Ab initio molecular dynamics (aiMD) simulations based on density functional theory (DFT) were performed on a set of five anionic platinum complexes in aqueous solution. (195)Pt nuclear magnetic shielding constants were computed with DFT as averages over the aiMD trajectories, using the two-component relativistic zeroth-order regular approximation (ZORA) in order to treat relativistic effects on the Pt shielding tensors. The chemical shifts obtained from the aiMD averages are in good agreement with experimental data. For Pt(II) and Pt(IV) halide complexes we found an intermediate solvent shell interacting with the complexes that causes pronounced solvent effects on the Pt chemical shifts. For these complexes, the magnitude of solvent effects on the Pt shielding constant can be correlated with the surface charge density. For square-planar Pt complexes the aiMD simulations also clearly demonstrate the influence of closely coordinated non-equatorial water molecules on the Pt chemical shift, relating the structure of the solution around the complex to the solvent effects on the metal NMR chemical shift. For the complex [Pt(CN)(4)](2-), the solvent effects on the Pt shielding constant are surprisingly small.

The influences of solvent effects and dynamic averaging on the (195)Pt NMR shielding and chemical shifts of cisplatin and three cisplatin derivatives in aqueous solution were computed using explicit and implicit solvation models. Within the density functional theory framework, these simulations were carried out by combining ab initio molecular dynamics (aiMD) simulations for the phase space sampling with all-electron relativistic NMR shielding tensor calculations using the zeroth-order regular approximation. Structural analyses support the presence of a solvent-assisted "inverse" or "anionic" hydration previously observed in similar square-planar transition-metal complexes. Comparisons with computationally less demanding implicit solvent models show that error cancellation is ubiquitous when dealing with liquid-state NMR simulations. After aiMD averaging, the calculated chemical shifts for the four complexes are in good agreement with experiment, with relative deviations between theory and experiment of about 5% on average (1% of the Pt(II) chemical shift range).

Electric field gradient (EFG) fluctuations for the monoatomic ions (7)Li(+), (23)Na(+), (35)Cl(-), (81)Br(-), and (127)I(-) in aqueous solution are studied using Car-Parrinello ab initio molecular dynamics (aiMD) simulations based on density functional theory. EFG calculations are typically performed with 1024 ion-solvent configurations from the aiMD simulation, using the Zeroth Order Regular Approximation (ZORA) relativistic Hamiltonian. Autocorrelation functions for the spherical EFG tensor elements are computed, transformed into the corresponding spectral densities (under the extreme narrowing condition), and subsequently converted into NMR quadrupolar relaxation rates for the ions. The relaxation rates are compared with experimental data. The order of magnitude is correctly predicted by the simulations. The computational protocol is tested in detail for (81)Br(-).

The characterization of hydrides on the surface of bulk metals and organometallic nanoparticles is of primary importance, especially in the context of catalysis, such as olefin hydrogenation. Although hydride titration as well as combination of nuclear magnetic resonance (NMR), and to a lesser extent infrared (IR) spectroscopies, confirm the presence and the potential reactivity of these simple species, the unambiguous characterization of their coordination on the surface and subsurface is still challenging owing to the number of degrees of freedom accessible at intermediate coverage rate. In this work, the influence of varying coordination modes (bridging, face capping, terminal, subsurfacic) and coverage upon the thermodynamic, kinetic, and spectroscopic properties have been investigated by means of density functional theory (DFT) calculations achieved on a ruthenium slab model. The predicted IR and 2H NMR observables reinforce the previous spectral assignment. Interestingly, following our seminal work for a monolayer coverage value (Φ) of 1/4 (Chem. Phys. Chem. 2009, 10, 2939), quadrupolar parameters obtained for higher Φ confirm that experimental 2H NMR spectra measured on Ru nanoparticles result from the presence of deuterides adsorbed on terminal, 2-fold and 3-fold symmetry sites. The kinetic and thermodynamic parametersnamely, diffusion barriers for the former and adsorption energies and phase diagrams for the lattershow that the probability of finding a subsurfacic hydride increases with Φ, whereas the saturation threshold on pristine Ru surfaces should stay close to the unityfor very low and moderate values of H2 pressurein order to ensure favorable thermodynamic conditions. Finally, this study partially opens the route to DFT studies of multistep hydrogenation reactions at the surface of ruthenium nanoparticles monitored by spectroscopic techniques. Conversely, it also demonstrates that the use of finite size models will be mandatory in the future if one wants to reach a reliable description of the metallic nanoparticle properties.

Quadrupolar NMR relaxation rates are computed for O andH nuclei of liquid water, and of Na, and Cl in aqueous solution via Kohn-Sham (KS) density functional theory ab initio molecular dynamics (aiMD) and subsequent KS electric field gradient (EFG) calculations along the trajectories. The calculated relaxation rates are within about a factor of 2 of experimental results and improved over previous aiMD simulations. The relaxation rates are assessed with regard to the lengths of the simulations as well as configurational sampling. The latter is found to be the more limiting factor in obtaining good statistical sampling and is improved by averaging over many equivalent nuclei of a system or over several independent trajectories. Further, full periodic plane-wave basis calculations of the EFGs are compared with molecular-cluster atomic-orbital basis calculations. The two methods deliver comparable results with nonhybrid functionals. With the molecular-cluster approach, a larger variety of electronic structure methods is available. For chloride, the EFG computations benefit from using a hybrid KS functional.