state two in the load register unaffected.

[0061] Figure 5Possible implementations of a load-well array arranged adjacent to a storage-well array according to various aspects of this disclosure are shown. Specifically, a 1040 nm fixed bowtie optical lattice, similar to that discussed above, can be used as a physical array 510 (with a grid of ax = 579 nm and ay = 1187 nm). The accessible region can be subdivided into a load region 520 and a storage region 530. The trapping sites of the optical lattice in the load region 520 can overlap with the fixed array of optical tweezers 525 (e.g., at a wavelength of 520 nm) to form a load-well array. As discussed in detail below, neutral atoms can be moved / transported using an acousto-optic deflector that guides a single optical tweezer beam 540 from the load region 520 to the storage region 530 in the lattice plane. The trapping sites of the optical lattice in the storage region 530 form a storage-well array, wherein a large array of neutral atoms can be iteratively assembled in single-atom control and maintained substantially indefinitely (e.g., Figure 8 (as shown in the image).

[0062] Figure 6 The typical trajectory of a neutral atom moving from the loading well array to the storage array is shown. The upper panel of panel a) shows the energy distribution of the optical lattice. The lower panel of panel a) shows the reduction in well depth modulation experienced by atoms traveling between lattice sites (line 610 in the upper panel) compared to traveling across lattice sites (line 620 in the upper panel).

[0063] Column b) illustrates an exemplary rearrangement sequence using AOD-controlled traveling rearrangement optical tweezers, which preferentially move horizontally along the channels between lattice sites to maintain the lowest possible heating.

[0064] Column c) shows the probability of atomic loss based on the distance between the occupied lattice sites and the traveling rearrangement optical tweezers. Perturbations leading to atomic loss are observed below 1 µm distance.

[0065] Column d) shows the change in the probability of successful atom movement with travel distance when performing a movement across lattice sites (trace 630) and when performing a movement between lattice sites (trace 640). The probability of successful atom movement can be defined, for example, as the probability of no atom loss during the entire movement operation.

[0066] Figure 7 Possible implementations of a load well array arranged at least partially co-located with a storage well array are shown. For example, in Figure 7In the diagram, grid 710 illustrates the optical lattice, with vortex 715 indicating multiple trapping sites of the optical lattice within the loading region. Black dots 720 indicate the locations of co-located optical tweezer arrays. In this configuration, the combined trapping potential of the optical lattice and optical tweezers forms a loading well array, and adjacent trapping sites of the optical lattice form a storage well array. As mentioned above, this configuration reduces the complexity of moving atoms from the loading array to the storage array.

[0067] Figure 8 The number of atoms in the continuous operation array in the storage area (trace 810) and the number of atoms in the load area (trace 820) are shown. Figure 8 The illustration shows the number of atoms during the initial assembly phase, from 0 minutes to ~2.0 minutes. (As can be seen in...) Figure 8 As seen in the figures, using the methods and apparatus disclosed herein, arrays with an average of more than 1,000 neutral atoms can be iteratively assembled and maintained substantially indefinitely. Illustrations I, II, and II show exemplary fluorescence images of neutral atoms in the storage trap array taken at three different time points.

[0068] The various devices disclosed herein for trapping and manipulating neutral atoms can be used as part of a quantum computing device, which includes: the devices discussed above; and a qubit-gate laser system for performing local single-qubit gate operations and local two-qubit gate operations on neutral atoms in a physical well array or physical trapping lattice. Similarly, an atomic clock device may include: the devices discussed above; and a clock laser system for performing clock spectral sequences on neutral atoms in a physical trapping lattice or physical well array. Similarly, a quantum simulation device may include: the devices discussed above, wherein the physical well array or physical trapping lattice can be operated such that neutral atoms can tunnel within the physical well array or physical trapping lattice; and optionally a many-body evolution laser system for performing many-body evolution on neutral atoms in the physical well array or physical trapping lattice.

[0069] Figure 9 A method 900 is shown for capturing and manipulating a plurality of neutral atoms, for example, using the apparatus described above and / or in References 1 and 2. Step 910 includes loading the plurality of neutral atoms into a loading well array arranged adjacent to a storage well array (see [reference]). Figure 5 ) or at least partially co-located with the storage trap array (see Figure 7In step 920, neutral atoms in the first electronic state |1> are selectively removed from the storage well array. In step 930, the well occupancy of neutral atoms in the second electronic state |2> in the storage well array is determined (i.e., indirectly determined by re-pumping atoms in state |2> to state |1> and then imaging state |1>) and the well occupancy of neutral atoms in the loading well array. Method 900 further includes step 940: based on the determined well occupancy of neutral atoms in the storage well array and the well occupancy of neutral atoms in the loading well array, moving neutral atoms from the loading well array to one or more unoccupied trapping sites in the storage well (e.g., as described above). Figure 6 (As shown). Typically, during detection, atoms in state |2> are rapidly repumped to state |1>.

[0070] In some implementations, the well depth of the loaded well array for neutral atoms differs from the well depth of the storage well array for neutral atoms. Preferably, the well depth of the loaded well is greater than the well depth of the storage well. For example, the well depth of the loaded well array may be at least 10% larger or smaller than the well depth of the storage well array, preferably at least 30%, and more preferably at least 30%. Using such loaded well arrays and storage well arrays allows for realization state and well-selective atom removal, as shown below. Figure 12 The discussion focuses on the state and the trap-selective heating pulse.

[0071] Typically, step 820, which selectively removes neutral atoms in the first electronic state |1> from the storage well array, may include selectively heating the neutral atoms in the first electronic state |1> in the storage well array. Alternatively or additionally, step 830 may include selectively ionizing the neutral atoms in the first electronic state |1> in the storage well array. Alternatively or additionally, step 830 may include selectively modifying the trapping potential of the neutral atoms in the first electronic state |1>; and / or selectively modifying the trapping potential of the neutral atoms in the loaded well array. Specifically, selectively heating the neutral atoms in the first electronic state |1> in the storage well array may include applying a trap and state selective heating laser pulse to the loaded well array and the storage well array. Preferably, the trap and state selective heating pulse may be a trap and state selective Sisyphus heating pulse (see below). Figure 12 In some embodiments, the application of the trap and state selective heating laser pulse may include: applying a sideband heating laser pulse whose frequency is detuned to correspond to the blue sideband of the neutral atoms in the loaded trap array.

[0072] Typically, determining the well occupancy of neutral atoms in a storage well array and the well occupancy of neutral atoms in a loaded well array can include performing fluorescence imaging on the neutral atoms while simultaneously laser cooling them. For example, laser cooling of neutral atoms can use sideband cooling, Raman sideband cooling, optical clump cooling, or Sisyphus cooling, or combinations thereof.

[0073] Step 840, which moves neutral atoms from a load well array to one or more unoccupied capture sites in a storage well array, may include determining one or more movement trajectories for one or more neutral atoms in the load well based on the determined well occupancy of neutral atoms in the storage well and the well occupancy of neutral atoms in the load well. Step 840 may also include controlling beam guiding units of one or more optical tweezers to execute the determined movement trajectories for the one or more neutral atoms. For example, the beam guiding unit may include a single-axis or multi-axis acousto-optic deflector or a digital mirror device. (See reference...) Figure 6 The determination of one or more movement trajectories may also include: determining the movement trajectory along the minimum energy surface in the trapping potential of the folded single-plane butterfly lattice as disclosed in references [1] and [2] above.

