The contribution of computational studies to organometallic catalysis: descriptors, mechanisms and models.

With the growing availability of cheap and fast computing resources, increasingly detailed and sophisticated computational studies of homogeneous organometallic catalysis have become feasible. In addition, databases of calculated descriptors can be used to analyse and interpret the key factors contributing to a wide range of experimental observations, facilitating the development of novel catalysts. Such descriptors, most frequently used to capture ligand properties, can also be used to quantify catalyst similarities and derive maps of ligand space which guide experimental screening. Recent publications have shown that descriptor databases can be combined with calculated mechanistic information to gain more detailed insight into catalytic cycles, useful for the design of catalysts. This perspective seeks to describe the contribution of computational studies to the field of homogeneous organometallic catalysis, and to summarise the challenges and opportunities for the future of computational prediction and evaluation of catalytically active transition metal complexes.