RESEARCH

Molecular processes in aqueous solutions are essential for life and for energy applications. These phenomena result from correlated electronic and atomic dynamics and span a vast range of size and time scales that are hard to model even by harnessing the power of modern supercomputers. The new Computational Chemical Sciences Center Chemistry in Solution and at Interfaces (CSI) unites scientists at four universities (Princeton, Temple, Rutgers (Newark) and Stony Brook) to meet the challenge of simulations of molecular processes in solution. The new algorithms and software developed at the CSI Center will make use of modern artificial intelligence techniques to learn from accurate but costly quantum mechanical calculations the complex way in which the energy depends on atomic and electronic coordinates. The energy function constructed in this way will allow scientists to model and predict chemical and structural transformations occurring in liquid environments with an unprecedented level of accuracy using codes that run efficiently on the most advanced computational platforms, such as the Exascale computers that will soon be available at the DOE national laboratories. The new algorithms and methods will be used by the CSI scientists to study concentrated electrolyte solutions, and to model a variety of phenomena associated with the distribution in space and the dynamics of electronic and ionic charges in aqueous environments, such as photocatalytic interfaces, electrodes for desalination, organic interfaces and nanodevices based on biological molecules. The ultimate goal of this effort will be to model and predict how external fields drive chemical reactions. The algorithms and codes developed at the CSI Center will be collected in a toolkit that will be made accessible to the scientific community on an open source platform. The synergy between senior and junior investigators realized at the CSI Center will boost scientific progress and will create the optimal conditions to mentor a new generation of scientific leaders.