Failure Analysis - AI-Readiness and Application

Purpose and goal

FA2IR builds on the Penta/Euripides project FA4.0 that has demonstrated the use of AI algorithms in Failure Analysis to improve the efficiency of analysis techniques. The FA2IR project will investigate important applications of Artificial Intelligence (AI) methods to databases in microelectronic failure analysis (FA). The main objective of FA2IR is to get FA databases AI-ready and to develop improved FA4.0-AI-based methods for image and measurement data analysis. Standardization efforts will push digitalization standards within the international semiconductor community.

Expected effects and result

Reducing the time it takes to analyse microelectronic failures, enables companies to respond more quickly to production and field problems. The average analysis time will decrease and a higher data standardization level will be achieved after the project ends. Because of the enhanced efficiency in microelectronic AI-driven failure analysis, Companies can seize a larger share of the microelectronics market due to an increase in precision in failure assessment within microelectronics production processes due to a consistent reduction in data errors.

Planned approach and implementation

17 partners represent the full FA value chain and 4 countries. Six work packages are defined: WP 1. Specification, Gap Analysis and Monitoring and WP 6. Use Cases, Validation and Performance Tests set-up the frame with specs and requirements and assessment, respectively. WP2 - WP5 reflect innovations where a state-of-the-art does not exist in microelectronics failure analysis: They are: WP 2. AI-ready Data Landscape and FA Tool Integration WP 3. AI-enhanced Data Analysis WP 4. Failure Analysis Data Environment Assessment Tools WP 5. Integration of Cloud Computing and external AI tools