3—
The Radio and the Cyclotron
Going on the Air
The distinctive feature of the Radiation Laboratory's approach to particle accelerators was reliance on radio technology. Many of the Laboratory's earliest workers, including Lawrence, had been radio hams in their youth; they continued the sport on a grand scale by a ham link between the Laboratory and one of its earliest satellites at the University of Michigan.[70] The heart of Livingston's first cyclotron was a pair of off-the-shelf Radiotrons, air-cooled, rated at 75 watts each and arranged in a standard radio circuit, with the single dee in place of an antenna. The analogy is not idle. The cyclotron could interfere with commercial broadcasting or police communications; the Laboratory listened to itself on the radio so as to correct spillage into others' air waves and "prevent investigations by the Radio Commission which may result in our having to put in elaborate and expensive protection devices." It is said that Lawrence used to tune his home radio to the cyclotron's operating frequency to monitor its, and his students', performance.[71] The second cyclotron required a 20-kW water-cooled oscillator in another standard radio circuit; the third cyclotron used two 20-kW tubes, which also drove Sloan's x-ray machine.[72] As we know, their cost provoked the Laboratory to make their own and to enter still more deeply into the art of radio engineering. This move so far from ordinary physics was a surprise even to people familiar with Lawrence's methods.[73] The Laboratory was so filled with radio waves that its members could light a standard electric bulb merely by touching it to any metallic surface in the building.[74] Many cyclotron laboratories were to eke out their resources by cannibalizing old radio parts.[75]
[70] Cork to Lawrence, 21 Aug 1936, and Lawrence to Cork, 26 Aug 1936 and 9 Feb 1937 (5/1).
[71] Lawrence, memo to Purchasing Dept., 31 Jul 1935; Childs, Genius , 251.
[72] Terman, Radio-engineering , 228, 253, 258–9; Livingston, Production , 13–4, and fig. 3; Lawrence and Livingston, PR, 40 (1932), 28, in Livingston, Development , 127; Lawrence, "Workbook," 139 (39/4).
[73] Cooksey to Lawrence, 6 May 1933 (4/19).
[74] Alvarez, Adventures , 42.
[75] E.g., disused tubes from a radio station, at Rochester, Dubridge to Lawrence, 20 Sep 1935 (15/26); an old radio transmitter from the navy, at Har-Lawrence, 20 Sep 1935 (15/26); an old radio transmitter from the navy, at Harvard, Hickman to J.R. Wier (GE), 1 May 1937 (UAV, 691/60/3); out-of-date radio tubes for ionization gauges at Berkeley and elsewhere, Lawrence to H.W. Edwards, 18 Jan 1934 (7/2).
The maximum energy given particles by a cyclotron depends not on the power but on the frequency of its oscillator. From the fundamental equation 2.1,
where R is the radial distance to the collecting cup. Livingston achieved million-volt protons with the second cyclotron with the cup at 11.5 cm and the oscillator at 20 MHz. This last number represented close to the practical limit on frequency: to have gone higher would have required, first, a more intense magnet (since the frequency is proportional to the field strength) and, second, a tube capable of changing polarity over twenty million times a second and delivering around twenty thousand watts of power. The first requirement would have pushed the art of magnet design, the second that of power oscillators, to or beyond the edge of available technology. The only immediate way to increase the energy of the protons by an order of magnitude was to increase R three-fold . Hence the great value to the Berkeley cyclotroneers of Federal's derelict magnet, which could be rebuilt to give a field of appropriate intensity between pole pieces 27.5 inches (70 cm) in diameter.
The Federal magnet, like the oscillator tubes, was a product of radio technology, the Poulsen generator, which had at its heart a periodic arc between electrodes in hydrogen. When the arc struck, it carried the oscillatory discharge from a large condenser, which fed an antenna; when the discharge current diminished sensibly, the arc went out and a battery recharged the condenser, which, when full, relit the arc (fig. 3.4). The magnet assisted the extinction of the arc: it made the ions carrying the current run in a curved path from one electrode to another; they therefore could not create by collisions any fresh ions along the straight line between the electrodes; and consequently not enough carriers were available there to continue or restart the arc when the potential across the gap at A (fig. 3.4) no longer sufficed to ionize the gas
Fig. 3.4
Schematic diagram of the electrical arrangement in the Poulsen arc. I1 , the
charging current; I , the rf current through the resonant circuit when the arc
strikes. The choke coils L keep the rf current from the battery circuit.
