Low field dc SQUID NMR on room temperature samples and single crystal UPt3

This thesis is an account of two distinct experiments with a common theme,
which is the technique of dc SQUID NMR. Firstly the application of the technique for broadband spectroscopy on room temperature samples is described. The motivation behind this work was to try to obtain SQUID NMR signals from liquid samples such as water and the amino acid glycine in order to demonstrate a few of the potential applications of this technique in the low field regime. These include increased frequency resolution, the possible detection of relatively small amounts of oil contamination in water samples and low field J-spectroscopy as a chemical bond detector. Measurements were performed on samples of water and oil-water mixtures on a dipper probe that provided a simple, compact shielding arrangement that uses a two-stage SQUID sensor as the front end amplier. With the introduction of a low frequency amplier and modications to the setup the SNR was increased by a factor of 4. However, the SNR became limited by flux trapping in the superconducting materials surrounding the sample, leading to the investigation of alternative materials and methods to maintain low field homogeneity at higher polarising fields.

The second part describes SQUID NMR measurements on a single crystal of the heavy fermion superconductor UPt3. Despite serious experimental and theoretical efforts, the symmetry and nature of the unconventional superconducting order parameter has not been resolved due to contradicting experiments, in particular that of the Knight shift into the superconducting state. Measurements were performed on a dilution refrigerator. A two-stage SQUID was mounted onto the fridge, a 3He marker was implemented for accurate local field determination and an overlapping superconducting shield was made to decrease the spectrometer deadtime. NMR measurements are presented for the static field parallel to the c-axis of the crystal from 600 mK down to 400 mK in a static field of 71.5 mT. The Knight shift is measured to decrease from -1.7 % in the normal state to -1.35 % at 400 mK. Together with the results from a Knight shift experiment at 30 degrees to the c-axis, it is argued that these results may provide evidence for the E2u model. Strategies for future measurements are addressed.