Search

Abstract Photonics offers a transformative approach to artificial intelligence (AI) and neuromorphic computing by enabling low-latency, high-speed, and energy-efficient computations. However, conventional photonic tensor cores face significant challenges in constructing large-scale photonic neuromorphic networks. Here, we propose a fully integrated photonic tensor core, consisting of only two thin-film lithium niobate (TFLN) modulators, a III-V laser, and a charge-integration photoreceiver. Despite its simple architecture, it is capable of implementing an entire layer of a neural network with a computational speed of 120 GOPS, while also allowing flexible adjustment of the number of inputs (fan-in) and outputs (fan-out). Our tensor core supports rapid in-situ training with a weight update speed of 60 GHz. Furthermore, it successfully classifies (supervised learning) and clusters (unsupervised learning) 112 × 112-pixel images through in-situ training. To enable in-situ training for clustering AI tasks, we offer a solution for performing multiplications between two negative numbers.

Abstract The continuous growth in data volume has sparked interest in silicon-organic-hybrid (SOH) nanophotonic devices integrated into silicon photonic integrated circuits (PICs). SOH devices offer improved speed and energy efficiency compared to silicon photonics devices. However, a comprehensive and accurate modeling methodology of SOH devices, such as modulators corroborating experimental results, is lacking. While some preliminary modeling approaches for SOH devices exist, their reliance on theoretical and numerical methodologies, along with a lack of compatibility with electronic design automation (EDA), hinders their seamless and rapid integration with silicon PICs. Here, we develop a phenomenological, building-block-based SOH PICs simulation methodology that spans from the physics to the system level, offering high accuracy, comprehensiveness, and EDA-style compatibility. Our model is also readily integrable and scalable, lending itself to the design of large-scale silicon PICs. Our proposed modeling methodology is agnostic and compatible with any photonics-electronics co-simulation software. We validate this methodology by comparing the characteristics of experimentally demonstrated SOH microring modulators (MRMs) and Mach Zehnder modulators with those obtained through simulation, demonstrating its ability to model various modulator topologies. We also show our methodology's ease and speed in modeling large-scale systems. As an illustrative example, we use our methodology to design and study a 3-channel SOH MRM-based wavelength-division (de)multiplexer, a widely used component in various applications, including neuromorphic computing, data center interconnects, communications, sensing, and switching networks. Our modeling approach is also compatible with other materials exhibiting the Pockels and Kerr effects. To our knowledge, this represents the first comprehensive physics-to-system-level EDA-compatible simulation methodology for SOH modulators.

Abstract Silicon photonics has developed into a mainstream technology driven by advances in optical communications. The current generation has led to a proliferation of integrated photonic devices from thousands to millions-mainly in the form of communication transceivers for data centers. Products in many exciting applications, such as sensing and computing, are around the corner. What will it take to increase the proliferation of silicon photonics from millions to billions of units shipped? What will the next generation of silicon photonics look like? What are the common threads in the integration and fabrication bottlenecks that silicon photonic applications face, and which emerging technologies can solve them? This perspective article is an attempt to answer such questions. We chart the generational trends in silicon photonics technology, drawing parallels from the generational definitions of CMOS technology. We identify the crucial challenges that must be solved to make giant strides in CMOS-foundry-compatible devices, circuits, integration, and packaging. We identify challenges critical to the next generation of systems and applications—in communication, signal processing, and sensing. By identifying and summarizing such challenges and opportunities, we aim to stimulate further research on devices, circuits, and systems for the silicon photonics ecosystem.

The widespread adaptation of systems relying on optically controlled quantum information will require reliable and efficient multi-channel fiber-to-chip connections that function at cryogenic temperatures. Here we demonstrate low loss (2 dB per channel) connections between a single mode fiber array and tapered silicon waveguides down to 5 K using polymer based photonic wire bonds (PWBs). A method is described for assembling the silicon chip and fiber array such that the PWB connections are robust to temperature cycling and cryostat bakeout. The threshold power handling capability of the PWBs is greater than 4 dBm, sufficient to demonstrate optical bistability in silicon microring resonators coupled to the waveguides at 5 K.

We present a new fab-in-the-loop reinforcement learning algorithm for the design of nano-photonic components that accounts for the imperfections present in nanofabrication processes. As a demonstration of the potential of this technique, we apply it to the design of photonic crystal grating couplers fabricated on an air clad 220 nm silicon on insulator single etch platform. This fab-in-the-loop algorithm improves the insertion loss from 8.8 to 3.24 dB. The widest bandwidth designs produced using our fab-in-the-loop algorithm can cover a 150 nm bandwidth with less than 10.2 dB of loss at their lowest point.

