2025 BIST Ignite Project: RADIANT

Project overview

Heat dissipation is an increasing challenge in modern electronics due to the miniaturization of devices. Promising materials such as graphene or carbon nanotubes have indicated enhanced heat spreading characteristics. Despite the proven potential, the characterization of heat spreading characteristics remains highly difficult, and time intensive given the complex behaviour at lower scale.

In this context, methods like Raman thermometry or scanning thermal microscopy have been employed to study different heat spreading related parameters such as the thermal conductivity of materials. Nonetheless, these methods are still time-intensive and lack accuracy. To gain faster data output, the RADIANT project will develop a machine learning (ML) model to predict the temperature of materials based on their Raman spectrum. Typical Raman fitting demands a measurement time that, with this new methodology, is expected to be reduced by a factor of 15. This time reduction will further enhance the accuracy of the temperature mapping, as one source of variation stems from changes over time. In that way, RADIANT will open a new direction for ML-assisted Raman spectroscopy studies which has so far been largely confined to the analysis of structural properties.

This project brings together the expertise of the Thermal Properties of Nanoscale Materials group at ICN2 in experimental thermal characterization techniques and of the Simulation in Catalysis group at ICIQ in machine learning approaches. This work will set the base for future collaborations at advancing ML assisted Raman spectroscopy extending its application for the study of heat dissipation and exploring further opportunities in fields such as in medical diagnostics and food science.