OPTICS 2025

Time: 9:00-16:00 (UTC, Coordinated Universal Time), October 29, 2025

Join us for free by registering here!

Organizers

Zefeng Xu, Hong Kong University of Science and Technology (Guangzhou)
Felipe Gohring de Magalhaes, Polytechnique Montréal

Final Program
Wednesday October 29, 2025, Coordinated Universal Time(UTC)

Openning by: Zefeng Xu, Hong Kong University of Science and Technology(Guangzhou)

Session #1
Photonic Computing and Optical MEMS Technology
Chair: Renjie Wang, Hong Kong University of Science and Technology(Guangzhou)

UTC 9:10-9:30 Advances in phase change photonics for accelerating AI
Dong Bowei, A*STAR Institute of Microelectronics, Singapore

Abstract: We will discuss what phase change photonics is, how phase change photonics is enabled by integrating novel functional phase change materials with conventional silicon photonics, what device functions can phase change photonics provide, and how phase change photonic circuits are designed and operated to realize new photonic in-memory computing functionalities. We will also discuss the limitation and provide an outlook of phase change photonics.

Biography: Dr. Dong Bowei is a Principal Scientist and Principal Investigator at the Institute of Microelectronics (IME) A*STAR, leading the development of photonic computing technology. He is a recipient of Singapore National Research Foundation Fellowship, A*STAR Young Achiever Award, and A*STAR International Fellowship.

UTC 9:30-9:50 Next generation of photonic integration: Photonic Wire Bonding and facet-attached micro-optical element
Mo Lu, Vanguard Automation, Germany

Abstract: Next-generation photonic computing and optical MEMS technologies demand precise, scalable integration across diverse material platforms to meet the performance requirements of AI accelerators, quantum processors, and high-bandwidth communication systems. Vanguard Automation’s Photonic Wire Bonding (PWB) and facet-attached micro-lens (FaML) technologies employ high-precision 3D nano-printing, machine vision, and passive alignment to fabricate freeform optical interconnects without active alignment steps. These techniques enable low-loss, mode-field-matched coupling between dissimilar waveguides and chiplets, supporting hybrid integration of InP, Si, SiN, and TFLN platforms. By bridging photonic and MEMS domains, PWB and FaML facilitate compact, thermally stable, and reconfigurable optical architectures. Together, they provide a scalable pathway from research prototyping to industrial manufacturing, accelerating the development of next-generation photonic computing and optical MEMS systems.

Biography: Dr. Mo Lu is a Staff Engineer at Vanguard Automation, where she leads process development activities within the R&D department. Her work focuses on advancing Photonic Wire Bonding (PWB) and facet-attached micro-lens (FaML) technologies for hybrid photonic integration, including process optimization, yield improvement, and reliability testing. She received her B.Sc. in Chemical Physics from the University of Science and Technology of China and her Ph.D. from the University of Cologne, where she studied light–matter interactions between optical antennas, quantum emitters, and two-dimensional materials. Dr. Lu is also actively engaged in several European collaborative projects, including ELENA and Agilight, driving innovation in scalable photonic manufacturing and integration technologies.

UTC 9:50-10:10 Optical MEMS technology in the future frontier research and development of micro/nano optics
Yu-Sheng Lin, Aerospace Information Technology University, China

Abstract: Micro-opto-electro-mechanical system (MOEMS) technology is a technique for actively tunable micro/nanostructures. Due to its ability to directly modify the micro/nanostructures, the optoelectronics could be performed to possess the ability to directly modify the optical characteristics, that can be integrated and implemented into metamaterial and nanophotonic fields. MOEMS tuning methods can provide an ideal platform for metadevices, which is not limited by the nonlinear characteristic of natural materials and exhibits unique optical performance. Recently, the actively MOEMS metadevice has become a hotspot in optoelectronics research as it provides the possibility to manipulate electromagnetic waves. It would have many potential opportunities to implant into current optoelectronics, optical communications, data analysis, and fundamental physic studies.

