Hardware-efficient neutral atom quantum computing method and device

Abstract

The present disclosure relates to an apparatus and method for quantum computing using a plurality of neutral atoms in an optical trap array wherein a first internal state of the neutral atoms is used as a qubit ground state 0gt; the second internal state is used as a quantum bit excitation state 1gt; . According to the present disclosure, a local single qubit gate operation can be performed on a qubit, comprising locally and selectively irradiating with a qubit addressing laser at a first qubit addressing laser frequency, which is prepared as a qubit ground state 0gt; and a quantum bit excited state 1gt; a superposition state sgt; the quantum bit is used for causing a quantum bit ground state 0gt; and a quantum bit excited state 1gt; and the differential stark displacement is realized. Furthermore, a local dual qubit gate operation may be performed on the pair of qubits, including locally and selectively irradiating with a qubit addressing laser at a second qubit addressing laser frequency prepared as a qubit ground state 0gt; a pair of qubits for a third internal state cgt, preferably via neutral atoms; coupling the quantum bit pair to a Rydberg state rgt of a neutral atom; a third internal state cgt; the method can be used as a quantum bit ground state 0gt; a Rydberg state rgt is obtained; and the intermediate state of two-photon transition is realized.

Description

introduce

[0001] Various aspects of this disclosure relate to quantum computing using, for example, neutral atoms in optical traps as qubits, and more specifically, to techniques, systems, and apparatuses for hardware-efficient quantum computing using, for example, neutral atoms in optical trap arrays.

[0002] For many years, quantum computers have been considered a computing platform capable of surpassing classical devices in certain applications. Quantum computers use physical qubits (qubits) as the basic units for storing information and execute quantum gates on these qubits to process the stored information, for example, according to the processing instructions of a quantum computing algorithm. Running quantum algorithms requires single-qubit gates and two-qubit gates, which are the basic computational operations performed on individual qubits and pairs of qubits, respectively.

[0003] Selective addressing of individual qubits and qubit pairs typically comes with significant technical overhead, complicating the scalability of quantum computing devices. Neutral atom-based quantum computers typically trap neutral atoms (i.e., electrically neutral atoms) in optical potentials (e.g., in optical dipole trap arrays or optical lattices), and the states of a qubit are formed by two long-lived (relative to operating time) internal (electronic) states or external (motion) states.

[0004] In such a system, transitions can occur between two stable or metastable internal states of a neutral atom, for example, via microwave transitions (e.g., for...). 87 Single-qubit gates can be realized by transitions (e.g., optical or microwave transitions) between two hyperfine states of a coupled electronic ground state (Rb = approximately 6.834 GHz). Two-qubit gates and multi-qubit gates, such as controlled phase gates, can be realized, for example, by optical transitions to highly excited Rydberg states that allow qubit-qubit interactions and entanglement in an optical trap array (for a full overview, see M. Morgado and S. Whitlock, Quantum simulation and computing with Rydberg-interacting qubits; AVS Quantum Sci. 3, 023501 (2021)). Summary of the Invention

[0005] One of the main challenges of neutral atom-based quantum computing devices is performing qubit gate operations on individual atoms with high fidelity. Typical single-qubit and two-qubit gate schemes require the use of lasers with local spatial control, and relative phase control between the atomic quantum state and the laser field at the neutral atom position used for the corresponding quantum gate operation. Known schemes for quantum computing using neutral atoms employ at least two different laser systems, each equipped with independent single-atom addressing units for operating local single-qubit and two-qubit gates.

[0006] Such addressing units for individual atoms in an optical trap array typically include specialized hardware (e.g., a combination of acousto-optic modulators (AOMs), acousto-optic deflectors (AODs), spatial light modulators (SLMs), and digital micromirror devices (DMDs)) for spatially controlling the addressing laser pattern, as well as additional hardware for rapidly turning the laser beam on and off to minimize error sources such as heating and qubit loss. Each laser control element also typically introduces power losses, which become critical for wavelengths where laser power is currently limited, such as for lasers operating in the UV or XUV range of the electromagnetic spectrum. These losses are particularly problematic for UV lasers that require the use of two-qubit gates for optical transitions to Rydberg states, where precise time switching and efficient local spatial control can be difficult, complex, and / or unreliable, thus limiting scalability and gate fidelity.

[0007] To address such and similar drawbacks of existing technologies, this disclosure provides quantum computing methods, apparatuses, and systems using neutral atoms trapped in optical potentials optimized for hardware efficiency.

[0008] This disclosure provides a method for quantum computing using multiple neutral atoms in an optical trap array, wherein a first internal state of the neutral atoms serves as the qubit ground state (|0>) and a second internal state serves as the qubit excited state (|1>). The method involves performing local single-qubit gate operations on the qubit, including locally and selectively irradiating a qubit prepared as a superposition state |s> of the qubit ground state |0> and the qubit excited state |1> at the first qubit-addressable laser frequency with a qubit-addressable laser to induce a differential Stark shift in the qubit ground state |0> and the qubit excited state |1>. The method also involves performing a local two-qubit gate operation on a pair of qubits, including locally and selectively irradiating the pair of qubits prepared as the ground state |0> of the qubits with a qubit-addressable laser at a second qubit-addressable laser frequency, for coupling the qubit pair to a Rydberg state |r> of the neutral atom, preferably via a third internal state |c> of the neutral atom, which serves as an intermediate state for the two-photon transition from the ground state |0> to the Rydberg state |r>. In some examples, a local one-qubit gate operation can be performed on multiple qubits forming a real subset of the set of qubits available in an optical trap array, wherein multiple qubits can be locally and selectively irradiated. Similarly, a local two-qubit gate operation can be performed on multiple qubit pairs forming a real subset of the set of qubits available in an optical trap array, wherein multiple qubit pairs can be locally and selectively irradiated.

[0009] For example, the first internal state can be the electronic ground state of a neutral atom (e.g., the spin singlet state of an alkaline earth atom in its electronic ground state). 1 S0). The second internal state can be a metastable state of a neutral atom or a long-lived excited state (e.g., the spin triplet state of an alkaline earth atom). 3 (P0). For example, the ground state of a qubit can be the S=0, L=0, J=0 state of the electronic ground state, the excited state of a qubit can be the S=1, L=1, J=0 state of the electronic excited state, and the third internal state can be the T=1, L=1, J=1 state of the electronic excited state, where S is the total spin quantum number, L is the orbital angular momentum quantum number, and J is the total angular momentum quantum number. Typically, the metastable, long-lived excited states of neutral atoms are excited states whose dipole matrix elements are essentially zero with respect to the electronic ground state. The long lifetime of such states results in very narrow linewidths for so-called clock transitions that can be used in optical atomic clocks, such as strontium optical lattice clocks.

[0010] Therefore, the quantum computing scheme disclosed in this paper is hardware efficient because the same addressing laser system can be used to implement both single-qubit and two-qubit gates. As mentioned above, a two-qubit gate can be implemented using the excited state |c> of a neutral atom as an intermediate state for two-photon coupling between the qubit ground state |0> and the highly excited, long-lived Rydberg state |r>. For example, in some implementations (see below) Figure 6 |c> can be used as a shelving state from which neutral atoms can be further excited into highly excited, long-lived Rydberg states |r>. Such shelving can be performed quickly and reliably because, for example, the effect of frequency shifts added to the laser on the two-qubit gate is negligible. In other implementations, |c> can be used as a so-called virtual state, for example, when the qubit ground state |0> is coupled to the Rydberg state |r> via a two-photon transition detuned relative to the intermediate state |c>. Furthermore, and importantly, the second photon used to couple to the Rydberg state |r> during the two-qubit gate can be obtained from a global laser source, thus avoiding the rapid local switching of UV sources, which can be technically challenging, expensive, and / or unreliable. In some implementations, to achieve a two-qubit gate, two independent addressing beams can be used to cancel the differential frequency shift between atoms caused by the local addressing hardware (e.g., AOD) of the qubit-addressed laser system. However, typically, the Rabi frequency of a qubit-addressable laser at or near the resonance of the |0> to |c> transition is much greater than the differential frequency shift between two adjacent qubits (e.g., about 1 MHz). Therefore, the effect of such a differential frequency shift has virtually no impact on gate fidelity.

[0011] Furthermore, qubit-addressable lasers can also be used to perform sideband cooling of neutral atoms during qubit initialization sequences and / or qubit readout sequences, for example, by globally irradiating an optical trap array with a qubit-addressable laser at a third qubit-addressable laser frequency (e.g., equal to the transition frequency of the |0> to |c> transition minus the trap frequency ω of the optical trap array), or by locally and selectively irradiating one or more qubits with a qubit-addressable laser at the third qubit-addressable laser frequency. For example, during qubit initialization or during qubit readout of the qubit array, global sideband cooling can be used to convert all qubits to the moving ground state and / or counteract the heating and losses caused by the imaging laser during qubit readout. Fluorescence emitted from atoms via global or local irradiation can also be directly collected for detecting qubit states.

