Chemometrics for complex data handling and process optimization in environmental sciences.

Scientists involved in environmental studies are called to investigate the effects of pollutants in environment compartments and on human beings. On one hand they are required to develop adequate environmental analytical methods, on the other hand they are charged with collecting and evaluating complex and often inexplicit data characterized by an extremely varying character. In this thesis successful applications of chemometric techniques in these field of environmental analysis are described. Chapter one focuses on the study of a complex groundwater system situated on the Lake Como area. Geochemical methods and exploratory Principal Component Analysis (PCA) were applied on the chemical data and their results were finally compared. In the second chapter the study concerning some of the main tropospheric pollutants in the Lake Como area from 1992 to 2007 is presented. The pros and cons of the use of unfolding PCA and multiway methods (Tucker3) were discussed. Chapter three and four show how PCA and PLS (Partial Least Square Discriminans Analysis) could be used in classification matters characterized by a limited samples number. Chapter three shows a new strategy for the unambiguous establishment of the link between the chemical and isotope fingerprint of historical copper mines and ancient copper artifacts. In chapter four multivariate analysis analysis was employed for the volatile profile discrimination of different Italian Pecorino cheese samples. Chapter five and chapter six show how multivariate regression could be employed in the process optimization using the “Experimental Design" strategy. The chapter five with the study of a new low cost analytical method for the platinum determination by Anodic Stripping Voltammetry (ASV) while the chapter six illustrates the optimization of a prevention strategy against Candida Albicans biofilm based on a molecule with antifouling effect.