Optoelectronic Computing Research Articles

Electro-optical logics by three-terminal quantum-well-light-emitting transistors integration

The three-terminal quantum-well-light-emitting transistors (QW-LETs) possess appealing characteristics, including multipath bidirectional electrical and optical inputs/outputs, transistor functionalities, and picosecond recombination lifetimes. This article presents the inaugural demonstration of electro-optical sequential logic circuits—first implementations of set-reset (SR) latches. These two latches feature both electrical and optical connectivities through a three-terminal QW-LET platform. Universal gates, such as NOR and NAND logic gates, have been manufactured and showcased, exhibiting accurate logic functionalities. Additionally, we propose two electro-optical SR latches based on NOR and NAND logic gates—the electro-optical registers, serving as fundamental building blocks for more intricate optoelectronic field-programmable gate arrays. The monolithically integrated QW-LET platform holds the potential to offer a comprehensive range of potent building blocks for optoelectronic chips and high-performance optoelectronic computing.

Dimension-Controlled VO2 Film for Optoelectronic Logic Gates and Information Encryption.

As the fields of photonics and information technology develop, a lot of novel applications based on VO2 material, such as optoelectronic computing and information encryption, have been developed. While the performance of these devices was not only closely associated with the VO2 phase transition properties but also depended on their dimensional characteristics. In the current study, we conducted the dimension-controlled vanadium dioxide (VO2) film growth, resulting in the epitaxial 2-dimensional (2D) VO2 film and well-distributed 3-dimensional (3D) VO2 crystal film deposition, respectively. It was revealed that, unlike the 2D film, the pronounced localized surface plasmon resonance dominated the near-infrared spectrum across the phase transition for the 3D VO2 film due to the naturally formed meta-surface structure, which showed a transmittance valley in the infrared spectrum after metallization. Based on this distinct infrared spectrum feature in the 3D VO2 film, we proposed an optoelectronic logic gate controlled by the input voltage and the probing Vis/IR light. By detecting the transmittance states of the probing light with different wavelengths, we achieved multistate encoding functions and demonstrated the information encryption application. This new conception device also showed great potential for some other applications such as optoelectronic coupled computing, information encryption, and optical near-field sensing computing.

Hyperspectral in-memory computing with optical frequency combs and programmable optical memories

The rapid rise of machine learning drives demand for extensive matrix-vector multiplication operations, thereby challenging the capacities of traditional von Neumann computing systems. Researchers explore alternatives, such as in-memory computing architecture, to find energy-efficient solutions. In particular, there is renewed interest in optical computing systems, which could potentially handle matrix-vector multiplication in a more energy-efficient way. Despite promising initial results, developing high-throughput optical computing systems to rival electronic hardware remains a challenge. Here, we propose and demonstrate a hyperspectral in-memory computing architecture, which simultaneously utilizes space and frequency multiplexing, using optical frequency combs and programmable optical memories. Our carefully designed three-dimensional opto-electronic computing system offers remarkable parallelism, programmability, and scalability, overcoming typical limitations of optical computing. We have experimentally demonstrated highly parallel, single-shot multiply-accumulate operations with precision exceeding 4 bits in both matrix-vector and matrix-matrix multiplications, suggesting the system’s potential for a wide variety of deep learning and optimization tasks. Our approach presents a realistic pathway to scale beyond peta operations per second, a major stride towards high-throughput, energy-efficient optical computing.

Broadband Achromatic Imaging of a Metalens with Optoelectronic Computing Fusion.