[0012] Therefore, according to various aspects of this disclosure, single-laser systems (e.g., including single-laser media and / or single-output), i.e., qubit-addressable laser systems, can also be used for other processes crucial to neutral atom-based quantum computing. Importantly, the high degree of spatial and temporal control of qubit-addressable lasers required to achieve high-fidelity local quantum gates can therefore also be used to improve qubit array initialization and qubit readout, for example, for selective defect repair, calibration, and / or quantum error correction schemes due to atomic loss, where selective and local lossless readout and / or selective and local qubit cooling may be useful. Thus, this disclosure also facilitates the use of advanced quantum computing schemes without significantly increasing hardware complexity.

[0013] In some implementations, performing local single-qubit gate operations using a qubit-addressable laser can further include globally irradiating the optical trap array with a first qubit state transition laser pulse (e.g., a π / 2 pulse) to coherently convert qubits prepared as the ground state |0> to a superposition state |s> of the ground state |0> and the excited state |1>, and globally irradiating the optical trap array with a second qubit state transition laser pulse (e.g., a π / 2 pulse) to coherently convert qubits prepared as the superposition state |s> to the ground state |0>, i.e., qubits not irradiated by the qubit-addressable laser pulse. For example, the 180° phase shift required by the global qubit state transition laser can be induced by an EOM. In other implementations, the qubit-addressable laser can be used to induce such a phase shift in atoms that should not participate in single-qubit gate operations, i.e., returning to the initial state after a complete pulse sequence.

[0014] Furthermore, in some implementations of this disclosure, performing a local two-qubit gate operation using a qubit-addressable laser may include globally irradiating the optical trap array with a Rydberg laser pulse (e.g., a π pulse) to coherently transform one or more qubit pairs from the third internal state |c> of a neutral atom to a Rydberg state |r>, or to a superposition of the Rydberg state and the third internal state. In some implementations, performing a local two-qubit gate operation using a qubit-addressable laser may include locally and selectively irradiating one or more qubit pairs with a qubit-addressable laser at a fourth qubit-addressable laser frequency or a first qubit-addressable laser frequency to induce a differential AC-Stark shift between the third internal state and the Rydberg state. In this way, even advanced two-qubit gates, such as the so-called Levine-Pichler gate (see H. Levine et al.: Parallel Implementation of High-Fidelity Multiqubit Gates with Neutral Atoms; Phys. Rev. Lett. 123, 170503) or the so-called time-optimal gate (see S. Jandura and G. Pupilo: Time-Optimal Two- and Three-Qubit Gates for Rydberg Atoms; Quantum 6712 (2022)), can be implemented without significantly increasing hardware complexity.

[0015] As disclosed herein, qubits used for quantum computing can therefore be realized using the ground and metastable excited states of neutral atoms (e.g., alkaline earth metal atoms such as Dy, Er, Sr, Yb, etc.). In such and similar quantum computing systems, qubit states |0> and |1> can be coupled via a single laser beam resonance on an ultranarrow atomic transition in a single-photon excitation scheme. In such and similar systems, arbitrary SU(2) rotations in the Hilbert space spanned by |0> and |1> can be realized in a single qubit gate, for example, via two global π / 2 laser pulses with a 180° phase difference provided by a global qubit state transition laser, combined with one or more selective and local Stark shift pulses with controllable effective durations or pulse areas provided by a qubit-addressed laser system before, between, and after the global rotation generated by the qubit state transition laser of a globally illuminated optical trap array. In this way, qubit state-transition lasers, which typically require narrow laser linewidths and ultra-low phase noise, do not require local position control and arbitrary pulse width or pulse shape control to achieve arbitrary single-qubit gates. Since the typical hardware used for controlling such laser parameters degrades laser phase noise and power, as well as frequency and / or amplitude stability, this disclosure substantially improves the single-qubit gate fidelity in such systems because the qubit state-transition laser can operate as cleanly and therefore robustly as possible without negatively impacting spatial control and gate speed.

[0016] For example, in some implementations, neutral atoms may include boson strontium isotopes. 88 For such physical qubits, the qubit-addressable laser can operate at a design wavelength of 689.4 nm, the Rydberg laser at 318.5 nm, and the qubit-state transition laser at 698.4 nm. As discussed above, neither the Rydberg laser nor the qubit-state transition laser requires local position control, arbitrary pulse width / pulse shape control, or tunable laser frequency control. Therefore, this disclosure enables a significant reduction in the hardware requirements of two laser systems that need to operate as robustly and reliably as possible in their design.

[0017] As disclosed herein, the hardware complexity required for position, pulse width / shape, and frequency control is thus transferred to qubit-addressable laser systems. For qubit-addressable laser systems, the effects of fluctuations in gate fidelity, losses, etc., caused by additional hardware (e.g., amplitude or laser frequency fluctuations caused by AOM, AOD, etc.) are far less severe than in ultra-narrow qubit state transition lasers or UV Rydberg lasers. For example, qubit-addressable laser systems can operate in the visible or near-infrared portion of the electromagnetic spectrum, where robust and reliable laser sources and fast laser control devices, including integrated optics, are readily available. In contrast, local addressing optics and spatial light modulation devices are difficult to implement in the UV spectral range. Furthermore, the third internal state |c> (as the initial state |c> for Rydberg transitions) for single-qubit gates (via AC Stark shift) and for two-qubit gates typically have much shorter lifetimes than the qubit excited state |1>. Therefore, the associated linewidth is also substantially larger than the linewidth of the qubit state transition from |0> to |1>. Therefore, the amplitude and / or laser frequency fluctuations of qubit-addressed lasers have a much smaller impact on gate fidelity compared to similar fluctuations in qubit state transition lasers.

[0018] To further reduce the hardware complexity of neutral atom-based quantum computing, qubit-addressable laser systems can also be used for sideband cooling of neutral atoms in optical trap arrays, for example, during qubit readout and / or for high-fidelity state preparation after initial trap loading and / or after qubit readout. Typically, pre-cooled neutral atoms (e.g., using laser cooling) are loaded into the optical trap array such that each atom is in an excited motion state of the optical trap array. However, for reliable quantum computing, all qubits should be cooled to or near the motion ground state of the optical trap array. According to this disclosure, the qubit-addressable laser system can operate at a third laser frequency that facilitates sideband cooling of each trapped atom. Therefore, using the same laser system required to perform arbitrary local quantum gates, globally illuminating the optical trap array with the qubit-addressable laser system, for example via an additional global illumination path that can be superimposed with a qubit state transition laser, can be used to initialize the qubit array to or near the motion ground state of the trap array.

[0019] Therefore, in some implementations, the methods disclosed herein may further include one or more of the following: shifting the laser frequency of the qubit-addressing laser to a first qubit-addressing laser frequency for controlling the phase of the qubit during a single-qubit gate, shifting the laser frequency of the qubit-addressing laser to a second qubit-addressing laser frequency for coherently transitioning the qubit to an initial state of a Rydberg transition, and shifting the laser frequency of the qubit-addressing laser to a third qubit-addressing laser frequency for sideband cooling (e.g., resolvable sideband cooling or Raman sideband cooling) during qubit initialization and / or readout.

[0020] For example, as follows Figure 5 and Figure 6 As shown, position control of a quantum gate laser can be achieved using an acousto-optic deflector (AOD), where an applied radio frequency (RF) driving tone causes an angular deflection of the laser beam transmitted through the AOD. Multiple RF tones can be applied to split a single beam into multiple beams, where the deflection angle can be configured by controlling the RF frequency of the driving tone. However, in such a locally addressed qubit system, the light passing through the AOD is also frequency-shifted by the RF tone, causing a spatially correlated laser frequency in the locally addressed beam, which can be a problem for conventional qubit gate schemes. Typically, this is addressed using additional frequency shifters for compensation, which further increases the hardware complexity and potential error sources of conventional quantum computing devices. In contrast, in the methods, apparatuses, and systems disclosed herein, the qubit-addressed laser can operate with respect to any atomic transitions far from neutral atoms, such that the frequency shift caused by the AOD or similar device has a negligible effect on, for example, gate fidelity, and therefore, no compensation hardware is required when this disclosure is applied to such a position control system.