Biography: Prof. Yu-Sheng Lin is a professor at the Aerospace Information Technilogy University in Jinan, China. His research focuses on the innovative design and application development of tunable and reconfigurable metadevices. He has published more than 160 SCI papers (H-index: 45), as well as be selected in the list of the world's top 2% scientists in 2025, which was jointly released by Stanford University and the international authoritative academic publisher Elsevier and received awards for his acclaimed research contributions, such as awarded Highly Ranked Scholar by ScholarGPS within 5 years ranking among the top 0.05% of scholars in the world, ranked 2nd in the world in the field of metamaterials in 2025, Danmark International Scientific Organization Vebleo Fellow in 2025, Sweden International Association of Advanced Materials Fellow in 2023, respectively.

UTC 10:10-10:30 Break

Session #2
Innovations in Photonic Integration and Device Design
Chair: Yeyu Tong, Hong Kong University of Science and Technology(Guangzhou)

UTC 10:30-10:50 Thin Film Barium Titanate-on-Insulator Optoelectronics
Aaron Danner, National University of Singapore, Singapore

Abstract: Barium titanate has an exceptionally strong Pockels parameter which may ultimately be exploited to create nonlinear chip-based devices. I will describe my group's recent efforts in growing single crystal barium titanate and its use in fabricating useful devices.

Biography: Aaron Danner is with the National University of Singapore. His main research area is chip-based photonics. He received his Ph.D. in Electrical Engineering from the University of Illinois at Urbana-Champaign.

UTC 10:50-11:10 Integrated optical phased array for free-space optical communications
Hao Hu, Technical University of Denmark, Denmark

Abstract: Conventional free-space and underwater optical communication systems rely on mechanical alignment and tracking, which are inherently slow, vibration-sensitive, and unsuitable for fast-moving platforms. Optical phased arrays (OPAs) provide a compact, solid-state alternative, enabling electronic beam steering at a high speed without moving parts. A long-standing challenge, however, has been achieving simultaneously wide field of view (FOV), high resolution, and low side-lobe levels (SLL).
In this talk, I will present our recent advances in integrated OPAs that overcome these trade-offs. We first demonstrate aliasing-free two-dimensional beam steering across the entire 180° FOV using a half-wavelength-pitch waveguide array combined with a trapezoidal slab grating, achieving high beam quality with side lobes below –19 dB. Building on this architecture, we scale to a 1000-channel OPA with optimized waveguide routing to suppress crosstalk and a novel matrix control scheme that reduces the complexity of driving 1000 thermo-optic phase shifters to only 70 control signals. The resulting device achieves grating-lobe-free 2D steering with fine angular resolution (0.07° × 0.17°) and SLL below –18.7 dB.
These results establish a new route toward scalable, low-cost, chip-based OPAs, opening opportunities for high-resolution, wide-angle beam steering in free-space optical communications, LiDAR, and beyond.

Biography: Hao Hu is a Senior Researcher and Head of the Photonic Integrated Circuit (PIC) Based Systems group at DTU Electro, Technical University of Denmark. He has coordinated and led multiple EU and Danish research and innovation projects, building extensive experience in collaborations with both academic and industrial partners.
Dr. Hu has authored more than 300 peer-reviewed publications, received over 7,000 citations, and holds four patents. His international experience includes visiting scientist positions at Fraunhofer HHI in Berlin (2007-2008) and at Alcatel-Lucent Bell Labs, Crawford Hill, New Jersey (2013). He is the recipient of the Independent Research Fund Denmark’s Research Talent Award, the VILLUM Young Investigator Award, and led the team that won the European Commission’s Horizon Prize in 2016.
His achievements include the world-record optical fiber transmission of 43 Tbit/s using a single laser source, ranked #7 among ‘The 20 Greatest Engineering Feats of 2014’ and, more recently, the development of integrated optical phased array (OPA) based beam steering technology that lays the foundation for photonic chip-based LiDAR. His research focuses on integrated photonics, optical communications, and beam steering technologies for next-generation communication and sensing systems.

