The growing importance of computations in materials science. Current capabilities and perspectives

Materials scientists are facing unprecedented challenges in many areas, such as energy conversion and storage, microelectronics, telecommunication, display technologies, catalysis, and structural materials. Experimental methods generate increasing amounts of data. New computational methods, highperformance computer hardware, and powerful software environments are evolving rapidly. As a result, the importance of computational materials science is growing. The following cases illustrate the current capabilities: computed thermochemical and mechanical properties of metal hydrides show trends in the heats of formation and the hydrogen-induced softening of elastic moduli; a study of the effect of impurities on the strength of a Ni grain boundary reveals hydrogen as an embrittler and boron as a strengthener; ab initio phonon calculations for hydrogen impurities in aluminum show a temperature-dependent site -preference; the screened-exchange approach predicts accurate energy band gaps of semiconductors; a computational screening of hydro-desulphurization catalysts points to new combinations. The major current challenges for computational materials science include more accurate total energies, unified methods to deal with multi-phase systems, e.g., solid/liquid, novel approaches to determine complex kinetic processes, and novel concepts to bridge the atomistic and the macroscopic scales.