The discovery of superconductivity

2011-04-25

Dirk van Delft and Peter Kes describe events surrounding the discovery that made possible the powerful magnets used today in particle accelerators such as the LHC.

One hundred years ago, on 8 April 1911, Heike Kamerlingh Onnes and his staff at the Leiden Cryogenic Laboratory were the first to observe superconductivity. In a frozen mercury wire, contained in seven U-shaped capillaries in series, electrical resistance suddenly seemed to vanish at 4.16 K.

The discovery of superconductivity may have been accidental, but nonetheless the experiment was part of a carefully considered research programme at Leiden. Studying the behaviour of the electrical resistance of metals (such as gold and platinum) at low temperatures was interesting from both a practical and a theoretical point of view. Practical, because the fact that metal resistors were dependent on temperature made it possible to use them as (secondary) thermometers – thereby raising the possibility of a welcome addition to the (primary) gas thermometer that, although accurate, was cumbersome to use and slow in response. Theoretical, because Paul Drude had already applied the kinetic theory of gases to an electron gas in a metal in 1900, and on that basis had deduced the linear decrease in resistance with absolute temperature, while William Thomson (Lord Kelvin) had predicted one year later that at extremely low temperatures, the conducting electrons would, in fact, become "frozen solid" to the atoms, such that at absolute zero, resistance would become infinite.

Using liquid hydrogen as a coolant, Jacob Clay and other students of Kamerlingh Onnes had succeeded in carrying out experiments down to 14 K (the freezing point of hydrogen) at the Leiden Physics Laboratory, starting in 1906. It was noted during these experiments that, although the resistance of gold and platinum wire did fall with decreasing temperatures, at the same time it started to level out. The successful liquefaction of helium on 10 July 1908 gave a massive boost to this research because, at a stroke, temperatures as low as 1 K had suddenly been made achievable. The result of these new measurements was that at such low temperatures, resistances reached a sort of residual value that became lower the purer the platinum or gold could be made.

The logical next step was the choice of mercury because, via distillation, the metal could be made extremely pure. The capillary construction, a masterpiece of the Leiden-based glass-blower Kesselring, was installed in the helium cryostat next to the liquefactor. The actual goal of the experiment was the test of the transfer system for liquid helium. During the decisive experiment on 8 April 1911, Kamerlingh Onnes and Gerrit Jan Flim, head of the cryogenic laboratory and master instrument maker, were responsible for the cryogenic installations. Measuring the temperature (using a gas thermometer) was the task of Cornelis Dorsman, while the resistance of the mercury wire (and of gold) was determined via an electrical bridge circuit with a mirror galvanometer. The galvanometer was placed in a room at a safe distance from the throbbing pumps, on a vibration-proof column, and was monitored by Gilles Holst (who communicated via a speaking tube). The result of these experiments was that the mercury resistance did, indeed, fall to zero.

In December 1912, mercury as a superconductor was joined by tin and lead, metals with a transition temperature of 3.8 and 7.2 K, respectively. From then on, there was no need to experiment with fragile mercury capillaries. Experiments could now be carried out with handy coils of wire.

The stakes were high: nothing less than a compact, powerful superconducting magnet. At the start of the century, Jean Perrin had already put forward the idea of a liquid-nitrogen-cooled magnet of copper wire, with a magnetic field of 100 kG.

At the third International Congress of Refrigeration in Chicago, in the autumn of 1913, Kamerlingh Onnes once again raised the issue of the super magnet. "The solution to the problem of obtaining a field of 100,000 gauss could be obtained by a coil of, say, 30 cm in diameter, and the cooling with helium would require a plant that could be realized in Leiden with a relatively modest financial support," he wrote in his summary of the cryogenic work in Leiden. "Since we may confidently expect an accelerated development of experimental science, this future ought not to be far away."

It was not until the 1960s that the powerful superconducting magnet was finally introduced, thanks to niobium-titanium wire. This is a conventional superconducting material with a high threshold field, a large current density and a transition temperature (TC) of 9 K. MRI scanners and deflection magnets in particle accelerators still make use of magnets of this kind.

Source: CERN Courier website