The heat transfer of the shell I to the lattice of core C can be made fairly high. That of the lattice to the protons of the core is very low in accordance with the very long lattice-nuclei relaxation time. This relaxation time, however, can be reduced by the incorporation of paramagnetic ions in the core C as indicated by Bloembergen, 9 by Rollin and Hatton, 1ยฐ and by Khutsishvili :x (either Cr a+ in the alum, or neodymium in the nitrate).
These paramagnetic contaminations will have a large effect on the susceptibility of the core. Kurti, 1~ however, showed that below about 0"02 ยฐ K the susceptibility of the impurities becomes independent of temperature, so that below that temperature any change in susceptibility can be ascribed to the nuclear spin system.
It is hoped that Kurti's conclusion will also remain Substantially correct in our experiments, although the amount of impurities will be much higher.
If the longitudinal relaxation time tx of protons or other nuclei in an anisotropic (eventually diluted) salt such as cerium or neodymium magnesium nitrate were of the order of several minutes at temperatures of about 0"01 ยฐ K, a much simpler procedure for the nuclear cooling could be followed. Then, after slow rotation cooling of the salt, one should have to wait only a reasonable time for the electron system to cool the proton system. Following this, the demagnetization would cool the nuclei to a very low temperature. The warming Up of the nuclear by the electronic system would be slow enough to allow the temperature measurement.