On 10 July 1908 , in his laboratory at Leiden University, the great Dutch physicist Heike Kamerlingh Onnes (1853-1926) experienced the most glorious moment of his career. That was the day he first liquefied helium and thus opened an entirely new chapter in low-temperature physics. (See the article in Physics Today, March 2008, page 36 ).
In a triumphant report to the Royal Netherlands Academy of Arts and Sciences (KNAW), Kamerlingh Onnes documented his achievement in great detail. Therefore it is remarkable that reliable details about his serendipitous discovery of superconductivity three years later have been hard to come by. Lack of information has led to speculations about the discovery. In particular, it has perpetuated an apocryphal tale about the role played by a sleepy young apprentice in Kamerlingh Onnes’s lab. That tale was treated as established fact in a September 1996 Physics Today article by Jacobus de Nobel ( page 40 ). There have even been rumors of the possible disappearance of Kamerlingh Onnes’s laboratory notebooks.
Enough reason, then, to have a close look at the Kamerlingh Onnes Archive, housed at the Boerhaave Museum in Leiden, to see whether any new clues could be found about the discovery of superconductivity — that most important consequence of the ability to reach liquid-helium temperatures.
Of course, it’s roughly known when the first two superconductivity experiments were carried out. Kamerlingh Onnes’s two earliest reports to the KNAW about zero resistance and “supraconductivity,” as he preferred to call it, are dated 28 April and 27 May 1911. Figure 1 shows his laboratory at about that time.
According to the archive’s inventory, two notebooks (numbers 56 and 57) should cover the period 1909-1912. 1 But on the cover of number 56 is written “1909-1910,” and 57 begins with an entry dated 26 October 1911. So it does indeed seem as if a crucial notebook is missing. That would explain why so many speculations began to circulate.
Another obscuring factor is Kamerlingh Onnes’s terrible handwriting. He wrote his lab notes, in pencil, in small household notebooks. They are very hard to read. After a few desperate hours trying, one tends to give up. And that’s a pity because, the cover notwithstanding, notebook 56 does indeed announce the 1911 discovery of superconductivity (see figure 2). Translated, the entry reads, “The temperature measurement was successful. [The resistivity of] Mercury practically zero.” A more literal rendering of the breezy Dutch tagline Kwik nagenoeg nul would be “Quick[silver] near-enough null.”
View largeDownload slide Figure 2. A terse entry for 8 April 1911 in Heike Kamerlingh Onnes’s notebook 56 records the first observation of superconductivity. The highlighted Dutch sentence Kwik nagenoeg nul means “Mercury[’s resistance] practically zero [at 3 K].” The very next sentence, Herhaald met goud, means “repeated with gold.” View largeDownload slide Figure 2. A terse entry for 8 April 1911 in Heike Kamerlingh Onnes’s notebook 56 records the first observation of superconductivity. The highlighted Dutch sentence Kwik nagenoeg nul means “Mercury[’s resistance] practically zero [at 3 K].” The very next sentence, Herhaald met goud, means “repeated with gold.” Close modal
When Kamerlingh Onnes took lab notes, he always started by writing down the date. For that entry he wrote 8 April, but not the year! He dated the second experiment on the resistivity of mercury 23 May, again without giving the year. It gets worse: Between those dates, he and Albert Perrier, a visitor from Lausanne, performed an entirely different experiment on the paramagnetism of liquid and solid oxygen. For that experiment, the written date—19 May 1910—did specify a year, but the wrong one! It should have been 1911.
Why did Kamerlingh Onnes make that mistake? It’s probably because an extensive series of similar experiments with Perrier had been carried out at the end of 1909 and during the first months of 1910. In any case, that little slip of the pencil has led many astray. It’s the most likely reason that researchers exploring the archive have, until now, not looked more closely at the lab notes. Had they made the effort, they would have found the excitement over the first successful transfer of liquid helium to a separate cryostat for measuring resistivity, the exact dates of the first superconductivity experiments, who was involved, and what their roles were. The notes also reveal that some nice, oft-retold stories about those events will always remain nice but will never become true.
“Mercury practically zero.” That penciled note heralded the birth of a new field. But probably at that moment Kamerlingh Onnes was simply thinking how right he had been to choose mercury. Zero resistance was what he expected to find in extremely pure metals at liquid-helium temperatures. 2 After he liquefied hydrogen in February 1906, he started a program to investigate the resistance of metals at low temperatures. He had a practical reason—thermometry. But he also had a purely scientific interest.
One of the issues in those days was the question of what happens to the resistivity of a metal as its temperature approaches absolute zero. 3 It was accepted that electrons were responsible for electric conductance and that resistance was due to the scattering of electrons by the ions of the metal crystal. Would the scattering amplitude decrease fast enough with falling temperature to yield zero resistance at zero temperature? Or would the mobility of the electrons also diminish at lower temperature, thus resulting in zero conductivity at absolute zero? If nature followed the latter prescription—put forward by Lord Kelvin in 1902—the resistance of a pure metal would first fall with decreasing temperature, then bottom out at some minimum, and finally climb to infinity at absolute zero.
In the earliest investigations at liquid-hydrogen temperatures in Leiden, Kamerlingh Onnes and his assistant Jacob Clay studied resistance R versus temperature T in very thin gold and platinum wires. 4 Before July 1908 the lowest available temperature was 14 K, at which solid hydrogen sublimates under reduced pressure. That was low enough to observe that the almost linear decrease of R with T at higher temperatures starts to level off to an almost constant value. In one of his KNAW reports, Kamerlingh Onnes even mentioned a trace of a minimum in the R(T) plot, which indicates that he originally believed in Kelvin’s model.
The almost linear R(T) behavior of platinum above 14 K made that metal suitable as a secondary thermometer. It was much more convenient than the helium gas thermometer Kamerlingh Onnes had been using. But a disadvantage was the platinum thermometer’s rather large size: 10 cm long and about 1 cm wide.
The resistance of the metal wires depended on the chemical and physical purity of the materials. For instance, Kamerlingh Onnes showed that the resistance increase due to adding small admixtures of silver to the purest available gold was temperature independent and proportional to the concentration of added silver. So, improving purity would yield metal wires of very low resistance that could serve as secondary thermometers at temperatures far below 14 K.