NUCLEAR MAGNETIC RESONANCE STUDIES OF CATALYSTS AND MODEL CATALYTIC SYSTEMS: BIOCHEMICAL AND INORGANIC APPLICATIONS (MINERALS, SILICON-29, OXYGEN-17, SHIELDING TENSORS, NMR)

A model for the prediction of silicon-29 nuclear magnetic resonance (NMR) chemical shifts in minerals, zeolite catalysts, ceramics and glasses is proposed based on a group electronegativity approach. Based on a total of 99 sites in 51 different compounds, the mean absolute deviation between theory and experiment is 1.96 ppm with correlation coefficient of 0.979. When all types of silicon are considered (Q('0)-Q('4)), this empirical approach is the most accurate method of predicting Si-29 chemical shifts to date, and much better than the best previous model (correlation coefficient 0.88) despite its relative simplicity.;The electronic environment of the mesomeric N-O moiety in a series of 4-substituted pyridine-N-oxides is probed by carbon-13 and oxygen-17 NMR in conjunction with a nitrogen-14 and oxygen-17 nuclear quadrupole resonance (NQR) study. Lineshape analysis of the C-13 resonances bonded to nitrogen facilitiates determination of the N-14 NQCC orientation. Bond orders determined by O-17 NQR are shown to parallel changes in O-17 and N-15 chemical shifts.;The carbon-13 NMR chemical shielding tensors of the amino acid L-threonine have been obtained. This represents the first unambiguous determination of the carbinol shielding tensor. A method for the unique determination of tensor orientation based on the heteronuclear dipolar interaction is developed and discussed.;Bonding models for silicates are assessed in relation to the local environment of oxygen, as determined by analysis of the oxygen-17 nuclear quadrupole coupling constants (NQCC), using Townes-Dailey methods. The experimental NQCC of the silica polymorph low cristobalite is indicative of covalent charge transfer from the oxygen lone pairs to silicon, and is consistent with Pauling's (d-p) (pi)-bonding model. Bonding models for both hybridized and unhybridized oxygen which exclude a (pi)-bonding mechanism are in poor agreement with the experimental results. The oxygen-17 NQCC of the bridging oxygen of diopside is shown to be in agreement with McDonald's (d=p) (pi)-bonding hypothesis. Calcium coordination to the diopside bridging oxygen is consistent with calcium acting as a charge acceptor. The use of Pauling ionicities in conjunction with the Townes-Dailey model gives good agreement with the experimental NQCC results for a variety of well-defined oxide and silicate systems, and further supports the (d-p) (pi)-bonding hypothesis in silicates.