The remarkable ultrathin ability of metalenses gives them potential as a next-generation imaging candidate. However, the inherent chromatic aberration of metalenses restricts their widespread application. We present an achromatic metalens with optoelectronic computing fusion (OCF) to mitigate the impact of chromatic aberration and simultaneously avoid the significant challenges of nanodesign, nanofabrication, and mass production of metalenses, a method different from previous methods. Leveraging the nonlinear fitting, we demonstrate that OCF can effectively learn the chromatic aberration mapping of metalens and thus restore the chromatic aberration. In terms of the peak signal-to-noise ratio index, there is a maximum improvement of 12 dB, and ∼8 ms is needed to correct the chromatic aberration. Furthermore, the edge extraction of images and super-resolution reconstruction that effectively enhances resolution by a factor of 4 are also demonstrated with OCF. These results offer the possibility of applications of metalenses in mobile cameras, virtual reality, etc.

Generation of orbital angular momentum hologram using a modified U-net

Orbital angular momentum (OAM) holography has become a promising technique in information encryption, data storage and opto-electronic computing, owing to the infinite topological charge of one single OAM mode and the orthogonality of different OAM modes. In this paper, we propose a novel OAM hologram generation method based on a densely connected U-net (DCU), where the densely connected convolution blocks (DCB) replace the convolution blocks of the U-net. Importantly, the reconstruction process of the OAM hologram is integrated into DCU as its output layer, so as to eliminate the requirement to prepare training data for the OAM hologram, which is required by conventional neural networks through an iterative algorithm. The experimental and simulation results show that the OAM hologram can rapidly be generated with the well-trained DCU, and the reconstructed image’s quality from the generated OAM hologram is significantly improved in comparison with those from the Gerchberg–Saxton generation method, the Gerchberg–Saxton based generation method and the U-net method. In addition, a 10-bit OAM multiplexing hologram scheme is numerically demonstrated to have a high capacity with OAM hologram.

All-analog photoelectronic chip for high-speed vision tasks

Photonic computing enables faster and more energy-efficient processing of vision data1–5. However, experimental superiority of deployable systems remains a challenge because of complicated optical nonlinearities, considerable power consumption of analog-to-digital converters (ADCs) for downstream digital processing and vulnerability to noises and system errors1,6–8. Here we propose an all-analog chip combining electronic and light computing (ACCEL). It has a systemic energy efficiency of 74.8 peta-operations per second per watt and a computing speed of 4.6 peta-operations per second (more than 99% implemented by optics), corresponding to more than three and one order of magnitude higher than state-of-the-art computing processors, respectively. After applying diffractive optical computing as an optical encoder for feature extraction, the light-induced photocurrents are directly used for further calculation in an integrated analog computing chip without the requirement of analog-to-digital converters, leading to a low computing latency of 72 ns for each frame. With joint optimizations of optoelectronic computing and adaptive training, ACCEL achieves competitive classification accuracies of 85.5%, 82.0% and 92.6%, respectively, for Fashion-MNIST, 3-class ImageNet classification and time-lapse video recognition task experimentally, while showing superior system robustness in low-light conditions (0.14 fJ μm−2 each frame). ACCEL can be used across a broad range of applications such as wearable devices, autonomous driving and industrial inspections.

Microcomb-Driven Optical Convolution for Car Plate Recognition

The great success of artificial intelligence (AI) calls for higher-performance computing accelerators, and optical neural networks (ONNs) with the advantages of high speed and low power consumption have become competitive candidates. However, most of the reported ONN architectures have demonstrated simple MNIST handwritten digit classification tasks due to relatively low precision. A microring resonator (MRR) weight bank can achieve a high-precision weight matrix and can increase computing density with the assistance of wavelength division multiplexing (WDM) technology offered by dissipative Kerr soliton (DKS) microcomb sources. Here, we implement a car plate recognition task based on an optical convolutional neural network (CNN). An integrated DKS microcomb was used to drive an MRR weight-bank-based photonic processor, and the computing precision of one optical convolution operation could reach 7 bits. The first convolutional layer was realized in the optical domain, and the remaining layers were performed in the electrical domain. Totally, the optoelectronic computing system (OCS) could achieve a comparable performance with a 64-bit digital computer for character classification. The error distribution obtained from the experiment was used to emulate the optical convolution operation of other layers. The probabilities of the softmax layer were slightly degraded, and the robustness of the CNN was reduced, but the recognition results were still acceptable. This work explores an MRR weight-bank-based OCS driven by a soliton microcomb to realize a real-life neural network task for the first time and provides a promising computational acceleration scheme for complex AI tasks.

