Abstract
Thermal diffusion, or the Ludwig‐Soret effect, can be described by adding, to the isothermal flux equation, a second term proportional to the product of the temperature gradient and Q*, the heat of transport. The magnitude of Q* for vacancies in zinc, was determined by placing markers in a piece of pure zinc, which was then held in a temperature gradient near its melting point. No marker movement was observed in twenty days. Thus the net flux of vacancies was effectively zero so that the effect of the strong concentration gradient was canceled by the thermal‐diffusion force. This result is rationalized in a discussion of Q*. A brief analysis of Q* for solutes, other than vacancies, shows that in general Qi* will be approximated by the partial molar heat of solution of the solute, i, where the pure solute is taken as the standard state. A survey of the available data for Q* shows that one obtains the experimentally observed sign for Q* in each case if one takes simply Qi*=—i.