Abstract We propose a photonic processing unit for high-density analog computation using intensity-modulation-based microring modulators (IM-MRMs). The output signal at the fixed resonance wavelength is directly intensity modulated by changing the extinction ratio (ER) of the IM-MRM . Thanks to the intensity-modulated approach, the proposed photonic processing unit is less sensitive to the inter-channel crosstalk. Simulation results reveal that the proposed design offers a maximum of 17-fold increase in wavelength channel density compared to its wavelength-modulated counterpart. Therefore, a photonic tensor core of size 512 $$\times $$ × 512 can be realized by current foundry lines. A convolutional neural network (CNN) simulator with 6-bit precision is built for handwritten digit recognition task using the proposed modulator. Simulation results show an overall accuracy of 96.76%, when the wavelength channel spacing suffers a 3-dB power penalty. To experimentally validate the system, 1000 dot product operations are carried out with a 4-bit signed system on a co-packaged photonic chip, where optical and electrical I/Os are realized using photonic and electrical wire bonding techniques. Study of the measurement results show a mean squared error (MSE) of 3.09 $$\times $$ × 10 $$^{-3}$$ - 3 for dot product calculations. The proposed IM-MRM, therefore, renders the crosstalk issue tractable and provides a solution for the development of large-scale optical information processing systems with multiple wavelengths.

One of the significant challenges in neuromorphic photonic architectures is the lack of good tools to simulate large-scale photonic integrated circuits. It is crucial to perform simulations on a single platform to capture the circuit’s behavior in the presence of both optical and electrical components. Here, we adopted a Verilog-A based approach to model neuromorphic photonic circuits by considering both the electrical and optical properties. Verilog-A models for the primary optical devices, such as lasers, couplers, waveguides, phase shifters, and photodetectors, are discussed, along with studying the composite devices such as microring resonators. Model parameters for different optical devices are extracted and tuned by analyzing the measured data. The simulated and experimental results are also compared for validation of Verilog-A models. Finally, a single photonic neuron circuit is simulated by implementing input, weight, and non-linear activation function by using lasers, microring resonators, and modulator, respectively. Electro-optical rapid co-simulation would significantly improve the efficiency of optimizing the devices and provide an accurate simulation of the circuit performance.

We demonstrate automated adaptation, and stabilization of a silicon photonic wavelength-division multiplexing (WDM), polarization-independent receiver. A two-channel, tunable WDM polarization-independent receiver is designed, and used to demonstrate automated WDM polarization control. Using a control algorithm based on Barzilai, and Borwein's two-point step size gradient descent method, we realize automated polarization adaptation, and wavelength stabilization for two arbitrarily polarized input data streams. 10 Gb/s on-off keying, and 20 Gb/s pulse-amplitude modulation 4-level formats are generated as the high-speed input data streams. In addition, we implement a long-duration experiment, in which we measure the bit-error-ratio for continuously varying polarization states, and changing chip temperatures. The experimental results show that, with the automated control, the WDM polarization-independent receiver can adapt, stabilize, and track the arbitrary input polarization states from a standard optical fiber into the transverse electric mode of a silicon waveguide, and simultaneously stabilize the transmitted wavelength channels at various chip temperatures. We also show how the presented WDM polarization-independent receiver scales with N channels, and propose an improved design for large-scale WDM applications.

Neural networks have enabled applications in artificial intelligence through machine learning, and neuromorphic computing. Software implementations of neural networks on conventional computers that have separate memory and processor (and that operate sequentially) are limited in speed and energy efficiency. Neuromorphic engineering aims to build processors in which hardware mimics neurons and synapses in the brain for distributed and parallel processing. Neuromorphic engineering enabled by photonics (optical physics) can offer sub-nanosecond latencies and high bandwidth with low energies to extend the domain of artificial intelligence and neuromorphic computing applications to machine learning acceleration, nonlinear programming, intelligent signal processing, etc. Photonic neural networks have been demonstrated on integrated platforms and free-space optics depending on the class of applications being targeted. Here, we discuss the prospects and demonstrated applications of these photonic neural networks.

