Anomalous behavior is observed in the time dependence of paramagnetic relaxation for dilute cerous magnesium nitrate at Zeeman frequencies of 11 GHz, temperatures 1.2-4.2°K, and nominal cerium concentrations of 0.2%. Immediately after excitation of the paramagnetic transition to a negative spin temperature, direct relaxation is accompanied by heating of the resonant lattice modes, and a "phonon avalanche" occurs. The decay profile of the center of the paramagnetic resonance line is distorted in a characteristic manner, and certain peculiarities in the shape of the line appear. Later in the decay, after the spin temperature has become positive, the decay approaches the behavior customarily identified with a phonon bottleneck, and the line shape becomes normal. The avalanche builds up on a time scale of microseconds, while the decay at positive spin temperature requires tens of milliseconds. Concentration, size, and surface effects are found. A simple theoretical model is developed and shown to agree qualitatively, but not quantitatively, with experiment. 157 3 J. A.

Some time ago, Van Vleck 1 predicted that, under certain conditions, an ensemble of paramagnetic ions present in a nonmagnetic crystal and in a magnetic field H should attain thermal equilibrium with a characteristic time 7\ that is inversely proportional to H 4 and to the ambient temperature T. The linear variation of relaxation rate with temperature has been observed by many authors, 2 " 4 but the dependence on magnetic field does not appear to have been given a clear experimental test. 5 Since the fourth-power variation rests on assumptions that are rather basic, it seemed worthwhile to study the field dependence in detail.As discussed by Orbach, 6 the fourth power is expected when the paramagnetism arises from just one pair of Kramers conjugate states, when the thermalizing interaction occurs through thermal modulation of the local crystalline electric field at the site of the paramagnetic ion, and when the temperature is low enough that the "direct" relaxation process, viz. the absorption and emission of phonons resonant with the Zeeman transition, prevails. If the last condition is not satisfied, the direct process is augmented by a "Raman" process, which carries the temperature variation T 9 and is independent of magnetic field.It has been pointed out 2 that 3 Fe + present as a dilute substitutional impurity in KgCo(CN) 6 furnishes an almost ideal vehicle for a test. Several workers 2 ' 7 have reported a temperature dependence consistent with l/7\ =AT + BT 9 , but the variation of A and B with field was not measured. In the present study, the relaxation rate was measured as a function of H and T.Measurements were made at Zeeman frequencies between 3.9 and 12 kMc/sec, corresponding to fields between 1100 and 3700 gauss, and at temperatures between 2.16 and 1.38°K. A sample having an iron concentration of 0. 24 at. % was used. The data were taken with an electron paramagnetic resonance spectrometer that has been described elsewhere. 2 For the present work, the sample cavity was replaced by a multimode cavity that displayed three resonances across S-band (2-4 kMc/sec) and ten in the Xband region (8-12 kMc/sec), and in addition, the microwave spectrometer components were replaced by their S-band counterparts for the low-field measurements.The results are summarized in Fig. 1. Since we were interested in the direct relaxation rate AT, the Raman rate BT g was subtracted from the observed rate, the value B having been taken from the measurements of Rannestad and Wagner. 2 It is satisfying to note that this Raman rate, measured at an altogether different value of magnetic field (about 600 gauss), agreed with the values found in the present work. It is to be noted that the higher temperature relaxation rates observed at the lowest value of the Zeeman field and the Raman contribution to those rates were of comparable magnitude. As a consequence, the calculated direct rate showed sub-141

No abstract

Measurements have been made of the electron spin-lattice relaxation of the -|, -f-| hne of Cr +++ in K 3 Co(CN) 6 at 9kMc/sec as a function of temperature, chromium concentration, and the proximity of the -f, -i line. The experimental procedure, involving inversion of the line, is capable of distinguishing a "bottleneck" relaxation time from a true spin-phonon relaxation time, T\. At Cr +++ concentrations up to 0.5%, the relaxation data are fitted well by single exponential functions of time. Between 1.3°K and 4.8°K, Ti varies approximately as T~1 2 indicating that the single phonon process is dominant. No phononbath bottleneck is observed, in agreement with calculations based on the measured parameters. A "proximity effect" is observed in which the relaxation rate of the -J, +J line is enhanced when the -f, -J line is within 20 linewidths. At one percent Cr +++ , the relaxation behavior is markedly different: the recovery is considerably faster and can no longer be described by a single time constant. This change and the proximity effect are interpreted qualitatively in terms of spin cross relaxation. The measured linewidth increases with concentration from 0.03% to 2% Cr +++ , even though the line is observed to be inhomogeneous at and below 0.5% Cr +++ .

An investigation has been made of the electrostatic charge produced in vacuum on several inorganic insulators when they are put into rolling contact with nickel, platinum, and copper surfaces. The primary role of relative motion is to increase the area on the insulator that ultimately undergoes contact, thereby increasing the total charge separated. With quartz, all evidence favors electron transfer which occurs because the nonconductor has a higher effective work function than the metal. Field emission-induced back leakage of electrons during separation of the charged surfaces is considered to reduce the charge originally separated to that actually measured. Little difference is seen between the charging properties of quartz cleaned chemically in room air and quartz subsequently outgassed at 700°C in vacuo better than 10−8 mm Hg. No difference in electron affinity is found between fused and monocrystalline quartz. No evidence directly supporting electron transfer is found for the other insulators tested, synthetic single crystals of Al2O3, MgO, NaCl, KCl, KBr, and KI; and the possibility of ion transfer is considered. All charge against nickel much less vigorously and less reproducibly than does quartz. The alkali halides and MgO which contains excess Mg or O as a bulk impurity charge positively, while stoichiometric MgO charges negatively. With Al2O3, there is a strong dependence of charge density on crystallographic orientation. Alteration of charging properties by surface and/or bulk conduction is found to be negligible for all insulators tested, within the reproducibility of the charging measurements.