With the development of higher field magnets and capabilities of spinning samples
at greater speeds, high resolution solid-state nuclear magnetic resonance (SSNMR)
spectra of quadrupolar nuclei resolving multiple spectroscopic sites in a sample are
now attainable, where just two decades ago it would not have been possible. With
these advances, the field of SSNMR has truly been opened to determination of structure in inorganic materials, which are mainly quadrupolar nuclei.
Although, many obstacles to obtaining structural information using SSNMR have
been overcome, there are still issues, such as small sample size, poor sensitivity (small
gyromagnetic ratios, low natural abundance) and long relaxation times. In this thesis,
I will discuss methods of obtaining high resolution spectra of quadrupolar nuclei using
a theoretical description of interactions to second-order that may contribute to the
NMR spectra of quadrupolar nuclei and provide a method to remove or isolate such
interactions. This description is not only useful for describing experiments currently
utilized, but lays the groundwork for the development of new experiments. With some
experiments and samples, a theoretical approach beyond second-order is required to
analyze spectra, an outline to use exact numerical calculations to simulate NMR
spectra is also given.
To combat the low sensitivity and to decrease experimental time, preparatory enhancement sequences that increase the sensitivity of spectra are necessary. A method
to ensure optimum enhancement when utilizing these experiments is given.
Using these techniques, I will give insight on the structure of densified amorphous
silica. I have measured the two-dimensional 17O dynamic-angle spinning solid-state
nuclear magnetic resonance spectrum of silica glasses produced from the melt and
densified in a multi-anvil device at pressures up to 15 GPa. From the spectra, two-dimensional histograms correlating Si-O-Si angle with Si-O distance, Si-O-Si angle
with Si-Si distance, and Si-O distance with Si-Si distance are derived.