Electron band theory 1952-1962

Work undertaken by the Theoretical Physics Division between 1952 and 1965 to obtain an understanding of electrons in metals, with uranium and the actinides and the structurally-important transition metals as the main targets is examined. A main result of that period was a conviction that the majority of the physical properties of all metals, except the 4f rare-earth series and the actinides beyond uranium, were dominated by band effects which could be described well enough for most purposes by simple one-electron calculations with simple self-consistent fields. The period from 1960 on showed increasingly clearly the necessity of incorporating relativistic spin-orbit coupling terms in the heavy metals, and some 'local exchange field' correction to the fields close to nuclei. The problems of the non-local interaction of spins - highly important for alloy theory and for antiferromagnetic instability -required the evolution of computers large enough to produce wave-functions at all wave-vectors for all bands so that the susceptibility at arbitrary wave-vector could be computed. This work has not proved to be very illuminating so far, and much interest again focusses today on heuristic arguments that give qualitative descriptions of band structures, such as canonical d-bands to account for crystal structure. (UK)