London, Heinz

views updated

London, Heinz

(b. Bonn, Germany, 7 November 1907; d. Oxford, England, 3 August 1970)


Like his brother, Fritz London, who exercised close parental influence on him after their father’s early death, Heinz received a classical education but was also interested in mathematics and science, especially chemistry. He attended the University of Bonn from 1926 to 1927 and after six months with the W. C. Heraeus chemical company completed undergraduate studies in Berlin and Munich. His graduate work on superconductivity was done at Breslau under F. E. Simon, and his Ph.D., issued early in 1934, was one of the last awarded to a Jew in Nazi Germany. Later in the same year he joined Simon and other members of the Breslau group at Oxford, where they were establishing the first center of low-temperature research in England since the time of James Dewar. Fritz London was already in Oxford, and Heinz lived there with the family for two years. In 1936 he moved to Bristol, where he remained until 1940, when he was for a brief time interned as an alien. After his release he worked with Simon and others on the British atomic bomb project and then spent two years at Birmingham before transferring to Harwell in 1946, where he continued to work until his death. In 1939 he married Gertrude Rosenthal, but the marriage broke up soon afterward. In 1946 he mania Meissner; they had four children.

The origin of London’s work on alternating-current losses in superconductors has been described in the preceding entry, along with his early contributions to the theory, culminating in the joint paper with Fritz London on the electrodynamics of the super-conducting state. An interesting feature of the theory outlined there is the indeterminate parameter μ in equation (7). Evidently one solution would be μ = 0, in which case the acceleration equation (5) remains valid, although its justification is different from that originally given. A superconductor would then, like any other metal, exhibit surface charges in an electric field. An equally plausible alternative, discussed by the Londons, was to make μ, = Λc2p, where p is charge density. This had the advantage of allowing a symmetrical four-dimensional representation of the equations; it implied, however, that the electric field penetrates a superconductor to the same depth as the magnetic field and therefore that the capacity of a condenser with super-conducting plates should change appreciably on cooling through its transition from the normal to the super-conducting state. In 1935 London tried this experiment at Oxford, showing conclusively that there is no change in capacity and hence that the solution μ= 0 must be accepted. About the same time he published an investigation of the equilibrium between super-conducting and normal phases of a metal, based on a thermodynamic analysis of the London equations. He found that the observed transition to the normal state in high magnetic fields depends on the existence of a threshold value for current density in the superconductor. He also discovered that unless a superconductor has a positive surface energy sufficient to counteract the field energy in the penetration layer, it will split into a finely divided mixture of normal and superconducting regions in high fields. Since the superconductivity of most pure metals disappears entirely above the critical field, London concluded that their surface energy is indeed positive. He conjectured however that some hysteresis effects just then observed in superconducting alloys might arise from a negative surface energy. These ideas were later entirely confirmed, and during the 1950’s they were elaborated into the distinction between type I and type II superconductors. The special properties of type II superconductors proved of great technical importance in obtaining high field superconducting magnets.

While at Oxford, London continued to search for alternating-current losses in superconductors, but had no success, even though he raised the operating frequency to 150 MHz. After his move to Bristol he took the subject up once more and finally in 1939 succeeded in demonstrating the occurrence of high-frequency losses at 1500 MHz. Figure 1 reproduces the experimental data from London’s paper, which shows significant alternating-current dissipation at points Well below the ordinary transition temperature. An unexpected by-product was the discovery of the “anomalous skin effect” in normal metals. At low temperatures the resistivity of tin in the normal state, as deduced from Joule heating at 1500 MHz, was several times larger than the measured direct-current resistivity, although at room temperatures the two figures were identical. London tentatively attributed the effect to the electrons’ having a mean free path considerably greater than the skin depth, so that relatively few of them contribute to conduction.

London’s work on high-frequency losses was terminated by World War II, but the subject was revived in 1946 by several other investigators, using resonant cavity techniques from radar research. A. B. Pippard in particular reached some highly significant results by measuring changes in resonant frequency of cavities above and below the transition temperature. He concluded that the wave functions were indeed coherent over distances greater than the penetration depth but that the coherence length was finite rather than infinite as in the London theory. Pursuing London’s idea about the electronic mean free path, he introduced impurities into the material, which acted as scattering centers to reduce both the free path and the coherence length. When this was done the penetration depth increased, the product of the Pippard coherence length , and the penetration depth λ being approximately constant. Other measurements led to the hypothesis of an energy gap in the distribution of conducting particles. These developments provided the basis of ideas for the Bardeen-Cooper-Schrieffer microscopic theory of superconductivity. The Pippard length was then identified with the pairing distance of the electrons.