[0021] As mentioned above, in some implementations, performing local two-qubit gate operations can also involve globally irradiating the optical trap array with Rydberg laser pulses to coherently convert one or more qubit pairs prepared as neutral atoms' third internal states |c> to neutral atoms' Rydberg states |r>, or to a superposition of Rydberg states and third internal states. Some two-qubit gates that use Rydberg blocking to generate entanglement require applying specific differential phase shifts to the third internal states |c> and Rydberg states |r>. Typically, such phase shifts require phase control of the Rydberg laser system (e.g., using an electro-optic modulator (EOM)), which further increases hardware complexity. According to this disclosure, such phase shifts required for some Rydberg blocking-based two-qubit gates can also be generated by a qubit-addressable laser system, for example, via inducing differential AC-Stark shifts for |c> and |r>, without significantly increasing hardware complexity. For example, the method described above may also include locally and selectively irradiating one or more pairs of qubits with a qubit-addressable laser at a fourth qubit-addressable laser frequency to induce a differential AC-Stark shift in the third internal state and the Rydberg state.

[0022] See below for reference. Figure 6 In detail, such a differential AC Stark shift for the third inner state and the Rydberg state can be applied between two consecutive Rydberg laser pulses or during a single Rydberg laser pulse. In both cases, the differential Stark shift can be controlled so that the qubit pair can be in an entangled state after the gate sequence is complete—essential for any useful quantum algorithm.

[0023] In some implementations, for example, when the excited qubit state is implemented as an optical clock state, for example... 88 Sr 3 In the P0 state, the method described herein can also include generating a magnetic field at the location of the optical trap array that defines the angular momentum quantization axis of the neutral atom and induces a perturbative admixture of the third internal state |c> with the qubit excited state |1>. In other implementations, states with different dipole matrix elements from the |0> state (dipole-allowed states) can also be perturbatively mixed into the |1> state. For example, such a magnetic field can be used to adjust the amount of perturbative mixture, thereby adjusting the lifetime of the qubit excited state |1>, which decreases linearly with the magnetic field strength. Since the linewidth of the qubit state transition from |0> to |1> essentially corresponds to the inverse lifetime of |1>, it makes it possible to maximize the normalized Rabi frequency of spontaneous emission from |1> with few constraints on the timescale required for quantum computing.

[0024] Various aspects of this disclosure also contribute to substantial improvements in gate timing and external noise suppression. For example, in some implementations, the methods disclosed herein may further include performing multiple single-qubit and two-qubit gate operations, and periodically driving a qubit-addressable laser to generate a first laser pulse train with a constant inter-pulse duration for performing multiple single-qubit and two-qubit gate operations, which improves reliability and repeatability. Furthermore, the electronic control hardware and circuitry required to drive optoelectronic pulse shaping devices (e.g., AOM, AOD, EOM, etc.) can be simplified. For example, a fixed RF pulse train can be used to drive the pulse shaping device of a locally qubit-addressable laser system, while the atomic population and coherent control required to perform arbitrary single-qubit and two-qubit gates can be achieved by controlling the laser amplitude and / or laser frequency. This allows for complete control of the pulse area even for high RF pulse repetition frequencies, because the normalized pulse shape can be fixed, and the peak pulse intensity and laser frequency can be controlled on a slower timescale corresponding to the inverse RF pulse repetition frequency, rather than on a faster timescale corresponding to the pulse width.

[0025] Various aspects of this disclosure also enable the reduction of the negative impacts of optical trap arrays on single-qubit and two-qubit gates. For example, in some systems, two-qubit gate fidelity benefits from low well depth because it reduces motion dephasing during Rydberg excitations. In contrast, single-qubit gates benefit from high well depths, such as wells located in the Lamb-Dicke region. Furthermore, as mentioned above, larger magnetic fields can be used to achieve faster single-qubit gates, for example, due to greater mixing of the third internal state |c> with the qubit excited state |1>. However, this typically also requires deeper optical traps to maintain high single-qubit gate fidelity.

[0026] To avoid or at least improve such a trade-off in optical trap depth between single-qubit gates and two-qubit gates, in some implementations of this disclosure, the optical trap array may include / present a trap frequency ω that is substantially the same for all qubits, and performing local single-qubit gate operations may include irradiating the optical trap array with a single-qubit state transition laser pulse or two consecutive qubit state transition laser pulses having a (combined) effective pulse duration equal to ω / Ω = 2n+1, n∈N, where Ω is the Rabi frequency associated with the qubit state transition laser and the |0> to |1> transition matrix element. As follows Figure 9As shown, adhering to this condition minimizes motion decoupling during single-qubit gates. Therefore, quantum computing devices can operate at lower well depths while maintaining single-qubit gate fidelity. Lower well depths improve two-qubit gate fidelity and reduce negative effects caused by high well laser intensity, such as non-resonant scattering, and also enable simplified, low-power design and operation of well laser systems. The same effect was observed for local single-qubit gates when an appropriate wait time was chosen between the two π / 2 pulses described above.

[0027] In some implementations, the methods described herein may also include synchronizing the periodic drive of the qubit-addressable laser system (and in some implementations, other laser systems described herein) with the characteristic frequencies of external noise sources that affect the qubit gate fidelity. For example, a pulse generator providing periodic drive for the qubit-addressable laser can act as the natural clock frequency of the quantum computing device. This allows the pulse generator capable of providing periodic drive to synchronize with (quasi-)periodic environmental effects, such as ambient electromagnetic field fluctuations caused by the line frequency of an AC power grid, which may cause transition frequency shifts, for example, via the Zeeman effect. This synchronization of qubit gate operation with external noise sources enhances the gate fidelity and repeatability of the quantum computing device.

[0028] In some implementations, the qubit-addressable laser can also be driven periodically to generate a second laser pulse train with a constant inter-pulse duration. For example, a single laser source can be fed into two identical opto-pulse devices, such as an AOM. In this way, the interval between the first and second laser pulse trains can be controlled, which enables phase control of the two-qubit gate for use in optimized gate sequences.

[0029] This disclosure also relates to qubit-addressable laser systems and quantum computing devices suitable for performing the methods disclosed herein. Such systems and devices are discussed below with reference to exemplary implementations illustrated in the accompanying drawings, as well as further details of the methods described above and related computer programs. The foregoing provides a broad overview of the features and technical advantages of examples according to this disclosure in order to better understand the following detailed description. 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 the associated advantages will be better understood from the following description when considered in conjunction with the accompanying drawings. Each drawing in the accompanying drawings is provided for illustrative and descriptive purposes and not as a definition of limitation of the claims. Attached Figure Description

[0030] To gain a more detailed understanding of the features described above, reference can be made to the aspects (some of which are shown in the accompanying drawings) for a more specific description summarized above. However, it should be noted that the drawings illustrate only some typical aspects of this disclosure and should not be construed as limiting its scope; other equally valid aspects may be acknowledged for this description. The same reference numerals in different drawings may identify the same or similar elements.

[0031] Figure 1 A neutral atom quantum computing device according to a possible implementation of the present disclosure is shown.

[0032] Figure 2 Position control of local single-qubit gates and local two-qubit gates according to possible implementations of this disclosure is shown.

[0033] Figure 3 A typical operating sequence of a neutral atom quantum device according to a possible implementation of this disclosure is shown.

[0034] Figure 4 Boson strontium isotopes shown can be used as physical qubits in some implementations of this disclosure. 88 Atomic-level scheme for Sr.

[0035] Figure 5 A block diagram of an exemplary qubit-addressable laser system according to a possible implementation of the present disclosure is shown.

[0036] Figure 6 This illustrates how a local single-qubit gate with arbitrary SU(2) rotations can be implemented according to possible implementations of this disclosure.

[0037] Figure 7 This illustrates how a local two-qubit gate using Rydberg blocking can be implemented according to possible implementations of this disclosure.

[0038] Figure 8 The following illustrates aspects of qubit initialization and readout according to some implementations of this disclosure.

[0039] Figure 9 This illustrates how the duration of a single-qubit pulse can be selected to minimize the nonfidelity of a single-qubit gate, according to possible implementations of this disclosure.

[0040] Figure 10 A processing diagram of a quantum computing method according to a possible implementation of the present disclosure is shown.

[0041] Figure 11 A processing diagram of a quantum computing method according to a possible implementation of the present disclosure is shown. Detailed Implementation

[0042] The various aspects of this disclosure will be described in more detail below 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 set forth 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, any number of aspects set forth herein may be used to implement an apparatus or system or a practice. Furthermore, the scope of this disclosure is intended to cover such apparatuses, systems, and methods practiced using other structures, functions, or structures and functions other than or different from the aspects of this disclosure 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 apparatus, system, method, and computer program 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 features, steps, components, and / or functional elements compatible with any other aspect of this disclosure.