UTC 11:10-11:30 Integrated Silicon Photonics with Quantum Dot On-Chip Lasers: Enabling Scalable and Energy-Efficient Photonic Systems
Yating Wan, King Abdullah University of Science & Technology, Saudi Arabia

Abstract: Integrated silicon photonic has emerged as a leading solution for scalable, power-efficient, and environmentally friendly applications. This talk will focus on the prospects and applications of on-chip lasers, a critical component driving the advancement of photonic integrated circuits (PICs). We will discuss various approaches to integrating lasers on silicon and their applications in optical communication, computing, and LiDAR, with a particular emphasis on heterogeneous integration of quantum dot (QD) lasers. QD lasers offer unique advantages, including high immunity to reflection, superior thermal stability, low threshold, and long-term reliability, making them ideal for high-speed optical interconnects, AI-driven computing, and quantum photonic systems. By leveraging the defect tolerance and temperature resilience of QDs, we achieve high-performance, energy-efficient integration with silicon photonics. This talk will highlight recent breakthroughs in QD-on-silicon integration, performance optimizations, and future directions toward heterogeneous photonic systems, paving the way for a new generation of high-speed, low-energy, and scalable optical circuits for next-generation applications.

Biography: Dr. Yating Wan is an Assistant Professor at KAUST, specializing in silicon photonics with a focus on integrating on-chip QD light sources. Before joining KAUST, she worked in Prof. John Bowers' group at UCSB (2017–2022), where she led Intel’s project on heterogeneously integrated QD lasers on silicon. She has published over 100 peer-reviewed papers, including 39 first-author publications (29 journals, 10 proceedings), 22 corresponding-author publications (7 journals, 15 conferences), and 10 journal cover features. She actively contributes to the scientific community as an Associate Editor for IEEE JSTQE, JQE, and Applied Optics, as a committee member of the IEEE Photonics Society Conference Council and CLEO (2022–2024), and as a referee for peer-reviewed journals across IEEE, OSA, and the Nature Publishing Group for over 100 times. For her pioneering work in on-chip laser integration on silicon, Dr. Wan has received numerous prestigious awards, including 2021 CLEO Tingye Li Innovation Prize (1 awardee per year), 2022 Rising Stars of Light by Light: Science & Applications (3 awardees per year), 2023 “35 Innovators Under 35 for China” by MIT Technology Review, 2025 Sony Women in Technology Award with Nature ($250,000 prize, 3 awardees per year), and 2025 IEEE Photonics Society Young Investigator Award (1 awardee per year).

UTC 11:30-11:50 Photonic Integrated Switches: from Homogeneous to Heterogeneous Integration
Richard Penty, University of Chambridge, UK

Abstract: This talk will review our work on large port count optical switches realized using photonic integrated circuits fabricated in both indium phosphide and silicon photonic generic foundries, using approaches. We will also review recent progress on heterogeneous integrations via micro-transfer printing, direct epitaxial growth and additive manufacturing such as the two-photon polymerization and discuss how these innovative methods can further enhance the integration of photonic switches.

Biography: Richard Penty is Professor of Photonics and Head of the School of Technology at the University of Cambridge, UK. His research interests currently include photonic integration, optical data communications and quantum communications.
Following his undergraduate degree in Engineering and Electrical Sciences at Cambridge Professor Penty completed his PhD in optical fibre nonlinearities at Cambridge, before taking up Lecturer posts at the University of Bath and the University of Bristol.
Returning to Cambridge, Professor Penty was appointed Assistant Director of Research in the Engineering Department in 2001, and then a Fellow in Engineering at Sidney Sussex in 2002. He was promoted to Professor in 2002 and became Master of Sidney Sussex in 2013. In 2012 he was elected a Fellow of the Royal Academic of Engineering and also of the IET.

UTC 11:50-12:10 Inverse Design in Silicon Nitride Photonics
Michaël Ménard, École de technologie supérieure, Canada

Abstract: Topology optimization has been successfully applied to create ultra-compact devices in silicon photonics, where the refractive index contrast between the core and the cladding is high. We recently demonstrated that this design technique can also produce highly compact and efficient devices on platforms with more moderate index contrast, such as silicon nitride. We will present the current state of the art in inverse-designed silicon nitride devices, including spatial and wavelength multiplexers as well as multimode edge couplers.

