In a new paper published in Light: Science & Applications, a team of scientists, led by Professor Yidong Huang from the Department of Electronic Engineering in Tsinghua University and their collaborators from Peking University, Berxel Photonics Company Ltd. and Shenzhen Technology University have proposed the SUANPAN architecture for optical inner product instead of optical matrix operations. Just like the transistors in an integrated circuit, the independent basic computing unit in such scheme contains only one emitter-detector pair and could be scaled up to form a photonic computing chip. The elemental values of two vectors are encoded on the output intensity of the light-emitters and the photoresponsivity of the photodetectors (PDs) by a brand-new Bit Encoding and Analog Detecting method without requiring large-scale ADC or DAC arrays. The photocurrent of the PD would be proportional to the multiplication of the light intensity and photoresponsivity, and the final result of the inner product can be obtained by the summation of all the photocurrents. Since there is no interaction among the propagating light beams of all emitter-detector pairs and only the output currents of all PDs are connected, such scheme is scalable by increasing the number of emitter-detector pairs with no additional loss or error as well as flexibly reconfigurable and programmable for different computational tasks.

As a proof of principle, the SUANPAN architecture is implemented by utilizing an 8×8 vertical cavity surface emission laser (VCSEL) array and an 8×8 MoTe2 two-dimensional (2D) material PD array. In experiment, the calculation fidelity of random vector inner product can be as high as >98% for various bit precisions (2-bit, 4-bit and 8-bit), and >95% for various vector dimensionalities (@4-bit precision). Furthermore, such implementation has been successfully reconfigured to perform two typical AI tasks, Ising machine and artificial neural network (ANN). A randomly generated 1024-dimensional Ising problem is successfully solved, which is the highest dimensionality of optical Ising machine with heuristic algorithm. Meanwhile, a competitive classification accuracy of 88% is achieved for ANN on MNIST handwritten digit dataset. It is believed that photonic SUANPAN is capable to serve as a fundamental linear vector machine and is potential to enhance the computing power for future various AI applications.

These scientists summarize the operational principles and advantages of SUANPAN:

“It breaks through the traditional mindset of obtaining optical matrix transformations through interaction of light beams. Instead, there is no interaction among those propagating light beams of all emitter-detector pairs. Therefore, the SUANPAN can be decomposed into emitter-detector pairs as independent computing units. The scalability, reconfigurability and programmability of the SUANPAN architecture are only based on the multiplication, recombination and modulation of emitter-detector pairs without any additional cost. Compared with optical matrix transformations through interaction between light beams, the SUANPAN possesses following advantages: (1) With massive and industrial multiplication of emitter-detector pairs, the SUANPAN can theoretically be infinitely scalable. (2) The SUANPAN can be flexibly reconfigured and programmed to perform various specific computing tasks. (3) Only correcting the intensity of light beam is required, and there is no requirement to correct the phase term. (4) Even if one emitter-detector pair is broken during fabrication or operation, other emitter-detector pairs would not be affected, and only the operating dimensionality would be decreased.”

“SUANPAN provides a promising solution for optoelectronic analog-digital hybrid computing. Large-scale DAC and ADC arrays are usually required in optoelectronic computing. However, with Bit Encoding and Analog Detecting paradigm, M-bit digital electronic signal is converted to analog with in a set of M emitter-detector pairs, while each emitter-detector pair only represents 1-bit information. Thus, no DAC is required. At the same time, only one ADC is required to convert the total photocurrent into electronic digital signal. Therefore, the Bit Encoding and Analog Detecting computing paradigm greatly reduces the heavy burden introduced by ADC and DAC. Actually, it is also an important issue for the scalability of the SUANPAN architecture.”

                        Journal

                        Light Science & Applications

                        DOI

                        10.1038/s41377-025-02059-7 
                        Article Title

                        SUANPAN: scalable photonic linear vector machine
            Media Contact
                                WEI ZHAO
                Light Publishing Center, Changchun Institute of Optics, Fine Mechanics And Physics, CAS
            zhaowei@lightpublishing.cn
]]>
					
		
		
		136986	
		
		
		https://scienmag.com/ustc-unveils-self-locking-broadband-raman-electro-optic-microcomb/
		
		
		Tue, 23 Sep 2025 14:15:09 +0000
				
		
		
		
		
		
		
		
		
		
		
		
		
		https://scienmag.com/ustc-unveils-self-locking-broadband-raman-electro-optic-microcomb/

					
										In a remarkable advancement poised to transform the landscape of photonic technology, a research team led by Professor Dong Chunhua from the University of Science and Technology of China (USTC), in collaboration with Professor Bo Fang’s group at Nankai University, has unveiled a breakthrough in the development of integrated microcombs. Their pioneering work, published recently in the renowned journal Nature Communications, details a self-locked Raman-electro-optic (REO) microcomb fabricated entirely on a single lithium niobate chip. This cutting-edge device harnesses the intricate interplay of electro-optic (EO), Kerr, and Raman nonlinear optical effects within a solitary microresonator, achieving unprecedented performance metrics without dependence on external electronic feedback systems.

