Multiscale computational frameworks for vibrational spectroscopy: integrating first-Principles DFT and finite element modeling.
Computational tools are essential for investigating physical and chemical phenomena that are challenging for experimental characterization, particularly in the prediction and interpretation of vibrational spectra. This thesis presents two complementary computational frameworks with direct relevance to environmental monitoring and human health: microplastic analysis and the optimization of Surface-Enhanced Raman Spectroscopy (SERS) substrates.For the microplastic study, density functional theory (DFT) utilizing the BLYP exchange-correlation functional and a Gaussian plane-wave (GPW) basis set was employed to predict the infrared (IR) spectra of six representative polymers. These theoretical results showed high agreement with experimental ATR-FTIR data collected from beach sediments in Bali, Indonesia. Multivariate statistical techniques, specifically principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA), identified polypropylene and polyvinyl alcohol as the dominant pollutants and established relevant spectral markers for environmental identification. Concurrently, the finite element method (FEM) was used to simulate electromagnetic scattering in realistic morphologies derived from atomic force microscopy (AFM) images. These simulations quantified electromagnetic enhancement factors and identified "hot-spot" regions associated with localized surface plasmon resonance (LSPR), with results validated against Mie theory and experimental measurements. Collectively, these methodologies establish a robust framework for both (IR) and (SERS) vibrational characterization, advancing the design of high-performance sensors for environmental monitoring and medical diagnostics.