Numerical simulations for industrial chemical research

Abstract

Large-scale numerical simulations are rapidly becoming an integral part of industrial chemical research in areas such as pharmaceuticals, agrochemicals, polymeric materials, and catalysis. The successess of this development hinges on the ability to calculate and accurately predict atomicscale properties for systems which are large enough to capture the complexity of technologically relevant compounds and materials. In addition, these simulations have to be performed within a reasonable time frame to have an impact on the industrial research activity. These three requirements, accuracy, complexity, and speed present a tremendous challenge for the theoretical formulation, the algorithmic choices, and the computational implementation. On May 24, 1991 Erich Wimmer from Cray Research reviewed the concepts, capabilities and present limitations of large-scale numerical simulation methods. He discussed highly accurate ab-initio and density functional methods, semiempirical quantum mechanical methods, and more approximative force-field approaches, which are especially useful in the study of the dynamics of proteins and DNA. He concluded with an outlook on future theoretical approaches and computational implementations.