Tailored Force Fields for Modelling Transition-Metal-Mediated Reactions

Transition metals play key roles in diverse processes of enormous chemical, industrial and biological significance. The understanding, and hence control, of these processes requires an intimate knowledge of the reaction mechanism - the pathway joining reactants to products. The key feature of this pathway is the species sitting at the mountain pass - the transition state (TS). If you know the structure and energy of the TS, you can predict the outcome of the reaction.However, the TS is a transient, unstable species which is difficult to characterise experimentally. In contrast, the theoretical definition of a TS is straightforward. It is just another point on the potential energy (PE) surface.Unfortunately, whilst the definition may be easy, locating the TS is much trickier but since the TS is so important, theoretical chemists are increasingly devoting themselves to developing new and better ways for finding them.Given that TSs often involve bond making/bond breaking processes, theoretical modelling has usually focussed on fully quantum methods which can treat such processes properly. However, all quantum chemical methods are relatively slow which severly limits the size of molecular system we can study in a reasonable timeframe. The alternative is to use fast classical modelling methods like molecular mechanics providing you can solve the issue that MM is not normally supposed for TS location and, in the context of this proposal, methods specifically targeted at transition-metal-mediated processes. However, the very features that make transition metal (TM) species ideal catalysts, also make them hard to calculate. The current proposal describes an alternative empirical approach: seam searching. The seam describes the intersection of the PE surfaces describing the reactant and product states. The enormous power of the approach is that:1. seam searching is minimisation: all the standard conformational search tools can be employed2. the seam depends only on a reasonable model for the ground states of the reactant and product3. the lowest energy point on the seam is not sensitive to the starting geometry4. the lowest energy point on the seam is a good approximation to the true TSSeam searching thus represents a powerful, general alternative to quantum chemical methods and represents the next leap forward in molecular modelling.