Development and implementation of chemometric and spectroscopic methods for process studies, with particular reference to the Near-Infrared radiation.

Near-infrared spectroscopy (NIRS) is widely used in industry for safety and quality control due to its rapid, non-destructive, and accurate analysis. However, challenges arise from overlapping absorption bands and background noise, necessitating the use of chemometric techniques such as principal component analysis (PCA) and partial least squares (PLS) regression. PCA simplifies spectral data while PLS regression builds predictive models to estimate component concentrations. This study integrates chemometric models with NIRS in the plastic additives (PA) industry, where PA enhance plastic properties like durability and thermal stability. Using PLS regression, eight offline and two online multivariate models were developed for quality control and real-time monitoring, achieving high R² values (>99%) and robust RPD indices (>10), reducing material consumption and enabling quantification of difficult analytes. Given NIRS's compatibility with online processes, models were also applied in polyurethane (PU) production. Specifically, a two-tiered NIR model quantified water content in the A-Component, achieving R² values of 99.8% and 99.7%, and another model monitored the free NCO number in the polyaddition reaction with R² of 99.7%. Additionally, quantum chemical calculations using density functional theory (DFT) and vibrational perturbation theory (VPT2) were conducted to simulate NIR spectra, aiding in spectral feature assignment for molecules like DEHA, DEAA, HOP, and TEMPO-H. These findings enhance the understanding of spectral transitions and support further investigations into PA and PU analysis.