[0074] Typically, method 900 may further include generating laser-cooled neutral atoms by operating a magneto-optical trap (MOT) for neutral atoms in a spatial region comprising at least a portion of a loading trap array and at least a portion of a storage trap array. Specifically, the MOT for neutral atoms can be operated without using a laser resonating with an optical transition involving a neutral atom in a second electronic state |2>. Furthermore, the neutral atom may be an alkaline earth-like atom having a dipole-forbidden transition coupled to a first electronic state |1> and a second electronic state |2>. For example, the neutral atom may be a strontium atom or a ytterbium atom or a similar element. For example, a strontium atom may be... 88 Sr atoms and the first electronic state |1> can be 1 The S0 state, and the second electronic state |2> can be a long-lived 3P0 state or a long-lived 3P2 state. 171 Yb atoms can also acquire similar states.

[0075] Typically, the loading trap array may include an array of optical tweezers that may overlap with a first subset of the trapping sites of the optical lattice (see [link to documentation]). Figure 5 and Figure 7For example, the optical lattice can be a folded optical lattice with a bow-tie configuration as discussed above and in references [1] and [2]. Furthermore, a second subset of the trapping sites of the optical lattice can form a storage trap array. Specifically, loading multiple neutral atoms into the loading trap array can include, preferably, increasing the laser intensity of the array of optical tweezers while operating the MOT for the neutral atoms. Additionally, the optical lattice can be manipulated to maintain the storage trap array for the neutral atoms in a second electronic state |2>. In some embodiments, determining the trap occupancy of neutral atoms in the loading trap array can include, for example, transferring neutral atoms from the loading trap array to trapping sites of a first subset of the trapping sites of the optical lattice. Optionally, determining trap occupancy can also include, for example, performing fluorescence imaging (e.g., as described in references [1] and [2]) on the neutral atoms in the first subset of the trapping sites of the optical lattice while simultaneously laser cooling the neutral atoms in the optical lattice. Typically, method 900 may also include: transferring a neutral atom in the storage trap from a first electronic state |1> to a second electronic state |2>, as shown in reference Figure 10 To be discussed in more detail.

[0076] Figure 10 Another method 1000 for capturing and manipulating neutral atoms is illustrated. Method 1000 includes step 1010 of transferring a plurality of neutral atoms captured in a storage well from a first electronic state |1> to a second electronic state |2> (e.g., placing the atoms in a long-term metastable resting state, such as 3P0 of 88Sr). In step 1020, a plurality of laser-cooled neutral atoms are loaded into a loading well array arranged adjacent to or at least partially co-located with a storage well array. In step 1030, well occupancy of the neutral atoms in the storage well and the loading well array in the second electronic state |2> is determined. In step 1040, based on the determination of well occupancy as discussed in more detail above, neutral atoms are moved from the loading well array to one or more unoccupied sites in the storage well array.

[0077] Specifically, method 1000 may further include: after transferring a plurality of neutral atoms trapped in the storage trap array from a first electronic state |1> to a second electronic state |2>, selectively removing neutral atoms in the first electronic state |1> from the storage trap array (as above and referred to) Figure 12 (As discussed). In this way, the suspended atoms are not removed from the storage trap array.

[0078] In some embodiments, the duration for loading a plurality of laser-cooled neutral atoms into a loading well array can be less than the lifetime of the second electronic state |2>. For example, the loading duration can be at least one-half, preferably one-quarter, more preferably one-tenth, and even more preferably one-hundredth of the lifetime of the second electronic state |2>. Typically, loading a plurality of laser-cooled neutral atoms into a loading array can include generating the laser-cooled neutral atoms by operating a MOT for neutral atoms. Preferably, the MOT is operated in a spatial region comprising at least a portion of the loading well array and optionally at least a portion of the storage well array. Typically, transferring a plurality of neutral atoms trapped in the storage well array from a first electronic state |1> to a second electronic state can include coherently transferring the plurality of neutral atoms from the first electronic state |1> to the second electronic state. Alternatively or additionally, transferring a plurality of neutral atoms can include incoherently transferring the plurality of neutral atoms from the first electronic state |1> to the second electronic state |2> using one or more repump lasers.

[0079] Figure 11 A method for trapping and manipulating neutral atoms is illustrated. The method includes step 1110 of loading a plurality of neutral atoms into a loaded well array adjacent to or at least co-located with a storage well array. The loaded well array is configured to include at least 1000 trapping sites, preferably at least 5000 trapping sites, and more preferably at least 10000 trapping sites, overlapping with a source region of neutral atoms for laser cooling. For example, the loaded well array can be formed from a plurality of trapping sites of an optical lattice, wherein, for example, the well depth of the optical lattice can be greater than 0.1 mK, preferably greater than 1.0 mK. Alternatively or additionally, the well frequency in the z-direction, substantially orthogonal to the xy-plane comprising the optical lattice, can be greater than 2π × 1 kHz, preferably greater than 2π × 2 kHz, more preferably greater than 2π × 4 kHz, and even more preferably greater than 2π × 8 kHz.

[0080] Method 1100 further includes step 1120 of determining the well occupancy of neutral atoms in the storage well array and the well occupancy of neutral atoms in the load well array. In step 1130, based on the determined well occupancy of neutral atoms in the storage well array and the well occupancy of neutral atoms in the load well array, neutral atoms are moved from the load well array to one or more unoccupied capture sites in the storage well array, as discussed in more detail above.

[0081] Another aspect of this disclosure relates to a method for continuous operation of a quantum computing, simulation, and / or metrology device. The method includes the step of repeatedly performing a method according to one embodiment of the above embodiments for capturing and manipulating neutral atoms to iteratively assemble and maintain an ensemble of neutral atoms in a storage trap array. The method also includes the step of using a subset of the ensemble of neutral atoms to execute a quantum computing sequence, a quantum simulation sequence, and / or a quantum metrology sequence. Specifically, the method may further include the step of receiving instructions from a remote computing device for executing the quantum computing sequence, the quantum simulation sequence, and / or the quantum metrology sequence. Furthermore, the method may include: executing the quantum computing sequence, the quantum simulation sequence, and / or the quantum metrology sequence based on the received instructions. Additionally, the method may include the step of transmitting the results of the executed quantum computing sequence, the executed quantum simulation sequence, and / or the executed quantum metrology sequence to the remote computing device.

[0082] On the other hand, it relates to a quantum computing, simulation, and / or metrology device, which includes means for performing a method for capturing and manipulating neutral atoms according to one embodiment of the above embodiments and a method for continuously operating the quantum computing, simulation, and / or metrology device according to one embodiment of the above embodiments.

[0083] On the other hand, it relates to a computer program for controlling quantum computing, simulation and / or metrology equipment to perform a method for capturing and manipulating neutral atoms according to one embodiment of the above embodiments, and a method for continuously operating quantum computing, simulation and / or metrology equipment according to one embodiment of the above embodiments.

