Answers are grounded in this paper and cite their sources.
The CMOS industry has been expecting silicon photonics to provide photonic and electro-photonic integrated circuits based on the CMOS processes and infrastructures for scalability of incumbent technology evolutions and creation of novel technologies. However, the compatibility with the legacy CMOS has been compromised with the development convenience of early silicon photonics in that the specialty silicon-on-insulator substrates have been widely used as integration platforms. Since this specialty substrate may hinder the photonics integration with legacy volume products later, a legacy-friendly integration platform with a generic bulk-silicon substrate has been developed for better compatibility. This paper overviews the bulk-silicon photonics platform born for DRAM integration, upgraded with III/V-on-bulk-Si lasers, and transplanted to LiDAR applications requiring the virtuous cycle of cost and volume. The photonics integration with DRAM was to resolve the speed-capacity trade-off in the DRAM interconnects, and technical feasibility as well as lessons learned from the integration attempt are reviewed. The bulk-silicon device performance approaches that of silicon-on-insulator devices with the thermal advantage of ∼40-% lower thermal impedance and the optical disadvantage of ∼0.4-dB/mm higher waveguide loss. In the LiDAR applications, detection performance up to ∼20 m at 20 fps by a single-chip scanner integrating tunable laser, semiconductor optical amplifiers, and optical phased array are presented with future outlooks.
No takes yet. Share an insight, caveat, or question.
Shin et al. (2022) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: