Skip to content
#edge computing Open access

Packaging Design and Characterization of a 16-Channel Silicon Photonic Transmitter Module for a 1.6 Tb/s-Class Optical Engine

Sep 2026 · Microelectronics · Vol 2, pp. 14 · 16 references
Photonic and Optical Devices

Abstract

The increasing bandwidth demand of artificial-intelligence computing and data-center interconnects has driven silicon photonic optical engines toward higher capacity, compact packaging, and improved scalability. This work presents the packaging design and characterization of a completed 16-channel silicon photonic transmitter module for a 1.6 Tb/s-class optical engine. The module integrates an IMEC-fabricated silicon photonic chip, fiber-array edge-coupling interfaces, wire-bond submounts, PAM4 driver circuits, per-channel MZM heaters, and module-level thermoelectric cooling. A custom submount reduced the highly nonuniform wire-bond distance of 255.98–1788.48 μm to approximately 275.26 μm and provided a simulated packaged-path −3 dB bandwidth of approximately 32 GHz. Fiber-array monitor-port scans showed transverse alignment tolerances of approximately 4–6 μm; after accounting for the nominal 1% monitoring tap and a separately calibrated 4 dB grating-coupler-to-MMF path, the effective packaged input-coupling estimate at 1310 nm was approximately 3.77 dB. All 16 channels were independently verified at 53.125 GBaud PAM4, corresponding to 106.25 Gb/s per lane, and the completed module supports concurrent activation of all lanes. Because the Keysight N1000A optical sampling oscilloscope (Keysight Technologies, Inc., Santa Rosa, CA, USA) accepts one optical input at a time, the eye diagrams and metrics were acquired by sequentially selecting the observed channel and do not constitute a simultaneous 16-channel crosstalk measurement. Across the 16 channels, the measured optical power, OMA, ER, TDECQ, and RLM ranged from −0.09 to 0.30 dBm, 1.141 to 1.267 dBm, 3.037 to 3.055 dB, 3.09 to 3.18 dB, and 0.956 to 0.981, respectively. The measured module-level TEC power was 1.2 W. Using the same nominal 1.6 Tb/s data-rate denominator, the estimated energy consumption is 3.95 pJ/bit for the transmitter electronics excluding the external lasers and TEC and 7.83 pJ/bit when the external laser operating power and measured module-level TEC power are included. These results demonstrate a practical package-level proof of concept for scalable 1.6 Tb/s-class silicon photonic transmitters.

View source

Similar papers

#computer vision Review Sep 2017

Agile Software Development Methods: Review and Analysis

This publication proposes a definition and a classification of agile software development approaches and analyses ten software development methods that can be characterized as being "agile" against the defined criterion.

P. Abrahamsson, O. Salo, Jussi Ronkainen et al. · 727 citations · ⚡54
#computer vision Jun 2008

The impact of agile practices on communication in software development

The study shows that agile practices improve both informal and formal communication, but indicates that, in larger development situations involving multiple external stakeholders, a mismatch of adequate communication mechanisms can sometimes even hinder the communication.

M. Pikkarainen, Jukka Haikara, O. Salo et al. · 401 citations · ⚡48
#machine learning Review Open access Oct 2014

Software development in startup companies: A systematic mapping study

The results indicate that software engineering work practices are chosen opportunistically, adapted and configured to provide value under the constrains imposed by the startup context.

Nicolò Paternoster, Carmine Giardino, M. Unterkalmsteiner et al. · 394 citations · ⚡54

Related blog posts

Microsoft Research Blog Oct 6, 2026

What AI gets wrong and what failure teaches us

Jennifer Neville did not want to go into computer science—but that’s exactly where she landed. Neville discusses the starts and stops that led to her professional sweet spot and her work identifying “surprising failures” making it hard for AI to handle complexity.  The post What AI gets wrong and what failure teaches us appeared first on Microsoft Research.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.