Hardware Error Correction for MZI-Based Matrix Computation.

With the rapid development of artificial intelligence, the electronic system has fallen short of providing the needed computation speed. It is believed that silicon-based optoelectronic computation may be a solution, where Mach-Zehnder interferometer (MZI)-based matrix computation is the key due to its advantages of simple implementation and easy integration on a silicon wafer, but one of the concerns is the precision of the MZI method in the actual computation. This paper will identify the main hardware error sources of MZI-based matrix computation, summarize the available hardware error correction methods from the perspective of the entire MZI meshes and a single MZI device, and propose a new architecture that will largely improve the precision of MZI-based matrix computation without increasing the size of the MZI's mesh, which may lead to a fast and accurate optoelectronic computing system.

Super-resolution orbital angular momentum holography

Computer-generated holograms are crucial for a wide range of applications such as 3D displays, information encryption, data storage, and opto-electronic computing. Orbital angular momentum (OAM), as a new degree of freedom with infinite orthogonal states, has been employed to expand the hologram bandwidth. However, in order to reduce strong multiplexing crosstalk, OAM holography suffers from a fundamental sampling criterion that the image sampling distance should be no less than the diameter of largest addressable OAM mode, which severely hinders the increase in resolution and capacity. Here we establish a comprehensive model on multiplexing crosstalk in OAM holography, propose a pseudo incoherent approach that is almost crosstalk-free, and demonstrate an analogous coherent solution by temporal multiplexing, which dramatically eliminates the crosstalk and largely relaxes the constraint upon sampling condition of OAM holography, exhibiting a remarkable resolution enhancement by several times, far beyond the conventional resolution limit of OAM holography, as well as a large scaling of OAM multiplexing capacity at fixed resolution. Our method enables OAM-multiplexed holographic reconstruction with high quality, high resolution, and high capacity, offering an efficient and practical route towards the future high-performance holographic systems.

Efficient light transfer in coupled nonlinear triple waveguides using shortcuts to adiabaticity

We use the method of shortcuts to adiabaticity to design the variable couplings in a three-waveguide directional coupler which may contain nonlinear elements, in order to accomplish efficient light transfer between the outer waveguides for shorter device lengths, despite the presence of nonlinearity. The shortcut couplings are obtained for the ideal case where all the waveguides are linear, for which a perfect transfer is guaranteed in theory, but are tested for various combinations of linear and nonlinear waveguides in the device. We show with numerical simulations that, in most configurations, high levels of transfer efficiency can be maintained even for large values of the input power, and for shorter lengths than those of conventional adiabatic devices. We also find that efficiency is improved for shortcut couplings with less spatial extent, since in this case the nonlinearity acts during a shorter range. The present work is expected to find application in research fields like optoelectronic computing and ultrafast light switching, where the fast and controlled light transmission inside a set of waveguides is a crucial task. Additionally, the reduction in the device size may be exploited for incorporating them in integrated optical systems, where a high density of waveguides is required.

Photonic reinforcement learning based on optoelectronic reservoir computing

Reinforcement learning has been intensively investigated and developed in artificial intelligence in the absence of training data, such as autonomous driving vehicles, robot control, internet advertising, and elastic optical networks. However, the computational cost of reinforcement learning with deep neural networks is extremely high and reducing the learning cost is a challenging issue. We propose a photonic on-line implementation of reinforcement learning using optoelectronic delay-based reservoir computing, both experimentally and numerically. In the proposed scheme, we accelerate reinforcement learning at a rate of several megahertz because there is no required learning process for the internal connection weights in reservoir computing. We perform two benchmark tasks, CartPole-v0 and MountanCar-v0 tasks, to evaluate the proposed scheme. Our results represent the first hardware implementation of reinforcement learning based on photonic reservoir computing and pave the way for fast and efficient reinforcement learning as a novel photonic accelerator.