[0043] Several aspects of quantum computing utilizing neutral atoms will now be presented with reference to various devices, 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 these elements are implemented as hardware or software depends on the specific application and design constraints imposed on the entire system.

[0044] Although the following uses boson strontium isotopes 88 Sr is used as a qubit to represent various aspects of this disclosure, but it should be understood that any other kind of neutral atom with a suitable internal energy level structure may also be used in other implementations of this disclosure.

[0045] Figure 1 A quantum computing device 100 according to a possible implementation of the present disclosure is shown. The quantum computing device 100 is configured to perform steps of a quantum computing method as disclosed herein (e.g., the method is described below). Figure 10 and Figure 11 Neutral atoms used as qubits are trapped in an optical trap array 110 within a vacuum chamber 120 (e.g., a glass vacuum cell) that allows optical access. The optical trap array 110 can be an optical dipole trap array, generated, for example, using a digital micromirror device (DMD), a spatial light modulator (SLM), or an integrated photonic unit 130 for local addressing, and imaged onto a plane within the vacuum chamber 120 by a high-resolution imaging system 140. In other implementations, the optical trap array 110 can also be formed from the individual planes of an optical lattice. To initialize the trapped qubit array that can be used for quantum computing, typically, the neutral atoms are first laser-cooled, for example via a magneto-optical trap (MOT) operating within the vacuum chamber 120 and loaded from an atomic source (not shown), such as a Zeeman decelerator. The laser-cooled neutral atoms can then be loaded into the optical trap array 110 and further cooled to or near the motion ground state of the optical trap array 100 to initialize substantially the same qubit array. (See below for reference.) Figure 3 Further details regarding the initialization sequence of the qubit array will be discussed.

[0046] As discussed in detail above, typically, the two internal states of a neutral atom serve as the ground state |0> and the excited state |1> of a qubit. For example, the spin singlet electronic ground state (e.g., 1 S0) can be used as |0>, and metastable excited spin triplet states (e.g., 3 P0 can be used as |1> (see details below) Figure 4 The one shown 88 (Sr energy level scheme). The qubits trapped in the optical trap array 110 thus form an atomic quantum register that can be used for quantum computing and quantum simulation. To realize the local single-qubit gate and local two-qubit gate required for quantum computing, the quantum computing device 100 uses several laser systems to control the internal states of neutral atoms. According to this disclosure, a local single-qubit gate with arbitrary SU(2) rotation can be realized using a (global) qubit state transition laser 150 combined with a qubit-addressable laser 160, which enables selective and local control of individual qubits in the optical trap array 110.

[0047] exist Figure 1In the example, the qubit-addressable laser system includes a laser source included in a laser system rack 170, and an integrated photonic qubit addressing unit 162 coupled to the laser source via an optical fiber 164. The qubit-addressable laser system also includes a high-resolution imaging system 166, and a laser frequency and pulse area (e.g., pulse shape and amplitude) controller (see [link to other details and options]) that may be included in the laser system rack 170 and / or the integrated photonic qubit addressing unit 162. Figure 5 The quantum computing device 100 also includes a modular control signal unit 180, which generates various analog signals, digital signals, and RF control and drive signals required for precise and full-timing control of the laser system, magnetic field system (not shown), quantum bit readout system (e.g., including a CMOS camera 195 for fluorescence imaging using a high-resolution imaging system 166 addressing the laser system with quantum bits).

[0048] As discussed above, a local two-qubit gate can be implemented using a global Rydberg laser 190 combined with a qubit-addressable laser 160. According to this disclosure, a qubit-addressable laser system (details see...) Figure 5 , Figure 6 and Figure 7 The laser is configured to locally and selectively irradiate each qubit with a qubit-addressable laser pulse at a first qubit-addressable laser frequency when the qubit is prepared as a superposition state |s> (e.g., via a previous global π / 2 pulse from the qubit state transition laser 150) to induce a differential Stark shift in the qubit ground state |0> and the qubit excited state |1>. For a local two-qubit gate acting on a single qubit pair, the qubit-addressable laser 150 can be controlled to locally and selectively irradiate one or more qubit pairs prepared as the qubit ground state |0> with a qubit-addressable laser pulse at a second qubit-addressable laser frequency to transition one or more qubit pairs to a third internal state |c> of a neutral atom, which can be used as the initial state for an optical transition to the Rydberg state |r> of the neutral atom and / or as a temporary state for selective readout of the state. Refer to the following... Figure 6 and Figure 7 Details are described regarding single-qubit and two-qubit gates using a qubit-addressable laser 150, a qubit state transition laser 150, and a Rydberg laser 190.

[0049] As disclosed herein, neither the qubit state transition laser system nor the Rydberg laser system requires local control for implementing arbitrary single-qubit and two-qubit quantum gates. Instead, both laser systems can be designed such that they only require global illumination of the optical trap array 110, while only the qubit-addressable laser system requires optoelectronic control devices for selectively and locally illuminating individual qubits and single-qubit pairs within the optical trap array. As discussed above, the qubit-addressable laser system can also be used for laser cooling (e.g., sideband cooling) of neutral atoms during qubit array initialization and / or qubit state readout to further reduce hardware complexity.

[0050] The modular control signal unit 180 and optionally the laser system rack 170 can obtain single-qubit gate operation instructions and two-qubit gate operation instructions of the quantum computing protocol (e.g., from the digital computer 195 or via an interface to a cloud computing system), and in response, control the operation of the qubit-addressable laser 150 according to the obtained qubit gate operation instructions to locally and selectively irradiate individual qubits and qubit pairs to perform single-qubit gate and two-qubit gate operations. Similarly, as disclosed herein, the modular control signal unit 180 can also obtain and process instructions for qubit array initialization and / or qubit readout.

[0051] Figure 2 An enlarged view of a portion of an optical trap array 110 with a uniform spacing of 3 µm is shown to conceptually illustrate how to utilize, for example... Figure 1 The quantum computing device shown implements local single-qubit gates (see sub-figure a) and local two-qubit gates (see sub-figure b). For the single-qubit gate, the optical trap array 110 is globally irradiated by a qubit transition laser 150 to transition the irradiated qubit, prepared as the qubit ground state |0>, to a superposition state |s> of the qubit ground state |0> and the qubit excited state |1>. The qubit addressing laser 160 can be controlled to selectively and locally irradiate individual qubits, for example via the following... Figure 5 The laser pulse sequences shown and described above generate arbitrary SU(2) rotations. For a two-qubit gate, the optical trap array 110 is globally illuminated by a Rydberg laser 190 to couple the third inner state |c> to the Rydberg state |r>. The qubit-addressing laser 160 can be controlled to selectively and locally illuminate individual qubits to perform the following... Figure 7 The diagram shows a two-qubit gate based on Rydberg blocking.

[0052] Typically, this disclosure also relates to a neutral atom quantum computing device 100, which may include a trapping laser system for creating an optical trap array 110 for neutral atoms used as qubits in the quantum computing device 100, wherein a first internal state of the neutral atom serves as the qubit ground state |0> and a second internal state serves as the qubit excited state |1>. The disclosed quantum computing device may also include a qubit state transition laser system configured to globally irradiate the optical trap array 110 to transition a qubit prepared as the qubit ground state |0> to a superposition state |s> of the qubit ground state |0> and the qubit excited state |1>. The device 100 may also include a qubit addressing laser system (e.g., see below). Figure 5 The qubit-addressable laser system is configured to locally and selectively irradiate two or more individual qubits at the first qubit-addressable laser frequency, causing a differential Stark shift in the qubits in a superposition state |s>; and preferably couples the qubit pair to a Rydberg state |r> of a neutral atom via a third internal state c> of the neutral atom, the third internal state c> serving as an intermediate state for a two-photon transition from the qubit ground state |0> to the Rydberg state |r> (for the use of bosons). 88 For details on possible implementations of Sr atoms, see [link to relevant documentation]. Figure 4 As discussed above, the quantum computing device 100 may also include a Rydberg laser system configured to globally illuminate the optical trap array 110 to couple a third internal state c> to a Rydberg state |r>, and a quantum bit readout system that uses fluorescence emitted by trapped atoms during the quantum bit readout sequence.

[0053] As disclosed herein, such a neutral atom quantum computing device 100 may include a device / can be configured to perform the quantum computing methods disclosed herein, for example, as referenced above and below. Figure 10 and Figure 11 The methods discussed. In particular, the neutral atom quantum computing device 100 may include those disclosed herein (e.g., those discussed above or referred to below). Figure 5 A quantum bit addressing laser system. For example, a neutral atom quantum computing device 100 may include a magnetic field system to generate a magnetic field at the location of an optical trap array 110, which defines the angular momentum quantization axis of the neutral atom and causes a perturbation mixing of dipole allowed states and quantum bit excited states.

