Condensed matter theory has provided us with powerful equations to predict the properties of solids. However, even approximate solutions can be found only for relatively simple systems. The distance to real materials science is huge. For example, there is currently no recipe available to predict the crystal structure of a solid given its chemical composition, let alone its properties. The presence of lattice defects and other nonequilibrium conditions, crucial for materials behavior, makes matters even worse. It is in this area that computational approaches have become of great help.
This talk will highlight the current power of a few effective methods of computational materials science: density functional theory, data mining, atomic environment methods, and molecular dynamics. Several illustrations of new insights obtained by these methods will be given, including alloy design, thin film formation, and nanowire failure.
