Kinetic model of nucleation and growth in silicate melts: Implications for igneous textures and their quantitative description

Abstract

We present a new high-resolution numerical model for the simulation of crystallization and texture evolution using arbitrary rates of crystal nucleation and growth. The algorithm models single or multiphase solidification in a three-dimensional domain and 17 simulations using constant, linearly increasing, exponential, and Gaussian functions for the rates of nucleation and growth yield equigranular to seriate textures. Conventional crystal size distributions of all textures are nearly linear to concave-down (previously interpreted as formed by equilibration coarsening), and identical distribution patterns can result from multiple non-unique combinations of nucleation and growth rates. The clustering index is always a non-monotonous function, which initially increases then decreases with increasing crystal fraction. For texture from random homogeneous nucleation the index is substantially lower than previous predictions based on a random sphere distribution line, hence, natural samples interpreted as clustered now have greater degrees of randomness or ordering. The average number of contact neighbors and the average neighbor distance of a crystal depend linearly on crystal size, but one of the two remains insensitive to nucleation and growth kinetics and represents potential indicator of other crystallization processes than random nucleation and crystal growth. Simultaneous comparison of size, spatial and clustering patterns and of their departures from expected values are suggested to allow for separation of effects of crystallization kinetics, melt–mineral mechanical interactions, suspension mixing, or postcrystallization re-equilibration and coarsening on natural igneous rocks.