Heterogeneous integration of spin-photon interfaces with a CMOS platform.

Colour centres in diamond have emerged as a leading solid-state platform for advancing quantum technologies, satisfying the DiVincenzo criteria1 and recently achieving quantum advantage in secret key distribution2. Blueprint studies3-5 indicate that general-purpose quantum computing using local quantum communication networks will require millions of physical qubits to encode thousands of logical qubits, presenting an open scalability challenge. Here we introduce a modular quantum system-on-chip (QSoC) architecture that integrates thousands of individually addressable tin-vacancy spin qubits in two-dimensional arrays of quantum microchiplets into an application-specific integrated circuit designed for cryogenic control. We demonstrate crucial fabrication steps and architectural subcomponents, including QSoC transfer by means of a 'lock-and-release' method for large-scale heterogeneous integration, high-throughput spin-qubit calibration and spectral tuning, and efficient spin state preparation and measurement. This QSoC architecture supports full connectivity for quantum memory arrays by spectral tuning across spin-photon frequency channels. Design studies building on these measurements indicate further scaling potential by means of increased qubit density, larger QSoC active regions and optical networking across QSoC modules.

Quantum network utility: A framework for benchmarking quantum networks.

The central aim of quantum networks is to facilitate user connectivity via quantum channels, but there is an open need for benchmarking metrics to compare diverse quantum networks. Here, we propose a general framework for quantifying the performance of a quantum network by estimating the value created by connecting users through quantum channels. In this framework, we define the quantum network utility metric [Formula: see text] to capture the social and economic value of quantum networks. The proposed framework accommodates a variety of applications from secure communications to distributed sensing. As a case study, we investigate the example of distributed quantum computing in detail. We determine the scaling laws of quantum network utility, which suggest that distributed edge quantum computing has more potential for success than its classical equivalent. We believe the proposed utility-based framework will serve as a foundation for guiding and assessing the development of quantum network technologies and designs.

Alignment-free coupling to arrays of diamond microdisk cavities with fabrication tolerant spin-photon interfaces.

We propose a design for an efficient spin-photon interface to a color center in a diamond microdisk. The design consists of a silicon oxynitride triangular lattice overlaid on a diamond microdisk without any aligmnent between the layers. This enables vertical emission from the microdisk into low-numerical aperture modes, with quantum efficiencies as high as 46% for a tin vacancy (SnV) center. Our design is robust to manufacturing errors, potentially enabling large scale fabrication of quantum emitters coupled to optical collection modes. We also introduce a novel approach for optimizing the free space performance of our device using a dipole model, achieving comparable results to full-wave finite difference time domain simulations with 7 · 106 reduction in computational time.

An optoacoustic field-programmable perceptron for recurrent neural networks.

Recurrent neural networks (RNNs) can process contextual information such as time series signals and language. But their tracking of internal states is a limiting factor, motivating research on analog implementations in photonics. While photonic unidirectional feedforward neural networks (NNs) have demonstrated big leaps, bi-directional optical RNNs present a challenge: the need for a short-term memory that (i) programmable and coherently computes optical inputs, (ii) minimizes added noise, and (iii) allows scalability. Here, we experimentally demonstrate an optoacoustic recurrent operator (OREO) which meets (i, ii, iii). OREO contextualizes the information of an optical pulse sequence via acoustic waves. The acoustic waves link different optical pulses, capturing their information and using it to manipulate subsequent operations. OREO's all-optical control on a pulse-by-pulse basis offers simple reconfigurability and is used to implement a recurrent drop-out and pattern recognition of 27 optical pulse patterns. Finally, we introduce OREO as bi-directional perceptron for new classes of optical NNs.

Engineering colloidal semiconductor nanocrystals for quantum information processing.

Quantum information processing-which relies on spin defects or single-photon emission-has shown quantum advantage in proof-of-principle experiments including microscopic imaging of electromagnetic fields, strain and temperature in applications ranging from battery research to neuroscience. However, critical gaps remain on the path to wider applications, including a need for improved functionalization, deterministic placement, size homogeneity and greater programmability of multifunctional properties. Colloidal semiconductor nanocrystals can close these gaps in numerous application areas, following years of rapid advances in synthesis and functionalization. In this Review, we specifically focus on three key topics: optical interfaces to long-lived spin states, deterministic placement and delivery for sensing beyond the standard quantum limit, and extensions to multifunctional colloidal quantum circuits.

Nanoelectromechanical Control of Spin-Photon Interfaces in a Hybrid Quantum System on Chip.