Antiferromagnetic spatial photonic Ising machine through optoelectronic correlation computing

Recently, spatial photonic Ising machines (SPIM) have been demonstrated to compute the minima of Hamiltonians for large-scale spin systems. Here we propose to implement an antiferromagnetic model through optoelectronic correlation computing with SPIM. Also we exploit the gauge transformation which enables encoding the spins and the interaction strengths in a single phase-only spatial light modulator. With a simple setup, we experimentally show the ground-state-search acceleration of an antiferromagnetic model with 40000 spins in number-partitioning problem. Thus such an optoelectronic computing exhibits great programmability and scalability for the practical applications of studying statistical systems and combinatorial optimization problems.

All-Optical Naked-Eye Ghost Imaging

Ghost imaging is usually based on the optoelectronic process and electronic computing. A new ghost imaging approach is put forward in the paper that avoids any optoelectronic or electronic process. Instead, the proposed scheme exploits all-optical correlation and the vision persistence effect to generate images observed by naked eyes. To realize high contrast naked-eye ghost imaging, a special pattern-scanning architecture on a low-speed light-modulation disk is designed, which also enables high-resolution imaging with lower-order Hadamard vectors and boosts the imaging speed. With this approach, we realize high-contrast real-time naked-eye ghost imaging for moving colored objects.

Optoelectronic Logic Gates Based on Photovoltaic Response of Bacteriorhodopsin Polymer Composite Thin Films

We present designs of optoelectronic OR, AND, NOR, and NAND logic gates with multiple pulsed pump laser beams based on the photovoltaic response of bacteriorhodopsin (BR) molecules embedded in a polyvinyl matrix coated on ITO. A detailed experimental study of the photovoltaic response reveals that continuous pulsed exposure to 532 nm and 405 nm laser light results in a large photocurrent/photovoltage, due to rapid reprotonation and chromophore reisomerization, taking BR to the ground state in hundreds of nanoseconds. It also helps in sustaining the photovoltage at higher frequencies and in maintaining the shape of the photovoltage. It is shown experimentally that for a pulsed laser beam at 532 nm with peak pump intensity of 1.19 W/cm (2), a photovoltage of 50 mV is generated. A detailed numerical simulation of the photovoltaic response of BR has been carried out taking into account all the six states (B, K, L, M, N, and O) in the BR photocycle to ascertain the effect of various parameters such as lifetime of the M-state, the pump pulse-width, pump intensity, lifetime of excited protons, and rate constant of excited protons. Experimental results are in good agreement with theoretical simulations. The present study opens up new prospects for protein-based optoelectronic computing.

Code conversion from signed-digit to complement representation based on look-ahead optical logic operations

We present, for the first time to our knowledge, a generalized lookahead logic algorithm for number conversion from signed-digit to complement representation. By properly encoding the signed-digits, all the operations are performed by binary logic, and unified logical expressions can be obtained for conversion from modified-signed-digit (MSD) to 2's complement, trinary signed-digit (TSD) to 3's complement, and quarternary signed-digit (QSD) to 4's complement. For optical implementation, a parallel logical array module using an electron-trapping device is employed and experimental results are shown. This optical module is suitable for implementing complex logic functions in the form of the sum of the product. The algorithm and architecture are compatible with a general-purpose optoelectronic computing system. (C) 2001 Society of Photo-Optical Instrumentation Engineers.

Trinary signed-digit arithmetic using an efficient encoding scheme

The trinary signed-digit (TSD) number system is of interest for ultrafast optoelectronic computing systems since it permits parallel carry-free addition and borrow-free subtraction of two arbitrary length numbers in constant time. In this paper, a simple coding scheme is proposed to encode the decimal number directly into the TSD form. The coding scheme enables one to perform parallel one-step TSD arithmetic operation. The proposed coding scheme uses only a 5-combination coding table instead of the 625-combination table reported recently for recoded TSD arithmetic technique.

