Understanding structure and bonding in transition metal hydrides
Abstract
Transition metals hydrides have long been important objects in chemical research, because of their relevance to many catalytic industrial processes as well as their fundamental significance for the theory of chemical bonding.
Iridium hydrides have been applied particularly as highly efficient catalysts for the dehydrogenation of alkanes to alkenes where we have contributed to catalyst development. For iridium hydrides there are several aspects that has not been fully studied including the nature of the bonding where there are conflicting reports between solution and solid state studies. The hydride systems are intermediate between classical and non-classical and to fully characterize them it is necessary to accurately determine the distance between the hydrogen atoms, which reflects the type of bonding that the hydrogens display towards the metal. X-ray diffraction rarely locates hydrides precisely primarily due to the low electron density on hydrogen atoms to scatter X-rays; this is particularly true when hydrogens are located close to electron rich atoms such as iridium. Several solution-state techniques have been developed that correlate H–H distance with NMR spectroscopy properties of compounds such as T1-times and 1J coupling constants. In the last year we have used neutron diffraction to accurately determine distances in iridium dihydrides and we have found that distances are substantially shorter than predicted from NMR spectroscopy.
We are now working towards a unified picture in solution and solid state and towards better models that can transfer solution-state data to reliable H–H distances. To do so, theoretical calculations are key since they will help us understand how distances translate into 1J coupling constants, both of which can be computed using DFT. However, difficulties are met at the geometry optimisation stage, since different DFT functionals produce very different geometries in terms of H-H distances. In order to be able to determine which DFT functional to use, a benchmark must be performed by using advanced double hybrids and WFT methods ( for example DLPNO-CCSD(T)) to map PES of H-H distances.
