To See the World in a Grain of Sand: Recognizing the Origin of Sand Specimens by Diffuse Reflectance Infrared Fourier Transform Spectroscopy and Multivariate Exploratory Data Analysis

The experiment outlined in this paper exploits basic multivariate analysis (MVA) to identify in a number of diffuse reflectance infrared Fourier transform (DRIFT) spectra those corresponding to samples with similar compositions. The object of this work is the spectroscopic study of sand samples from different locations, whose DRIFT spectra exhibit features reflecting their composition. MVA has been shown to effectively sort subsets of homogeneous specimens collected in the same area, as well as pointing out similarities in composition among sands of different origin. This experiment is designed for graduate students and provides a practical application of both solid state spectroscopy and chemometrics. It is helpful to students already acquainted with MVA and willing to develop some spectroscopic skills or merely aiming to collect some data sets to work with. Conversely, the technique can be used by students familiar with spectroscopy who wish to investigate multivariate exploratory data analysis. The spectra can also illustrate a discussion of the spectroscopy of minerals (ad-hoc flow-charts for the identification of the main constituents of sands through their spectral features are proposed).