[0084] Figure 12 Possible implementations of trap and state selective removal of neutral atoms from a storage trap array according to various aspects of this disclosure are shown.

[0085] Column a) shows 88 Correlated states and energy levels of Sr.

[0086] Column b) shows the well filling variation with the applied laser frequency detuning (in MHz) after applying a well-selective 689 nm Sisyphus heating laser pulse to atoms trapped in the storage well array (trace 1210) and the loaded well array (trace 1220). The traces are normalized relative to the loading fraction of the lattice without the heating pulse and relative to the number of optical tweezers in the combined potential of the loaded well array. When a frequency detuning of 1.28 MHz (dashed line) is selected, 5 × 10−4 extinction of the ground state (|1> state) atoms in the storage well array (lattice potential only) is achieved, while the atoms in the combined optical tweezers-lattice potential are almost unaffected.

[0087] Column c) shows the round-trip hold fidelity as the hold duration varies, reaching 3% (dashed line) after 10 ms.

[0088] Column d) shows the shelving lifetime in a storage trap array at a trap depth of ~200 µK, reaching a 1 / e lifetime of 13 seconds. The dashed line indicates the exponential fit of the final fit for the last four data points. The inset shows a magnified view of the region with hold times below one second.

[0089] The foregoing disclosure provides illustrations and descriptions, but is not intended to be exhaustive or to limit the aspects to the exact forms disclosed. Modifications and variations can be made based on the foregoing disclosure, or modifications and variations can be derived from practice in various aspects. As used herein, the term "component" is intended to be broadly interpreted as hardware, firmware, or a combination of hardware and software. As used herein, a processor is implemented as hardware, firmware, or a combination of hardware and software.

[0090] It will be apparent that the systems and / or methods described herein can be implemented in various forms of hardware, firmware, or combinations of hardware and software. The actual dedicated control hardware or software code used to implement these systems and / or methods does not limit these aspects. Therefore, the operation and behavior of the systems and / or methods are described herein without reference to any specific software code—it should be understood that software and hardware can be designed to implement the systems and / or methods based on the descriptions herein.

[0091] Although specific combinations of features are listed in the claims and / or disclosed in the specification, these combinations are not intended to limit the disclosure of the aspects. In fact, many of these features can be combined in ways not specifically stated in the claims and / or not disclosed in the specification. Although each appended dependent claim may directly depend on only one claim, the disclosure of the aspects includes each dependent claim combined with each other claim in the claims. The phrase “at least one” in the list of references refers to any combination of these items, including single members. As an example, “at least one of the following: a, b, or c” is intended to cover a, b, c, ab, ac, bc, and abc, as well as any combination having multiple identical elements (e.g., aa, aaa, aab, aac, abb, acc, bb, bbb, bbc, cc, and ccc, or any other order of a, b, and c).

[0092] Unless explicitly described herein, no element, action, or instruction used herein should be construed as critical or essential. Furthermore, as used herein, unmodified terms are intended to include one or more items and may be used interchangeably with "one or more." Additionally, as used herein, the terms "set" and "group" are intended to include one or more items (e.g., related items, unrelated items, combinations of related and unrelated items, etc.) and may be used interchangeably with "one or more." Where only one item is referred to, the phrase "only one" or similar language is used. Furthermore, as used herein, the terms "have," "possess," "own," etc., are intended to be open-ended terms.

[0093] As used in this article, the phrase “based on” should not be interpreted as referring to a closed set of information, one or more conditions, one or more factors, etc. In other words, unless specifically stated differently, the phrase “based on A” (where “A” can be information, conditions, factors, etc.) should be interpreted as “based on at least A”.

[0094] As used herein, unless restrictive language is used relative to the listed alternatives, the term "or" is inclusive. For example, mentioning "X is based on A or B" should be interpreted as including X based on A, X based on B, and X based on both A and B. In this respect, mentioning "X is based on A or B" means "at least one of A or B" or "one or more of A or B," because "or" is inclusive. Similarly, mentioning "X is based on A, B, or C" should be interpreted as including X based on A, X based on B, X based on C, X based on both A and B, X based on both A and C, X based on both B and C, and X based on both A, B, and C. In this respect, mentioning "X is based on A, B, or C" means "at least one of A, B, or C" or "one or more of A, B, or C," because "or" is inclusive. As an example of restrictive language, mentioning "X is based only on one of A or B" should be interpreted as including X based on A and X based on B, but excluding X based on both A and B.

[0095] In addition, such as Figure 9 , Figure 10 and Figure 11 The process diagram does not necessarily indicate a specific order or sequence of steps. For example, steps may be performed in a different order, or simultaneously, if hardware capabilities permit, without departing from the scope of this disclosure.

[0096] Related original research

[0097] To illustrate further concepts, advantages, and exemplary implementation details of this disclosure, the original research paper entitled Continuous operation of large-scale atom arrays in optical lattices is further described in its entirety in which it is incorporated herein by reference and thus forms an integral part of this disclosure. It should be understood that, due to their specific legal function in the context of a patent application, terms such as “consisting of” or “comprises of” used in the original research paper will be replaced by appropriate terms such as “comprising” and “including” or similar terms.

[0098] appendix

[0099] Continuous operation of large-scale atomic arrays in optical lattices

[0100] Flavien Gyger, 1,2 Maximilian Ammenwerth, 1,2 Renhao Tao 1,2,3 Hendrik Timme, 1,2 Stepan Snigirev, 4 Immanuel Bloch, 1,2,3 and Johannes Zeiher 1,2,3,

[0101] 1 Max Planck Institute for Quantum Optics, Garching 85748, Germany

[0102] 2 Munich Center for Quantum Science and Technology (MCQST), 80799 Munich, Germany

[0103] 3 Ludwig Maximilian University of Munich, Department of Physics, 80799 Munich, Germany

[0104] 4 PlanQC GmbH, 85748 Garching, Germany

[0105] (Date: February 5, 2024)

[0106] Expanding the size of assembled neutral atom arrays trapped in optical lattices or optical tweezers is crucial for many applications, from quantum simulation to quantum metrology. However, preparation time increases with system size and constitutes a significant bottleneck in the bottom-up assembly of large ordered arrays from randomly loaded optical traps. Here, a novel approach is presented that circumvents this bottleneck by continuously reloading atoms and adding them to the array while cycling atoms from one experimental run to the next. Using this approach, dense arrays storing over 1000 atoms in an optical lattice are realized, continuously refilled with a net cycle time of 2.5 seconds, and approximately 200 atoms are reloaded during each cycle. Furthermore, it is shown that such large arrays can be sustained continuously by simply reloading atoms lost from one cycle to the next. The method presented in this paper paves the way for quantum science with large, continuously running ordered atom arrays containing thousands of atoms.

