Computational materials design

Full text: Trial and error experimentation is an extremely expensive route to the development of new materials. The coming age of reduced defense funding will dramatically alter the way in which advanced materials have developed. In the absence of large funding we must concentrate on reducing the time and expense that the R and D of a new material consumes. This may be accomplished through the development of computational materials science. Materials are selected today by comparing the technical requirements to the materials databases. When existing materials cannot meet the requirements we explore new systems to develop a new material using experimental databases like the PDF. After proof of concept, the scaling of the new material to manufacture requires evaluating millions of parameter combinations to optimize the performance of the new device. Historically this process takes 10 to 20 years and requires hundreds of millions of dollars. The development of a focused set of computational tools to predict the final properties of new materials will permit the exploration of new materials systems with only a limited amount of materials characterization. However, to bound computational extrapolations, the experimental formulations and characterization will need to be tightly coupled to the computational tasks. The required experimental data must be obtained by dynamic, in-situ, very rapid characterization. Finally, to evaluate the optimization matrix required to manufacture the new material, very rapid in situ analysis techniques will be essential to intelligently monitor and optimize the formation of a desired microstructure. Techniques and examples for the rapid real-time application of XRPD and optical microscopy will be shown. Recent developments in the cross linking of the world's structural and diffraction databases will be presented as the basis for the future Total Pattern Analysis by XRPD. Copyright (1999) Australian X-ray Analytical Association Inc