Crystallite thickness and defect density of phyllosilicates in low-temperature metamorphic pelites: a TEM and XRD study of clay-mineral crystrallinity index standards

Abstract

The thickness distribution and defect density of illite crystallites (both illite–smectite and illite–muscovite phases) and chlorite are presented for a set of clay-mineral crystallinity-index standards, as determined by TEM and XRD methods for both the whole rock and disaggregated clay-sized fractions. The pelitic standards, selected from the Variscan rocks of southwestern England, display the full array of typical microstructural transformations that characterize the prograde transition from diagenesis (zeolite facies) to epizonal (greenschist facies), with increasing thickness of crystallites and decreasing concentration of lattice imperfections. Crystallite thickness, measured from TEM lattice-fringe images in the c* direction of both illite and chlorite, commonly deviate from lognormal distributions, with positive skewness toward small sizes. These distributions are considered to reflect complex histories of nucleation and growth, as well as structural modifications due to rock strain. Crystal damage caused by rock disaggregation, in the form of splitting along crystallite boundaries and other planar defects, can be observed in all clay-sized separates. The degree of damage increases as a function of increasing size of crystals, and clearly enhances the differences in mean thickness of the crystallites between the whole rock and clay-sized fractions, particularly in anchizonal and epizonal slates. Although clay separates yield smaller mean thicknesses of crystallites and show thickness distributions with lower degrees of skewness and standard deviation than measured from matrix of the pelitic rocks, the aaa and bbb2 shape parameters (the mean and variance of the natural logarithms of the sizes) for both grain-size fractions define a common crystal-growth path when plotted in aaa – bbb2 space. This feature shows that ultrasonic disaggregation has not modified the shape of the thickness distributions in a way that has influenced the interpretation of the mechanisms of crystal growth and deformation. NEWMOD modeling of XRD profiles based on TEM constraints show a relatively good correspondence between the size of X-ray-scattering domains and crystallite thicknesses in most cases. XRD-determined area-weighted mean thicknesses of crystallites calculated using an integral peak-width method yield sizes that correspond to those measured directly from TEM measurements. However, the refined Warren–Averbach method of MUDMASTER, adopted for the analysis of PVP-saturated clays, produced profiles that do not match well the TEM-determined distributions. Refinements to the method are required before accurate thickness-distributions can be extracted from the narrower XRD reflections of metamorphic phyllosilicates.