It was recalled that a recent experimental measurement of the width of a twin wall had suggested that the wall-width could vary considerably as a function of position. This contradicted a previously-held assumption that the wall-width was a constant, having the same value in adjacent walls and in different parts of the same wall. The possibility was explored that this variation in wall-width was the result of an interaction of the order parameter with point defects. This was done in the context of a 2-dimensional model, on a square grid, in which a continuous order parameter interacted with a lattice of Ising spins which represented the presence or absence of point defects. It was shown that the key to reproducing the experimentally observed range of wall-widths involved 2 interactions. Firstly, an interaction between the order parameter and the point defects had to cause the accumulation of point defects within the wall. Secondly, an interaction between point defects had to cause them to form clusters. When both of these effects were present, the simulations produced distributions of wall-widths which were similar to those observed experimentally.

Influence of Point Defects on the Distribution of Twin Wall Widths. W.T.Lee, E.K.H.Salje, U.Bismayer: Physical Review B, 2005, 72[10], 104116 (4pp)