At Bristol, London also investigated superconductivity of thin metallic films in collaboration with E. T. S. Appleyard, A. D. Misener, and J. R, Bristow. From measurements of critical fields they were able to determine the penetration depth experimentally. Measurements on the magnetic susceptibilities of superconducting particles by D. Shoenberg about the same time gave comparable results. This work on thin films was another example, like the experiments on alternating-current losses, of work begun by London which became a major field of research studied by

hundreds of other investigators. A third subject of study from the same period was his work on helium, which included an experiment that helped establish the two-fluid model, and the thermodynamic analysis mentioned in the previous entry, which proved that the superfluid component carries no entropy. A by-product of the analysis was London’s prediction of the mechanocaloric or inverse fountain effect, a temperature difference generated by the moving superfluid.

Following the work of Fritz London and Tisza much effort was applied to the theory of liquid helium II. In 1941 Lev Landau formulated the problem in quite a different way, describing the normal fluid in terms of quantized excitations, phonons and rotons, and giving new hydrodynamical equations to describe superfluidity. One of his conclusions was that the superfluid is subject to the condition curl vs = 0, making it incapable of rotation. Actually the same condition for the superfluid was also strongly advocated by Fritz London in the London-Tisza model, along with the analogous condition on curl ps = 0 in a superconductor, as evidence that each phenomenon constitutes a macroscopic quantum state. In 1946 Heinz London proposed an experiment to distinguish the behavior of a fluid subject to constraints of this kind from mere absence of viscosity. The idea was to set a cylindrical container of helium I in slow uniform rotation about its axis and then cool the system slowly to a temperature well below the lambda point. If the superfluidity is simply absence of viscosity, the angular momentum of the fluid should be unaffected by the transition; but if the equilibrium state is one of constrained vorticity, the superfluid created by cooling should stop rotating and the container has to rotate more rapidly to conserve the total annular momentum. London then made a further conjecture. In 1941 A. Biji, J. de Boer, and A. Michels had pointed out that the critical velocity vc(the maximum velocity for superfluid flow) in a film of helium 11 seems to obey the condition vc ~ h/4πmd, where m is the atomic mass of helium and d the film thickness. London suggested that the angular velocity Q of the helium II. might be subject to a similar quantum Condition, so that in a container of radius R, rotation stops only if

For a vessel one cm. in diameter the rotational period obtained from (I) is about one day, so the experiment was a difficult one. Experiments at higher speeds tried several times from 1950 onward disclosed no difference between the rotating superfluid and classical rigid-body rotation. Various suggestions about the state of motion were advanced by Fritz London and others; then Lars Onsager and later, independently* Richard Feynman assumed that the circulation around an arbitrary path in the fluid is a multiple of h/m consistent with London’s condition (1), but that the rotating fluid as it gains more angular momentum will break up into an array of vortices, each with quantized circulation h/m. Thus by a different path the suggestion of a macroscopic quantum condition entered the theory of helium II as well as the theory of superconductivity; it is remarkable that one suggestion originated with Heinz London and the other with Fritz London. The existence of individual quantized vortices in helium II was first demonstrated experimentally by W. F, Vinen in 1961 through the coupling of vorticity to a vibrating wire;1 since then they have been observed in several other ways. The rotating-bucket experiment suggested by London was finally performed successfully in 1965,2 The measurements confirmed that at low angular velocities the superfluid stops rotating and the bucket rotates faster on cooling through the lambda point. Thus the conjecture that curl v8 — 0 is an equilibrium state for a superfluid was confirmed. A notable feature was that while quantized vortices of the predicted size were indeed observed, the angular velocity below which rotation ceased was about fifteen times higher than condition (1). The last result is explained by the large amount of energy required to form a single vortex core., it is the equivalent for the superfluid of the Ochsenfeld-Meissner effect in a superconductor.

London’s work on the atomic bomb project had been on methods of separating uranium 235 by ionic migration and liquid thermal diffusion. At Harwell after the war he continued to work on isotope separation, concentrating on the production of carbon 13 as a stable tracer element for medical research. He developed a method of low-temperature distillation and designed a fractionating column using carbon monoxide for the enrichment of C[13 and O18. This machine has operated successfully for many years and currently supplies all the C13 used in Great Britain and the United States.

