AI designs ultracompact photonic components 500x smaller than engineers imagined

TL;DR Summary
AI-driven inverse design shrank three photonic-chip components—mirrors, a wavelength splitter and a spatial mode sorter—by up to 500×, producing devices about 11 μm long for mirrors and ~5 μm for the splitter, made from silicon nitride (400–800 nm thick). The AI-optimized designs deliver strong light confinement and up to 98.5% reflectivity, demonstrating manufacturable, compact components that could boost on-chip performance for fiber-optic communications, data centers, AI, LiDAR and quantum computing.
Topics:science#artificial-intelligence#inverse-design#nanophotonics#photonic-chips#silicon-nitride#technology
Reading Insights
Total Reads
0
Unique Readers
7
Time Saved
137 min
vs 138 min read
Condensed
100%
27,422 → 70 words
Want the full story? Read the original article
Read on Live Science