The Path from Electrons to Photons: Shining a Light on the Fabrication Challenges of Silicon Photonics

Rapid advances in Artificial Intelligence (AI) are already impacting our lives at an almost unbelievable pace. The addition of situational context to AI, and real time access to AI analysis through augmented reality and spatial computing, will only stand to further increase this impact. But, despite the incredible leaps in processing power that have enabled these advances in AI, memory and interconnect bandwidth have not scaled at the same rate. This leaves systems developers facing the need to overcome these performance barriers and in parallel deliver significant improvements in the energy efficiency of an AI gigafactory infrastructure. The convergence of these technology trends is ushering in a new era of next-generation optoelectronic and photonic solutions that will be at the forefront of how AI is enabled, experienced and made more efficient in the future. In this presentation we will talk to this convergence and the role that optoelectronics and photonics will play in it. We will then take a more detailed look at silicon photonics and its potential for improving future AI systems. We will discuss how, even though such silicon photonics systems may be manufactured using 45nm or even 65nm CMOS design rules, some of the fabrication challenges associated with them are among the most demanding facing the semiconductor industry today. This will be highlighted by considering examples such as silicon and silicon nitride waveguides, optical fiber coupling technologies and next generation materials for modulators and photodetectors.

Speakers: Michelle Bourke & Dr. David Haynes, Lam Research Corporation