The atomic structure of a large angle [001] twist boundary in gold determined by a joint computer modelling and X-ray diffraction study

Abstract

Experimental techniques involving X-ray diffraction have been developed which provide information, through the structure factor, on the local atomic arrangements in the vicinity of large angle grain boundaries. Theoretical techniques involving computer modelling have also been developed to determine directly the structure of grain boundaries, but which rely to a large extent on the chosen interatomic potential and computational procedure. In order to obtain the actual grain boundary structure a joint study was undertaken involving both X-ray diffraction and computer modelling techniques. By comparing the computed structure via the structure factor to X-ray observations on the same boundary it is possible to obtain detailed information on the boundary structure and also to check the validity of the computer calculation. X-ray experiments were performed on the θ − 22.6° (∑ = 13) [001] twist boundary in gold. The atomic structure of the ∑ = 13 boundary was simulated using potentials representing four f.c.c. metals and was found to be strongly potential dependent. The correlation between the computer generated structure obtained using one particular gold potential and the X-ray observations was excellent. Additional X-ray diffraction information on a θ = 23.8° boundary was also correlated with the computer generated structure. The most striking feature of this structure is the puckered nature of the atomic planes in the vicinity of the boundary, where within a plane one atom which is nominally in ‘good’ match is displaced a large amount toward the boundary, while the remaining atoms are displaced a smaller amount away from the boundary. The distortions in the regions of ‘good’ match are only slightly smaller than those in the regions of ‘bad’ match. Since this structure is the result of both computer modelling and X-ray diffraction observations, considerable confidence can be placed in it being an accurate representation of the actual boundary structure in Au. It is believed that progress in the immediate future in determining the structure of grain boundaries will be made by joint studies, such as the one described here.