Multivariate analysis in the study of the chemical processes

The classical problem of the determination of the spectral components from a data set is very often a badly conditioned problem without univocal solutions. Very often one deals with experimental data which are a superposition, or mixture, of the pure spectra of the individual components and their associated proportions. When dealing with evolving systems such as chemical reactions several compounds are coexisting and if one collects UV-Vis or any other kind of absorption spectra they will be the overimposition of the singles spectra associated with the chemical species present in the reaction bath with an intensity that depends on their concentration profiles. In this case it is very important to be able to extract each single component that is the absorption spectrum associated with each species, and the concentration evolution with time. More than often, this decomposition is aimed at situations for which little a priori information is available. Several theoretical approaches have been developed to achieve this goal and the multivariate analysis (MA) is one of the most promising approach. The multivariate curve resolution (MCR) is the generic denomination of a family of MA methods meant to solve the mixture analysis problem that is able to provide a chemically meaningful additive bilinear model of pure contributions from the sole information of an original data matrix including a mixed measurement. MA methods are key-tools in order to extract the pure component information (pure component spectra and the concentration profiles) from the chemical mixture (spectroscopic) data. The problem is to compute: 1) the number of independent components s and 2) the pure component factors C (concentration profiles) and A (spectra). An actuality of the problem of the determination of the spectral components from a chemical-physical data is driven by its broad range of applications in situations where a reasonable approximation of the bilinear model, or any other fundamental basic equation that has the same mathematical structure, holds. Also, the bilinear model can be extended for the analysis of multiple data sets that are meant to connect different experiments together. The main goal of the thesis is to develop and apply an innovative MCR approach for the multivariate analysis of chemical systems. In particular, this method has been developed to have an ad hoc tool to be used for chemical reactions for which a combined UV-Vis and X-ray absorption (XAS) investigation is carried out. The final goal is to extract the UV-Vis and XAS spectra associated with all the intermediates that are formed during bimolecular chemical reactions occurring in the ms time range. In particular, in the first part of the thesis I have implemented the non negative matrix factorization (NMF) method to a newly written code, PyFitIt, that has been specifically developed for the multivariate analysis of XAS spectra. This method has been then applied to different sets of experimental data collected on different bimolecular chemical reactions in solution for which UV-Vis and XAS spectra were collected simultaneously. By using this new approach it was possible to extract, from the initial experimental data sets, both the XAS spectra and the concentration profiles of the intermediate species that were formed during the reaction. From the analysis of the X-ray absorption near edge structure (XANES) spectra it will be possible to determine the oxidation state of the absorption element, the nature of the short-lived intermediates that are formed during the reaction, and their three-dimensional structure with a picometric accuracy. To test the reliability of the entire procedure ten different systems have been investigated and the principal components and concentration profiles have been successfully extracted for all the investigated systems. The new procedure and code that has been developed in this thesis will be a valid tool that can be applied to different XAS data sets in the future.