Computer simulation of random and non-random second-phase distributions in two-phase materials

Abstract

The mechanical properties of any material with a discontinuous second phase dispersed in a matrix are recognized to be influenced by the distribution of the second-phase particles. Current models for the prediction of material properties from particle distributions are based on the assumption of a random particle distribution. Through computer simulation, nearest-neighbor particle spacings have been calculated for random and non-random distributions. For low fractions, random distributions approach the theoretical spacing predicted from consideration of random, infinitesimal points. For finite sized particles, increasing fraction results in larger spacings than predicted for infinitesimal points. For very high fractions, the spacing approaches that for regular (crystalline) arrays. Also, metal matrix composites initially possess clustered particle distributions. Upon processing, such distributions can be transformed into banded distributions with areas of both high and low density. With sufficient processing, random distributions can be attained. Non-random (banded) distributions were simulated. Sufficient banding results in reduced average particle spacing.