The dominant factor in determining the atomic structure of grain boundaries is the crystal structure of the material, e.g. FCC vs. BCC. However, for a given crystal structure, the structure of grain boundaries can be influenced by electronic effects, i.e. by the element comprising the crystal. Understanding and modeling the influence of electronic structure on defect structures is a key ingredient for successful atomistic simulations of materials with more complicated crystal structures than FCC. We have found that grain boundary structure is a critical test for interatomic potentials. To that end, we have fabricated the identical x5 (3 lO)/[OOl] symmetric tilt grain boundary in three different BCC metals (Nb, MO, and Ta) by diffusion bonding precisely oriented single crystals. The structure of these boundaries have been determined by high resolution transmission electron microscopy. The boundaries have been found to have different atomic structures. The structures of these boundaries have been modeled with atomistic simulations using interatomic potentials incorporating angularly dependent interactions, such as those developed within Model Generalized Pseudopotential Theory. The differing structures of these boundaries can be understood in terms of the strength of the angular dependence of the interatomic interaction. We report here the results for Ta.
INTRODUCTIONAtomistic simulations are an increasingly important means of understanding the behavior of materials under a variety of conditions. With this technique, an assembly of thousands, or even millions, of atoms is defined in a computer simulation and allowed to interact according to certain rules and boundary conditions. The boundary conditions include temperature and states of stress, allowing the calculation of such properties as the equation of state or unstable stacking fault energy. The structure and properties of crystal defects can also be predicted, such as the stressed configuration of a dislocation core or the interaction energy of an interstitial with a vacancy (see e.g. [l]). The rules of interaction are often very simple in order to speed computation. This simplification requires approximations to be made about the physics of the interacting atoms. Hence, in the development of models of interatomic interactions, an evaluation is necessary of whether the essential physics have been incorporated in the model. The models are tested by comparing their predic'tions with experimental observations.Receqtly developed models of interatomic interactions incorporate angularly dependent contributions to model materials with directional bonding [2-71, such as the body centered cubic transition metals in which the d -bands participate in bonding. The strength of the directional component of the bonding has a major influence on the structure of crystal defects. The model of interatomic interactions with angular dependence that we use is the Model Generalized Psuedopotential Theory (MGPT) [5]. We have applied it to modeling the x5 (3 1 O)/[OO 1 ] symm...