Chemometric applications for innovative analytical methods in food, environmental, and pharmaceutical analysis

This thesis develops and applies chemometric methods for exploratory analysis, supervised classification, quantitative calibration, mixture resolution, and experiment optimization across heterogeneous chemical datasets. The workflow integrates PCA for dimensionality reduction and diagnostics; LDA/PLS-DA for pattern recognition; PLS/SO-PLS/MLR and NAS-SAM for quantification; MCR-ALS for curve resolution under meaningful constraints; and Design of Experiments for factor screening and response optimization. The case studies span food, materials, and pharmaceutical contexts, with signals from GC-FID/GC-IMS, AF4–MALS–UV/Vis, NIR and ATR-IR spectroscopy, PXRD, and UV/Vis DRS, emphasizing low-waste, interpretable, and transferable procedures. On complex food fingerprints, headspace GC-based profiles enabled rapid screening pipelines for authenticity and class differentiation. For wines, AF4 fractograms, detected with MALS and UV/Vis, were unmixed by MCR-ALS into elution profiles and pure spectral/angle responses, and the resolved information improved subsequent PLS-DA discrimination among varieties. In honey, visible DRS and color features supported robust botanical classification with clear diagnostic structure. For quantitative problems, recycled-PET content in bottles was predicted from NIR and ATR-IR using PLS and SO-PLS, with performance stable in the central composition range and transparent error diagnostics near class boundaries. PXRD profiles were modeled via multivariate calibration to quantify solid solutions without full profile refinement, demonstrating a generalizable alternative when structural priors are limited. Finally, a Cu-MOF enhanced luminol–H₂O₂ chemiluminescence system was optimized by DoE, clarifying factor effects and interactions and yielding a calibrated response surface suitable for routine use. Overall, chemometrics converted instrumental signals into reliable decisions on real samples and produced reusable workflows that transfer across matrices, instruments, and objectives. This thesis is organized into seven chapters. The first provides a general overview of chemometrics and its most innovative applications. The following five chapters are dedicated to presenting the main tasks that chemometrics is capable of addressing, the last one draws the overall conclusions.