A Preference for High Tension

The major laboratories for nuclear physics in France and Britain had declined the opportunity to clone the 27-inch cyclotron from the obsolescent Poulsen-arc magnets decommissioned by the

French radio service in 1932. Their disinclination had both negative and positive causes. On the minus side, the cyclotron of 1932 ran unsteadily, could not accelerate electrons, and when operating delivered only a fraction of a µA to the target. It retained its reputation for unreliability after it had attained the ability to work an eight-hour day, perhaps because visitors who came when it was under repair advertised their disappointment.^[9] As late as the winter of 1935/36, the cyclotroneers themselves acknowledged that their machines did not function most of the time. At Princeton they expected to operate only two hundred hours a year. Making his usual virtue of necessity, Lawrence reassured Chadwick that he need not worry much about the power bill for his planned cyclotron since it would run so infrequently. We may recall Lawrence's inability to supply $2,000 worth of radiosodium to clients of the Macy Foundation in 1935.^[10]

By the winter of 1937/38 a well-made cyclotron could work steadily if not driven at its maximum capabilities. In the spring of 1938 the 37-inch ran at an average of eight hours a day, seven days a week, and Princeton's cyclotron went on for three months without serious mishap. Nonetheless the impression remained abroad that, as Lawrence wrote to an English physician trying to promote a medical cyclotron in Britain, "the cyclotron is still a very unreliable apparatus;" and occasionally he had to reassure American inquirers that "the cyclotron is no longer a capricious laboratory device. . . . [but] an efficient and thoroughly rugged and reliable apparatus." And yet, between the desire to improve performance and the need to repair and replace parts, cyclotroneers often had their machines apart. In July 1938, in a tour of all the cyclotron laboratories in the East, Livingston saw nothing to see. "Don't let this get out," he wrote Lawrence, "but I did not find a single cyclotron operating."^[11]

In addition to their unreliability, the early Berkeley machines were depreciated in Europe for their finicky and "empirical" character. The Europeans had the "general impression [Lawrence acknowledged] that the cyclotron is a very tricky and difficult apparatus to operate." As for empiricism, the missionary cyclotroneers freely admitted the charge, and even gloried in what elevated a possible science into an actual art. In describing the Cornell cyclotron, Livingston pointed to the size of the gap between pole faces, the height of the dee aperture, the position of the source filament, and the shimming of the magnet as parameters that could only be fixed by "experimental maneuvers;" at Princeton, Henderson and White admitted to the method of cut and try in setting the dimensions of their magnet, the position of their filament, and so on, and to their inability to justify many design details, "except to say that [they are] known to work."^[12] Nor did the mother church affect to know the principles of its practice. In describing the definitive version of the 27-inch cyclotron late in 1936, Lawrence and Cooksey recommended maximizing the beam by ad hoc adjustments to, among other things, the deflecting potential, the shimmed magnetic field, and the positions of the dees, deflecting plate, and source filament.^[13]

All this summed to trop d'empirismes , too much tinkering, according to Nahmias, who thus depreciated the cyclotron after a visit to Princeton. The criticism was almost out of date when Nahmias made it in the spring of 1937; cyclotrons then being built would operate with fewer empirismes as well as with greater regularity. Here one story is worth a thousand words. Cooksey and Kenneth Bainbridge of Harvard visited the Bartol cyclotron in April 1938. Its creator, Alexander Allen, threw the switches; a beam came immediately, without fiddling. Bainbridge, who had never seen a cyclotron work, cried in astonishment, "Why, he just turned it on!"^[14]