Heilbron, Museoscienza, 22 (1983), 16.
near the electrodes. In a word, the field insured that the arc went out without flickering (plate 3.5). The objective was to produce an undamped signal of almost constant frequency rather than the broadband output characteristic of the spark transmitters of early wireless.[76]
We are familiar with Federal's growth under navy contracts, and with the culmination of their relations in the commissioning of four 1,000-kW generators, two for each end of a radio link between the United States and its expeditionary forces in France. Their magnets could deliver 18,000 gauss.[77] The war ended before the huge antenna towers—second only in height to the Eiffel tower—could be completed in France. The navy withdrew, leaving France with half a radio station and Federal with four 80-ton magnets. In 1919 the French government decided to proceed with the Lafayette station, as they called it in memory of the American alliance, in order to communicate with its empire in Southeast Asia. Its signal received in San Francisco was four to eight times stronger than the signals from other major European transmitters. That did not recommend completion of the American end of the
[76] Heilbron, Museoscienza, 22 (1983), 13–24.
[77] Howeth, Communications , 241, 243, 246, 253; Anon., Radio review, 2 (1921), 85–90.
link to the navy, however, which now favored development of vacuum-tube oscillators. That left Federal with two war-surplus magnets. Everyone switched to tubes except the Dutch, who constructed a 3,600-kW Poulsen generator to drive an antenna stretched between two hills in Java, which gave the Dutch East Indies a voice audible in Amsterdam.[78]
Lawrence's big magnet, being a piece of high technology, required professional engineering help in its metamorphosis into a tool of science. Fuller advised about renovating the pole pieces and about windings and power supplies; and he procured the assistance of an employee at Federal, Gilbert W. Cattell, who inquired into all sorts of details: the best sort of paper insulator for the windings, of oil for cooling, of cables for connecting, and so on.[79] Cattell wound and insulated the coils at Federal before the magnet went to Pelton for machining (plate 3.6). Toward the beginning of February 1932, the foreman at Pelton's machine shop, Henry Nelson, carted its handiwork across the Bay and erected it in the Radiation Laboratory. Pelton did an excellent job, the pole faces parallel to four-thousandths of an inch, the field homogeneous up to 18 kG.[80] Lawrence hoped to have the cyclotron itself in operation at the end of February and shortly thereafter to pass "the next milestone," protons with energies above 3 MeV.[81] In March he worked "night and day" with Livingston and a graduate student, James Brady, on the machine. "I have neglected everything else—even my fiancée has suffered."[82] Molly's suffering did not make the machine go. In April they thought that they had cured the leaks that the huge magnetic forces kept springing in the cyclotron tank. Still no results. Lawrence went East in May, to Molly and marriage. Livingston
[78] Ibid., 93, 579; Howeth, Communications , 136, 208–9; Hooper, Electr., 18 (1921), 1112–3.
[79] Correspondence between Lawrence and Cattell, Nov and Dec 1931 (25/2); Fuller, interview, 72, 142–5 (TBL).
[80] Pelton to Lawrence, 14 Sep 1931, 29 and 30 Oct 1931, and Lawrence to Pelton, 2 Feb 1932 (25/2); Lawrence and Livingston, PR, 45 (1934), 608.
[81] Lawrence to Poillon, 9 Jan 1932 (15/16), to Tuve, 13 Jan 1932 (17/34), to Livingood, 18 Feb 1932 (12/11), and to Buffum, 21 Jan 1932 (46/15R); cf. Lawrence to Tuve, 10 Nov 1931 (17/34), expecting to have the machine going before Christmas.
[82] Lawrence to Haupt, 11 Mar 1932, and reply, 15 Mar 1932 (9/2).
labored on; the chairman of the Physics Department, Elmer Hall, alarmed at his appearance, counselled a long rest.[83]
Vacation cured what slavery had not. By mid September Livingston had overcome his various difficulties and produced hydrogen-molecule ions at 1.6 MeV. "When he gets up in the 3,000,000 volt range [Lawrence wrote Cottrell] he intends to stop and bombard various elements with them before going to higher energies." A week later he had reached 3.6 MeV at about one µA, always with
Stopping for physics was perhaps a pleasure. It was certainly a financial necessity. Lawrence had started planning for a larger apparatus long before Livingston had got a beam. He did not plan to carry off the Poulsen arc from Java, but to enlarge the pole pieces and vacuum chamber of the new cyclotron from 27.5 to 37.5 inches. He floated this bubble in April 1932. Poillon indicated that he would incline toward granting the $1,100 needed when asked for it. But Lawrence had overreached. Poillon had the same month pledged $2,500 to the Laboratory to enable it to make much needed detectors, a cloud chamber with cinema camera ($1,500) and a magnet to analyze particle beams ($1,000). Somehow Lawrence thought that the Research Corporation and the Chemical Foundation would provide another $800 for instruments and supplies; but when he asked for it in August, he was told that the University should pay for such things. Lawrence already had spent almost the entire allocation for expenses and power for AY 1932/33. It was only August! He would have to close down the Laboratory, he said, unless his hard-fisted backers allowed him to pay for his electricity from the $1,000 granted for
[83] Lawrence to Poillon, 12 Apr 1932 (3/28, 5/16); Hall to Lawrence, 16 June 1932 (8/8).
