Correlation between chemical and mineralogical characteristics and permeability of phyllite clays using multivariate statistical analysis

• Application of multivariate statistical analysis to the study of a group of phyllite clays
• Correlations have been found between chemical and mineralogical characteristics of phyllite clays and permeability.
• Deduction of regression equations from the results with high confidence levels

Phyllite clays are applied as a layer on a surface to be waterproofed and subsequently compacted. For this purpose, phyllite clays deposits can be grouped by their chemical and mineralogical characteristics, and these characteristics can be connected with their properties, mainly permeability, in order to select those deposits with the lowest permeability values. Several deposits of phyllite clays in the provinces of Almería and Granada (SE Spain) have been studied. The results of applying a multivariate statistical analysis (MVA) to the chemical data analysed from 52 samples determined by XRF, mineralogical analysis by XRD and permeability are reported. Permeability, a characteristic physical property of phyllite clays, was calculated using the results for experimental nitrogen gas adsorption and nitrogen adsorption-desorption permeability dependence. According to the results, permeability values differentiated two groups, i.e. group 1 and group 2, with two subgroups in the latter. The influence of chemical as well as mineralogical characteristics on the permeability values of this set of phyllite clays was demonstrated using a multiple linear regression model. Two regression equations were deduced to describe the relationship between adsorption and desorption permeability values, which support this correlation. This was an indication of the statistical significance of each chemical and mineralogical variable, as it was added to the model. The statistical tests of the residuals suggested that there was no serious autocorrelation in the residuals.