-
- News
- Books
Featured Books
- pcb007 Magazine
Latest Issues
Current IssueVoices of the Industry
We take the pulse of the PCB industry by sharing insights from leading fabricators and suppliers in this month's issue. We've gathered their thoughts on the new U.S. administration, spending, the war in Ukraine, and their most pressing needs. It’s an eye-opening and enlightening look behind the curtain.
The Essential Guide to Surface Finishes
We go back to basics this month with a recount of a little history, and look forward to addressing the many challenges that high density, high frequency, adhesion, SI, and corrosion concerns for harsh environments bring to the fore. We compare and contrast surface finishes by type and application, take a hard look at the many iterations of gold plating, and address palladium as a surface finish.
It's Show Time!
In this month’s issue of PCB007 Magazine we reimagine the possibilities featuring stories all about IPC APEX EXPO 2025—covering what to look forward to, and what you don’t want to miss.
- Articles
- Columns
Search Console
- Links
- Media kit
||| MENU - pcb007 Magazine
Imec Demonstrates Compact 32-Channel Silicon-Based Wavelength Filter with Low Loss and High Tuning Efficiency
April 1, 2024 | ImecEstimated reading time: 2 minutes
In a top-scored paper at the OFC Conference (San Diego), imec, a world-leading research and innovation hub in nanoelectronics and digital technologies, is presenting a major advancement in silicon-based wavelength-division multiplexing (WDM) capability. A compact 32-channel silicon-based wavelength filter is demonstrated for the first time with low loss and high tuning efficiency, enabling a 4x increase in the number of transmitted and received wavelength channels compared to today’s commercial transceivers. The demonstrated capability will enable continued scaling of the bandwidth density and power efficiency of next-generation silicon-photonics-based transceivers, addressing the requirements for short-reach optical interconnects in high-performance AI/ML compute clusters.
Continued scaling of optical interconnects is essential for achieving further performance and efficiency gains in AI/ML compute clusters. Silicon photonics is a key technology platform for cost-effective integration of Tbps-scale transceivers, but the demanding requirements for bandwidth density, power efficiency, and latency in AI/ML system interconnects require substantial improvements at the process, device, and circuit level. For example, scaling up link bandwidth by adding more wavelength channels is typically associated with significant optical losses and excessive tuning power when using silicon-integrated components.
By combining the advanced Si patterning capability in imec’s 300mm silicon photonics platform (iSiPP300) with innovative and highly optimized designs of Si ring resonator and interleaver components, imec has overcome these challenges, realizing a WDM filter with 32 wavelengths at 100GHz channel spacing in the O-band, matching the specifications of the CW-WDM multi-source agreement (MSA). Owing to the careful design optimization, low optical insertion losses of about 2dB and crosstalk levels better than 16dB are maintained, even at this high channel count. By leveraging imec’s local Si substrate undercut process and optimized heater designs, the thermo-optic tuning efficiency is among the best-reported values at 290GHz/mW.
Joris Van Campenhout, fellow and R&D program director at imec: “Imec’s optical I/O R&D program aims at disruptive scaling of both bandwidth density and power consumption of silicon-integrated optical interconnects, which often present conflicting requirements. Our new 32-channel WDM filter enables Tbps-scale aggregate bandwidth per optical fiber using individual lane rates of only 32Gbps, enabling error-free link operation without power-hungry digital signal processing. Leveraging its compact footprint, low optical losses, and high wavelength tuning efficiency, the demonstrated WDM filter opens a pathway to Tbps/mm-scale optical interconnects with overall power efficiencies of just 1pJ/bit.”
Suggested Items
Koh Young Installs 24,000th Inspection System at Fabrinet Chonburi
04/23/2025 | Koh YoungKoh Young, the global leader in True 3D measurement-based inspection and metrology solutions, proudly announces the installation of its 24,000th inspection system at Fabrinet Chonburi in Thailand. This advanced facility is operated by Fabrinet Co., Ltd., a global provider of advanced manufacturing services, specializing in complex optical, electro-optical, and electronic products
Lam Research Donates Leading-Edge Etch System to Accelerate Nanofabrication R&D at UC Berkeley
04/17/2025 | PRNewswireLam Research Corp. announced the donation of its innovative multi-chamber semiconductor etching system to the Marvell Nanofabrication Laboratory at the University of California, Berkeley to advance research and development (R&D) for next-generation chip technologies.
Real Time with... IPC APEX EXPO 2025: Emerging Trends in Design and Technology
04/16/2025 | Real Time with...IPC APEX EXPOAndy Shaughnessy speaks with IPC design instructor Kris Moyer to discuss emerging design trends. They cover UHDI technology, 3D printing, and optical data transmission, emphasizing the importance of a skilled workforce. The role of AI in design is highlighted, along with the need for understanding physics and mechanics as designs become more complex. The conversation concludes with a focus on enhancing math skills for better signal integrity.
TDK Demonstrates the World's First ‘Spin Photo Detector’ Capable of 10X Data Transmission Speeds for the Next Generation of AI
04/16/2025 | PRNewswireThis new device is expected to be a key driver for implementing photoelectric conversion technology that boosts data transmission and data processing speed, particularly in AI applications, while simultaneously reducing power consumption.
IEEE Study Leverages Silicon Photonics for Scalable and Sustainable AI Hardware
04/14/2025 | PRNewswireThe emergence of AI has profoundly transformed numerous industries. Driven by deep learning technology and Big Data, AI requires significant processing power for training its models. While the existing AI infrastructure relies on graphical processing units (GPUs), the substantial processing demands and energy expenses associated with its operation remain key challenges.