[0107] I. Introduction

[0108] Atom arrays stored in optical lattices or optical tweezers are promising platforms for quantum simulation, quantum computing, and quantum metrology [1-6]. The typical experimental sequence for controlling atoms in optical lattices or optical tweezers begins with the preparation of an ensemble, followed by a sequence of simulations, computations, or metrologies. Finally, a destructive measurement of the system state is performed, which typically makes it impossible to recover atoms from one cycle to the next. The subsequent preparation of new atomic ensembles takes significantly more time than the actual experimental sequence, resulting in significantly longer dead times for large arrays. This naturally necessitates different operating modes in each cycle, where only the lost atoms are prepared and replaced. While this has been demonstrated in recent pioneering work in bulk gases [7], the reuse of atoms and the cyclic operation with microscopic control are challenging and require a combination of nondestructive detection and rearrangement to replenish lost atoms [8,9]. For several types in optical lattices and optical tweezers, high-fidelity and low-loss detection of individual atoms is now routinely achieved [10-15]. Although recent first steps toward extended operation toward atomic arrays have been demonstrated in small-scale systems with finite storage [16,17], true continuous operation requires a scheme for reloading new atoms without affecting the atomic arrays already present in the system. Recently, an efficient strategy for “hiding” stored atoms during reloading of new atoms has been demonstrated in bi-element arrays of two base atom species [18,19], where forming a magneto-optical trap (MOT) of one element has minimal impact on the other, enabling continuous operation by preparing arrays of each element in alternation. Alternative approaches are provided by utilizing the more complex energy level structures available in alkaline earth metal (or similar) atoms such as strontium or ytterbium. Here, two independent optical series with different total spins and metastable states exist, which have proven useful for a variety of applications when combined with microscopic control [17,20-30]. In particular, metastable states can also be used to efficiently hide stored atoms while forming an MOT for ground-state atoms. Although such energy level structures have been shown to be well-suited for preparing one-dimensional atomic arrays with near-1 filling based on dark-state enhanced loading combined with site control using two cross acousto-optic deflectors (AOD)

[29] , continuous loading has so far remained a difficult goal to achieve.

[0109] Here, a novel scheme is presented that combines the aforementioned aspects to achieve continuously operating large-scale atomic arrays with atomic numbers consistently exceeding 1000 and reaching up to 1247 atoms. This scheme relies on a continuously operating storage region within an optical lattice, periodically replenished from the loading region and the MOT (Mean Interchange of Tolerance). Using a two-color combination of the loading and storage arrays, excellent spatial control of the loading region is achieved, strongly suppressing loading of sites in the storage register. Loading approximately 200 new atoms for each cycle constructs and then sustains an atomic array with over 1000 atoms within the optical lattice—five times the number of atoms loaded during each cycle. The results mark a paradigm shift from the operation of neutral atom-based quantum simulators and quantum computers to iterative assembly and continuously operating arrays.

[0110]

[0111] Figure 1 The concept and demonstration of continuous operation. Key features of our experimental instrument. We use a 1040 nm fixed bowtie-shaped optical lattice as our physical array (grey grid, where...). and We subdivide the accessible region into loading and storage regions. The loading region overlaps with a 520 nm fixed optical tweezers array. Atoms are transported via AODs, which guide single beams of light from the loading region to the storage region within the lattice plane. b. Experimental sequence of our continuous loading scheme. c. Exemplary single image of an array assembled iteratively with an average of over 1000 atoms at different time points. Bottom chart: Number of atoms in the continuously running array (blue) and number of atoms in the loading region (red). Inset: Enlarged view of the number of atoms in the accumulation phase.

[0112] II. Assembly of large arrays

[0113] The architecture of the device for continuous operation

[13] is in Figure 1 As shown in the figure. Spanning approximately 130 μm Experiments were conducted in a 130 μm region, corresponding to the area currently capable of being processed by AOD. This region is located at the center above the objective lens and contains approximately 24,000 well sites within a bowtie-shaped lattice. This lattice region is divided into two sub-regions: a loading region and a storage region, see [link to documentation]. Figure 1 In section a, the lattice loading region is supplemented by a reservoir with 342 optical tweezers, which overlap three-dimensionally with the lattice sites. These optical tweezers themselves are loaded with data from the bilevel MOT. 88 Sr atoms, this bipolar MOT is based on a width at 461 nm 1 S0- 1 P1 transition and narrow line based on 689 nm1 S0- 3 P1 transition

[31] . For high-fidelity detection, atoms are transferred from an optical tweezers array to an optical lattice

[13] and fluorescence imaging is performed therein. The cyclic sequence is presented in Figure 1 b, and the loop time is 2.5 seconds (excluding data processing).

[0114] In the first iteration, N was loaded from MOT on average. L Atoms are loaded into optical tweezers. The atoms are then transferred to the lattice and high-fidelity, low-loss imaging is performed to detect the position of the loaded atoms in the loading region

[13] . The detected atoms are then placed from the loading region to the storage region as needed by moving optical tweezers controlled by cross-AOD. After rearrangement, the atoms in the storage region are placed into long-lived metastable states. 3 P0, and then optical tweezers refill the loading region from the MOT generated at the lattice site. Holding in a magnetically insensitive clock state protects the stored atoms from loss during the MOT. Subsequent fluorescence imaging of both the storage and loading regions will be performed in... 3 P0's atoms pump back to the ground state 1 S0 reveals unoccupied sites in the storage area that need to be refilled. The possibility of reusing atoms between experimental runs provides a significant scalability advantage for the achievable array size. The maximum array size is reached when the number of atoms lost during the previous cycle is precisely balanced with the number of atoms replenished in the current cycle. This condition limits the maximum achievable number of atoms to [value missing]. amplification factor

[0115] (1)

[0116] With the probability of successful transmission Proportional to, and with cyclical loss Inversely proportional, this cycle loss quantifies the proportion of atoms lost from one cycle to the next. The effective number of atoms that can be added to the array in each cycle is... Due to atomic losses during transport, this value is less than the number of atoms loaded. This is less than the cycle loss typically achieved in this work. The number of saturated atoms has significantly exceeded the number of loaded atoms per cycle, see [link / reference]. Figure 1 d. We note that, in principle, by staggering the loading and storage sites, the loading region can completely overlap with the storage region. Such a configuration would be beneficial because it increases the available space in the storage region and shortens the movement distance between the loading and storage regions. However, in our configuration, the 520 nm optical tweezers light causes the storage region to... 3 A large number of atoms in the P0 state are lost, preventing the overlap of these two regions. We found evidence that...3 Photoionization of the P0 state is the cause of this loss, see Appendix A

[26] , which can be mitigated in the future using alternative optical tweezer wavelengths (e.g., 813 nm). All steps involved in the continuous operation of the array are characterized in detail below.

[0117]

[0118] Figure 2 Atomic state preparation. 88 The energy levels of Sr. b. The filling variation with applied frequency detuning after applying a capture-selective 689 nm Sisyphus heating pulse to atoms trapped in the lattice (blue) and atoms trapped in the combined optical tweezers + lattice potential (red). The trace for the lattice is normalized to the loading ratio without the heating pulse, and the trace for the combined potential well is normalized to our optical tweezers number (342). When a frequency detuning of 1.28 MHz (green dashed line) is chosen, 5n lattice atoms are achieved for the ground state. 10 -4 The extinction occurs, while the atoms in the combined optical tweezers-lattice potential remain almost unaffected. c. The round-trip shelving disfidelity changes with shelving duration, reaching 3% after 10 ms (green dashed line). d. In our lattice at 200 μK, the shelving lifetime reaches 13 seconds. The dashed line is an exponential fit to the last four data points. Inset: Enlarged view of the shelving time region below one second. Red dots indicate the proportion of actual recovery when MOT occurs during the shelving time.

