ABSTRACTMajor developments in materials characterization instrumentation over the past decade have helped significantly to elucidate complex processes and phenomena connected with the microstructure of materials and interfacial interactions. Equally remarkable advances in theoretical models and computer technology also have been taking place during this period. These latter now permit, for example, in selected cases the computation of material structures and bonding and the prediction of some material properties. Two assessments of the state of the art of instrumental techniques and theoretical methods for the study of material structures and properties have recently been conducted. This paper will discuss aspects from these assessments of computational theoretical methods applied to materials. In addition, an approach will be presented which uses advanced instrumentation and complementary theoretical computational techniques in tandem in an effort to construct and verify hierarchies of models to translate engineering materials performance requirements into microscopic-level and atomic-level materials specifications (composition, structure, and bonding). Areas of practical interest include catalysis, tribology (contacting surfaces in relative motion), protective coatings, and metallurgical grain boundaries. A first attempt involving modeling of grain boundary adhesion in Ni3Al with and without boron additions will be discussed.
Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems
