The Seebeck coefficient is the differential drop in voltage per unit temperature across a conductor. Goldsmid says that there must be a temp. gradient across a junction for a seebeck coefficeint to be defined. This makes sense if the difference in the fermi level energy between two dissimilar contacts that causes a change in energy of the electron as it jumps from one metal to another. But I usually hear that the Seebeck coefficient is the property of the metal itself. Any comments to clear this up would be received.