Low-Cost Automated Integrated Circuit Analysis through AI-assisted Laser-Scanning-Microscopy
Abstract
A laser beam in the near-infrared (NIR) range can pass through the silicon material of an integrated circuit (IC) and interact withelectrical charges on the device. This allows one to indirectly study the structure and data on a silicon-based IC with lower invasiveness and cost compared to scanning-electron-microscopy (SEM). The family of techniques based on this principal can find defense related applications in the realm of electronic hardware security: reverse engineering and data extraction from enemy electronics, and integrity checking of friendly devices in the presence of untrusted supply chains. This project is requesting funding for equipment acquisition that will help build a low-cost NIR laser-scanning microscopy (LSM) system to allow for the development and deploymentof novel open-source tools that can automate various IC analysis tasks. The ultimate goal is to have tools that can control the instrument safely by themselves, take given a priori user information about theIC design, fuse it with power consumption measurements and collected NIR laser images, and in an ``AI-and-IC-in-the-loop fashion extract maximal information from the device. This is in line with the goal of automated hardware security analysis of the Cyber Security and Complex Software Systems program of ONR.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- May 15, 2023
- Source ID
- N000142312411
Entities
People
- Kaveh Shamsi
Organizations
- Office of Naval Research
- United States Navy
- University of Texas at Dallas