Some experiments on thermal diffusion

It has been shown theoretically by Dr. S. Chapman that a temperature gradient applied to a uniform mixture of two gases will tend to produce non-uniformity of composition, the heavier and larger molecules diffusing towards the cooler side, and the smaller and lighter molecules diffusing towards the hotter side. This phenomenon was termed “thermal diffusion.” The difference in composition due to thermal diffusion increases until it is balanced by the opposite effects of ordinary diffusion, when a steady state will be reached. The effect will be greatest when the gases are mixed in nearly equal proportions by volume, and will be greater the more unequal are the masses and diameters of the gas molecules. It was also shown that the extent of the effect would vary with the character of the gas molecules, being at a maximum when the molecules behave like rigid elastic spheres. In the case where molecules behave like fifth power centres of force the effect would disappear entirely. Experiments made by Chapman and Dootson established the existence of the phenomenon. The results were chiefly qualitative, the nature and order of magnitude being in agreement with the theory. In no case were the differences equal to the theoretical values obtained on the assumption that gas molecules behave like rigid elastic spheres. Although the results are not claimed to be exact, the general effect could be regarded as rather less than a half of this calculated effect. The character of the gas molecule required to give this result would, however, be in agreement with that obtained for the actual gas molecules from the investigation of the variation of viscosity with temperature.