[0119] III. Reload the storage

[0120] A key step in array cyclic operation involves transferring atoms from the memory before reloading the storage. 1 S0 is transferred to 3 P0. Currently driven by a combination of 689 nm and 688 nm light. 1 S0→ 3 P1→ 3 S1 achieves population shelving. An additional repump beam at 707 nm enables... 3 The P0 state becomes the only dark state; see [link / reference] Figure 2 The 'a' in the example. Figure 2 As shown in c, after 10 ms of pumping, the proportion of stalled atoms reached 97%. Note that a higher stalling ratio can be achieved by combining coherent stalling and incoherent pumping, which will directly reduce cycle loss

[32] . At a lattice well depth of 200 μK, 3 The lifetime of an atom in the P0 state is 13 s. See [reference needed]. Figure 2In the sequence, this lifetime significantly exceeds the lifetime of the suspended atoms. 3 The total duration of P0 is 115 ms, making the loss caused by holding the suspended atoms themselves negligible. However, a small increase in the loss of suspended atoms was observed due to the presence of MOT, resulting in... 3 A total of 6% shelving loss occurs during the P0 preparation and atom reloading steps; see [link to relevant documentation]. Figure 2 The inset in Figure d shows the important point: the MOT is generated without the 679 nm repump light, which would otherwise deplete the light in the MOT. 3 The suspended atoms of P0. This significantly reduces the number of loaded atoms.

[0121] However, by optimizing the parameters, a filling ratio as high as 45% atomic loading probability was achieved in the loaded optical tweezers + lattice register sites after parity projection. (Missing information) 3 An additional consequence of the P0 repump light is that a small fraction of atoms pass through... 1 D2- 3 P2- 3 S1 decays into 3 P0, in 3 At P0, they become indistinguishable from atoms placed in the storage array. To counteract this effect, 1 The D2 state can be re-pumped to a higher position via transitions at 716 nm or 448 nm. 1 The P1 state

[33] will increase the MOT loading ratio and eliminate the accidental abandonment of atoms during the MOT phase. 3 The drawback of P0 is that, importantly, the MOT loads the entire lattice, including both the loading and storage regions. To remove ground-state atoms from any location in the lattice except at sites overlapping with the optical tweezers in the loading region, a necessary trap-selective heating pulse is then applied. This heating pulse is optimized to remove ground-state atoms from the lattice while preserving ground-state atoms in the optical tweezers and in the lattice. 3 Both P0 atoms were intact; see [link / reference] Figure 2 b. To achieve this selective removal of atoms, a beam at 689 nm is used, tuned to a repulsive Sisyphus heating state. This heating feature is narrow enough to be spatially resolvable. 1 S0 state and 3 The transitions between P1 states at differential trap depths are unique. For the chosen detuning, a net kinetic energy gain is achieved between subsequent excitation-decay cycles, thus leading to a rapid, highly parallel, and controllable heating mechanism for the ground-state atoms in the lattice. Importantly, atoms in the loaded region undergoing the combined two-color lattice-optical tweezers potential are shielded from the effects of heating resonances and are therefore kept trapped, see [link to relevant documentation]. Figure 2b in the equation. Therefore, the selective heating pulse effectively removes the state of the crystal that has only undergone lattice potential. 1 All atoms of S0, and extinction greater than 5 10 −4 This prevents atoms loaded directly from MOT from being stored uncontrollably in the storage area. Alternatively, site-selective parallel addressing can also be used to hide already loaded sites to prevent them from being loaded further, as effectively shown in 1d

[29] , and to directly extend using state-selective parallel addressing in a higher dimension

[34] .

[0122]

[0123] Figure 3 Atom rearrangement. a. Top image: The energy landscape of our folded lattice. Bottom image: Reduced well depth modulation experienced by atoms traveling between lattice sites (red line in the top image) compared to traversing lattice sites (blue line in the top image). b. The rearrangement algorithm favors horizontal movement along channels between lattice sites. c. The probability of atom loss varies with the distance between the occupied lattice site and the traveling rearrangement optical tweezers. Interference leading to atom loss is observed at distances below 1 μm. d. The probability of successful atom movement varies with travel distance when movements are performed traversing lattice sites (red dots) and when they are performed between lattice sites (blue dots). The probability of successful atom movement is defined as the probability of not losing atoms during the entire movement operation.

[0124] IV. Rearrangement in optical lattices

[0125] The next important step in the cyclic sequence is to rearrange the newly loaded atoms into vacant sites in the already arranged array. Such movements are performed using a pair of AODs similar to those in previous work [14,35,36]. Compared to optical tweezers arrays, the movement of atoms across lattice sites undergoes large periodic well depth modulation, potentially leading to strong heating. In the special case of bowtie lattices, this modulation can be significantly reduced by moving atoms between lattice sites, thus mitigating any thermal effects; see [link to relevant documentation]. Figure 3 In line with this expectation, long-distance movements between lattice sites were observed to have a significantly higher success rate than movements across lattice sites. See also a. Figure 3 In this measurement and during continuous loading operations, all movements were performed at a peak velocity of 54 μm / ms. In particular, for the long-distance movements considered in this work, movement between sites is crucial. The spacing between sites along the x-axis and y-axis of the lattice are respectively... and And therefore particularly well-suited for horizontal movement, see Figure 3a. To best utilize the favorable geometry of the lattice, the rearrangement process is designed to move primarily between lattice sites in a five-stroke movement mode; see also a. Figure 3 In step b, the first stroke removes the atom from its lattice site and carries it between lattice sites. The second stroke removes the atom from the loading region. The third and fourth strokes adjust the vertical (y-axis) and horizontal (x-axis) positions, respectively, to nearly match the position of its final destination, and finally, the last stroke inserts the atom into its final position, see [link to previous section]. Figure 3 b. Each move consists of (i) slowly and gradually increasing the optical tweezers potential depth to approximately ten times the lattice depth to extract the target atom from the lattice, (ii) a series of parameterized frequency chirps encoding the velocity distribution and orientation, and (iii) finally gradually decreasing the optical tweezers depth to release the atom into the desired target lattice site in the storage region. The initial and final intensity gradient durations are each 400 μs to ensure thermal adiabatic properties

[37] . The imperfect rearrangement process affects sequential loading in two ways: first, it reduces the number of effective atoms N that can be added to the storage array in each cycle when atoms are lost during the transfer process. L,eff Secondly, by traveling at very close distances to already stored atoms, unintended perturbations to the trap site can lead to the loss of stored atoms, thus directly increasing cycle loss. A corresponding limit value for the minimum distance between the stored atoms and the trajectory of the moving optical tweezers was observed, approximately equal to 1 μm; see [link to relevant documentation]. Figure 3 c. The minimum distance sets a limit on the minimum feasible spacing between atoms in the storage register. The total rearrangement duration for each cycle is approximately 700 ms. This duration can be significantly reduced by implementing more complex parallel rearrangement schemes [8,14,38,39].