Biography: Michaël Ménard (Member, IEEE) was born in Québec City, QC, Canada. He received the B.Eng. and Ph.D. degrees in electrical engineering from McGill University, Montreal, QC, Canada, in 2002 and 2009, respectively. From 2009 to 2011, he was a Postdoctoral Fellow with the Cornell Nanophotonics Group under the supervision of Prof. Michal Lipson. In 2011, he joined the Microelectronic Program with UQAM till 2021, when he started to work with the Department of Electrical Engineering, École de Technologie Supérieure. He jointly manages the Microtechnology and Microsystems Laboratory. In 2019, he was a Visiting Researcher with the Department of Applied Physics, University of Campinas, São Paulo, Brazil. He has authored or coauthored more than 100 publications, and holds five issued patents and three pending patent applications. His research interests include integrated optics, silicon photonics, nonlinear optics, micro-opto-electro-mechanical systems (MOEMS), optomechanics, and microfabrication. He is an Active Member of NanoQAM, the Research Center on Nanomaterials and Energy. He is a Member of the Quebec Order of Engineers. He holds or has held financial support from the Microsystems Strategic Alliance of Quebec, Center for Optics, Photonics, and Lasers, Quebec Fund for Research in Nature and Technology, Prompt Québec, PRIMA Québec, and Natural Sciences and Engineering Research Council of Canada.

UTC 12:10-12:40 Break

Session #3
Next-Gen Photonic Solutions for AI and Quantum Applications
Chair: Zefeng Xu, Hong Kong University of Science and Technology(Guangzhou)

UTC 12:40-13:00 Integrated Photonic Neural Processors: Potential and Pitfalls
Mahdi Nikdast, Colorado State University, US

Abstract: Integrated photonic neural processors offer the promise of ultrafast, energy-efficient computing, but their practical realization faces significant hurdles. This talk will explore architectural approaches for photonic neural processors, highlighting how fabrication variations, optical loss, and crosstalk impact accuracy, performance, and scalability. I will also discuss emerging strategies to mitigate these challenges and improve robustness, with the goal of enabling scalable and reliable photonic accelerators for next-generation machine learning systems.

Biography: Mahdi Nikdast is an Associate Professor in the Department of Electrical and Computer Engineering at Colorado State University (CSU), Fort Collins, where he directs the Electronic-PhotoniC System Design (ECSyD) Laboratory. His research interests are at the intersection of integrated photonics, high-performance computing (HPC), and artificial intelligence (AI).

UTC 13:00-13:20 Integrated frequency combs for quantum technology
Xu Yi, University of Virginia, US

Abstract: Optical frequency combs serve as a coherent gearbox bridging the electrical and optical frequency domains, unifying the electromagnetic spectrum in a remarkable way. A compact realization—the microcomb, based on integrated microresonators—has the potential to revolutionize classical instrumentation such as timekeeping and navigation, while also opening new avenues in quantum photonics. In this talk, I will present our recent demonstration of a record-low noise mmWave oscillator achieved with integrated photonic components, highlighting the transformative impact of bridging electronic and photonic frequencies.

Biography: Dr. Xu Yi is an Associate Professor in the Department of Electrical and Computer Engineering at the University of Virginia, with a courtesy appointment in the Department of Physics. He earned his Ph.D. in Applied Physics from the California Institute of Technology in 2017 and joined the University of Virginia faculty in 2018. His research interests focus on integrated photonics, optical microresonators, quantum optics, and optical frequency combs. Dr. Yi is the recipient of the 2017 NASA group achievement award, 2021 Air Force Young Investigator Program award, and 2023 NSF CAREER award.