[84] Lawrence to Cottrell, 22 Sep 1932 (5/3), and to Cooksey, 29 Sep 1932 (4/9); Livingston, PR, 42 (1933), 441–2 (3 Oct 1932). Parameters: R = 10 inch (25.4 cm), H = 15,250 gauss, f = 11.6 MHz. Cf. Lawrence to Barton, 29 Sep 1932 (2/25), and to Boyce, 1 Oct 1932 (3/8).
the analyzing magnet. As an inducement to Poillon, Lawrence dropped—only temporarily—the unfunded proposal to enlarge the pole pieces and vacuum chamber. Poillon allowed the redirection of funds.[85]
Broadcasting Upon the Waters
The creation of the 27-inch cyclotron called for an unusual blend of faith, energy, and entrepreneurism. How unusual the combination was may be gathered from the reluctance of other physicists to follow Lawrence's lead. The nuclear physicists in and around the Cavendish at first considered the cyclotron to give too small a current to be useful.[86] They next depreciated it as "ticklish to adjust," a view common among English physicists as late as 1935, although Cockcroft had seen for himself, in 1933, that "all the trick lies in correcting for inhomogeneities of field around the gap by inserting 'shims' or pieces of sheet iron." He had witnessed the operation; the trickster was Livingston, "who does most of the work."[87] A closer observer saw the same thing. "Lawrence does no actual experimental work anymore," Birge wrote another member of the Department. "It keeps him busy just bossing all the men working with him!"[88]
The first Englishman to see the 27-inch run, Ralph Fowler, came to Berkeley in January 1933, when the machine was producing 2.5 MeV hydrogen molecule ions. Neither the machine nor its products interested Fowler. "Probably only trivial stuff," he wrote Rutherford. Six months later Cockcroft did not see that the cyclotron had opened up any important areas of investigation not accessible to the Cockcroft-Walton accelerator. "We can get in long before California in this field [he wrote his co-inventor] and there are a lot of points to be cleared up [about nuclear
[85] Poillon to Leuschner, 6 Apr 1932, and to Lawrence, 30 Aug and 8 Sep 1932; Lawrence to Poillon, 26 Aug and 5 Sep 1932 (46/25R, 15/16). The cost of enlargement was projected at over $2,000 when Lawrence next mentioned it to Poillon, in a letter of 23 Jan 1934.
[86] Cockcroft to Gamow, 29 Sep 1932 (CKFT, 20/10), and to Lawrence, 17 Sep 1932 (5/4); cf. McKerrud (Metropolitan-Vickers) to Cockcroft, 27 Feb 1932 (CKFT, 20/60).
[87] Cockcroft to Walton and Dee, [24 June 1933], "ticklish," and to Rutherford, 22 Jul 1933, "trick" (ER); Lawrence? to Chadwick, 31 Jan 1936 (3/34).
[88] Birge to Jenkins, 18 Sep 1932 (Birge P, 33).
processes]."[89] There is no doubt that the 27-inch was a temperamental machine and that the maintenance of its vacuum occupied and irritated many members of the laboratory. But where visitors saw unreliability, and nonvisitors doubted that cyclotrons worked at all, the natives appreciated difficulties overcome and augured a rich harvest when they turned their machine to physics.[90]
Ever zealous and generous in his cause, Lawrence told Cockcroft that he could easily reproduce the Berkeley machine with a magnet that could be had for the asking. He referred to the Lafayette arcs, then about to be junked in favor of vacuum-tube oscillators and to decommissioned 500-kW arcs procurable from navy surplus. "[They] could be obtained for a song and transport," Cockcroft wrote Rutherford, "if Oliphant [Marcus Oliphant, a prominent nuclear physicist at the Cavendish] shows any enthusiasm in this direction." Oliphant did not show enthusiasm, preferring to go to a million volts in the old, direct, dependable, one-step way. Not until 1936 did the Cavendish decide to create a cyclotron, which it did from scratch, following the plans of Berkeley's then newly designed 37-inch machine. The delay put it two generations of accelerators behind Berkeley at the end of World War II.[91]
While the Cavendish was losing its opportunity, Frédéric Joliot, son-in-law and heir apparent to Mme Curie, inquired of Lawrence what it might cost to build a cyclotron. The expense of the large magnet worried him. Lawrence replied with news about the Lafayette monsters and the necessary modifications of the pole pieces.[92] Joliot applied to the engineer at the station and received all the consideration he could have wished. One of the magnets was being dismounted. "It is only a matter of hauling it to the Ecole Normale Supérieure." The refurbishing could be done in the station's shop, which could also supply 20-kW water-cooled
[89] Fowler to Rutherford, 15 Jan 1933, and Cockcroft to Walton [24 June 1933] (ER).
