RESUMEN
Nanoscale zinc-oxide doped with aluminum ZnO:Al is studied by different techniques targeting surface changes induced by the conditions at which ZnO:Al is used as support material in the catalysis of methanol. While it is well established that a variety of 1H and 27Al resonances can be found by solid-state NMR for this material, it was not clear yet which signals are related to species located close to the surface of the material and which to species located in the bulk. To this end, a method is suggested that makes use of a paramagnetically impregnated material to suppress NMR signals close to the particle surface in the blind sphere around the paramagnetic metal atoms. It is shown that it is important to use conditions that guarantee a stable reference system relative to which it can be established whether the coating procedure is conserving the original structure or not. This method, called paramagnetically assisted surface peak assignment, helped to assign the 1H and 27Al NMR peaks to the bulk and the surface layer defined by the blind sphere of the paramagnetic atoms. The assignment results are further corroborated by the results from heteronuclear 27Al{1H} dipolar dephasing experiments, which indicate that the hydrogen atoms are preferentially located in the surface layer and not in the particle core.
RESUMEN
The convergence of simulated C-REDOR curves of (infinitely) large spin systems is investigated with respect to the number of spins considered in the calculations. Taking a sufficiently large number of spins (>20,000 spins) into account enables the simulation of converged C-REDOR curves over the entire time period and not only the initial regime. The calculations are based on an existing approximation within first order average Hamiltonian theory (AHT), which assumes the absence of homonuclear dipole-dipole interactions. The C-REDOR experiment generates an average Hamiltonian close to the idealized AHT behavior even for multiple spin systems including multiple homonuclear dipole-dipole interactions which is shown from numerically exact calculations of the spin dynamics. Experimentally it is shown that calculations accurately predict the full, experimental C-REDOR curves of the multi-spin systems (31)P-(19)F in apatite, (31)P-(1)H in potassium trimetaphosphimate and (1)H-(31)P in potassium dihydrogen phosphate. We also present (13)C-(1)H and (15)N-(1)H data for the organic compounds glycine, l-alanine and l-histidine hydrochloride monohydrate which require consideration of molecular motion. Furthermore, we investigated the current limits of the method from systematic errors and we suggest a simple way to calculate errors for homogeneous and heterogeneous samples from experimental data.