UTC 13:20-13:40 SCARLET: A Hybrid Photonic Architecture for Multi-Billion Parameter LLM Inference
Sina Karimi, Boston University, US

Abstract: Transformer-based large language models (LLMs) have set new benchmarks across various AI tasks but face significant computational challenges during inference, especially with multi-billion-parameter models. While photonic accelerators—particularly those using phase-change memory (PCM)-based crossbars—offer promising advantages in speed and energy efficiency, existing designs fall short in supporting the precision, dynamism, and storage demands of LLM models with large parameters. This work introduces SCARLET, a hybrid photonic architecture that addresses these challenges through two synergistic components. First, we design a high-density optical PCM (OPCM) crossbar tailored for static matrix multiplications, achieving 5.6x higher bit density, 10.96x lower latency while reducing energy consumption by 86.43% for FFN layers with large sequence length compared to prior designs. Second, to support the dynamic matrix operations along with quantization/dequantization steps found in quantized LLM decoder layers, we introduce an approximate photonic floating-point multiplier that eliminates the need for frequent PCM reprogramming and dedicated floating-point units. This architecture approximates floating-point operations using weighted integer sums, enabling efficient, low-overhead computation in the photonic domain. Our evaluation shows that, compared to electronic implementation, we achieve 69.2% lower energy consumption for computing the attention score. Together, these components form a complete and energy-efficient photonic backend for LLM inference, achieving scalability and precision without sacrificing performance. Evaluation on OPT, Llama-2, and GPTNeo models with a range of 2.7B to 13B parameters shows SCARLET achieves up to 17.15x and 8.45x lower latency during the prefill phase and generation phase, respectively, compared to GPUs.

Biography: Sina Karimi is a third-year Ph.D. student in Electrical and Computer Engineering at Boston University, working with Professor Ayse Coskun and Professor Ajay Joshi. His research focuses on memory management and accelerator design for ML and LLM models, with particular emphasis on Optical Phase-Change Memory (OPCM) crossbars.

UTC 13:40-14:00 Photonic Non-Volatile Memory Devices
Aaron Thean Voon Yew, National University of Singapore, Singapore

Abstract: With the ability to convert electrical to optical signals, Electro-Optic Modulator (EOM) as an electronic device is an increasingly important component for emerging technologies that includes optical sensors, photonic neural networks, quantum-information processing, and future possibilities of inter-chiplet optical interconnect networks through a photonic interposer. With high-conversion-efficiency due to Pockel’s effect, electro-optic materials like Lithium Niobate (LN) and Barium Titanate (BTO), the intimate integration of photonics and electronics will enable exciting functional and compact devices by heterogeneous integration, beyond co-packaged optics. Among such new class of devices are non-volatile memory functional photonic devices.
We have proposed new devices that combine high-efficiency thin-film LN EOM and Ferroelectric HZO Non-Volatile Memory (NVM) Capacitor. The resultant integrated Electro-Optic Modulator and Memory (EOMM) allows electronically encoded analog memory states to be translated into high-speed optical transmission and phase modulations in a single compact device. Such a device would enable new possibilities for dynamically reconfigurable photonic systems. Furthermore, can photonics enable Compute-in-memory (CIM) architectures and what would be the benefits? For resistive-based CIM systems, the array scalability is inevitably limited by IR losses with increasing error accumulation due to the increasing wire resistance as arrays grow. We have recently proposed an electro-optic memory array with an optical bitline (BL) that circumvents the BL IR loss and capacitive loading issue. By implementing a two-transistor-one-modulator (2T1M) memory cell, the matrix dot-products can be performed by FeFET memories, operated in sub-threshold region, while the accumulation is summed through phase modulation of an optical signal through a LN capacitive modulator. By eliminating IR loss on the BL, we can enable up to 3750kb array size and achieving up to 45% inference accuracy improvement even on a large-scale transformer model compared to conventional CIM arrays.