Color centers (CCs) in nanostructured diamond are promising for optically linked quantum technologies. Scaling to useful applications motivates architectures meeting the following criteria: C1 individual optical addressing of spin qubits; C2 frequency tuning of spin-dependent optical transitions; C3 coherent spin control; C4 active photon routing; C5 scalable manufacturability; and C6 low on-chip power dissipation for cryogenic operations. Here, we introduce an architecture that simultaneously achieves C1-C6. We realize piezoelectric strain control of diamond waveguide-coupled tin vacancy centers with ultralow power dissipation necessary. The DC response of our device allows emitter transition tuning by over 20 GHz, combined with low-power AC control. We show acoustic spin resonance of integrated tin vacancy spins and estimate single-phonon coupling rates over 1 kHz in the resolved sideband regime. Combined with high-speed optical routing, our work opens a path to scalable single-qubit control with optically mediated entangling gates.

Synchronous micromechanically resonant programmable photonic circuits.

Programmable photonic integrated circuits (PICs) are emerging as powerful tools for control of light, with applications in quantum information processing, optical range finding, and artificial intelligence. Low-power implementations of these PICs involve micromechanical structures driven capacitively or piezoelectrically but are often limited in modulation bandwidth by mechanical resonances and high operating voltages. Here we introduce a synchronous, micromechanically resonant design architecture for programmable PICs and a proof-of-principle 1×8 photonic switch using piezoelectric optical phase shifters. Our design purposefully exploits high-frequency mechanical resonances and optically broadband components for larger modulation responses on the order of the mechanical quality factor Qm while maintaining fast switching speeds. We experimentally show switching cycles of all 8 channels spaced by approximately 11 ns and operating at 4.6 dB average modulation enhancement. Future advances in micromechanical devices with high Qm, which can exceed 10000, should enable an improved series of low-voltage and high-speed programmable PICs.

Direct imaging of electron density with a scanning transmission electron microscope.

Recent studies of secondary electron (SE) emission in scanning transmission electron microscopes suggest that material's properties such as electrical conductivity, connectivity, and work function can be probed with atomic scale resolution using a technique known as secondary electron e-beam-induced current (SEEBIC). Here, we apply the SEEBIC imaging technique to a stacked 2D heterostructure device to reveal the spatially resolved electron density of an encapsulated WSe2 layer. We find that the double Se lattice site shows higher emission than the W site, which is at odds with first-principles modelling of valence ionization of an isolated WSe2 cluster. These results illustrate that atomic level SEEBIC contrast within a single material is possible and that an enhanced understanding of atomic scale SE emission is required to account for the observed contrast. In turn, this suggests that, in the future, subtle information about interlayer bonding and the effect on electron orbitals could be directly revealed with this technique.

Selective and Scalable Control of Spin Quantum Memories in a Photonic Circuit.

A central goal in many quantum information processing applications is a network of quantum memories that can be entangled with each other while being individually controlled and measured with high fidelity. This goal has motivated the development of programmable photonic integrated circuits (PICs) with integrated spin quantum memories using diamond color center spin-photon interfaces. However, this approach introduces a challenge into the microwave control of individual spins within closely packed registers. Here, we present a quantum memory-integrated photonics platform capable of (i) the integration of multiple diamond color center spins into a cryogenically compatible, high-speed programmable PIC platform, (ii) selective manipulation of individual spin qubits addressed via tunable magnetic field gradients, and (iii) simultaneous control of qubits using numerically optimized microwave pulse shaping. The combination of localized optical control, enabled by the PIC platform, together with selective spin manipulation opens the path to scalable quantum networks on intrachip and interchip platforms.

Large-scale optical characterization of solid-state quantum emitters.

Solid-state quantum emitters have emerged as a leading quantum memory for quantum networking applications. However, standard optical characterization techniques are neither efficient nor repeatable at scale. Here we introduce and demonstrate spectroscopic techniques that enable large-scale, automated characterization of colour centres. We first demonstrate the ability to track colour centres by registering them to a fabricated machine-readable global coordinate system, enabling a systematic comparison of the same colour centre sites over many experiments. We then implement resonant photoluminescence excitation in a widefield cryogenic microscope to parallelize resonant spectroscopy, achieving two orders of magnitude speed-up over confocal microscopy. Finally, we demonstrate automated chip-scale characterization of colour centres and devices at room temperature, imaging thousands of microscope fields of view. These tools will enable the accelerated identification of useful quantum emitters at chip scale, enabling advances in scaling up colour centre platforms for quantum information applications, materials science and device design and characterization.

