Magnetic Phenomena Accompanying the Transition to the Superconducting State

Prof. W. Meissner

Already in our first communication on the new magnetic phenomena accompanying the transition to the superconducting state⁸, we pointed out the following: only in the case of a solid cylinder (i.e., one having no internal cavities), made of a pure metal and cooled in the absence of a magnetic field, can it be said that the body behaves as though its permeability, on transition to the superconducting state, had decreased to zero. If, however, a hollow cylinder is cooled below the transformation point in an applied homogeneous magnetic field, the phenomenon becomes considerably more complicated. This is confirmed by our subsequent experiments with a cylindrical tin crystal bored along its axis⁹. Gorter and Casimir¹⁰ proposed the following explanation of this phenomenon: a superconductor below the transformation point possesses nonsuperconducting regions into which the lines of magnetic induction are drawn, as a result of which the field here exceeds the critical value. It is precisely in these regions that the lines of induction can penetrate into the substance. In this way Gorter and Casimir were also able to explain why a solid sphere, when cooled below the transformation point, retains its magnetic moment despite all these new phenomena. The existence of such a residual magnetic moment was demonstrated by the old experiments of Onnes and his collaborators, as well as by the recent experiments of Mendelssohn and Babbit¹¹. F. and H. London¹², adopting the hypothesis of Gorter and Casimir, have created a very interesting phenomenological theory of superconductivity, explaining both these new magnetic effects and the very phenomenon of superconductivity itself. Naturally, the following important question arises: whether the Gorter–Casimir hypothesis agrees with all experiments. If so, does the theory of the Londons follow exactly, or must it also be modified?

This question can be considered in the light of the experiments carried out by Dr. Heidenreich in the cryogenic laboratory of the Berlin Reichsanstalt under my supervision after my move to Munich in April 1934. Dr. Heidenreich once again undertook the study of the magnetic field both in the inner cavity of a tin crystal and around its outer surface, making special new measurements of the residual field. Already in our previous experiments we found that, both in a lead tube and in a crystalline tin tube, after lowering the temperature below the transformation point and removing the homogeneous external field, a residual field can be observed in the inner cavity. Now we have found that not only in the case of a lead tube (as was already mentioned in my first communication), but also in the case of a crystalline tube made of tin, the residual field is retained also near the outer surface.

The details of the results are as follows: when the temperature is lowered below the transition point, the intensity of the magnetic field in the cavity of the tin crystal increases at all points, regardless of how the crystal is cooled—externally or internally. On average the magnetic field in the cavity increases by approximately 10%. The direction of this magnetic field at all points deviates from the direction of the initial homogeneous field by no more than a few degrees. The fluctuations of the magnetic-field intensity from point to point are of the order of 30% and do not change with the change in the orientation of the crystal relative to the initial homogeneous field. The field near the outer surface of the crystalline tube, before we remove the external field, does not exactly coincide with the value expected for a permeability equal to zero.

In a hollow cylinder the concentration of magnetic-induction lines is smaller than in a solid crystalline cylinder, and the fluctuations of the magnetic-field intensity from point to point near the surface are more complicated.

After removal of the external homogeneous field, a field remains in the internal drilled-out region and near the outer surface. If the external field is then switched on again, then near the outer surface of the cylinder there appears almost the same field as was present there before the removal of the external field; however, in the internal cavity the residual field increases by only 3%. This increase disappears if the external field is switched off again. With repeated removal and application of the external field, these changes are always reproduced. The direction of the residual field in the internal space is the same at all points and, as before, deviates by a few degrees from the direction of the initial field. The intensity of the residual field fluctuates from point to point by approximately 20% and on average amounts to 35% of the original. Near the outer surface the intensity of the residual field is about 20% of the initial one, and the fluctuations from point to point reach 100%. The changes in the direction of the field from point to point are very complicated. One may also gain the impression that at many points of the outer surface the normal component of the residual field is different from zero.

These experiments of Dr. Heidenreich confirm the statements contained in our first communication, namely that the hypothesis proposing a decrease of the magnetic permeability to zero does not make it possible to explain the results of the experiments with a hollow cylinder. I also believe that the latter experiments compel one to doubt the validity of the Gorter and Casimir hypothesis for all experiments in general. Of particular importance, it seems to me, is the conclusion that can be drawn from consideration of the most recent diagrams: that the normal component of the residual field at almost all points of the inner and outer surfaces of the crystalline tube is different from zero. This fact, it seems to me, speaks against the universal applicability of the Gorter and Casimir hypothesis and, consequently, of the London theory as well.

A good test of the Gorter and Casimir hypothesis could be an experiment based on the following consideration: the critical field that destroys superconductivity is a function of temperature; consequently, as the temperature is lowered, the induction lines should become more and more concentrated. Therefore the distribution of the external residual field should change as the temperature is lowered. It is very remarkable that the residual field is significantly stronger in the case of the lead tube than in the case of the crystalline tin one. However, our original idea, according to which the residual field is caused by irregularities in the crystal lattice and therefore should vanish for an ideal crystalline tube, agrees poorly with the independence of the orientation of the induction lines from a change in the orientation of the tube by approximately 20%.