We report a comprehensive density functional theory (DFT) study of the mechanism of the methanol dehydrogenation reaction catalyzed by [RuH 2 (H 2 )(PPh 3 ) 3 ]. Using the B97-D dispersion-corrected functional, four pathways have been fully characterized, which differ in the way the critical β-hydrogen transfer step is brought about (e.g. by prior dissociation of one PPh 3 ligand). All these pathways are found to be competitive (∆G ‡ = 27.0 to 32.1 kcal/mol at 150 °C) and strongly interlocked. The reaction can thus follow multiple reaction channels, a feature which is expected to be at the origin of the good kinetics of this system. Our results also point out to the active role of PPh 3 ligands, which undergo significant conformational changes as the reaction occurs, and provide insights into the role of the base, which acts as a "co-catalyst" by facilitating proton transfers within active species. Activation barriers decrease on going from methanol to ethanol and isopropanol substrates, in accord with experiment.

We report molecular dynamics (MD) simulations of the aqueous interface of the hydrophobic [BMI][Tf2N] ionic liquid (IL), composed of 1-butyl-3-methylimidazolium cations (BMI+) and bis(trifluoromethylsulfonyl)imide anions (Tf2N-). The questions of water/IL phase separation and properties of the neat interface are addressed, comparing different liquid models (TIP3P vs TIP5P water and +1.0/-1.0 vs +0.9/-0.9 charged IL ions), the Ewald vs the reaction field treatments of the long range electrostatics, and different starting conditions. With the different models, the "randomly" mixed liquids separate much more slowly (in 20 to 40 ns) than classical water-oil mixtures do (typically, in less than 1 ns), finally leading to distinct nanoscopic phases separated by an interface, as in simulations which started with a preformed interface, but the IL phase is more humid. The final state of water in the IL thus depends on the protocol and relates to IL heterogeneities and viscosity. Water mainly fluctuates in hydrophilic basins (rich in O(Tf2N) and aromatic CH(BMI) groups), separated by more hydrophobic domains (rich in CF3(Tf2N) and alkyl(BMI) groups), in the form of monomers and dimers in the weakly humid IL phase, and as higher aggregates when the IL phase is more humid. There is more water in the IL than IL in water, to different extents, depending on the model. The interface is sharper and narrower (approximately 10 A) than with the less hydrophobic [BMI][PF6] IL and is overall neutral, with isotropically oriented molecules, as in the bulk phases. The results allow us to better understand the analogies and differences of aqueous interfaces with hydrophobic (but hygroscopic) ILs, compared to classical organic liquids.

The binding enthalpy of a triphenylphosphine ligand in Ru(CO)Cl(PPh 3 ) 3 (CH=CHPh) is studied with "standard" (BP86 and B3LYP), dispersion-corrected (B3LYP-D and B97-D), and highly parametrized (M05 and M06 series) density functionals. An appropriate treatment of non-covalent interactions is mandatory as these turn out to account for a large fraction of the metal-ligand interaction energy. Among the tested methods, B97-D and the M06 series of functionals best reproduce the experimental binding enthalpy value of Sponsler et al. (Inorg. Chem. 2007, 46, 561).Phosphine ligands are of key importance in homogenous catalysis, 1 and their interaction with metal centers has stimulated numerous experimental and modeling studies. 2, 3The latter usually apply density functional theory (DFT) in one of its many flavors. The accurate prediction of thermodynamic parameters for metal-ligand bond formation/breaking processes is of key interest and remains a challenging task for modern DFT. Dictated by computational cost, early computational studies usually used simplified model ligands, e.g. PH 3 instead of the widely used PPh 3 . Evidently, reaction channels that involve phosphine coordination or dissociation are difficult to assess with such models. Now that "real" systems with bulky ligands have become amenable to DFT calculations, a critical evaluation of the corresponding thermodynamic driving forces is possible.As the systems under study become larger, long-range (non-covalent) interactions tend to become more important. Such interactions have emerged as rather notorious problems for most common DFT methods. 4 Thus, new functionals have been developed that are able to describe long-range dispersion forces, either by specifically adding an empirical R -6 term, 5 or by massive parametrization against experiment. 6 In the quest for a reliable protocol to compute transition-metal/ligand binding energies, we have now tested these new functionals for the binding of PPh 3 in a sterically encumbered metal complex prototypical for many homogeneous catalysts.We chose the binding of PPh 3 (P) to the five-coordinate complex Ru(CO)Cl(PPh 3 ) 2 (CH=CHPh) (1; see Scheme 1) as test case, because this is a rare example of an equilibrium apparently unperturbed by competing coordination of the solvent, and for which reliable thermodynamic parameters have been measured. 7 This reaction represents a typical case where a bulky ligand binds to a highly coordinated metal complex, and in which multiple non-covalent interactions can take place, e.g. between neighboring phenyl rings of the ligands.

We report a molecular dynamics study of ionic liquids (ILs) at the R-quartz(001) surface. The studied ILs are based on the 1-butyl-3-methylimidazolium (BMI + ) cation and different anions Y -(Y ) Cl, BF 4 , PF 6 , and Tf 2 N) of increasing size and hydrophobicity. Two chemically different quartz(001) model surfaces were compared: a fully hydrolyzed surface covered by silanol groups, and a more "apolar" surface, covered by silane groups. All studied ILs are found to be well-ordered at the solid/liquid interfaces, in a different manner, depending on the solid surface. Interactions with the Si(OH) 2 surface are mainly determined by the H-bonding attractions of solvent anions with silanol groups. The BMI + cations are oriented more or less parallel to the surface, depending on the nature of Y -. This contrasts with the SiH 2 surface that displays repulsive interactions with all Yanions (excepted Tf 2 N -) and is mainly solvated by BMI + cations, oriented parallel to the surface. For the [BMI][Tf 2 N] and [BMI][Cl] ILs, the comparison of dry versus humid ILs and of "real" (polar) versus all-neutral quartz surfaces reveal small perturbations of the cations orientation at the interface, indicating that their orientation is mainly determined by their interactions with the anions and, to a lesser extent, by the IL/surface Coulombic + van der Waals interactions. We finally simulated more amphiphilic ILs, one being composed of the 1-octyl-3-methylimidazolium (OMI + ) cation and the Tf 2 Nanion, the other of BMI + cations and octylsulfate (OSF -) anions. In both cases, the octyl chains are mainly parallel to the Si(OH) 2 and SiH 2 surfaces, and the orientations of the imidazolium rings are similar. In the different systems, anisotropic distribution of ions is not only observed at the contact surface, but also more deeply, and markedly differs from that observed at aqueous or "air" interfaces.