The thermal conductivity of air

Abstract

When Miss Nelson and the writer prepared in 1929 an article for the 'International Critical Tables' on the thermal conductivity of gases, we found that the value for air had been measured by 19 observers and that the mean departure from the mean was 7%. Further the values obtained by the hot wire method by Weber, Gregory and Archer, and Schneider were higher than the value (5.40 X 10 -5 cal. cm. -1 sec. -1 deg. -1 ) which Hercus and the writer had found by a parallel plate method, and higher than 13 of the 14 determinations (including hot wire ones) made previous to 1918. In view of these facts it was desirable to repeat the parallel plate method and to obtain evidence as to whether or not there was a systematic difference between the two methods mentioned. The hot wire method, as used by the experimenters named, has the practical advantages measured, and it is convenient and simple. As carried out in the experiments referred to in which fine wires were used it has the disadvantages that the elimination of the convection effects is not attained with certainly, the temperature gradient in the gas is large (which introduces both theoretical and practical difficulties) and there is a temperature discontinuity at the surface of the wire which has to be determined. Hercus and Sutherland have nearly completed in this laboratory a measurement of the thermal conductivity of air with a parallel plate apparatus. This method has the inherent advantages that there are no convection currents in the horizontal lamina of gas used, that the temperature gradient may be made small, and the temperature disments now in progress radiation is eliminated by using the metal plates at two different separations and considerable improvements have been made us compared with the experiment of Hercus and in their measurement.