[0126] V. Continuous operation

[0127] Finally, combining all the steps, the ability to build and maintain a large-scale, dense optical tweezers array for over an hour is demonstrated. See [link to documentation]. Figure 4 After the initial loading phase, the array maintained more than 1000 atoms for 97% of the time, and a maximum of 1247 atoms was recorded in the array. Normalized cycle loss was extracted from the atom loss within and across different cycles. and the fidelity of rearrangement Both of these parameters are incorporated into the number N of stored atoms. i In a simple model of accumulation and saturation at each loop i,

[0128] (2).

[0129] Figure 4 The dashed line in 'a' is calculated using the time averages of the measured cycle loss and rearrangement loss according to these equations, and it is very consistent with the measured number of atoms. Four parameters of interest are extracted from the occupancy matrix and plotted on... Figure 4 In b: (i) the loading ratio, defined as the normalized value of the number of loaded atoms in the loading region relative to the total number of optical tweezer sites (342); (ii) the success probability of rearrangement movement, defined as the proportion of loaded atoms that are moved to the storage region after rearrangement; (iii) the shelving survival ratio, defined as the proportion of atoms that survive the shelving, holding, and re-pumping operations between two cycles; and (iv) the rearrangement storage survival ratio, defined as the proportion of atoms in the storage array that survive the rearrangement process and are not moved themselves. After 80 cycles, the rearrangement operation is disabled, and the storage array is allowed to decay naturally. From this point on, it can be observed that: (i) the loading curve rises because loaded atoms are no longer removed and accumulate in the loading region; and (ii) the survival ratio of stored atoms increases because no rearrangement operation interferes with the storage array.

[0130] A more detailed correlation analysis, linking the measured final atom count with the synchronously extracted cycle loss, rearrangement nonfidelity, and reloaded atom count, directly reveals that fluctuations in the final atom count are most strongly correlated with the cycle loss (see Appendix B). This behavior is expected due to the inverse scaling of the steady-state atom count with the cycle loss and highlights the large potential gains in further optimized sequences.

[0131]

[0132] Figure 4 Continuous operation. The number of atoms in storage region changes with each cycle. After 80 cycles, we disable rearrangement and allow the array to decay naturally. The black dashed line is our model using the measured average parameters according to Equation 2, while the red line is the exponential fit, which allows us to measure the cycle loss. (The effect of rearrangement is ignored in this case). b. Evolution of loading ratio (orange triangle), rearrangement move success probability (red square), shelving survival ratio (blue circle), and rearrangement storage survival ratio (green pentagon) with cycle. c. Single-shot image of a storage array containing 1230 atoms. d. Average image of the storage array. d. Array continuously running for more than 1.5 hours.

[0133] VI. Conclusion and Outlook

[0134] In summary, we present the first realization of a densely packed, sequentially loaded array of atoms stored in an optical lattice. Looking ahead, we anticipate that our technique could allow the assembly of atomic arrays much larger than the atomic matrices we have previously demonstrated. In previous work, atoms were shelved to... 3 The efficiency of P0 has been shown to reach 99.7%

[32] . For clock-state lifetimes on the order of 100 s at reduced lattice potentials

[40] and typical MOT phases lasting 100 ms, the loss of stalled atoms during the MOT phase can be reduced to 0.1%, resulting in a total stall loss as low as 0.4%. Assuming a vacuum lifetime of 273 s that we measured, the typical vacuum confinement loss is also approximately 0.4% for experiments with a cycle time of 1 s, resulting in a total cycle loss of approximately Furthermore, the atomic loss achieved due to rearrangement shifts is as low as...

[14] , Equation (1) predicts that the achievable amplification factor can reach At such a large amplification, we can reach approximately 10,000 atoms in a single array by loading 100 atoms in each cycle, provided that sufficient storage area and high-fidelity detection are available

[13] . Deterministically loaded arrays [21,41] or directly loaded lattices

[13] as loading regions can potentially increase the number of achievable steady-state atoms by several orders of magnitude. We would like to emphasize that the advantage of sequential loading is that, compared to directly assembling such a large number of atoms in a single experimental cycle, only the newly loaded atoms are moved for each cycle, thereby reducing both the loss caused by movement and the rearrangement time overhead by the amplification factor β. Furthermore, maintaining coherence during the reloading phase, for example for atoms placed in specially shielded physical arrays, will open up exciting new prospects for quantum metrology or quantum information tasks [17,19]. Such large, sequentially maintained arrays of atoms, combined with recently demonstrated fast, high-fidelity quantum gates

[42] and basic logic quantum circuits

[35] , make neutral atoms a promising platform for large-scale quantum computing and quantum simulation.

[0135] Additional notes: During the preparation of the manuscript, we became aware of the relevant work, which involved the use of 171 Similar results were reported for arrays of Yb

[43] .

[0136]

[0137] Figure 5 From 520 nm optical tweezers 3 P0 ionization. 3 Measurement and parabolic fitting of P0 lifetime as a function of optical tweezers trap depth. The extracted quadratic term contributing 250 s to the loss rate. -1 (mK) -2 .

[0138] Acknowledgments

[0139] We thank the Max Planck Society (MPG), the German Research Foundation (DFG) for their funding under the German Excellence Strategy – EXC-2111 – 390814868, the Munich Quantum Valley Initiative as part of the Bavarian High Technology Agenda Plus, and the BMBF for their funding through the MUNIQC-Atoms and MAQCS programs. This publication also received funding from the Horizon Europe program HORIZON-CL4-2022-QUANTUM-02-SGA through project 101113690 (PASQuanS2.1). JZ thanks the BMBF for its support through the project “Quantum Technologies – From Basic Research to the Market” (SNAQC, grant number 13N16265). MA and RT thank the International Max Planck Institute for Quantum Science and Technology (IMPRS) for their funding. MA thanks the Hector Institute for its scholarship support.

[0140] Appendix A: 520 nm optical tweezers 3 P0 ionization

[0141] In our experiments, we found that for optical tweezers with a well depth of 300 μK, in 520 nm optical tweezers... 3 The lifetime of the P0 atom is limited to 40 ms and is quadratic with respect to the well depth, such as... Figure 5 As shown. This indicates that 520 nm light can be ionized and stored under two-photon excitation. 3 Atoms of P0. In our continuous loading experiments, we observed a significant reduction in the shelving lifetime of atoms in the loading region, even when atoms were trapped in lattice sites that did not overlap with the optical tweezers. As a result of this observation, we spatially separated our loading and storage regions to mitigate the cycling loss caused by this additional optical tweezers. We anticipate that using a different wavelength (e.g., 813 nm) for the optical tweezers will circumvent this problem and increase the available space in both the loading and storage regions.

