No abstract

Restilts obtained in the temperature range 0.3~1.2°K give direct evidence that electronelectron scattering in the transition metal Pt is at least an order of magnitude greater than in Ag. Our results do not confirm the anomalous behavior in thermal conductivity which has been reported for numerous metals below 1°K.Theoretically, the thermal conductivity of normal metals is expected to vary linearly with temperature in the limit of very low temperatures. In recent years, however, several experimental papers have reported oscillatory deviations from this behavior for a variety of metals in the temperature range 0.3 -1.0°K.^"^ No attendant anomalies were observed in the electrical conductivity, nor in the thermal conductivity of the superconducting state/ One might therefore infer that the phenomenon is not found in the thermal transport contributed by phonons, but rather by electrons. In addition, the effect would be associated with small-angle electron scattering which is of importance in thermal conductivity as opposed to electrical conductivity. The fact that the magnitude of the anomalies was found to scale roughly with the magnitude of the thermal conductivity, i.e., with the purity of the metal,^'^ is suggestive of an error in thermometry, but thermal-conductivity data for the superconductors and for sapphire which varied as T^ appeared to rule this out.^ Nevertheless, in view of the unusual nature of the above-mentioned implications and since all previous measurements were evidently made in the same cryostat, we decied to make similar measurements on Ag and Pt. This also allowed us, in a temperature range where electron scattering by phonons should contribute less than 1:10^ to the total scattering probability, to compare electron scattering in a noble metal with that in a transition metal. The latter has a greater density and variety of states at the Fermi surface because of a more significant overlap of s and d bands.The technique was to establish a known thermal flux Q in a sample of cross sectional area A and length Z, measure the (time-independent) temperatures TQ and Tjj at points C and/fa distance I apart on the sample, and compute the thermal conductivity K from K = lQ/A{Tjj -TQ),where Tff--Tc^0.ITff. An uncalibrated resistance thermometer clamped to C maintained that point at a fixed temperature (TQ) by means of an electrical heater and electronic (or manual) regulation. With Q=0, TQ was measured with a cerium magnesium nitrate (CMN) magnetic thermometer attached to H. Turning on Q then permitted the same magnetic thermometer to measure Tjj while T^ remained constant. Since the thermometric parameter of CMN-namely the inverse susceptibility-varies linearly with T, errors in calibration could at most produce a slight curvature in a plot of K vs T. The CMN was calibrated against the vapor pressure of liquid He^ in the range 1-2°K. No corrections were required. The calibration reproduced to 1:10^ from run to run, and was frequently checked during a run against a calibrated resistance thermometer. In me...

Application of magnetization measurements in iron to high temperature thermometry . The paper ?escribes a highly stable ac mutual inductance bridge suitable for low temperature use. The applica-tIon of the bndge to magnetic thermometry below 3 K is discussed. *

No abstract

No abstract