Estudos dos métodos de análise multivariada em espectroscopia vibracional Raman

Abstract

The study of advance materials at the atomic scale has made it possible to understand certain phenomena that occur in this region and how they influence the macroscopic world. To achieve this understanding using techniques that facilitate and help in deepening this knowledge. One can hightlight three very spectroscopy methods disseminated for this purpose: infrared spectroscopy, luminescence spectroscopy and Raman spectroscopy. However the focus of the work lies on the Raman scattering spectroscopy. For experimental procedures it is common to have a large amount of data about the sample, therefore, we need techniques to help interpret these data. The equipment (LabRam) gives us four techniques Principal Component Analysis (PCA) Multivariate Curve Resolution (MCR), Hierarchical Clustering Analysis (HCA) and Divisive Clustering Analysis (DCA). Where the objective is to undertand and explain the first two methods applied to Raman spectroscopy. As for the measures are generated too many results (spectra), depending on what we are looking for we can use the PCA or MCR to treat these spectra. The treatment consists basically in algebra with matrices, which are determinated by the spectra. And as a result is termed some specific loadings spectra also from the treatment used to determine the scores are associated with loadings. Taking as an example the MCR by means of loadings which are obtained that method we determine wich components make up the region to delimit the sample we studied.