Ultra-Low-Loss Silicon Nitride Photonics Based on Deposited Films Compatible with Foundries
Xingchen Ji, Yoshitomo Okawachi, Andres Gil-Molina, Mateus, Corato-Zanarella, Samantha Roberts, Alexander L. Gaeta, Michal Lipson

TL;DR
This paper demonstrates ultra-low-loss silicon nitride photonic devices fabricated with foundry-compatible, low-temperature processes, achieving losses comparable to high-temperature methods and enabling scalable integrated photonics for various applications.
Contribution
It introduces a low-loss fabrication process for silicon nitride photonics compatible with foundries, using surface roughness reduction and post-processing techniques.
Findings
Propagation loss as low as 0.06 dB/cm with post-processing
Devices suitable for linear and nonlinear applications
Demonstrated low-threshold parametric oscillation and frequency combs
Abstract
The fabrication processes of silicon nitride photonic devices used in foundries require low temperature deposition, which typically leads to high propagation losses. Here, we show that propagation loss as low as 0.42 dB/cm can be achieved using foundry compatible processes by solely reducing waveguide surface roughness. By post-processing the fabricated devices using rapid thermal anneal (RTA) and furnace anneal, we achieve propagation losses down to 0.28 dB/cm and 0.06 dB/cm, respectively. These low losses are comparable to the conventional devices using high temperature, high-stress low-pressure chemical vapor deposition (LPCVD) films. We also tune the dispersion of the devices, and proved that these devices can be used for linear and nonlinear applications. Low threshold parametric oscillation, broadband frequency combs and narrow-linewidth laser are demonstrated. Our work…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
