Discrete Atomic Simulation of Fracture in Iron

Abstract

<p>The thesis is concerned with improving the understanding of materials at the atomic level by computer simulation. The work has centered on three areas: a study of the efficiency of various computer algorithms used to carry out these simulations; the development of a new and very versatile boundary condition scheme for such problems; and the application of the results of these two studies to the simulation of (001) plane fracture in α-iron.</p> <p>Tests were performed comparing various solution methods used for atomic level computer modelling. The results are presented and indicate the most efficient method to be chosen for various problems. This may allow a reduction in computer cost by factors of two or three over that of other, often used, methods.</p> <p>The new boundary scheme which was developed involves the use of the finite element method. This offers several advantages over previous methods.</p> <p>The new boundary scheme was applied to the (001) plane crack in α-iron. Two dimensional cracks with crack line directions of [010] and [110] were modelled. Significant differences in lattice trapping limits and crack propagation speeds were observed between these two cases. Information on the magnitude and shape of the non-linear component of the crack tip displacement field is provided for both cases. No dislocations were emitted at low temperatures from these models, but warming the [010] model to 400 K apparently resulted in the emission of a dislocation from the crack tip. Experimental information available in the literature pertaining to these points is discussed and possible future work is described.</p>