Raman

RamanFormer: A Transformer-Based Quantification Approach for Raman Mixture Components

ABSTRACT: Raman scattering can be employed as vibrational fingerprints of the molecules. Thus, materials can be identified by utilizing their Raman spectra. However, it is challenging to identify the components in a mixture spectrum due to the overlapping of individual spectra of components, especially if the components have similar spectral features. Moreover, quantification of the components is a more complicated problem, since individual spectra are included in the mixture spectrum at different rates for different wavenumbers. In this paper, we propose RamanFormer, a transformer-based method for the identification and quantification of Raman mixture components. Results indicate that our approach can capture intricate patterns in the Raman spectrum and outperforms other studies, where deep learning-based approaches and conventional methods are applied.

This project is funded by ASELSAN, their generous financial support and invaluable contributions have played a pivotal role in the successful execution of this research endeavor. The authors would like to thank Esra Ayantuna for the efforts on the early optical measurements and analysis, Bilal Kızılelma for the support on the sample preparations, and anonymous reviewers for improving the quality of the paper.