Abstract: Finding new materials with novel properties is one of the "grand challenges" of science. Examples abound: The discoveries of materials with special properties such as high temperature superconductors and giant magnetoresistive materials have been recognized by Nobel Prizes. The discovery of special catalytic, superhard, photovoltaic or "green magnetic" materials (to name only a few) would surely have profound scientific and technological significance. Unfortunately, finding new materials is in general tedious and laborious at best. Consider the number of yet to be discovered materials. While the number of experimentally known binary materials is fairly complete, of the roughly 160,000 possible ternary materials only about 5% are known and of the possible 4 million quaternary materials less than 1% are known. The potential for the discovery of a new material in this realm is great, but an efficient search in this myriad of possible combinations is a daunting task. A successful search using the usual "trial and error" method all too often relies on an enormous dose of luck and serendipity.
Research in this grand challenge area will focus on developing computational tools that will expedite the discovery and design of materials. In particular, efficacious searching procedures coupled with computational methods to evaluate the properties of a candidate material could place theoretical methods on par with experiment. Materials scientists would be able to examine postulated materials on a routine basis and predict their properties without resort to experiment. Synthesis of novel materials would be greatly enhanced and inefficient search activities via experimental means would be eliminated, i.e., the Edisonian trial and error procedure used in designing and discovering materials would become a paradigm of the past.