Nmr Study of the Electronic Structure of Solid and Liquid Metals

NMR was used to study the effect of melting on the electronic structure of copper and aluminium. The Knight shift and spin—lattice relaxation time were measured as a function of temperature in the solid and in the liquid state. From these measurements the temperature dependence of K(x). the reciprocal enhancement factor of the Korringa relation, is obtained. In the case of copper it is shown that the main temperature effect is indirectly through thermal expansion. It is shown that the conduction electrons density of states and spin density at the nucleus are strongly influenced by sd hybridization. Their temperature dependence is explained as due to the volume dependence of the hybridization. In aluminium, in contrast to copper, the results of K(oc) as a function of temperature cannot be explained as due to volume change, but rather a direct temperature effect. The explanation is based on the strong mixing of states near the Brillouin zone. This mechanism should prevail only in polyvalent metals such as aluminium, in contrast to monovalent metals such as copper, where this effect can be neglected. The behaviour of other metals upon melting is discussed. About twenty years ago, W. D. Knight discovered that the nmr line of a nucleus in a metal is shifted relative to its line in a nonmetal. This is the wellknown Knight shift. This discovery made the nmr technique an important research tool in the study of the electronic structure of metals. The study of nmr in metals has become a very broad and active field. We would like to discuss one problem in this field: the effect of melting on the electronic structure of metals. This is a problem to which nmr has made a significant contribution. When a metal is heated, two major processes occur. The first is an increase in the lattice volume, which is gradual in the solid state and changes abruptly on reaching the melting point. The second is an increase in the amplitude of atomic vibrations, which can be looked upon as a gradual deviation from the order typical of the crystalline state, till a completely disordered state is reached at the melting point. We will discuss the effect of each of these two processes. In fact, a good test for any theory of electronic structure is its ability to explain the effect of a change of atomic volume and of order on