[0054] Figure 3 The following illustrates a neutral atom quantum device (e.g., according to possible implementations of this disclosure). Figure 1 The exemplary operation sequence of the quantum computing device 100 shown is as follows. The exemplary operation sequence begins with laser cooling of neutral atoms. In the example shown, it first uses, for example, a wide...1 S0 to 1 Blue MOT neutrality during P1 transition operation 88 Laser cooling of Sr atoms (see below) Figure 4 ), followed by the red MOT (which is also used by quantum-addressable lasers in a narrower region). 1 S0 to 3 Laser cooling is performed on the P1 transition (operation). Due to the lower Doppler cooling limit, which is proportional to the natural linewidth of the laser-cooled transition, 1 S0 to 3 The relatively narrow linewidth of the P1 transition allows for further reduction 88 The temperature of Sr atoms.

[0055] The laser-cooled neutral atoms are then loaded into an optical trap array to form a qubit array, and then imaged, for example, via fluorescence imaging including sideband cooling as described below, to detect potential defects in the qubit array. These potential defects can then be removed during a subsequent atom sorting process, resulting in a qubit array with unit filling. Next, sideband cooling (e.g., using a qubit-addressable laser system) cools the neutral atoms to or near the motion ground state of the optical trap array to prepare substantially the same qubit array, which can then be used to perform quantum computing protocols that typically involve executing multiple local single-qubit gates and local two-qubit gates. See below for reference. Figure 8 As discussed, at the end of the quantum computing phase, the internal states of qubits can be read out, for example, via lossless and state-sensitive fluorescence imaging. During and / or after the qubit readout sequence, a qubit-addressable laser system can be used, for example (see below). Figure 5 Using a global illumination path to perform sideband cooling enables the reuse of the qubit array to execute subsequent quantum computing protocols without requiring laser cooling and optical trap loading to initialize the qubit array.

[0056] Figure 4 Boson strontium isotopes shown can be used as physical qubits in some implementations of this disclosure. 88 The atomic-level scheme for Sr. In this implementation, the ground state |0> of the qubit can be obtained through the spin singlet state of the electronic ground state. 1 S0 can be achieved, and the excited state |1> of the qubit can be obtained from the first electronic excited state. 3 The P0 state is realized. As is known in the art, the dipole transition matrix elements of both states vanish due to the angular momentum selection rule, and |1> is a metastable excited state with a long lifetime compared to the typical timescale of quantum computing protocols.

[0057] As disclosed in this article, qubit-addressable laser systems can be used with approximately The natural line width 1 S0 to 3 The P1 transition corresponds to the operation at the design wavelength of 698.4 nm. 88 Sr 3 The P1 state is an example of an inner state (specified above as the third inner state |c>), which can be used to achieve two-photon coupling between the qubit ground state |0> and the Rydberg state |r> required for a two-qubit gate. For example, the state |c> can be used as a reference to... Figure 7 The initial state of the optical transition of the highly excited Rydberg state |r> in the described two-qubit gate scheme is given. When the laser frequency of the qubit-addressable laser resonates with the transition from |0> to |c>, the qubit-addressable laser system can selectively and locally transition one or more individual qubits to a third inner state |c>, a process sometimes referred to as shelving. Therefore, |c> is also designated as the shelving state herein. For example, a qubit can be shelved in |c>, as discussed below.

[0058] When with 1 S0 to 3 When operating with slight detuning of the P1 resonance (e.g., within 100 kHz), the qubit-addressable laser can also be used, for example, for sideband cooling of neutral atoms during qubit array initialization and / or qubit state readout. For local single-qubit gates with arbitrary SU(2) rotations, the qubit-addressable laser can also be used with... 1 S0 to 3 Operating under further detuning of the P1 resonance (e.g., about 1 GHz), such that when the qubit is prepared in a superposition state |s> (e.g., 1 / √2(0>+|1>)), for example via a laser from the qubit state transition, 1 S0 to 3 The resonance of the P0 transition, preceded by a global π / 2 pulse, induces differential AC Stark shifts in the |0> and |1> states. Due to detuning, this can essentially be achieved without population... 3 Applying AC Stark displacement in the P1 state.

[0059] Switch to 3Atoms in the P1 transient state |c> can also transition to the Rydberg state |r>, for example, via a global π pulse generated by a Rydberg UV-laser system operating at a designed wavelength of 318.5 nm. Therefore, operating the qubit-addressable laser at two different laser frequencies enables the realization of local single-qubit gates and local two-qubit gates without requiring selective position control of the qubit-state transition laser system and the Rydberg laser system. As mentioned above, the qubit-state transition laser system and the Rydberg laser system can thus operate at fixed laser frequencies without requiring complex optoelectronic hardware for position control and arbitrary pulse area control.

[0060] As mentioned above, this is permissible in wide dipole ranges. 1 S0 to 1 The P1 transition is performed at a design wavelength of approximately 460.9 nm. 88 Initial laser cooling of Sr atoms (e.g., for manipulating Zeeman slowers or blue MOTs). The same transition can also be used for fluorescence imaging during qubit state readout (see below). Figure 8 ).

[0061] Figure 5 A functional block diagram of a qubit-addressable laser system 500 according to a possible implementation of this disclosure is shown. The qubit-addressable laser system 500 includes a laser source 510, such as an extended cavity diode laser amplified by a cone amplifier and frequency-stabilized to a high-fine-precision cavity using the Pound–Drever–Hall technique. In other implementations, the laser amplifier may be injection-locked to an extended cavity diode laser whose frequency can be stabilized to a frequency comb. In the example shown, the output of the laser source 510 may be split into two paths by a beam splitter 520 (e.g., a polarization beam splitter or similar optical element), wherein a first path may be used for selective and local illumination of a single qubit, and a second path may be used for global illumination of the qubit array. In other implementations, instead of using a beam splitter, the laser source 510 may include two or more separate lasers employing relative frequency stabilization mechanisms (e.g., injection locking, optical beat locking, etc.) to provide two independent illumination paths with offset laser frequencies.

[0062] The first output of beam splitter 520 can be controlled by a laser frequency shifter 520, such as an acousto-optic frequency shifter (AOFS), to modify the frequency of the beam via the Doppler effect. Acoustic waves traveling through the device crystal cause the frequency of the laser beam to change proportionally to the RF drive frequency. For example, a dual-pass AOM or fiber-coupled laser frequency shifter can be used to apply a controllable frequency shift up to 1.5 GHz while substantially maintaining the pointing stability of the laser beam.

[0063] For precise pulse width and / or pulse shape control, an optical pulse width / pulse shape controller 530, such as an AOM driven by a controllable RF signal, can be employed. To improve stability and ensure independent parameter control, it is often meaningful to use a separate device for frequency shifting and another device for precise pulse width and / or pulse shape control. However, pulse width / shape control can also be performed via a frequency shifter 520, for example, by driving the frequency shifter with a pulsed RF signal of controllable pulse width / shape. To completely suppress stray light, both paths may also include an optomechanical switch / shutter (not shown).

[0064] To achieve selective and local addressing of each qubit in the optical trap array 395, such as Figure 1 As shown in the example, the qubit-addressable laser system 500 also includes a 2D focus position controller 550, followed by a high-resolution imaging system 560. Combined, the optical elements 520 to 560 enable selective addressing of one or more qubits in the optical trap array 595 with a controllable laser frequency and a controllable pulse area. For example, to perform four identical single-qubit gates, the frequency shifter 520 can be controlled, for example, via controlling the frequency of an RF driver (not shown), to shift the laser frequency relative to the transition from |0> to |c> by a red detuning, thereby causing a differential AC Stark shift in the qubits in the superposition state |s>. The phase difference between |0> and |1> required for arbitrary SU(2) rotation can be controlled using a pulse width / pulse shape controller 530 to adjust the pulse area of ​​the laser pulse causing the differential Stark shift, while selective qubit addressing is provided by the 2D focus position controller 550.

[0065] The 2D focus position controller 550 may include two orthogonal AODs in series, driven by multiple RF tones having controllable RF frequencies. In other implementations, it may also use... Figure 1 The DMD, SLM, or integrated photonic addressing unit shown is used to achieve focus position control. The controllable laser focusing pattern generated by the 2D focus position controller 550 can then be imaged onto the optical trap array by the high-resolution imaging system 560 (see...). Figure 2 In this context, addressing and / or selective storage are required during local single-qubit or two-qubit gates. 3 Each qubit of the P1 state.