Microcombs, versatile light sources generating a series of discrete, evenly spaced frequency lines, have revolutionized fields ranging from precision metrology and telecommunications to quantum computing and spectroscopy. Traditional approaches to microcomb generation often demand complex auxiliary electronic feedback to stabilize their output, thereby increasing system complexity, size, and power consumption. The novel REO microcomb presented by this collaborative effort transcends these limitations by self-locking its frequency comb output intrinsically through the dynamics of combined nonlinear processes native to the lithium niobate platform.
At the heart of this innovation is the lithium niobate chip, a material long celebrated for its exceptional electro-optic properties. By ingeniously exploiting the synergistic effects of the electro-optic effect, Kerr nonlinearity, and Raman gain within a single microresonator, the researchers achieved an extraordinarily wide spectral coverage, exceeding 300 nm, all while maintaining a stable repetition rate of 26.03 GHz. This bandwidth surpasses many traditional microcomb devices and opens new vistas for dense wavelength division multiplexing in optical communications, high-resolution spectroscopy, and ultrafast optical signal processing.
The electro-optic effect intrinsic to lithium niobate allows rapid, voltage-controlled modulation of the refractive index, enabling fine tuning of the optical modes within the microresonator. Meanwhile, the Kerr effect—a nonlinear optical phenomenon where intense light induces an intensity-dependent refractive index shift—facilitates the generation of new frequency components, effectively broadening the comb spectrum. The inclusion of stimulated Raman scattering, a nonlinear process whereby light interacts with vibrational modes of the medium to produce frequency-shifted photons, complements these mechanisms by providing energy transfer pathways that reinforce the comb stability and spectral extension.
What sets this REO microcomb apart is its self-locking behavior. Conventional microcomb systems require external feedback loops, involving sophisticated electronic circuitry to lock the frequency comb’s repetition rate and phase coherence. Such complexity not only limits integration and scalability but also imposes constraints on the operational stability under varying environmental conditions. The intrinsic self-locking enabled by the interplay of EO, Kerr, and Raman effects bypasses these obstacles, yielding a fully integrated, compact photonic chip solution with robust and repeatable performance.
The fabrication of the microresonator on lithium niobate represents a significant engineering feat. Lithium niobate’s excellent optical transparency and strong nonlinearities make it ideal for integrated photonic applications but pose challenges for microfabrication due to its chemical and mechanical properties. The team overcame these hurdles using advanced lithography and etching techniques to fabricate high-quality, low-loss microresonators with precise control over their geometry, which is critical for achieving the desired resonance conditions and phase matching necessary for multi-effect nonlinear interactions.
Experimentally, the REO microcomb was pumped using a continuous-wave laser source coupled into the lithium niobate microresonator. The interplay of the electro-optic modulation, Kerr nonlinearity, and Raman scattering within the resonator not only generated a broad comb spectrum but also stabilized it through a feedback mechanism embedded in the device physics. Optical measurements confirmed a repetition rate of 26.03 GHz with a spectral bandwidth stretching beyond 300 nm, parameters that underscore the device’s suitability for high-speed optical communication systems and precision timekeeping.
Beyond its immediate capabilities, the REO microcomb platform presents several compelling prospects for future applications. Its integration on a chip scale paves the way for mass-manufacturable photonic devices tailored for next-generation optical networks, frequency synthesis, and even on-chip quantum entanglement sources. The self-locking characteristic enhances robustness against environmental perturbations, reducing the need for bulky stabilization hardware and enabling deployment in compact, portable setups.
Moreover, the researchers’ success showcases the potential of lithium niobate as a powerhouse material for nonlinear optics in integrated photonics. Recent advances in thin-film lithium niobate technology have unlocked the ability to engineer complex photonic circuits with low insertion loss and high electro-optic efficiency, catalyzing a new wave of devices—from modulators to frequency combs—that leverage multifaceted nonlinear effects. The REO microcomb is a prime example, tying together multiple nonlinear phenomena in a seamless and scalable fashion.
The implications of integrating Raman processes into microcomb generation are particularly exciting. Raman gain can help suppress noise and boost the power of certain frequency lines, thereby improving the overall signal-to-noise ratio of the microcomb output. Additionally, Raman nonlinearity extends the comb’s spectral reach into wavelength regions that might otherwise be inaccessible solely through Kerr-based comb generation, providing greater versatility for multiplexed optical functions.
This research underscores a broader trend in the photonics community: leveraging material properties and nonlinear physics not just to create new device functionalities but to streamline photonic circuits toward compactness, stability, and multifunctionality. The REO microcomb encapsulates this philosophy by merging multiple nonlinear effects within a monolithic microresonator, turning what were once discrete, external control functions into inherent properties of the device itself.
The study’s publication in Nature Communications marks a significant milestone, drawing attention from the global scientific and engineering communities focused on cutting-edge integrated photonics technology. As optical systems demand increasing speed, bandwidth, and integration, innovations like the self-locked REO microcomb on lithium niobate chips provide promising avenues toward next-generation optical architectures that are scalable, efficient, and robust.
In conclusion, the collaborative work by Professor Dong Chunhua’s and Professor Bo Fang’s groups exemplifies the power of interdisciplinary innovation combining photonic materials science, nonlinear optics, and microfabrication. The resulting self-locked Raman-electro-optic microcomb extends the frontier of microcomb technology through a unique amalgamation of nonlinear effects within a single chip-scale device, opening exciting possibilities for ultra-broadband, high-speed photonics applications without the baggage of cumbersome external stabilization systems.
As this technology matures, it is expected to significantly impact fields spanning from optical frequency metrology and coherent communications to quantum information processing, further propelling the miniaturization and integration of complex photonic systems. The lithium niobate-based REO microcomb stands as a beacon of future photonics—where materials, physics, and device engineering converge to redefine the limits of light manipulation on a chip.