[0142] Appendix B: Correlation Analysis of Atom Number Fluctuations in Memory Arrays

[0143] To understand the origin of atom number fluctuations during the continuous operation of the memory array, we investigated the correlation between a set of parameters related to atom occupancy during various stages of the sequence and the fluctuations. In this analysis, we only consider atoms placed at sites within the memory array defined by a custom objective; all atoms not at sites within the memory array, even if they are placed in memory regions (defects), are ignored. We define the survival rate s.mn (in ) represents the summation and normalization of the occupancy of all storage array positions filled in both images m and n to the occupancy of image n. Therefore, s mn The atoms that survive from image m to image n in the storage array are quantized. Similarly, we will quantize the gain. Defined as the sum of all storage array positions in image m that are empty but filled in image n, normalized back to the occupancy number of image m. Therefore, (in The number of atoms appearing in the storage array during the transition from image m to image n is quantified. For the survival rate s... mn and gain We use enhanced image indices m,n = 0,1,2 to label the images, where indices 0 and 1 refer to the first and second images in loop i, respectively, and index 2 refers to the first image in the next loop i+1. See [link to documentation]. Figure 6 b. Fluctuations are caused by quantity Quantization, where N s It is the number of filled sites in our storage array, and It is defined as the difference in the number of filled sites in the storage array between the two images under consideration.

[0144] Atomic wave With s mn and The correlation summary is in Figure 6 c. Figure 6 d. Therefore, it can be seen that... With s 12 The strongest correlation was with (Pearson coefficient of 0.5), and the second strongest was with s. 02 The correlation coefficient was 0.3. The absolute values ​​of the Pearson coefficients for other correlations were all below 0.15. This indicates that within our parameter range, atomic number fluctuations are most sensitive to shelf fidelity. We expect... It is always 0 because no atoms are intentionally added to the array between the second image of cycle i and the first image of cycle i+1. However, we observe that some atoms are added in each cycle. This is due to the fact that during MOT... 3 The contributions from both the defects caused by the suspended atoms of P0 and the imperfect heating pulses are too small, and we attribute them to misclassification in our deconvolution algorithm.

[0145]

[0146] Figure 6 Correlation analysis of atom number fluctuations. Image labeling during continuous operation of the array. Each image can be indexed cyclically. and image index To identify. The truth table for b explains how to calculate s. mn and A Boolean operation is performed on the occupancy count between selected image pairs. 1 represents an occupied site, and 0 represents an unoccupied site. For example, s 01 Returns the number of lattice sites occupied in both image (i, 0) and image (i, 1) in the storage region, normalized to the number of sites already loaded in image (i, 1). Similarly, Returns the number of target lattice sites in the memory region that are occupied in image (i+1, 0) but not occupied in image (i, 0), normalized to the number of loaded sites in image (i, 0). The survival rate of c atoms varies with the number of atoms when moving from one image to another. 12 The strongest correlation is with fluctuations in the number of atoms in the storage area. The gain of d atoms varies with the number of atoms from one image to another.

[0147]

[0148]

[0149]

[0150]

Claims

1. A method for capturing and manipulating neutral atoms, comprising: Multiple neutral atoms are loaded into a loading well array, which is arranged to be adjacent to or at least partially co-located with a storage well array; Selectively remove neutral atoms in the first electronic state |1> from the storage trap array; Determine the well occupancy of neutral atoms in the second electronic state |2> in the storage well array and the well occupancy of neutral atoms in the loading well array; as well as Based on the determined well occupancy of neutral atoms in the storage well array and the well occupancy of neutral atoms in the loading well array, neutral atoms are moved from the loading well array to one or more unoccupied capture sites in the storage well array.

2. The method according to claim 1, in, The loading well array has a different well depth for the neutral atom than the storage well array, preferably at least 10%, more preferably at least 30%, and more preferably at least 75% greater than the storage well array for the neutral atom.

3. The method according to claim 1 or 2, wherein, Selectively removing neutral atoms in the first electronic state |1> from the storage trap array includes one or more of the following: Selectively heat the neutral atoms in the storage trap array that are in the first electronic state |1>; Selectively ionize neutral atoms in the storage trap array that are in the first electronic state |1>; Selectively modify the trapping potential of neutral atoms in the first electronic state |1> in the storage trap array; as well as The trapping potential of neutral atoms in the loaded trap array can be selectively modified.

4. The method according to claim 3, wherein, Selectively heating the neutral atoms in the storage trap array that are in the first electronic state |1> includes: A trap-selective heating laser pulse is applied to the loading trap array and the storage trap array, preferably a trap-selective Sisyphus heating laser pulse.

5. The method according to claim 4, wherein, Applying the trap-selective heating laser pulse includes: applying a sideband heating laser pulse whose frequency is detuned to the blue sideband of the neutral atoms in the loaded trap array.

6. The method according to any one of claims 1 to 5, wherein, Determining the well occupancy of neutral atoms in the storage well array and the well occupancy of neutral atoms in the loading well includes: Fluorescence imaging is performed on neutral atoms in the storage trap array and neutral atoms in the loading trap array or neutral atoms in the trapping sites of the optical lattice co-located with the loading trap array. At the same time, the neutral atoms are laser-cooled, preferably using sideband cooling, Raman sideband cooling, optical agglomerate cooling or Sisyphus cooling or combinations thereof to laser-cool the neutral atoms.

7. The method according to any one of claims 1 to 6, wherein, Moving the neutral atom from the load trap array to one or more unoccupied capture sites in the storage trap array includes: Based on the determined well occupancy of neutral atoms in the storage well array and the well occupancy of neutral atoms in the loading well array, determine one or more movement trajectories of one or more neutral atoms in the loading well array; and A beam guiding unit that controls one or more optical tweezers to execute a determined movement trajectory of the one or more neutral atoms, wherein the beam guiding unit preferably comprises a single-axis or multi-axis acousto-optic deflector or a digital mirror device.

8. The method according to any one of claims 1 to 7, further comprising: Laser-cooled neutral atoms are generated by operating a magneto-optical trap (MOT) for the neutral atoms in a spatial region comprising at least a portion of the loading trap array and at least a portion of the storage trap array.

9. The method according to claim 8, wherein, The MOT for the neutral atom is operated without using a laser that resonates with the optical transition of the neutral atom involving the second electronic state |2>.

10. The method according to any one of the preceding claims, wherein, The neutral atom is an alkaline earth-like atom that exhibits a dipole forbidden transition coupling the first electronic state |1> and the second electronic state |2>.

11. The method according to claim 10, wherein, The neutral atom is a strontium atom, preferably. 88 Sr atoms, and / or those thereof, The first electronic state is 1 The S0 state and the second electronic state is 3 P0 state.

12. The method according to any one of claims 1 to 11, in, The loading trap array includes an array of optical tweezers, the array of optical tweezers overlapping a first subset of the trapping sites of an optical lattice, the optical lattice preferably being a bow-shaped folded optical lattice; and The second subset of the capture sites of the optical lattice forms the storage trap array.

13. The method according to claim 12, in, Loading the plurality of neutral atoms into the loading trap array includes: preferably increasing the laser intensity of the optical tweezers array while operating the MOT for the neutral atoms; and The optical lattice is operated to maintain the storage trap array for neutral atoms in the second electronic state |2>.

