From Discovery to Data: What Must Happen for Molecular Simulation to Become a Mainstream Chemical Engineering Tool

Molecular simulation is a vibrant and growing area of research within the chemical engineering community. Its use is pervasive within the academic community. The most common form of use is in discovery-driven research, where new phenomena are predicted and explained. To become a mainstream computational tool, however, non-expert users must be capable of carrying out data-driven molecular simulations to compute properties of interest or develop and test new engineering models. Such data-driven calculations are already routinely performed for quantum chemical simulations, and to some extent for molecular dynamics simulations in the biological community. Monte Carlo and molecular dynamics simulations are not yet a mainstream tool in the chemical field, however, because several barriers are preventing their adoption. The recommendations made here will help overcome these barriers. Many of these suggestions are already being undertaken by researchers around the world, but in an ad hoc manner. An agency or multinational body needs to help coordinate these efforts by establishing goals and guidelines, and providing funding to support the development of the tools, software and databases required. In the U.S., NIST is the most logical agency, given their mission of supporting commerce and their expertise in related matters. Other organizations such as the Centre Europeen de Calcul Atomique et Moleculaire (CECAM), or a consortium of companies and universities could also provide the necessary organization and coordination. When harnessed, the user community at large can accomplish much more than a single institution or company, as amply seen in the overwhelming success of open source and free software. Over the last 50 years, molecular simulation has grown into a powerful tool for discovery-driven research, and now makes up a significant fraction of the total research enterprise. These advances are the result of twin developments in methods and computing power. While these advances continue, the adoption of molecular simulation as a mainstream tool for data-driven research will require the integration of disparate methods and the harnessing of information technology tools such as databases and web-based applications. Ultimately this is an engineering problem, and chemical engineers are poised to play a key role in this challenge in the coming decade. © 2009 American Institute of Chemical Engineers.