Transferable learning on analog hardware.

While analog neural network (NN) accelerators promise massive energy and time savings, an important challenge is to make them robust to static fabrication error. Present-day training methods for programmable photonic interferometer circuits, a leading analog NN platform, do not produce networks that perform well in the presence of static hardware errors. Moreover, existing hardware error correction techniques either require individual retraining of every analog NN (which is impractical in an edge setting with millions of devices), place stringent demands on component quality, or introduce hardware overhead. We solve all three problems by introducing one-time error-aware training techniques that produce robust NNs that match the performance of ideal hardware and can be exactly transferred to arbitrary highly faulty photonic NNs with hardware errors up to five times larger than present-day fabrication tolerances.

Single-shot optical neural network.

Analog optical and electronic hardware has emerged as a promising alternative to digital electronics to improve the efficiency of deep neural networks (DNNs). However, previous work has been limited in scalability (input vector length K ≈ 100 elements) or has required nonstandard DNN models and retraining, hindering widespread adoption. Here, we present an analog, CMOS-compatible DNN processor that uses free-space optics to reconfigurably distribute an input vector and optoelectronics for static, updatable weighting and the nonlinearity-with K ≈ 1000 and beyond. We demonstrate single-shot-per-layer classification of the MNIST, Fashion-MNIST, and QuickDraw datasets with standard fully connected DNNs, achieving respective accuracies of 95.6, 83.3, and 79.0% without preprocessing or retraining. We also experimentally determine the fundamental upper bound on throughput (∼0.9 exaMAC/s), set by the maximum optical bandwidth before substantial increase in error. Our combination of wide spectral and spatial bandwidths enables highly efficient computing for next-generation DNNs.

Individually addressable and spectrally programmable artificial atoms in silicon photonics.

A central goal for quantum technologies is to develop platforms for precise and scalable control of individually addressable artificial atoms with efficient optical interfaces. Color centers in silicon, such as the recently-isolated carbon-related G-center, exhibit emission directly into the telecommunications O-band and can leverage the maturity of silicon-on-insulator photonics. We demonstrate the generation, individual addressing, and spectral trimming of G-center artificial atoms in a silicon-on-insulator photonic integrated circuit platform. Focusing on the neutral charge state emission at 1278 nm, we observe waveguide-coupled single photon emission with narrow inhomogeneous distribution with standard deviation of 1.1 nm, excited state lifetime of 8.3 ± 0.7 ns, and no degradation after over a month of operation. In addition, we introduce a technique for optical trimming of spectral transitions up to 300 pm (55 GHz) and local deactivation of single artificial atoms. This non-volatile spectral programming enables alignment of quantum emitters into 25 GHz telecommunication grid channels. Our demonstration opens the path to quantum information processing based on implantable artificial atoms in very large scale integrated photonics.

Highly-twisted states of light from a high quality factor photonic crystal ring.

Twisted light with orbital angular momentum (OAM) has been extensively studied for applications in quantum and classical communications, microscopy, and optical micromanipulation. Ejecting high angular momentum states of a whispering gallery mode (WGM) microresonator through a grating-assisted mechanism provides a scalable, chip-integrated solution for OAM generation. However, demonstrated OAM microresonators have exhibited a much lower quality factor (Q) than conventional WGM resonators (by >100×), and an understanding of the limits on Q has been lacking. This is crucial given the importance of Q in enhancing light-matter interactions. Moreover, though high-OAM states are often desirable, the limits on what is achievable in a microresonator are not well understood. Here, we provide insight on these two questions, through understanding OAM from the perspective of mode coupling in a photonic crystal ring and linking it to coherent backscattering between counter-propagating WGMs. In addition to demonstrating high-Q (105 to 106), a high estimated upper bound on OAM ejection efficiency (up to 90%), and high-OAM number (up to l = 60), our empirical model is supported by experiments and provides a quantitative explanation for the behavior of Q and the upper bound of OAM ejection efficiency with l. The state-of-the-art performance and understanding of microresonator OAM generation opens opportunities for OAM applications using chip-integrated technologies.

Field programmable spin arrays for scalable quantum repeaters.

