Molecular-modeling Methods and Use for Product and Process Design

Significant progress has been made in recent years on development of molecularly based methods for accurate prediction of chemical and physical properties relevant for product and process design. A number of ideal-gas properties are routinely predictable. However, despite progress in computing hardware and simulation methodologies, prediction of condensed-phase properties of interest to industrial applications is not currently practiced on a routine basis. Present status of theory, molecular models, and industrial practice will be described. Molecular simulation refers to computational statistical mechanics based on force fields. Most existing force fields have been optimized to the configurational properties of isolated molecules and thermodynamic or structural properties of liquids near room temperature. Recently developed force fields have become available that also reproduce phase coexistence properties and critical parameters for selected systems, but they are not yet generally applicable to many systems of interest. Simulation methodologies for rapid determination of intermolecular potential parameters from experimental data are discussed. Two key unresolved questions remain, namely how to incorporate polarizability and other non-additive interactions, and the logistics of large-scale efforts to obtain parameters for broad classes of components. Computational quantum chemistry does not require predetermined force fields. Instead, it predicts energy and related properties from the nuclear (geometric) structure and the electronic orbital structure using the Schrödinger equation. Solution time and storage requirements increase rapidly with the number of atoms, but high accuracy can be achieved. At present, most results are for the zero-K, isolated-molecule (ideal-gas) case, but methods are advancing rapidly, including for solvation.