Subject of Research: Integrated photonics; microcombs; nonlinear optics; lithium niobate microresonators
Article Title: (Not provided)
News Publication Date: (Not provided)
Web References: (Not provided)
References: (Not provided)
Image Credits: University of Science and Technology of China (USTC)
Keywords
lithium niobate, microcomb, electro-optic effect, Kerr nonlinearity, Raman scattering, integrated photonics, microresonator, self-locked frequency comb, optical communications, nonlinear optics, photonic chip, spectral broadening
]]>
					
		
		
		80977	
		
		
		https://scienmag.com/nonlinear-optoelectronic-engine-powers-integrated-photonic-computing/
		
		
		Thu, 04 Sep 2025 08:49:37 +0000
				
		
		
		
		
		
		
		
		
		
		
		
		
		https://scienmag.com/nonlinear-optoelectronic-engine-powers-integrated-photonic-computing/

					
										In a groundbreaking advancement poised to revolutionize the landscape of photonic computing, researchers Zhu S. and Zhu N.H. have unveiled a nonlinear optoelectronic engine capable of driving monolithic integrated photonic computation. This innovation, detailed in the recent publication in Light: Science & Applications, ushers in a new era where optical and electronic components seamlessly interplay within a unified chip architecture, overcoming longstanding challenges in scalability, speed, and energy efficiency in optical computing technologies.