14. The method according to any one of claims 12 or 13, in, Determining the well occupancy of neutral atoms in the loaded well array includes: Neutral atoms are transferred from the loaded trap array to trap sites in a first subset of the trap sites of the optical lattice; and optionally, While laser cooling the neutral atoms in the optical lattice, fluorescence imaging is performed on the neutral atoms in a first subset of the trapping sites of the optical lattice.

15. The method according to any one of claims 1 to 14, further comprising: Neutral atoms in the storage trap array are transferred from the first electronic state |1> to the second electronic state |2>.

16. A method for continuous operation of a quantum computing, simulation, and / or metrology device, comprising: Repeat the method according to any one of claims 1 to 15 to iteratively assemble and maintain the ensemble of neutral atoms in the storage trap array; A subset of the ensemble of the neutral atoms is used to perform quantum computing sequences, quantum simulation sequences, and / or quantum metrology sequences.

17. The method of claim 16, further comprising: Receive instructions from a remote computing device for executing the quantum computing sequence, the quantum simulation sequence, and / or the quantum metrology sequence; Based on the received instructions, execute the quantum computing sequence, the quantum simulation sequence, and / or the quantum metrology sequence; as well as The results of the executed quantum computing sequence, the executed quantum simulation sequence, and / or the executed quantum metrology sequence are sent to the remote computing device.

18. A quantum computing, simulation and / or metrology device, comprising means for performing the method according to any one of claims 1 to 17.

19. A computer program for controlling quantum computing, simulation and / or metrology devices to perform the method according to any one of claims 1 to 17.

20. A method for capturing and manipulating neutral atoms, comprising: Multiple neutral atoms trapped in the storage trap array are transferred from the first electronic state |1> to the second electronic state |2>; Multiple laser-cooled neutral atoms are loaded into a loading well array, which is arranged to be adjacent to or at least partially co-located with the storage well array; Determine the well occupancy of neutral atoms in the storage well array and the well occupancy of neutral atoms in the loading well array; as well as Based on the determination of well occupancy, neutral atoms are moved from the loaded well array to one or more unoccupied sites in the storage well array.

21. The method of claim 20, further comprising: After the plurality of neutral atoms trapped in the storage trap array are transferred from the first electronic state |1> to the second electronic state |2>, the neutral atoms in the first electronic state |1> are selectively removed from the storage trap array.

22. The method according to any one of claims 20 or 21, in, The loading duration of the plurality of laser-cooled neutral atoms into the loading trap array is at least one-half, preferably one-quarter, more preferably one-tenth, and even more preferably one-hundredth of the lifetime of the second electronic state |2>.

23. The method according to any one of claims 20 to 22, wherein, Loading the plurality of laser-cooled neutral atoms into the loading trap array includes: The laser-cooled neutral atoms are generated by operating a magneto-optical trap (MOT) for neutral atoms in a spatial region comprising at least a portion of the loading trap array and at least a portion of the storage trap array, or in a spatially separated region from the spatial region comprising at least a portion of the loading trap array.

24. The method according to any one of claims 20 to 23, wherein, Transferring the plurality of neutral atoms trapped in the storage trap array from the first electronic state |1> to the second electronic state |2> includes: The plurality of neutral atoms are coherently transferred from the first electronic state |1> to the second electronic state |2>; and / or The plurality of neutral atoms are incoherently transferred from the first electronic state |1> to the second electronic state |2> using one or more repump lasers.

25. A method for continuous operation of a quantum computing, simulation, and / or metrology device, comprising: Repeat the method according to any one of claims 20 to 24 to iteratively assemble and maintain the ensemble of neutral atoms in the storage trap array; A subset of the ensemble of the neutral atoms is used to perform quantum computing sequences, quantum simulation sequences, and / or quantum metrology sequences.

26. The method of claim 25, further comprising: Receive instructions from a remote computing device for executing the quantum computing sequence, the quantum simulation sequence, and / or the quantum metrology sequence; Based on the received instructions, execute the quantum computing sequence, the quantum simulation sequence, and / or the quantum metrology sequence; as well as The results of the executed quantum computing sequence, the executed quantum simulation sequence, and / or the executed quantum metrology sequence are sent to the remote computing device.

27. A quantum computing, simulation and / or metrology device, comprising means for performing the method according to any one of claims 20 to 26.

28. A computer program for controlling quantum computing, simulation and / or metrology devices to perform the method according to any one of claims 20 to 26.

29. A method for capturing and manipulating neutral atoms, comprising: Multiple neutral atoms are loaded into a loading well array, which is adjacent to or co-located with a storage well array. The loading well array includes at least 1,000 trapping sites, preferably at least 5,000 trapping sites, and more preferably at least 10,000 trapping sites overlapping with the source region of the laser-cooled neutral atoms. The loading well array is formed by multiple trapping sites of an optical lattice. Determine the well occupancy of neutral atoms in the storage well array and the well occupancy of neutral atoms in the loading well array; and Based on the determined well occupancy of neutral atoms in the storage well array and the well occupancy of neutral atoms in the loading well array, neutral atoms are moved from the loading well array to one or more unoccupied capture sites in the storage well array.

30. The method according to claim 29, in, The well depth of the optical lattice is greater than 0.1 mK, preferably greater than 1.0 mK, and / or The well frequency in the z direction, which is substantially orthogonal to the xy plane including the optical lattice, is greater than 2π × 1 kHz, preferably greater than 2π × 2 kHz, more preferably greater than 2π × 4 kHz, and even more preferably greater than 2π × 8 kHz.

31. The method according to claim 29 or 30, in, The depth of the loading well array is at least twice, preferably ten times, the depth of the storage well array; and / or The lifetime of the neutral atoms in the storage trap array is at least twice, preferably ten times, the lifetime of the neutral atoms in the loading trap array.

32. The method according to any one of claims 29 to 31, wherein, The neutral atoms are alkaline earth atoms of the same atom type; and The storage trap array operates at magic wavelengths for the electronic ground state and long-lived metastable electronic excited states of neutral atoms.

33. A method for continuous operation of quantum computing, simulation, and / or metrology devices, comprising: The method according to any one of claims 1 to 15 is repeated to iteratively assemble and maintain the ensemble of neutral atoms in the storage trap array; A subset of the ensemble of the neutral atoms is used to perform quantum computing sequences, quantum simulation sequences, and / or quantum metrology sequences.

34. The method of claim 33, further comprising: Receive instructions from a remote computing device for executing the quantum computing sequence, the quantum simulation sequence, and / or the quantum metrology sequence; Based on the received instructions, execute the quantum computing sequence, the quantum simulation sequence, and / or the quantum metrology sequence; as well as The results of the executed quantum computing sequence, the executed quantum simulation sequence, and / or the executed quantum metrology sequence are sent to the remote computing device.

35. A quantum computing, simulation and / or metrology device, comprising means for performing the method according to any one of claims 29 to 34.

36. A computer program for controlling quantum computing, simulation and / or metrology devices to perform the method according to any one of claims 29 to 34.

Images