Presentation description

Temperature-dependent Raman spectroscopy, otherwise known as Raman thermometry, is a valuable characterization technique for analyzing thermal properties in nanostructured and 2D materials. It is an optical, non-destructive characterization technique that provides an alternative to traditional methods that suffer from size effects, thermal rectification, and ballistic transport at the smaller scale. Raman thermometry has gained popularity for its excellent spatial resolution and lack of contact error; however, Raman thermometry remains limited by its temperature resolution. Our work focuses on improving temperature resolution by training neural networks to simultaneously track the locations, widths, and heights of multiple Raman peaks within a material's spectra to reinforce temperature predictions. Additionally, we employ image processing techniques to assess the laser spot diameter, necessary for thermal conductivity calculations.