During the last fifteen years of his life London worked in collaboration with various colleagues on three main topics, neutron production and neutron-scattering experiments in liquid helium, techniques for producing high field superconducting magnets, and, most important, the He3-He4 dilution refrigerator, one of the most ingenious and useful contributions to cryogenic technology in many years. He had suggested the essential idea of the dilution refrigerator in 1951 at the Oxford Conference on Low-Temperature Physics. He pointed out that since He3 behaves as a Fermi gas below l° K., and He4 as a Bose gas, then if He3 is allowed to mix with He 4 it will in effect expand. Now in this region He4, being practically pure super-fluid, has negligible entropy; whereas He3has entropy proportional to n-2/3T, where n is the atomic concentration. Dilution to one part in a thousand would therefore cool a bath of helium to 10-2° K., while still retaining appreciable specific heat. In 1951 the idea seemed quite impractical in view of the scarcity of He3, but in 1955 London revived it in collaboration with E. Mendoza and G. Clarke. They first measured the osmotic pressure of He3 dissolved in He4 and attempted adiabatic dilution starting at 0.8° K. but obtained no cooling. After G. K. Walters and W. M. Fairbank had discovered that He3 and He4 separate at 0.87° K. into two phases, one concentrated and one dilute, London and Mendoza at first concluded that dilution cooling would be ineffective below about 0.8° K. They later realized that it could be achieved in another way, by exploiting the latent heat of transition from the concentrated to the dilute phase.3 The essential principle of the dilution refrigerator as conceived by London and Mendoza is illustrated in Figure 2. Refrigeration occurs in the pot A, where the lighter He3-rich phase floats on top of the dilute

phase and expansion proceeds downward. The diluted mixture is then taken out to a still, C, where the He3 evaporates and is recycled through heat exchangers B, B’. The effectiveness of the cooling cycle depends on the fortunate coincidence of two special properties of He3. First, the phase boundary for the mixtures has a shape such that significant proportions of He3 enter the dilute solution down to the absolute zero. Second, He3 has a higher vapor pressure than He3 and is therefore evaporated preferentially in the still. It is customary to use the refrigerator with conventional He3 cooling as a first stage. After various experiments by Mendoza at Manchester, a working model was built in 1965 by H. E. Hall, P. J. Ford, and K. Thompson.4Improved versions with operating temperatures down to 0.005° K. have since been made and marketed in several countries, and the device is now widely used.

The two London brothers form a study of unusual psychological interest. Fritz’s influence on Heinz in the critical years after their father’s death was strong, yet despite their close professional relationship they are in many ways contrasting figures. Fritz was well-organized, Heinz shockingly untidy. It is customary to think of Fritz as a theoretician and Heinz as an experimentalist, but actually neither fits exactly into ordinary categories. Of Heinz, D. Shoenberg has said: “Though he spent much time on experiments his most valuable contributions have been ideas and inventions. Perhaps he might be described as a cross between a theoretical physicist and an inventor.”5 In their theoretical contributions Fritz’s genius consisted in identifying the deep conceptual issues, Heinz’s in the vital and far from simple task of bringing experiment into fruitful contact with theory. As an experimentalist he belonged to that not undistinguished class of workers, among whom was J. J. Thomson, whose junior colleagues wisely protect them by various subterfuges from too close contact with apparatus. His clumsiness was a byword. Both brothers brought a remorseless thoroughness to everything they did. Heinz’s attitude to physics had also an element of spiritual passion. “For the second law of thermo dynamics,” he proclaimed vehemently to one colleague, “I would die at the stake.”6


1. W, F. Vinen, “The Detection of Single Quanta of Circulation in Liquid Helium II,” in Proceedings of the Royal Society, 260(1961), 218.

2. G. B. Hess and W. M. Fairbank, “Measurement of Angular Momentum in Superfluid Helium,” in Physical Review Letters, 19 (1967), 216.

3. Many interesting details are given in an unpublished paper by E. Mendoza, “Early History of the He3-He4 Dilutions Refrigerator: Some Personal Impressions.”

4. H. E. Hall, P. J. Ford, and K. Thomson, “A Helium-3 Dilutions Refrigerator,” in Cryogenics, 6 (1966), 80.

5. D. Shoenberg, obituary of H. London, in Biographical Memoirs of Fellows of the Royal Society, 17 (1971), 441–461.

6. H. Montgomery, quoted in 5.


A complete bibliography of London’s published papers with an admirable biographical sketch is D. Shoenberg, in Biographical Memoirs of Fellows of the Royal SocietyM, 17 (1971), 441–461. His unpublished MSS are in Mrs. London’s possession. See also K. Mendelsohn, in Review of Modern Physics, (1964), 71.

For further sources on particular areas of low-temperature physics, see the bibliography for Fritz London.

C. W. F. Everitt

W. M. Fairbank