Silicon photonic (SiP) sensors offer a promising platform for robust and low-cost decentralized diagnostics due to their high scalability, low limit of detection, and ability to integrate multiple sensors for multiplexed analyte detection. Their CMOS-compatible fabrication enables chip-scale miniaturization, high scalability, and low-cost mass production. Sensitive, specific detection with silicon photonic sensors is afforded through biofunctionalization of the sensor surface; consequently, this functionalization chemistry is inextricably linked to sensor performance. In this review, we first highlight the biofunctionalization needs for SiP biosensors, including sensitivity, specificity, cost, shelf-stability, and replicability and establish a set of performance criteria. We then benchmark biofunctionalization strategies for SiP biosensors against these criteria, organizing the review around three key aspects: bioreceptor selection, immobilization strategies, and patterning techniques. First, we evaluate bioreceptors, including antibodies, aptamers, nucleic acid probes, molecularly imprinted polymers, peptides, glycans, and lectins. We then compare adsorption, bioaffinity, and covalent chemistries for immobilizing bioreceptors on SiP surfaces. Finally, we compare biopatterning techniques for spatially controlling and multiplexing the biofunctionalization of SiP sensors, including microcontact printing, pin- and pipette-based spotting, microfluidic patterning in channels, inkjet printing, and microfluidic probes.

Silicon photonic (SiP) evanescent-field biosensors aim to combine the information-rich readouts offered by lab-scale diagnostics, at a significantly lower cost, and with the portability and rapid time to result offered by paper-based assays. While SiP biosensors fabricated with conventional strip waveguides can offer good sensitivity for label-free detection in some applications, there is still opportunity for improvement. Efforts have been made to design higher-sensitivity SiP sensors with alternative waveguide geometries, including sub-wavelength gratings (SWGs). However, SWG-based devices are fragile and prone to damage, limiting their suitability for scalable and portable sensing. Here, we investigate SiP microring resonator sensors designed with SWG waveguides that contain a “fishbone” and highlight the improved robustness offered by this design. We present a framework for optimizing fishbone-style SWG waveguide geometries based on numerical simulations, then experimentally measure the performance of ring resonator sensors fabricated with the optimized waveguides, targeting operation in the O-band and C-band. For the O-band and C-band devices, we report bulk sensitivities up to 349 nm/RIU and 438 nm/RIU, respectively, and intrinsic limits of detection as low as 5.1 × 10−4 RIU and 7.1 × 10−4 RIU, respectively. This performance is comparable to the state of the art in SWG-based sensors, positioning fishbone SWG resonators as an attractive, more robust, alternative to conventional SWG designs.

Universal quantum computing holds the promise to fundamentally change today’s information-based society, yet a hardware platform that will provide a clear path to fault-tolerant quantum computing remains elusive. One recently proposed platform involves the use of circuit-bound photons to build cluster states and perform one-way measurement-based quantum computations on arrays of long-coherence-time solid-state spin qubits. Herein, we discuss the challenges that are faced during any practical implementation of this architecture by itemizing the key physical building blocks and the constraints imposed on the spin qubits and the photonic circuit components by the requirements of fault-tolerant performance. These considerations point to silicon as a leading candidate to host such a platform, and a roadmap for developing a silicon photonic circuit-based platform for measurement-based, fault-tolerant universal quantum computing is offered.

Miniaturization of optical components could give way to dense photonic-integrated circuits. Here, the authors demonstrate the control of evanescent waves using all-dielectric metamaterials and show that they can reduce cross-talk and bending loss, which limit the integration density in photonic circuits.

We demonstrate a polarization-independent mode-evolution-based coupler for the silicon-on-insulator platform. The measured coupler has negligible insertion loss over a bandwidth of about 100 nm, i.e., from 1500 to 1600 nm. The measured maximum power imbalances for the polarization-independent coupler are 1.2 and 0.2 dB for the fundamental transverse electric (TE00) mode and the fundamental transverse magnetic (TM00) mode, respectively. Our coupler also has a compact design footprint with mode-evolution region not more than 75-μm long.

Side and top silicon nitride stressors were proposed and shown to be effective in reducing the threshold current Ith and in improving the wall-plug efficiency ηwp of Ge-on-Si lasers. Side stressors only turned out to be a more efficient way to increase ηwp than using the top and side stressors together. With the side stressors and geometry optimizations, a ηwp of 34.8% and an Ith of 36 mA (Jth of 27 kA/cm2) can be achieved with a defect limited carrier lifetime (τp,n) of 1 ns. With τp,n = 10 ns, an I th of 4 mA (Jth of 3 kA/cm2) and a ηwp of 43.8% can be achieved. These are tremendous improvements from the case with no stressors. These results give strong support to the Ge-on-Si laser technology and provide an effective way to improve the Ge laser performance.

Photonic-based instantaneous frequency measurement systems are usually bulky or require high optical power. Here, Burla et al. demonstrate an integrated measurement system that can identify frequency-varying signals in a dynamic fashion, without the need for fast measurement instrumentation.