Free-space interconnects for high-performance optoelectronic switching

The authors describe the two architectures they have developed that use and demonstrate free-space optical interconnects for digital logic: a high-performance optoelectronic computing module and a second-generation digital optoelectronic computer. The low-power, high-performance optoelectronic computing (HPOC) module was designed for switching and data processing applications. Its architecture uses global, free-space, smart optical interconnects rather than electronic interconnects. The HPOC modules, which incorporate arrays of microlasers, diffractive optical interconnect elements and detectors, are currently undergoing prototype fabrication. By extending the high fan-in and fan-out capabilities of free-space optical interconnects, HPOC modules-when properly configured-offer a significant reduction in power, while maintaining good algorithmic efficiency and high noise margin.

Development of free-space digital optics in computing

To someone with an electronics or computer science background, many of optical computing's concepts may seem outlandish. Optics grew out of applied physics and still retains many aspects of that heritage. This is in contrast to digital computing's roots in electronics. Recent efforts have been made to bring optical computing more in line with microelectronic engineering. Perhaps that will speed the acceptance of opto-electronic technology. The authors describe their research into optical devices for data communication. They are investigating free-space optics, the propagation of optical signals through the air using lenses and mirrors to focus and redirect the beams. The advantages of free-space optics, derive from their large spatial bandwidth and physical channel density. Like the human eye, which takes in an enormous amount of information in parallel, a low-cost lens can provide more than a million independent connections. The authors aim to exploit optoelectronic computing's capability for such massively parallel data transfers.

Tolerance analysis for three‐dimensional optoelectronic systems packaging

A series of optoelectronic processing systems is under development at the Optoelectronic Computing System Center (OCSC) at the University of Colorado. The demonstrations consist of two facing optoelectronic modules that communicate with each other using bidirectional free-space optical channels. An analysis of the fabrication tolerances required for correct operation of these systems is presented. An overview of the system and the techniques used for fabrication is given, along with the tolerances achieved in practice. A comparison with theoretical results indicates the critical alignments within the system. Methods to obtain the alignments required for larger systems are discussed.

Ferroelectric Liquid Crystals for Nonlinear Optics: Orientation of the Disperse Red 1 Chromophore along the Ferroelectric Liquid Crystal Polar Axis

ADVERTISEMENT RETURN TO ISSUEPREVCommunicationNEXTFerroelectric Liquid Crystals for Nonlinear Optics: Orientation of the Disperse Red 1 Chromophore along the Ferroelectric Liquid Crystal Polar Axis†David M. Walba, Daniel J. Dyer, Teresa Sierra, Peter L. Cobben, Renfan Shao, and Noel A. ClarkView Author Information Department of Chemistry and Biochemistry and Optoelectronic Computing Systems Center Campus Box 215, University of Colorado Boulder, Colorado 80309-0215 Cite this: J. Am. Chem. Soc. 1996, 118, 5, 1211–1212Publication Date (Web):February 7, 1996Publication History Received19 July 1995Published online7 February 1996Published inissue 1 January 1996https://doi.org/10.1021/ja952387pCopyright © 1996 American Chemical SocietyRIGHTS & PERMISSIONSArticle Views538Altmetric-Citations38LEARN ABOUT THESE METRICSArticle Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated. Share Add toView InAdd Full Text with ReferenceAdd Description ExportRISCitationCitation and abstractCitation and referencesMore Options Share onFacebookTwitterWechatLinked InReddit Read OnlinePDF (157 KB) Get e-AlertsSupporting Info (2)»Supporting Information Supporting Information SUBJECTS:Absorption,Chromophores,Liquid crystals,Mixtures,Polarization Get e-Alerts