ABSIL Matthieu

Abstract: Semiconductor technologies are at the core of modern digital infrastructures, including communications, mobility, healthcare, energy, and industrial automation. However, the fabrication of integrated circuits requires ultra-pure water, large energy consumption, and critical raw materials, contributing significantly to environmental impact. At end-of-life, electronic systems are predominantly shredded for material recovery. While precious metals such as gold, silver, palladium, platinum, and copper can be partially recovered, the overwhelming majority of functional value embedded in integrated circuits is destroyed. Functional reuse remains marginal due to the absence of reliable qualification methodologies for second-life components. RF integrated circuits (RFICs) and RF modules represent a particularly promising target for reuse. Communication standards such as Wi-Fi, Bluetooth, LTE, and 5G exhibit long-term stability, and RF system-on-chip architectures are highly integrated and robust. Many discarded consumer devices (smartphones, routers, tablets) contain RF modules that remain electrically operational after first life. Yet, today no standardized method exists to assess their residual quality, grade their health state, or predict their Remaining Useful Life (RUL) in second-life applications such as IoT or edge devices. Reliability science has traditionally focused on device-level degradation mechanisms, including hot-carrier effects, charge trapping, electromigration, and thermal stress. Advanced DC and RF characterization techniques allow detailed analysis of transistor aging. However, system-level reliability assessment of used integrated circuits remains largely unexplored. Current methodologies extrapolate system reliability from component-level models without direct health evaluation of aged modules. No grading protocol or warranty-compatible scoring framework exists for reused RF systems. Recent advances in non-destructive imaging, including high-energy X-ray tomography and laminography, enable three-dimensional structural inspection of electronic systems without physical destruction. In parallel, thermal mapping, electromagnetic emission analysis, and advanced RF fingerprinting offer promising electrical-level indicators of degradation. However, these approaches have not yet been integrated into a unified, system-level diagnostic framework. The central scientific bottleneck is therefore clear: There is currently no fast, non-invasive, system-level methodology capable of grading used RF integrated circuits and predicting their remaining useful life. This PhD project addresses this bottleneck by developing a multi-modal reliability framework combining electrical RF fingerprinting, thermal analysis, and non-destructive structural imaging. The project leverages the complementarity between TIMA (system-level modeling, test methodologies, design technology co-optimization) and CEA-Leti (RF device expertise, advanced characterization platforms), with interaction with ESRF imaging capabilities. UCLouvain with the RF-SOI group managed by Prof. Jean-Pierre Raskin will participate actively to the supervising of the PhD student. RF-SOI group has a strong expertise in the field of RF and millimeter-wave ICs including simulation and characterization of advanced technologies. Moreover, Prof. Jean-Pierre Raskin is the holder of the Chair of Excellence in eco-innovation at CEA-Leti & UGA since January 2024.