[0066] In the illustrated example, the second output of beam splitter 520 can be used for global illumination of optical trap array 595, for example, for sideband cooling during initialization sequences and / or qubit readout sequences, as discussed above. The global illumination path may include a frequency shifter 570 that can operate at a substantially fixed frequency shift required for sideband cooling, and an optical switch 580, which may be a pulse width / shape controller as discussed above for local and selective paths, or may be implemented using simpler hardware. Clearly, a 2D focus position controller and a high-resolution imaging system are not necessary for the global illumination path. Alternatively, global illumination optics such as optical fibers and beam collimators may be used for global illumination of optical trap array 395.

[0067] Therefore, this disclosure also typically relates to a qubit-addressing laser system 500 for a quantum computing device 100, which uses neutral atoms trapped in an optical trap array 110 as qubits. The qubit-addressing laser system 500 may include a laser source 510 that provides laser light with a designed wavelength corresponding to the transition between a first internal state and a second internal state of the neutral atom used as qubit states |0> and |1> as discussed above. The qubit-addressable laser system 500 may include a laser frequency controller / shifter 530 for setting the frequency of the laser provided by the qubit-addressable laser system 500 to a first laser frequency and a second laser frequency, and a pulse width or pulse shape controller 540 for generating laser pulses with controllable pulse width or pulse shape / area, such that for the first laser frequency, the phase difference between the qubit ground state |0> and the qubit excited state |1> of the qubit superposition state |s> can be controlled between 0° and 180°, and such that for the second laser frequency, the qubit ground state |0> can be transitioned to a third internal state |c> of a neutral atom, the third internal state |c> being used as an intermediate state for the two-photon transition of a neutral atom to the Rydberg state |r>. The qubit-addressable laser system 500 may also include a laser focusing position controller 550, which controls the focusing position of the laser provided by the qubit-addressable laser system 500, such that at least two neutral atoms trapped in the optical trap array 110 can be selectively and locally irradiated simultaneously, for example, via a high-resolution imaging system 560.

[0068] In some implementations of this disclosure, the laser focusing position controller 550 may include one or more of an acousto-optic deflector, a digital micromirror device, a spatial light modulator, an electro-optic deflector, and an integrated photonic unit 162 for local addressing. Furthermore, the laser frequency controller 530 may include one or more of an acousto-optic modulator, an acousto-optic frequency shifter, and an optical phase-locked loop, and in some implementations, the pulse width or pulse shape controller 540 may include one or more of an acousto-optic modulator, an acousto-optic deflector, an electro-optic modulator, and an electro-optic deflector.

[0069] Furthermore, the laser frequency controller 530 or an additional laser frequency controller, such as a fixed frequency shifter 570, can be configured to set the frequency of the laser provided by the qubit-addressed laser system 500 to a third laser frequency. For example... Figure 5 As shown, the qubit-addressable laser system 500 may also include a global illumination path to globally illuminate the optical trap array 110 with a laser at a third laser frequency to cool the trapped atoms via sideband cooling.

[0070] Figure 6 This conceptually illustrates how a local single-qubit gate can be implemented according to the present disclosure. Figure 6 In subfigure a), a combination of one-dimensional selective qubit addressing by a qubit-addressing laser beam 610 and global illumination by a qubit state transition laser beam 620 is shown. As described above, the qubit-addressing laser system may include an AOD 630 and a large NA imaging objective 640, which focuses the output beam of the AOD 630 onto an optical trap array 650 that traps neutral atoms used as physical qubits. As mentioned above, the focal position of the qubit-addressing laser beam 610 can be controlled by changing the frequency of the RF signal driving the AOD 630. For example, multiple qubits can be addressed simultaneously by driving the AOD 630 with RF drive signals comprising two or more different RF frequencies (e.g., via multi-tone RF signals), where each tone corresponds to a different deflection angle of the AOD 630. Two-dimensional position control can be achieved, for example, by using two orthogonal AODs 630 in series.

[0071] Figure 6 Subgraph b) shows a configuration in which all atoms are simultaneously irradiated by a global qubit transition laser beam 620 and a local qubit addressing laser pulse 610, and the local qubit addressing laser pulse 610 locally irradiates all qubits to be addressed simultaneously, for example, via an AOD 630 driven by a four-tone RF drive signal.

[0072] Figure 6Subgraph c) illustrates an exemplary sequence of pulses for performing a single-qubit gate on all four atoms previously prepared as the qubit ground state |0> (e.g., via a qubit array initialization process) shown in subgraph b). The single-qubit gate operation shown includes a first global qubit state transition laser pulse 650 (e.g., a π / 2 pulse) that transitions the four qubits prepared as state |0> to a superposition state |s>, e.g., |s> = 1 / √2 (0> + |1>). Each atom can then be locally and selectively irradiated by a separate qubit-addressed laser pulse 660, causing a differential AC Stark shift, resulting in a phase difference between |0> and |1>. ,For example, .

[0073] like Figure 6 As shown by numbers 1, 2, 3, and 4, the AC Stark shift pulses 660 can be applied sequentially to each individual qubit, or they can be applied simultaneously to all addressed qubits. The relative phase difference between 0> and |1> is controlled by adjusting the pulse area of ​​the qubit addressing laser pulses 660, for example by controlling the pulse duration, amplitude, or shape, and / or by applying multiple pulses. It can be adjusted to any value between 0° and 180°.

[0074] Next, a second global qubit transition laser pulse 670 is applied, which, for example, uses a second π / 2 pulse with a 180° phase difference relative to the first qubit state transition laser pulse 650, to convert all atoms not irradiated by the local AC Stark shift pulse back to |0>. Since the phase difference between the addressed qubits' |0> and |1> can be controlled between 0° and 180°, such a single-qubit gate pulse sequence enables arbitrary SU(2) rotations to be performed on atoms addressed by the qubit-addressed laser.

[0075] In some implementations, Figure 6 The single-qubit gate sequence shown in sub-figure b) may include additional AC-Stark shift pulses 680 and 690 that may be required by the quantum computing protocol, wherein the addressed qubit is not initially prepared as the qubit ground state |0>, but is prepared as a general superposition of the qubit ground state |0> and the qubit excited state |1>.

[0076] Figure 7This illustration demonstrates how a local two-qubit gate required to generate entanglement (e.g., Bell states, W states, GHZ states, etc.) during quantum computing can be implemented according to the present disclosure. In subfigure a), one-dimensional selective addressing of a single qubit pair by a qubit-addressing laser beam 710 is shown combined with global illumination by a Rydberg transition laser beam 720. As described above, the qubit-addressing laser system may include an AOD 730 and a large NA imaging objective 740 that focuses the output beam of the AOD 730 onto an optical trap array 750 that traps neutral atoms as physical qubits. As mentioned above, the focal position of the qubit-addressing laser beam 710 can be controlled by changing the frequency of the RF signal driving the AOD 730. For example, multiple qubits can be addressed simultaneously by driving the AOD 730 with a multi-tone RF drive signal.

[0077] To execute a local two-qubit gate, in a possible implementation of this disclosure, two adjacent atoms, such as atoms 2 and 3 in sub-figure b), are first converted to a temporary state |c> via a selectively and locally qubit-addressable laser pulse 760. Next, the optical trap array 750 is globally illuminated by a Rydberg laser pulse 770, which interacts only with atoms prepared as either the temporary state |c> or the Rydberg state |r>. Due to Rydberg blocking caused by the dipole interaction between atoms 2 and 3 in the Rydberg state |r> as indicated by circle 755 in sub-figure b), even a global Rydberg laser π pulse will not convert both atoms 2 and 3 from the temporary state |c> to the Rydberg state |r>, but instead generate a two-qubit entangled state, for example, |e> = 1 / √2(|c>|r> + |r>|c>), where only one qubit is in the Rydberg state |r>, and the other remains in the temporary state |c>. Typically, more general superposition states can also be generated when performing two-qubit gates based on Rydberg blocking. As is known in the art, Rydberg rotation 770 can also be accomplished from resonant detuning, for example, to implement Lavine-Pichler gates or time-optimal gates as mentioned above.

[0078] To transition the entanglement between atoms 2 and 3 from states |c> and |r> to qubit states |0> and |1>, a qubit-addressable laser can be used to generate controlled phase shifts for |c> and |r> of the two atoms during a two-qubit gate sequence. For example, as shown in sub-figure c), phase-shift pulse 780 can be interleaved with two global Rydberg laser pulses 770 and 790. To maintain the entanglement of atoms 2 and 3 after the reverse buffer pulse 795, the pulse duration of phase-shift pulse 780 and the detuning of the Rydberg laser can be controlled to achieve a controlled phase gate.

