scholarly journals The theory of the anomalous skin effect in metals

The problem of metallic conduction at high frequencies and low temperatures, recently discussed by Pippard, is reformulated using the general methods of the theory of metals, and exact solutions are obtained which are valid for all frequencies and temperatures. It is shown that, for large values of the free path of the conduction electrons, the electric field is propagated through the metal as a ‘surface wave’ which differs considerably from the classical exponential solution. The temperature variation of the surface impedance in the microwave region is considered in detail. Pippard’s simplified theory is shown to be qualitatively correct, and a quantitative discussion of his experimental results is given. The frequency variation of the surface impedance at low temperatures is also discussed, and it is shown that relaxation effects are negligible in the microwave region but become important in the infra-red and eventually restore the validity of the classical theory. The theory predicts that, as the frequency is increased, the reflexion coefficient of metals passes through a minimum in the far infra-red.

The anomalous skin effect arises in good conductors at low temperatures and high frequencies when the electronic mean free path becomes comparable with or greater than the classically calculated skin depth. Measurements have been made on a number of metals at frequencies of 1200 and 3600 Mc/s, and the form of variation of r. f. surface conductance with d. c. conductivity agrees well with that predicted theoretically by Reuter & Sondheimer, assuming that the electrons are scattered diffusely when they hit the surface of the metal. From the results, estimates are made of the effective value of σ/ l , the ratio of d. c. conductivity to mean free path, and hence of the free surface area of the occupied region of k -space. The estimate for copper agrees well with that expected theoretically; those for silver and gold are rather lower than the theoretical values. For the other metals investigated, tin, cadmium, lead and aluminium, no theoretical estimates are available. The results are very sensitive to the presence of surface imperfections; the effect of these is discussed.


The measurements described in the earlier papers of this series have been extended to 9400 Mc. /sec., a resonance technique being employed to determine the surface resistance of normal and superconducting tin, and the difference between the surface reactances of the material in the two states. Measurements on single crystals of different orientations have brought to light a marked anisotropy of all these quantities, of a kind which shows clearly the non-tensorial nature of the fundamental equations relating the field vectors. The prediction of the theory of the anomalous skin effect in normal metals, that the surface resistance should vary with frequency as ω ⅔ , is confirmed. The temperature variation of the resistance and reactance of superconducting tin has been studied in detail for a number of specimens of different orientations, and it has been found that over certain ranges of temperature the shapes of corresponding curves for different specimens are similar, apart from scaling factors depending on the orientation; the values of these scaling factors are used to characterize the surface impedance of each orientation.


The theory of the anomalous skin effect in metals is extended to a uniaxial metal crystal containing two energy bands in each of which the energy surfaces are ellipsoids of revolution about the crystal axis. Explicit formulae are obtained, for the extreme anomalous limit, giving the dependence of the surface impedance on the orientation of the crystal axis, both for a plane metal surface and for a circular wire. The form of the anisotropy of the surface impedance is found to depend upon the axial ratios of the spheroidal energy surfaces and upon the ratio of the electron free paths in the two bands. Wide variations in behaviour are possible, and the surface impedance may show a high degree of anisotropy even when the d.c. conductivity is almost isotropic (as with tin at low temperatures). The results are evaluated numerically for tin, and the surface conductivity of a circular wire is found to show the minimum observed by Pippard (1950); the parameters can be chosen to give reasonable agreement with Pippard’s results.


This paper is the first in a series in which the behaviour of the electrical impedance of metals at low temperatures and very high frequencies will be considered from experimental and theoretical standpoints. The technique of resonator measurements at 1200 Mcyc. /sec. is described in detail, and experimental curves are given showing the variation with temperature of the r. f. resistivities of superconducting tin and mercury. In contrast to the behaviour of superconductors in static fields, a finite resistance is present at all temperatures, tending as the absolute zero is approached to a very low value, which is probably zero for mercury but not for tin. The experimental results are in good agreement with London’s measurements on tin by a different method. The latter’s observation, that the r. f. resistance of normal tin above the transition point is much greater than that predicted from the classical skin-effect theory, is confirmed, and a similar, though less marked, effect is found also for mercury.


Sign in / Sign up

Export Citation Format

Share Document