In comparison with cyclotrons, again according to Nahmias's survey of March 1937, Tuve's improved two-meter Van de Graaff generator, which operated with great reliability at 1,000 kV and even at 1,200 kV, "under good conditions, compounded of low humidity, good fortune, and infinite other ingredients," had the advantage of few empirical adjustments to achieve a strong, homogeneous beam. High-tension apparatus not only produced better beams for exact work, but it did so by scaling up devices familiar to physicists. The cyclotron could not make headway in Europe until it could demonstrate advantages so decisive that physicists there would undertake to master the alien field of radio technology. After a visit home to the Cavendish in 1934, Bernard Kinsey succinctly explained his countrymen's reluctance to make cyclotrons: "They are all scared stiff at the thought of setting up an oscillator."^[15] Lawrence himself pointed to the radio engineering as the most challenging part of his operation: "The difficulties encountered [in making a new cyclotron] are similar to those found when a new radio broadcasting station of design and power that's never been used before is first constructed." The high-frequency oscillator for Bohr's cyclotron was to have an energy and to present a difficulty larger than the shortwave transmitter of the Danish state radio.^[16] Why trouble with a finicky machine and unfamiliar technology when other sorts of accelerators seemed capable of doing the same or similar jobs?

Tuve's group had advertised their technique by veiled reference to the superiority of the point of view of the investigator in a high-tension laboratory to his counterpart in a (or rather in the ) cyclotron laboratory. Their call for detailed exploration of reactions initiated by homogeneous beams before subliming to high voltages met with widespread agreement outside Berkeley.^[17] A clear conscience and a pure beam did not exhaust the advantages

of high-tension over magnetic-resonance acceleration. As Tuve told Nahmias, a Van de Graaff version of high-tension accelerators could work at continuously variable voltages, could push electrons as well as positive ions, could make x rays, might achieve as much as 10 MeV, and would do it all at less expense than any other model. Hence, he said, Westinghouse was building a Van de Graaff with two concentric spheres, the larger 10 meters in diameter, large enough to enclose an ordinary laboratory space, and capable of maintaining a pressure of 10 atmospheres. If all went well, and, as we know, it did not, Westinghouse would get 10 MeV for $15,000.^[18]

Decisions taken at the Cavendish in 1935 and 1936, when the laboratory recognized the need to go to higher energies, indicate the considerations then at work against cyclotrons. In May 1935 its newly appointed building committee for a high-tension installation approved a report drawn up by Cockcroft and Oliphant, who stressed the "immediate importance to develop apparatus for accelerating charged particles by at least two million volts." They recommended "an extension of the [Cockcroft-Walton] method of producing high steady potentials which has been in successful operation for two [more accurately three] years." In reaching this conservative conclusion they had the assistance of Philips of Eindhoven, which they had visited the preceding January. The Philips laboratories were "an eyeopener" to Oliphant, so he wrote Rutherford, "and in the opinion of Cockcroft . . . far superior to any in America." Oliphant eyed a million-volt high-tension set and desired to build a similar one, at twice the voltage, at the Cavendish.^[19] In explaining their decision to Lawrence, Cockcroft pointed to the Cavendish's interest in x rays as well as positive ions and to the existence of "plenty of [American] laboratories who would be capable of using the cyclotron method."^[20]

Cockcroft and Oliphant calculated that their laboratory would require a building sixty feet long, forty feet high, and forty feet wide, with concrete walls a foot and a half thick and a mobile

crane to install and repair the large tubes, transformers, and generators. They estimated the cost of the building at 4,000 pounds and that of the apparatus at under 3,000 pounds, over twice the price of the cyclotron then nearing completion at Princeton. They underestimated by much more than a factor of two.^[21] "The approximate cost of 15,000 pounds [for the building] staggered me, as I imagine it will you," Oliphant wrote Rutherford. "I can see the new laboratory receding into the distance if we are not careful. . . . It is a thing we need urgently, and not in some distant future when all the cream has been scooped off by folks whose results we dare not trust too deeply."^[22] From which it appears that Oliphant had in mind to do physics and further battle with Berkeley with his machine.

By July 1935 Rutherford had accepted these objectives and promoted Oliphant and the building. The former he made assistant director of research in place of Chadwick, who left the Cavendish for a professorship at Liverpool. The latter he lobbied for so effectively before the council of the senate of Cambridge University that it recommended proceeding at once with university funds to be repaid by proceeds from an outside appeal for the 250,000 pounds the laboratory deemed necessary to meet all its research needs. Rutherford's reluctance to ask for money has been overestimated.^[23]

