Without a doubt, nuclear magnetic resonance (NMR) has been an indispensable tool forstructural determination in chemistry, molecular biology, and more. The power of NMR to
elucidate positions and distances of atoms within molecules has allowed generations of chemists,
chemical biologists, and molecular biologists to develop strategies for synthesizing new
medicines, advanced materials, and to verify their syntheses. To this day, the core strategy from
which this tool draws its power remains a drive to higher and higher magnetic field. For the
purposes described above, it has been a necessary drive. A higher field allows for better signal to
noise ratio (S/N) and narrows relative spectral linewidths, resulting in faster data acquisitions and
for acquired data to be more granulated. There are, however, side effects that accompany the
drive to higher field. Extremely strong magnetic fields are physically dangerous if any
magnetizable materials come close enough to magnets. Equipment, tools, and some people
themselves must therefore be kept a safe distance away, and instances where failures in this
regard have led to deaths are not unheard of. The superconducting magnets that generate these
intense fields are also quite large, are limited to tiny sample volumes, and require special
maintenance, which can be expensive. Specifically, they require a constant supply of liquid
helium in order to keep the coils at temperatures low enough to maintain their superconducting
characteristics. This helium is not a renewable or ubiquitous resource and it gets more expensive
by the year, as challenges to exploration, capture, and even political upheaval destabilize the
global market for liquid helium. Meanwhile, advances in amplifier and NMR spectrometer
technology have resulted in better control and sensitivity in low-field experiments. For these
reasons, and more, low-field NMR has become an ever-more attractive area of research for the
advancement of NMR and magnetic resonance in general.
This work describes a body of research focused on design and data analysis for low-fieldNMR, which is loosely defined for the purposes of this manuscript as NMR performed in a static
magnetic field, B0, less than 0.1 T, or a Larmor frequency lower than about 4.4 MHz. Because
NMR at low-field does not generally produce chemical shift-resolved spectra, data acquisition
and analysis instead tends to focus on bulk magnetic relaxation rates and imaging. Furthermore,
these parameters are often difficult to understand using direct mathematical models because
samples investigated at low-field are often inhomogeneous and possess geometries that are not
always amenable to symmetry-based mathematical simplifications. It is often better to instead
use statistical correlation and categorization to understand experimental results from a higher,
less granular level. This fact motivates an initial exploration into the use of partial least squares
as a method to determine the electrical permittivity of aqueous electrolyte samples at extremely
high pressures. While this is not a study of magnetic resonance, it serves as an example for the
application of this technique to similar problems that might present themselves in NMR. The
next chapter describes Matlab codes that have been written for simulating magnetic fields for
arbitrary coil and magnet shapes. The final two chapters are an exploration into novel rf coil
designs for use in single-sided NMR, and specifically in the presence of shielding due to
conductive magnet material. Multiple coil designs are described and their performances
characterized. Ultimately an unexpected “fringe coil” geometry is shown to be far superior in
theoretical models and practical testing with a variety of sample types and geometries.