Biography: Aaron Voon-Yew Thean is the GlobalFoundries Professor of Electrical and Computer Engineering at the National University of Singapore (NUS). He is currently the Deputy President (Academic Affairs) and Provost of NUS. He was the Founding Dean of NUS College of Design and Engineering (2022), and the Dean of the Faculty of Engineering (2019). Aaron founded the $70M Applied Materials-NUS Corporate Laboratory at NUS in 2018, a major research collaboration between US-based Applied Materials and the University to co-innovate on next-generation semiconductor materials. Aaron currently directs the Singapore Hybrid-Integrated Next-Generation μ-Electronics (SHINE) Centre, a National Medium-Sized Research Center funded by the Singapore National Research Foundation to innovate on Heterogeneous Integration. Prior to joining NUS in 2016, Aaron was the Vice President of Logic Technologies at IMEC. Working with Semiconductor Industry leaders like Intel, TSMC, Samsung, Global foundries, Apple, and Sony, where he directed the research and development of next generation semiconductor technologies and emerging nano-device architectures. Prior to IMEC, he had careers at major US Semiconductor companies like Qualcomm, IBM, and Motorola/Freescale. Aaron graduated from University of Illinois at Champaign-Urbana, USA, where he received his B.Sc. (Highest Honors), M.Sc., and Ph.D. degrees in Electrical Engineering (Edmund J. James Scholar). He has published over 400 technical papers and holds more than 50 US patents. Aaron was recognized as Singapore’s National Research Foundation’s Returning Singapore Scientist and more recently, he has been recognized as a fellow of the US National Academy of Inventors.

UTC 14:00-14:20 Break

Session #4
Advancements in Photonic Interconnect and Computing Systems
Chair: Felipe Gohring de Magalhaes, Polytechnique Montréal

UTC 14:20-14:40 High-performance and diverse optical I/O solutions on a 300-mm monolithic CMOS-SiPh platform
Yusheng Bian, Global Foundries, US

Abstract: On this invited talk, we will present the latest advancements in high-performance and versatile I/O solutions on the FotonixTM platform—a 300-mm monolithic CMOS-integrated silicon photonics foundry. We will highlight significant achievements in three areas: high-power V-groove fiber attach, alternative pluggable couplers, and hybrid laser attach. Additionally, we will present an outline of the future roadmap.

Biography: Dr. Yusheng Bian is a Distinguished Member of Technical Staff and Director at GlobalFoundries in Malta, NY, where he leads the silicon photonics device team, driving innovation in passive/active components, optical I/O, and monolithic integration on the Fotonix™ platform. His team delivered key technologies including CMS90WG and 45SPCLO, and played a pivotal role in qualifying GlobalFoundries’ 300mm SiPh platform.
Dr. Bian has authored over 150 publications, two book chapters, and holds 260+ granted patents with 300+ pending. He has received multiple honors, including the SRC Mahboob Khan Outstanding Liaison Award, inaugural GFX prize, GlobalFoundries Master Inventor, Triple Diamond Inventor Award, Diversity in Inventorship Champion Award, and the USPTO Appreciation Award. He earned his Ph.D. in Optical Engineering from Beihang University and conducted postdoctoral research at Penn State and Peking University.

UTC 14:40-15:00 Hybrid Integration for Data Center Applications
Lukas Chrostowski, The University of British Columbia, Canada

Abstract: Photonic integrated circuit technology is used for many applications including optical data communications, optical and quantum computing, and sensing including LiDAR, biomedical and environmental. A major packaging challenge facing the industry is optical coupling between multiple integrated photonic components, with low insertion loss, in a cost effective manner, into a package suitable for commercialization. A promising approach is to use additive manufacturing of 3D-printed optics to create optical connections. Capabilities and advantages of the technique include gain chip integration with existing ’known good die’, dense optical I/O connections to the chip, scalability from prototyping to high-volume, and interconnects that are not possible with other standard photonic packaging techniques.