We experimentally demonstrate compact broadband directional couplers using subwavelength gratings for silicon-on-insulator wafers with silicon layers of 220 nm. The dispersion properties of the optical modes are engineered using subwavelength gratings, which allow broadband operation. Finite-difference time-domain (FDTD)-based band structure calculations, with significantly reduced simulation time, were used to analyze the design, which included both the structure and material dispersions. Compact broadband direction couplers, with device lengths shorter than 14 $\mu\text{m}$, which cover a bandwidth of 100 nm for power splitting ratios of 50/50, 40/60, 30/70, and 20/80, are designed and fabricated for the fundamental transverse electric mode with a central operating wavelength of 1550 nm.

We designed and fabricated a polarization splitter-rotator (PSR) on a silicon-on-insulator (SOI) platform based on an asymmetric directional coupler. The asymmetric directional coupler consists of a regular strip waveguide and a sub-wavelength grating (SWG) waveguide. The SWG waveguide used in the design provides the freedom to engineer the dispersion properties of the supermodes, which allow the design to be fabrication tolerant to the waveguide width variations. A measured peak polarization conversion efficiency of -0.3 dB and crosstalks below -10 dB were achieved. The designed SWG PSR has a compact size of 35 μm x 5μm.

Control of oxygen over cell cultures in vitro is a topic of considerable interest, as chronic and cyclic hypoxia can alter cell behaviour. Both static and transient hypoxic levels have been found to affect tumour cell behaviour; it is potentially valuable to include these effects in early, in vitro stages of drug screening. A barrier to their inclusion is that rates of transient hypoxia can be a few cycles/hour, which is difficult to reproduce in traditional in vitro cell culture environments due to long diffusion distances from control gases to the cells. We use a gas-permeable three-layer microfluidic device to achieve spatial and temporal oxygen control with biologically-relevant switching times. We measure the oxygen profiles with integrated, ratiometric optical oxygen sensors, demonstrate sensor and system stability over multi-day experiments, and characterize a pre-bleaching process to improve sensor stability. We show, with both finite-element modelling and experimental data, excellent control over the oxygen levels by the device, independent of fluid flow rate and oxygenation for the operating flow regime. We measure equilibration times of approximately 10 min, generate complex, time-varying oxygen profiles, and study the effects of oxygenated media flow rates on the measured oxygen levels. This device could form a useful tool for future long-term studies of cell behaviour under hypoxia.

We experimentally demonstrate apodized focusing subwavelength grating couplers for both the fundamental transverse electric (TE00) mode and the fundamental transverse magnetic (TM00) mode. A measured insertion loss of 3.2 dB with a 1-dB bandwidth of 36 nm has been obtained for the TE00 mode, and a measured insertion loss of 3.3 dB with a 1-dB bandwidth of 37 nm has been obtained for the TM00 mode. Back reflections of -24 dB and -21 dB have been obtained for the TE00 and TM00 modes, respectively.

The authors wish to make the following corrections in their published paper in Sensors [...]

Thanks to advanced semiconductor microfabrication technology, chip-scale integration and miniaturization of lab-on-a-chip components, silicon-based optical biosensors have made significant progress for the purpose of point-of-care diagnosis. In this review, we provide an overview of the state-of-the-art in evanescent field biosensing technologies including interferometer, microcavity, photonic crystal, and Bragg grating waveguide-based sensors. Their sensing mechanisms and sensor performances, as well as real biomarkers for label-free detection, are exhibited and compared. We also review the development of chip-level integration for lab-on-a-chip photonic sensing platforms, which consist of the optical sensing device, flow delivery system, optical input and readout equipment. At last, some advanced system-level complementary metal-oxide semiconductor (CMOS) chip packaging examples are presented, indicating the commercialization potential for the low cost, high yield, portable biosensing platform leveraging CMOS processes.

The presence and concentration of oxygen in biological systems has a large impact on the behavior and viability of many types of cells, including the differentiation of stem cells or the growth of tumor cells. As a result, the integration of oxygen sensors within cell culture environments presents a powerful tool for quantifying the effects of oxygen concentrations on cell behavior, cell viability, and drug effectiveness. Because microfluidic cell culture environments are a promising alternative to traditional cell culture platforms, there is recent interest in integrating oxygen-sensing mechanisms with microfluidics for cell culture applications. Optical, luminescence-based oxygen sensors, in particular, show great promise in their ability to be integrated with microfluidics and cell culture systems. These sensors can be highly sensitive and do not consume oxygen or generate toxic byproducts in their sensing process. This paper presents a review of previously proposed optical oxygen sensor types, materials and formats most applicable to microfluidic cell culture, and analyzes their suitability for this and other in vitro applications.