3D grain growth in nanocrystalline Al via molecular dynamics: Influence of size, and mean curvature on kinetics

Grain growth in nanosized polycrystalline Al was studied via molecular dynamics. In particular, the volumetric growth rate of grains as a function of size, topology and mean curvature was analyzed. To this aim, an algorithm for the identification of the topological features and calculation of the mean curvature was developed. It was found that in average grains with approximate to 15 faces have zero integral mean curvature and show no change in volume implying that the critical number of face (Fcr) is approximate to 15. Furthermore, integral mean curvature of grains was found to be linearly correlated with the difference between the number of faces grains have and the average number of faces of their neighbors, F- < FNN >. That is, a grain with F- < FNN >= 0 tends to have zero integral mean curvature and neither shrinks nor grows whereas grains with F- < FNN > greater than 0 grow and those with F- < FNN > less than 0 shrink. This agrees with the theoretical expectations from MacPherson-Srolovitz relationship and topology based descriptions of grain growth. Nonetheless, considering individual grain kinetics, the comparison of theoretical predictions (MacPherson-Srolovitz relationship, F- < FNN > and topological class) with measured growth rates showed more scatter. In addition, the deviation between the models and the measure values occurred for grains having a substantial relative size difference with the average grain size (|Rgrain- < R > |) and grains with relatively high |F - Fcr|. These grains also had average turning angles at triple lines (13TJ), which differed from the average equilibrium value expected in isotropic metals, rr/3, but were consistent with grain growth under finite triple line mobility. This seems to confirm the existence of triple junction drag at the nanocrystalline length scale, which is not accounted for in classical theoretical models.