The large scale control over thousands of quantum emitters desired by quantum network technology is limited by the power consumption and cross-talk inherent in current microwave techniques. Here we propose a quantum repeater architecture based on densely-packed diamond color centers (CCs) in a programmable electrode array, with quantum gates driven by electric or strain fields. This 'field programmable spin array' (FPSA) enables high-speed spin control of individual CCs with low cross-talk and power dissipation. Integrated in a slow-light waveguide for efficient optical coupling, the FPSA serves as a quantum interface for optically-mediated entanglement. We evaluate the performance of the FPSA architecture in comparison to a routing-tree design and show an increased entanglement generation rate scaling into the thousand-qubit regime. Our results enable high fidelity control of dense quantum emitter arrays for scalable networking.

Ultra-low loss quantum photonic circuits integrated with single quantum emitters.

The scaling of many photonic quantum information processing systems is ultimately limited by the flux of quantum light throughout an integrated photonic circuit. Source brightness and waveguide loss set basic limits on the on-chip photon flux. While substantial progress has been made, separately, towards ultra-low loss chip-scale photonic circuits and high brightness single-photon sources, integration of these technologies has remained elusive. Here, we report the integration of a quantum emitter single-photon source with a wafer-scale, ultra-low loss silicon nitride photonic circuit. We demonstrate triggered and pure single-photon emission into a Si3N4 photonic circuit with ≈ 1 dB/m propagation loss at a wavelength of ≈ 930 nm. We also observe resonance fluorescence in the strong drive regime, showing promise towards coherent control of quantum emitters. These results are a step forward towards scaled chip-integrated photonic quantum information systems in which storing, time-demultiplexing or buffering of deterministically generated single-photons is critical.

Photonic Indistinguishability of the Tin-Vacancy Center in Nanostructured Diamond.

Tin-vacancy centers in diamond are promising spin-photon interfaces owing to their high quantum efficiency, large Debye-Waller factor, and compatibility with photonic nanostructuring. Benchmarking their single-photon indistinguishability is a key challenge for future applications. Here, we report the generation of single photons with 99.7_{-2.5}^{+0.3}% purity and 63(9)% indistinguishability from a resonantly excited tin-vacancy center in a single-mode waveguide. We obtain quantum control of the optical transition with 1.71(1)-ns-long π pulses of 77.1(8)% fidelity and show it is spectrally stable over 100 ms. A modest Purcell enhancement factor of 12 would enhance the indistinguishability to 95%.

Asymptotically fault-tolerant programmable photonics.

Component errors limit the scaling of programmable coherent photonic circuits. These errors arise because the standard tunable photonic coupler-the Mach-Zehnder interferometer (MZI)-cannot be perfectly programmed to the cross state. Here, we introduce two modified circuit architectures that overcome this limitation: (1) a 3-splitter MZI mesh for generic errors, and (2) a broadband MZI+Crossing design for correlated errors. Because these designs allow for perfect realization of the cross state, the matrix fidelity no longer degrades with increased mesh size, allowing scaling to arbitrarily large meshes. The proposed architectures support progressive self-configuration, are more compact than previous MZI-doubling schemes, and do not require additional phase shifters. This removes a key limitation to the development of very-large-scale programmable photonic circuits.

Two-layer integrated photonic architectures with multiport photodetectors for high-fidelity and energy-efficient matrix multiplications.

Photonic integrated circuits (PICs) are emerging as a promising tool for accelerating matrix multiplications in deep learning. Previous PIC architectures, primarily focusing on the matrix-vector multiplication (MVM), have large hardware errors that increase with the device scale. In this work, we propose a novel PIC architecture for MVM, which features an intrinsically small hardware error that does not increase with the device scale. Moreover, we further develop this concept and propose a PIC architecture for the general matrix-matrix multiplication (GEMM), which allows the GEMM to be directly performed on a photonic chip with a high energy efficiency unattainable by parallel or sequential MVMs. This work provides a promising approach to realize a high fidelity and high energy efficiency optical computing platform.

Delocalized photonic deep learning on the internet's edge.

Advanced machine learning models are currently impossible to run on edge devices such as smart sensors and unmanned aerial vehicles owing to constraints on power, processing, and memory. We introduce an approach to machine learning inference based on delocalized analog processing across networks. In this approach, named Netcast, cloud-based "smart transceivers" stream weight data to edge devices, enabling ultraefficient photonic inference. We demonstrate image recognition at ultralow optical energy of 40 attojoules per multiply (<1 photon per multiply) at 98.8% (93%) classification accuracy. We reproduce this performance in a Boston-area field trial over 86 kilometers of deployed optical fiber, wavelength multiplexed over 3 terahertz of optical bandwidth. Netcast allows milliwatt-class edge devices with minimal memory and processing to compute at teraFLOPS rates reserved for high-power (>100 watts) cloud computers.