On the characterization of size and shape of irregular particles

Abstract

The size and shape characterization of irregular particles is a key issue in many fields of science, which is often associated with large uncertainties. We assess existing protocols and introduce new strategies for the study of size and shape of irregular particles by performing a comprehensive characterization of 127 volcanic clasts with diameters between 155μm and 37mm. Methods include caliper measurements, image analysis, laser scanning and scanning electron microscope micro-computed tomography. Volume, surface area and various shape descriptors including form factors (e.g. flatness, elongation), circularity measures and sphericity are analyzed. First, existing procedures commonly applied by caliper and image analysis to determine 1D (i.e. particle lengths in three dimensions) and 2D variables (e.g. particle projection perimeter and area) have been revised. A new procedure based on particle projection area (PA protocol) for measuring particle lengths in three dimensions is also proposed that is associated with the lowest operator-related errors with respect to existing protocols. In addition, the effect of number of particle projections on the variables obtained through image analysis is investigated. It was found that two to three perpendicular projections can be used to characterize 2D variables with a maximum error of <10%. Second, 1D and 2D variables calculated based on the new PA protocol and image analysis are used to derive shape descriptors and investigate their variability and correlations. Finally, both existing and new empirical correlations for the estimation of 3D particle parameters (i.e. volume, surface area, sphericity) based on 1D and 2D variables are presented and benchmarked. It was found that correlations that are based on 2D variables are associated with the lowest average error (~2.6–4.6%).