[0079] Figure 7Subgraph d) shows an alternative two-qubit gate scheme in which a phase-shift pulse 780 is applied during a single global Rydberg laser pulse 770 to maintain entanglement after a reverse temporary pulse 790. (See above for...) Figure 1 In other implementations, different techniques can be used, such as... Figure 1 The integrated photonic addressing units 130 and 162 shown realize selective laser position control of the qubit-addressable laser 710.

[0080] Figure 8 This demonstrates how a qubit-addressable laser can also be used for sideband cooling during a qubit readout sequence based on an imaging laser with a design wavelength of 461 nm. 88 Sr 1 S0 to 1 P1 transition is used for fluorescence imaging. During readout, the optical trap array is illuminated by an imaging laser, causing it to be excited to... 1 Atoms in the P1 state can be detected by a CMOS camera via emission (see...). Figure 1 The detected fluorescent photons spontaneously decayed to 1 The S0 state. Due to the cyclic nature of this transition, it can be considered as a state in... 1 Each qubit in the S0 state collects multiple fluorescent photons to detect which qubits are in the |0> state at the end of the ongoing quantum computing sequence. To avoid heating and loss during fluorescence imaging, the optical trap array can be simultaneously illuminated by a qubit-addressable laser at a third laser frequency to perform sideband cooling that cancels out the heating caused by the imaging laser and the recoil from the spontaneous emission of fluorescent photons.

[0081] In this way, qubits are not lost during imaging and can be reused for additional quantum computing sequences. The qubit-addressable laser can also be used to detect which qubits are in the |1> state without losing those in the |0> state. To do this, qubits in the |0> state are converted to a temporary state |c> before being converted to the |0> state and imaging as described above. Therefore, compared to conventional systems, the qubit-addressable laser also enables state-selective and essentially lossless qubit readout, significantly improving cycle time and reducing hardware complexity.

[0082] Figure 9This disclosure illustrates how, according to the present disclosure, the effective (combined) pulse duration of a single π pulse or two consecutive π / 2 pulses of a (global) qubit-state transition laser is selected to minimize gate nonfidelity caused by motion decoupling. As mentioned above, an optical trap array is illuminated with a single qubit-state transition laser pulse or two consecutive qubit-state transition laser pulses having an effective (combined) pulse duration equal to ω / Ω = 2n + 1, n ∈ N, where Ω is the Rabi frequency associated with the qubit-state transition laser and the |0> to |1> transition matrix elements under a given magnetic field. From Figure 9 The solid line graph of the π pulse clearly shows that, during the simulation, Ω is set to 50 kHz. When the sink frequency ω is equal to (2n+1)Ω, for example, for ω = 150 kHz (n=1) and ω = 250 kHz (n=2), the gate nonfidelity exhibits a significant minimum (note that the y-axis has a logarithmic scale). Furthermore, as the sink frequency increases (i.e., a smaller Lamb-Dicke factor), the nonfidelity is generally expected to decrease.

[0083] For such pulses with a specifically chosen duration, a Rabi frequency of at least one-third of the well frequency can be achieved without fidelity loss due to motion dephase. This behavior indicates that motion excitation is minimized after X-gate completion when the qubit state transition laser pulse satisfies the above-mentioned conditions of appropriately choosing the well frequency ω and the Rabi frequency Ω. For example, it can be expected that n = 1 may be the optimal trade-off between the single-qubit gate fidelity and the two-qubit gate fidelity mentioned above, which increases as the well frequency decreases.

[0084] As mentioned above, this disclosure enables the cleanest possible operation of qubit-state transition lasers, thereby promoting low phase noise and increasing frequency and amplitude stability. Therefore, the present invention enables the effective Rabi frequency Ω to be kept as stable as possible, which typically depends on phase noise, detuning relative to the |0> to |1> transition, and the laser pulse area. Thus, this disclosure also increases the probability that a given π pulse or two consecutive π / 2 pulse sequences of a qubit-state transition laser will satisfy the above relationship, and therefore minimizes motion dephase.

[0085] Figure 10A processing diagram of method 1000 for quantum computing is shown. In step 1010, as described above, a local single-qubit gate operation may be performed, including locally and selectively irradiating the qubit prepared as a superposition state with a qubit-addressable laser at a first qubit-addressable laser frequency, and in step 1020, also as described above, a local two-qubit gate operation is performed, including locally and selectively irradiating the qubit pair prepared as the qubit ground state with a qubit-addressable laser at a second qubit-addressable laser frequency.

[0086] Figure 10 The method may also include performing sideband cooling using a qubit-addressable laser during the qubit initialization sequence and / or the qubit readout sequence, including globally irradiating the optical trap array with a qubit-addressable laser at a third qubit-addressable laser frequency, or locally and selectively irradiating the qubits with a qubit-addressable laser at a third qubit-addressable laser frequency.

[0087] Furthermore, in some implementations, the first internal state is the electronic ground state of a neutral atom, and the second internal state is a metastable, long-lived excited state of a neutral atom. Additionally, performing the local single-qubit gate operation 1010 may further include globally irradiating the optical trap array with a first qubit state transition laser pulse to convert a qubit prepared as the qubit ground state |0> to a superposition state |s> of the qubit ground state |0> and the qubit excited state |1>, and globally irradiating the optical trap array with a second qubit state transition laser pulse to convert a qubit prepared as the superposition state |s> to the qubit ground state |0>. Furthermore, performing the local two-qubit gate operation 1020 may further include globally irradiating the optical trap array with a Rydberg laser pulse to convert one or more qubit pairs to a Rydberg state |r> of a neutral atom, or to a superposition of a Rydberg state |r> and a third internal state |c>.

[0088] Figure 10 The method may also include locally and selectively irradiating one or more qubit pairs with a qubit-addressable laser at a fourth qubit-addressable laser frequency or a first qubit-addressable laser frequency to induce a differential AC-Stark shift in the third internal state |c> and the Rydberg state |r> (see details). Figure 7 ), and generate a magnetic field at the location of the optical trap array, which defines the angular momentum quantization axis of the neutral atom and causes perturbation mixing of dipole-allowed states (e.g., |c>) to qubit excited states |1>.

[0089] Additionally, execute the qubit initialization sequence (see...) Figure 11 Step 1110) may include loading a plurality of neutral atoms into an optical trap array and performing sideband cooling to convert the neutral atoms into the motion ground state of the optical trap array.

[0090] Furthermore, the method disclosed herein may also include performing a qubit readout sequence (see step 1140), and may also include performing sideband cooling while globally illuminating the optical trap array with an imaging laser, and imaging the qubits using fluorescence induced by the imaging laser. In some implementations, performing a qubit readout sequence ( Figure 11 Step 1140 may also include, prior to imaging the qubit with an imaging laser, selectively and locally irradiating the qubit with a qubit address laser pulse at the second qubit address laser frequency to convert the qubit to a third internal state |c>.

[0091] Additionally, the optical trap array can include a trap frequency ω that can be substantially the same for all qubits, and performing local single-qubit gate operations can include irradiating the optical trap array with a single qubit state transition laser pulse or two consecutive qubit state transition laser pulses having an effective pulse duration equal to ω / Ω = 2n+1, n∈n, where Ω is the Rabi frequency associated with the qubit state transition laser and the |0> to |1> transition (see details). Figure 9 In some implementations, neutral atoms may include bosons. 88 The method discloses Sr atoms, and the qubit-addressable laser can have a design wavelength of 689.4 nm, the Rydberg laser can have a design wavelength of 318.5 nm, and the qubit state transition laser can have a design wavelength of 698.4 nm. Furthermore, the method disclosed herein may also include shifting the laser frequency of the qubit-addressable laser to a first qubit-addressable laser frequency, shifting the laser frequency of the qubit-addressable laser to a second qubit-addressable laser frequency, and shifting the laser frequency of the qubit-addressable laser to a third qubit-addressable frequency.

[0092] In some implementations, the methods disclosed herein may include performing multiple single-qubit gate operations and two-qubit gate operations, and periodically driving a qubit-addressable laser to generate a first laser pulse train with a constant inter-pulse duration for performing the multiple single-qubit gate operations and two-qubit gate operations. Furthermore, the methods disclosed herein may also include periodically driving a qubit-addressable laser to generate a second laser pulse train with a constant inter-pulse duration, controlling the interval between the first and second laser pulse trains, and synchronizing the periodic driving of the qubit-addressable laser with the characteristic frequency of an external noise source affecting the fidelity of the qubit gates.

