Application of the Method of Statistical Comparison of XRD-and XRF-Data for Identification of the Most Representative Rock Samples: A Case Study of an Extensive Collection of Carbonatites and Aluminosilicate Rocks of the Kontozero Alkaline Complex (Kola Peninsula, NW Russia)

Abstract

Abstract We investigated carbonatites and aluminosilicate rocks from the Kontozero Devonian carbonatite paleovolcano complex (198 samples). Some specific features complicate the geological exploration of this object: a) the rocks of the Kontozero complex are predominantly volcanic and therefore exhibit small dimension of mineral grains and the diversity of their structural relationships; b) because breccias are common in many parts of the complex, the rocks are mostly inhomogeneous; c) Kontozero belongs to the alkaline-carbonatite formation, which is typically characterised by mineral diversity and the presence of rare minerals. The purpose of this study was to develop an algorithm for selecting from an extensive collection of rock material the most informative for mineralogical and geochemical studies samples. As a tool for this selection, we chose an original method of statistical comparison of XRD and XRF data using factor analysis (FA). This methodological approach enables mathematical identification of all major, minor, and several accessory minerals and a rough estimation of their contents (Fomina et al., 2019). We carried out the mineralogical interpretation of the factors according to the peak positions on the graphs of factor loadings and qualitative analysis of diffraction data of rock samples with maximum factor scores. For the studied rock collection, this approach allowed us to identify more than 20 rock-forming minerals based only on XRD data. Also, we found about ten mineral phases, the lines of which are low-intensity and/or overlap by more intense peaks of other minerals in the diffraction patterns. The mineralogical interpretation of the factors of these hidden minerals requires verification by an electron microscope investigation of the samples selected with FA. Based on the results of this study, we developed an algorithm that facilitates choosing rock samples that are most contrasted in mineral and chemical composition and contain the entire set of mineral phases characteristic of the rocks of the study object with minimal effort. In the case of our Kontozero rock collection, 30 representative samples were selected, which is about 15% of the initial sample set.