Thermoelectric power in metals at normal temperatures-reply to a query

Two questions relating to this subject were raised by MacDonald and Pearson in 1961. The first concerns the need to take specific account of normal and Umklapp processes even at high temperatures in calculating the phonon drag thermoelectric power S g . The other discusses why the phonon drag contribution to the thermopower is always negligible compared with the diffusion contribution at ordinary temperatures. The first question is considered by both Bailyn and Ziman. The answer to the second is that theoretical estimates based on an assumed Debye spectrum very grossly overestimate both normal S g N and Umklapp S g U contributions to S g at ordinary temperatures although at low temperatures such estimates are probably reasonably good. It is concluded that the unexpectedly small values of S g at room temperature are, therefore, not the result of some curiously general and nearly precise cancellation between S g N and S g U .