[0093] Furthermore, single-qubit gate operation instructions and two-qubit gate operation instructions of quantum computing protocols can be obtained, and the operation of a qubit-addressable laser can be controlled according to the obtained qubit gate operation instructions to locally and selectively irradiate multiple qubits and multiple qubit pairs to perform single-qubit gate and two-qubit gate operations.

[0094] Figure 11 Including: those designated as steps 1120 and 1130 Figure 10 Steps 1010 and 1020; and step 1110, in which, as discussed above, the qubit initialization sequence can be executed using a qubit-addressable laser at a third qubit-addressable laser frequency; and step 1140, in which, also as discussed above, the initialization sequence can be executed by a quantum computing device, such as... Figure 1 Possible implementations include using qubit-addressable lasers to perform qubit readout sequences at a third qubit-addressable laser frequency.

[0095] The foregoing disclosure provides illustrations and descriptions, but is not intended to be exhaustive or to limit these aspects to the exact forms disclosed. Modifications and variations can be made based on the foregoing disclosure, or from practice in these 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 in hardware, firmware, or a combination of hardware and software.

[0096] 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, this document describes the operation and behavior of the systems and / or methods 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 description herein.

[0097] 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 listed in the claims and / or not disclosed in the specification. Although each of the appended dependent claims may directly depend on only one claim, the disclosure of the aspects includes every dependent claim in combination with every other claim in the claim set. The phrase “at least one” in the list of items refers to any combination of these items, including single members. As an example, “at least one of: 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).

[0098] Unless explicitly described as such, no element, action, or instruction used herein should be construed as critical or necessary. Furthermore, as used herein, the articles “a” and “one” are intended to include one or more items and are interchangeable 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 are interchangeable with “one or more.” Where only one item is intended, the phrase “only one” or similar language is used. Furthermore, as used herein, the terms “has,” “have,” “having,” etc., are intended to be open-ended terms.

[0099] As used in this article, the phrase “based on” should not be interpreted as a reference to information, one or more conditions, one or more factors, or a closed set of such information. In other words, unless expressed in a particularly different way, the phrase “based on A” (where “A” can be information, conditions, factors, etc.) should be interpreted as “based on at least A”.

[0100] As used herein, the term "or" is inclusive unless restrictive language is used relative to the listed alternatives. 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, because "or" is inclusive, mentioning "X is based on A or B" means "at least one of A or B" or "one or more of A or B". 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, because "or" is inclusive, 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, and C". 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.

[0101] In addition, such as Figure 10 and Figure 11 The processing diagram does not necessarily indicate a specific order or sequence of steps. For example, without departing from the scope of this disclosure, Figure 10 Steps 1010 and 1120 or Figure 11 Steps 1120 and 1130 can also be executed in a different order, or simultaneously if hardware capabilities allow.

Claims

1. A method (1000) for quantum computing using a plurality of neutral atoms in an array of optical traps (110), wherein, The first internal state of the neutral atom serves as a qubit ground state |0> and the second internal state serves as a qubit excited state |1>, the method comprising: performing a local single-qubit gate operation on the qubit (1010), including: addressing the qubit with a qubit-addressing laser (160) to locally and selectively illuminate the qubit prepared in a superposition state |s> of the qubit ground state |0> and the qubit excited state |1> at a first qubit-addressing laser frequency to cause differential Stark shifts of the qubit ground state |0> and the qubit excited state |1>; and performing a local two-qubit gate operation on the qubit pair (1020), including: addressing the qubit pair with the qubit-addressing laser to locally and selectively illuminate the qubit pair prepared in the qubit ground state |0> at a second qubit-addressing laser frequency for coupling the qubit pair to a Rydberg state |r> of the neutral atom, preferably via a third internal state |c> of the neutral atom serving as an intermediate state for a two-photon transition from the qubit ground state |0> to the Rydberg state |r>.

2. The method of claim 1, further comprising: performing sideband cooling using the qubit-addressing laser during a qubit initialization sequence and / or during a qubit readout sequence, including: globally illuminating the array of optical traps with the qubit-addressing laser at a third qubit-addressing laser frequency, or locally and selectively illuminating a qubit with the qubit-addressing laser at the third qubit-addressing laser frequency.

3. The method of claim 1 or 2, wherein, the first internal state is an electronic ground state of the neutral atom and the second internal state is a metastable, long-lived excited state of the neutral atom.

4. The method of any one of claims 1 to 3, wherein, performing the local single-qubit gate operation further includes: globally illuminating the array of optical traps with a first qubit-state transition laser pulse to convert a qubit prepared in the qubit ground state |0> to a superposition state |s> of the qubit ground state |0> and the qubit excited state |1>; and globally illuminating the array of optical traps with a second qubit-state transition laser pulse to convert a qubit prepared in the superposition state |s> to the qubit ground state |0>; and / or wherein performing the local two-qubit gate operation further includes: globally illuminating the array of optical traps with a Rydberg laser pulse to convert the one or more qubit pairs to the Rydberg state |r> of the neutral atom, or to a superposition of the Rydberg state |r> and the third internal state |c>.

5. The method of claim 4, further comprising: locally and selectively illuminating the one or more qubit pairs with the qubit-addressing laser at a fourth qubit-addressing laser frequency or at the first qubit-addressing laser frequency to cause differential AC Stark shifts of the third internal state |c> and the Rydberg state |r>.

6. The method of any one of claims 1 to 5, further comprising: A magnetic field is generated at the location of the optical trap array, the magnetic field defining an angular momentum quantization axis of the neutral atoms and causing a perturbative mixing of a dipole allowed state with the quantum bit excited state.

7. The method of any one of claims 2-6, further comprising: performing the quantum bit initialization sequence, comprising: loading the plurality of neutral atoms into the optical trap array; performing the sideband cooling to convert the neutral atoms to a motional ground state of the optical trap array; and / or further comprising: performing the quantum bit readout sequence, comprising: performing the sideband cooling while globally illuminating the optical trap array with an imaging laser; and imaging the quantum bits using fluorescence induced by the imaging laser.

8. The method of claim 7, wherein, performing the quantum bit readout sequence further comprises: prior to imaging the quantum bits with the imaging laser, illuminating one or more quantum bits selectively and locally with a quantum bit addressing laser pulse at the second quantum bit addressing laser frequency to convert the one or more quantum bits to the third internal state |c>.

9. The method of any one of claims 1-8, wherein the optical trap array comprises a substantially identical trap frequency ω for all quantum bits, and wherein performing the local single quantum bit gate operation comprises: with single-qubit state transition laser pulses having an effective pulse duration equal to irradiating the array of optical traps with either a single-qubit state transition laser pulse or two consecutive qubit state transition laser pulses, where Ω is the Rabi frequency associated with the qubit state transition laser and the |0> to |1> transition.

10. The method of any one of claims 1-9, wherein The neutral atoms include bosons 88 Sr atoms; wherein the quantum bit addressing laser has a design wavelength of 689.4 nm, the Rydberg laser has a design wavelength of 318.5 nm, and the quantum bit state transition laser has a design wavelength of 698.4 nm; and / or wherein the method further comprises: offsetting a laser frequency of the quantum bit addressing laser to the first quantum bit addressing laser frequency; offsetting a laser frequency of the quantum bit addressing laser to the second quantum bit addressing laser frequency; and offsetting a laser frequency of the quantum bit addressing laser to the third quantum bit addressing laser frequency.

11. The method of any one of claims 1 to 10, further comprising performing a plurality of single-qubit gate operations and two-qubit gate operations; and periodically driving the quantum bit addressing laser to generate a first string of laser pulses having a constant inter-pulse duration for performing the plurality of single quantum bit gate operations and two-qubit gate operations; and / or wherein, the method further comprises: periodically driving the quantum bit addressing laser to generate a second string of laser pulses having a constant inter-pulse duration; and controlling a spacing between the first string of laser pulses and the second string of laser pulses.

12. The method of claim 11, further comprising: synchronizing the periodic driving of the quantum bit addressing laser with a characteristic frequency of an external noise source affecting quantum bit gate fidelity.

13. The method of any one of claims 1-12, further comprising: obtaining single quantum bit gate operation instructions and two-qubit gate operation instructions of a quantum computing protocol; and controlling operation of the qubit-addressed laser to locally and selectively illuminate the plurality of qubits and the plurality of qubit pairs to perform the single-qubit gate operations and the two-qubit gate operations in accordance with the obtained qubit gate operation instructions.

14. A neutral atom quantum computing device (100) comprising means for performing the steps of the method according to any one of claims 1 to 13.

15. A computer program comprising instructions for causing a processor to generate control signals to control a neutral atom quantum computing device (100) to perform the steps of the method of any one of claims 1 to 13.

Images