The crux of this pioneering work lies in the development of a nonlinear optoelectronic engine that synergistically combines nonlinear optical processes with electronic control to perform complex computational tasks entirely on a monolithically integrated photonic platform. Traditionally, photonic computing systems have faced significant hurdles due to the difficulty of integrating nonlinearity, a critical element for computation, directly on-chip without bulky discrete components. The authors’ innovative approach deftly navigates these constraints, enabling nonlinear functionalities intrinsic to the device’s architecture.
At the heart of this technological advance is the exploitation of inherent nonlinear optical responses within integrated photonic materials, modulated and enhanced through electronic circuitry embedded onto the same chip. This hybrid platform leverages both light’s ultrafast signaling capabilities and electronics’ precise control to realize computational operators traditionally reserved for electronic processors but now embedded within optical circuits. By interlacing these two domains, the system surpasses prior speed and power consumption limitations, a milestone that bears significant implications for the future of computation.
One of the most compelling features of this nonlinear optoelectronic engine is its monolithic integration, a design philosophy that consolidates all functional elements into a single photonic chip. This integration eliminates parasitic losses and delays caused by inter-device coupling, leading to minimized latency and maximized energy efficiency. The monolithic approach also paves the way for mass producibility using established semiconductor fabrication techniques, thereby promising scalable manufacturing of high-performance computing photonic chips.
The researchers meticulously demonstrated that this optoelectronic engine supports a variety of nonlinear operations central to computational tasks, including intensity-dependent modulation, all-optical switching, and pattern recognition. These operations are executed at speeds unattainable by traditional electronic processors, facilitated by the ultra-high bandwidth intrinsic to photonic components. The nonlinearities embedded within the device enable complex interactions between light waves, essential for advanced computational algorithms like neuromorphic processing and machine learning.
Crucially, this work addresses one of the major bottlenecks in photonic computing: the efficient generation and control of nonlinearity on-chip. Prior efforts often resorted to external nonlinear elements or inefficient materials, resulting in prohibitive power consumption and integration complexity. Zhu and Zhu’s approach circumvents these difficulties by engineering the device’s material properties and electronic control circuits to amplify nonlinear effects without compromising signal fidelity or chip-scale integration.
The implications for artificial intelligence and edge computing are profound. As the demand for instantaneous data processing accelerates, especially in applications involving vast sensor arrays and real-time analytics, the need for low-latency, energy-efficient computing rises accordingly. The nonlinear optoelectronic engine represents a leap forward in meeting these demands by delivering computation speeds orders of magnitude higher than conventional electronics while drastically reducing power footprints. This makes the technology particularly well-suited for deployment in compact, mobile, or remote devices where energy constraints dictate operational viability.
Delving deeper, the researchers showcased the engine’s versatility by implementing a suite of benchmark computational tasks encompassing matrix multiplications, nonlinear activation functions, and even decision-making operations intrinsic to neural networks. Each of these functions was executed within the photonic domain, underpinned by the nonlinear mechanisms fostered by the integrated design. This not only validates the engine’s computational fidelity but also highlights how complex algorithmic operations can be transposed from electronic to photonic frameworks.
Moreover, the study highlights the seamless interface between the nonlinear photonic components and their electronic counterparts, orchestrated to perform dynamic feedback control that fine-tunes system performance in real-time. This coalescence of optics and electronics within a monolithic platform offers unprecedented levels of adaptability and precision, allowing for error correction, signal regeneration, and state reconfiguration through electronic tuning, which is essential for robust and reliable computing systems.
The fabrication techniques employed to realize the nonlinear optoelectronic engine are rooted in mature semiconductor processing technologies, ensuring compatibility with existing foundry infrastructures. This facet is critical for transitioning the technology from laboratory demonstrators to commercially viable products at scale, facilitating rapid adoption across various sectors. By leveraging well-understood lithographic and doping processes, the researchers ensured that the nonlinear elements could be reliably produced with high yield and uniformity.
From a physical standpoint, the nonlinear interactions capitalize on resonant photonic structures embedded within the chip, such as micro-ring resonators and waveguide couplers, which enhance light-matter interactions. These structures are carefully engineered to increase the effective nonlinear coefficients and maintain low propagation losses, thereby enabling the high-speed, low-power nonlinear phenomena essential for computation. The synergy between these photonic architectures and the electronic drivers manifests as a finely balanced optoelectronic system optimized for performance.
In addition to performance metrics, the reliability and stability of the nonlinear optoelectronic engine under varying environmental conditions were tested extensively. The results indicate robustness against thermal fluctuations and fabrication-induced imperfections, attesting to the device’s practical viability. The integration of electronic feedback loops plays a pivotal role in this context, dynamically compensating for any performance drifts, ensuring consistent operation crucial for critical applications.
Looking forward, the legacy of this research is poised to redefine the roadmap for photonic computing. By overcoming the entrenched barriers of on-chip nonlinearity and achieving full monolithic integration, the nonlinear optoelectronic engine sets a new benchmark. The technique’s scalability and versatility hint at a future where entire computing architectures, from data storage to logical processing units, could migrate to photonic platforms, dramatically reshaping the computational paradigm.
Importantly, this advancement opens exciting avenues in quantum information processing as well, where nonlinear optics plays an indispensable role in generating and manipulating quantum states of light. The monolithic integration demonstrated here lays the groundwork for hybrid quantum-classical photonic processors that could harness the nonlinear optoelectronic engine for enhanced operation speed and reduced decoherence, crucial for practical quantum technologies.
The societal impact of such transformative technology cannot be overstated. As data demands skyrocket and electronic processors edge closer to physical limits imposed by heat dissipation and electron mobility, the nonlinear optoelectronic engine offers a sustainable alternative path forward. It melds the unparalleled speed of photons with the flexible processing capabilities of electronics, delivering a hybrid compute engine capable of underpinning the next generation of smart devices, autonomous systems, and intelligent infrastructure.
In summary, the nonlinear optoelectronic engine reported by Zhu and Zhu embodies a seminal leap in integrated photonic computing. The seamless fusion of nonlinearity and monolithic integration not only addresses pivotal challenges but also propels photonic technology into realms previously dominated by silicon electronics. As this technology matures, it is likely to spawn an ecosystem of applications and innovations that will redefine computing, communication, and beyond.
Subject of Research: Photonic computing, nonlinear optoelectronic devices, monolithic integration, integrated photonics.
Article Title: Nonlinear optoelectronic engine drives monolithic integrated photonic computing.
Article References:

Zhu, S., Zhu, N.H. Nonlinear optoelectronic engine drives monolithic integrated photonic computing.

Light Sci Appl 14, 302 (2025). https://doi.org/10.1038/s41377-025-01970-3
Image Credits: AI Generated
]]>
					
		
		
		75361