Shape, electronic structure and steric effects of organometallic nanocatalysts: relevant tools to improve the synergy between theory and experiment.

Working closely with experimentalists on the comprehension of the surface properties of catalytically active organometallic nanoparticles (NPs) requires the development of several computational strategies which significantly differ from the cluster domain where a precise knowledge of their optimal geometry is a mandatory prerequisite to computational modeling. Theoretical simulations can address several properties of organometallic nanoparticles: the morphology of the metal core, the surface composition under realistic thermodynamic conditions, the relationship between adsorption energies and predictive descriptors of reactivity. It is in such context that an integrated package has been developed or adapted in our group: (i) one tool aims at building a wide variety of the typical shapes exhibited by nanoparticles. Using Reverse Monte Carlo modeling, a given shape can be optimized in order to fit pair distribution function data obtained from X-ray diffraction measurements; (ii) trends in density functional theory (DFT) adsorption energies of surface species can be rationalized and predicted by making use of simple descriptors. This is why we have proposed an extension of the d-band center model, that leads to the formulation of a generalized ligand-field theory. A comparison between cobalt and ruthenium is proposed in the case of a 55-atoms nanocluster. The accuracy of the generalized coordination number [Angew. Chem., Int. Ed., 2014, 53, 8316], a very simple coordination-activity criterion, is also assessed; (iii) the builder package is completed by the steric-driven grafting of ligands on the surface of metal NPs. It easily generates structures with adjustable surface composition values and coordination modes; (iv) after a local optimization at the DFT level of theory, DFT energies and normal modes of vibration can feed a general tool based on the ab initio thermodynamics method. This method aims at easily calculating an optimal surface composition under realistic temperature and pressure conditions. In addition to that, we also show to what extent knowledge of the density of states (DOS) and of the crystal overlap Hamilton population (COHP), both projected from a plane-wave basis set to a local basis set, sheds light on metal core-ligand chemical bonding.