The great cost of the building, which was completed in 1937,^[24] and technical difficulties precluded going directly to 2 MeV. Again the Cavendish had an opportunity to consider a cyclotron and a more modest establishment. Again Oliphant went to Philips and again returned inspired. He decided to buy a copy of the 1.2 MV installation he saw at work. "After a stiff fight with some of

my colleagues and with Philips themselves over the price [just under 6,000 pounds] I have persuaded the Prof. [Rutherford] to invest in one of these sets." This trouble-free, ready-made instrument arrived in its impressive building ("a cinema outside, a cathedral within") after Christmas 1936 and worked as advertised.^[25] By then its promoters had forsaken it. Oliphant had been appointed professor in Birmingham the previous June, effective October 1937. Cockcroft had at last pronounced in favor of a cyclotron, and returned from a visit to Berkeley in 1937 intending to recommend selling Cambridge's high-tension equipment. That pleased Lawrence immensely. "The Cavendish Laboratory has expended large sums of money in installing high voltage equipment," he wrote in 1937, in a puff of Berkeley. "Although the Cavendish Laboratory pioneered with high voltage methods the distinguished scientists there have come to the conclusion that the cyclotron is superior, and are adopting it."^[26] The Philips accelerator, which had seemed essential to the Cavendish's place in nuclear physics, fell eventually under the management of a Swiss physical chemist, Egon Bretscher, who used it as a neutron source to make radioisotopes for chemists and biologists.^[27]

Several other British institutions followed Cambridge in preferring high tension to the cyclotron in the second generation of particle accelerators. At Bristol, for example, two members of the staff each tried to make a Van de Graaff generator, while a third, Cecil Powell, built a Cockcroft-Walton machine. In a few years, when they wished for a cyclotron, they judged that they could not acquire one on their own resources.^[28] At Oxford, the New Clarendon was designed for a Van de Graaff or a Cockcroft-Walton, although the professor, F.A. Lindemann, doubted that his university would put up the money for such expensive apparatus. As for

the cyclotron, "an instrument most popular abroad," it had the disadvantages, according to Lindemann, of not being cheap either and of not accelerating electrons. When in 1938 he decided that he wanted a cyclotron and proposed that the Ministry of Health pay for it, the ministry declined on the ground that physicists had not yet "decided on the scientific basis of the cyclotron." Lindemann rejected this opinion, which he supposed to derive from a prominent British physicist, as the nonsense it then was: "There is no possible doubt about the scientific basis of the cyclotron in Oxford and still less in California. . . . I have no doubt that if one liked to spend the money one could buy a ready-made cyclotron from America which would function quite satisfactorily."^[29] One did not like to spend the money, and Oxford got no cyclotron.

High-tension machines were also the preferred second-generation accelerators on the Continent. In Germany, where university physics had been all but destroyed by the implementation of Nazi racial laws, the leading centers of experimental nuclear physics sheltered in the institutes of the quasi-independent Kaiser-Wilhelm-Gesellschaft. In 1934 Walther Bothe, whose work had set up the discovery of the neutron, became head of the physics department of the Kaiser-Wilhelm-Institut für Medizinische Forschung in Heidelberg. He had considerable experience in high-voltage technique. With the help of Wolfgang Gentner, who had worked with Joliot and who was to become the first German cyclotroneer, he set up a Van de Graaff at 950 kV, which began to operate in 1937. Despite the name of their institute, Bothe and Geiger's work with high tension centered on basic physics, for example, photoinduced nuclear reactions, work that in Joliot's judgment was for a time the most productive "in radioactivity and nuclear physics."^[30] That made a curious reversal of the situation of cyclotron laboratories in physics departments in the United

States, which then were beginning to produce radioactive materials for biological work.

Also in 1937 the Kaiser-Wilhelm-Institut für Physik, a brand new institute built with a gift from the Rockefeller Foundation, officially opened its doors. Designed to emphasize nuclear physics, it had a tower fifteen meters in diameter and fifteen meters high to house a cascade generator of 2 MV. The Kaiser-Wilhelm-Gesellschaft's two high-tension installations were the only machines in Germany in 1937 capable of furnishing particle beams of a million volts or more. Not until 1938 did Bothe put forward a proposal to erect a small cyclotron, at a cost of 20,000 RM, which Gentner was to complete during the war.^[31]