Biography: Lukas Chrostowski is a Professor of Electrical and Computer Engineering at the University of British Columbia, and co-founder of Dream Photonics Inc. Through his research in silicon photonics, optoelectronics, high-speed laser design, fabrication and test, for applications in optical communications, biophotonics, and quantum photonics, he has published more than 300 journal and conference publications. He co-authored the book “Silicon Photonics Design” (Cambridge University Press, 2015). Dr. Chrostowski was the co-director of the Advanced Materials and Process Engineering Laboratory (AMPEL) Nanofabrication Facility (ANF), 2008-2016. Dr. Chrostowski was the Program Director of the NSERC CREATE Silicon Electronic-Photonic Integrated Circuits (Si-EPIC) training program in Canada, and has been teaching numerous silicon photonics workshops and courses since 2008, which continue today as the SiEPICfab consortium. Chrostowski received the Killam Teaching Prize at the University of British Columbia in 2014, IEEE Photonics Society Technical Skills Educator Award in 2021, and IEEE Canada's J.M Ham Outstanding Engineering Educator Award in 2021. He was an elected member of the IEEE Photonics Society 2014-2016 Board of Governors. He was elected to the college of the Royal Society of Canada in 2019, and Fellow of Optica 2024. Chrostowski is the Program Director for the NSERC CREATE 2020-2026 Quantum Computing program (Quantum BC), co-leading the Quantum Silicon Photonics design-fabricate-test workshop. Chrostowski co-founded Dream Photonics Inc., a company focused on silicon photonics IP and integration technology.

UTC 15:00-15:20 Ferroelectrics for Emergent Silicon-integrated Optical Computing
Alexander Demkov, The University of Texas, US

Abstract: Traditional computing based on CMOS technology is nearing physical limits in terms of miniaturization, speed, and power consumption. Consequently, alternative approaches are under investigation. The most promising is based on a “brain-like” or neuromorphic computation scheme, another is optical quantum computing. These approaches can be realized using silicon photonics (SiPh), and at the heart of both technologies is an efficient, ultra-low power broad band optical modulator. A complete or partial switch from electrons to photons would be revolutionary, the technology ultimately requires integration of active and passive photonic elements on a single chip. As silicon modulators suffer from relatively high-power consumption and large size, materials other than silicon are considered for the compact energy-efficient modulator. I will discuss recent progress in integrating ferroelectric oxides with SiPh for the purpose of fabricating modulators exploiting the linear electro-optic effect. These will enable neuromorphic circuit architectures that exploit shifting computational machine learning paradigms, while leveraging current manufacturing infrastructure. This will result in a new generation of computers that consume less power and possess a larger bandwidth.

Biography: Alexander Demkov is professor of physics at The University of Texas at Austin. He earned his Diploma in materials science from the Moscow Institute of Steel and Alloys in 1986, and PhD degree in theoretical condensed-matter physics from Arizona State University in 1995. He joined The University of Texas in 2005, after working in Motorola’s R&D organization, where he investigated materials-related problems of electronic devices making significant contributions to the understanding of high dielectric constant materials. His current research is focused on transition-metal oxides integrated with semiconductors. He researches both theory and oxide molecular beam epitaxy.

UTC 15:20-15:40 Next-Generation Silicon Photonics for AI Interconnects
Yuan Yuan, Northeastern University, US

Abstract: Silicon has long served as the foundational platform for photonic integrated circuits. However, its intrinsic optoelectronic limitations have historically constrained device performance, hindering broader adoption in AI interconnects and processing. In this seminar, we present our research efforts aimed at systematically unlocking the full potential of silicon photonics by advancing key components, including modulators, photodetectors, and non-volatile optical memory elements. These innovations collectively enable a highly scalable and cost-effective silicon-based solution for large-scale photonic systems. We will conclude with a forward-looking discussion on the future of heterogeneous silicon photonic architectures for AI interconnects and processing, highlighting the opportunities and challenges that lie ahead.

Biography: Yuan Yuan is an Assistant Professor in the Department of Electrical and Computer Engineering at Northeastern University. Prior to joining Northeastern, he was a Senior Research Scientist at Hewlett Packard Labs. He earned his Ph.D. from the University of Virginia, and his research focuses on optoelectronic devices and large-scale photonic integrated circuits. Yuan has authored over 80 journal and conference papers, including publications in Nature Photonics and Nature Communications, and holds 9 U.S. and international patents. His work has been recognized with the ACP 2021 and OECC 2023 Best Paper Awards for Industry Innovation. He is a member of the technical program committees of several international conferences and serves as a member of the Publications Council of the IEEE Photonics Society.

Sponsor