Nuclear Magnetic Resonance Saturation in NaCl and CaF2

Experiments were performed to test the hypothesis due to Redfield, that a nuclear spin system in a sufficiently large rf field, $H_1\left(\nu \right)$, is properly described by a spin temperature referred to a frame of reference rotating about the Zeeman field with the frequency $\nu$ of the rf field. Measurements were made on the ${\mathrm{Na}}^{23}$ spins in NaCl and the ${\mathrm{F}}^{19}$ spins in Ca${\mathrm{F}}_2$. A combination of steady-state and pulse techniques was used to measure the magnetization $M_z$ as a function of the frequency and amplitude of $H_1\left(\nu \right)$. When $H_1$ is sufficiently large, the data show that Redfield's theory is correct within 10% for NaCl but appreciably in error for Ca${\mathrm{F}}_2$. The same type of measurement performed at low amplitudes of $H_1$ show that the theory of Bloembergen, Purcell, and Pound is in agreement with the NaCl measurements but measurably in error for Ca${\mathrm{F}}_2$. It is suggested that neither theory is correct for Ca${\mathrm{F}}_2$ because the fluorine spins relax by means of paramagnetic impurities, and, therefore, do not relax independently with characteristic time $T_1$, as is assumed in both theories. In an intermediate range of saturating field amplitudes, neither theory is expected to apply; this was found to be the case experimentally.