Toward an Efficient Transformer-based Architecture for Multi-Agent Reinforcement Learning
Abstract
germanium (Ge) and silicon (Si) substrates. This has led to much new understanding of defective growth and how to minimize it. ii) For the first time we have prepared GeSn films by Sn ion implantation into Ge followed by pulsed laser melting (PLM) to produce unstrained GeSn alloy layers with attractive material and optical properties. iii) We also initiated the first investigation into producing GePb alloys by implantation and PLM and found that Pb concentrations up to a few atomic percent are feasible. In terms of these previous preliminary but encouraging results, a focus in the current project on defect reduction and high quality superlattices in GeSn layers on Si, and the fabrication of novel GePb alloys, should provide substantially improved materials. New studies include- electrical doping of the GeSn by ion implantation to facilitate a necessary p-n junction for photodetector fabrication; using Raman mapping to measure the defect density of GeSn films and understand the defect evolution with depth; utilize PLM on CVD-grown GeSn films as another novel method to reduce defects. If successful, this project could lead to the first efficient high-resolution mid-IR imaging array integrated onto Si that operates at room temperature.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Feb 05, 2025
- Source ID
- FA23862414012
Entities
People
- Cuong Ta
Organizations
- Air Force Office of Scientific Research
- United States Air Force
- University of Engineering and Technology