A new perspective on self-similar grain growth

Mean-field theories of grain growth have been established for more than 70 years. They rely on the concepts of average grain size and self-similar size and edge distributions. In contrast, mesoscopic computer simulations enable the observation of a large number of grains over long annealing time spans. Thus, they allow the detailed surveillance of the behavior of individual grains. Hence, while the average grain size is an important construct for mean-field theories, using computer simulations we show that individual grains with the same number of edges from one polycrystalline structure evolve by distinct paths. As a result, the self-similar grain size and number of edges distributions form in a complex manner.