[90] Cf. Varney, PT, 35:10 (1982), 27; Rasetti, Viaggi, 3 (1936), 78, complains about the irregular functioning of the 27-inch; Kurie to Cooksey, 4 Mar 1934 (10/21), says that "no one really believes [Lawrence's] cyclotron works."
[91] Cockcroft to Walton, [24 June 1933], and to Rutherford, 22 Jul 1933 (ER); Oliphant to ER, 7 Jan 1934 (ER); Crowther, 186, 230; infra, §7.2.
[92] Joliot to Lawrence, 14 June 32, and Lawrence to Joliot, 20 Aug 1932 (10/4).
oscillator tubes. The director of the French wireless service authorized the visit.[93] Then Joliot dropped the initiative, because, he later said, he could not get permission to rebuild the magnet. But it is probable that the size of the task, the demand for expertise in radio engineering, and the prospect of dragging an 80-ton magnet through the narrow streets of the Latin Quarter combined to defeat his interest.[94] Like his Cambridge colleagues, Joliot waited until 1936 to begin constructing a cyclotron, and he then required help from Berkeley to correct the many mistakes in its design.
No European center made use of Poulsen arc magnets except the Centre anticancéreux in Marseilles, which made a cyclotron from a small unit decommissioned from the telegraph service in Lyon.[95] The Dutch were the last to consider the option, in 1940, when they asked Lawrence for advice on cannibalizing their Batavian transmitter, which could attain 20 kG.[96] The war put an end to the plan. Cyclotroneering qualities were more easily found in the New World than in the Old. In 1932 New York University inquired into the fate of the 500-kW arcs at Annapolis and reserved one at its decommissioning in June 1934. In 1933 Stanford obtained the mate to Lawrence's magnet from Federal on the understanding that it would be released to the first institution that succeeded in raising the money needed for conversion. In 1934 Cornell tried to obtain one of the decommissioned Annapolis magnets. Columbia, too, was interested; the navy had but one magnet to give, the one that NYU had reserved but had since relinquished; Columbia, which had been frustrated by the navy's sale as scrap of another Poulsen magnet it had coveted, won the prize.[97]
[93] Ingénieur en chef, Station radiotélégraphique, Croix d'Hirns, to Joliot, 29 Oct 1932, 4 and 13 Jan 1933; Directeur du service de la TSF to Joliot, 17 Jan 33 (JP, F28).
[94] Joliot, preface to Nahmias, Technique . Possibly also Joliot could not acquire the pole piece from the second magnet, which he would have needed to make the first symmetrical, and could not raise the money to construct another.
[95] Nahmias, Machines , 27. Segrè proposed to utilize a magnet from a Poulsen installation in Italy, but it had been dismantled by the time he asked about it (personal communication).
[96] J. Clay to Lawrence, 23 Jan 1940 (3/20).
[97] Breit to Lawrence, 6 May 1932 (2/15), re NYU; Lawrence to J.E. Henderson, 14 Apr 1933 (9/5), re Stanford; Livingston to Lawrence, 20 Sep 1934 (12/12),re Cornell; Science service , 25 Jan 1939, and other documents (Pegram P, 27/"cyclotrons"), and infra, §6.2, re Columbia.
The last cyclotrons built around a Poulsen arc came into existence in 1951. The circumstances were unusual. The Japanese had begun to make cyclotrons as early as the British and the French. At the end of World War II, they had three. The U.S. Army, newly afraid of nuclear physics, threw them all into the ocean. Lacking resources to buy or build a substitute, physicists at the Nishina laboratory in Tokyo scrounged for parts to reconstruct what they had lost. "Fortunately we have a magnet which was originally used for a Poulsen arc generator."[98]