Making of the Atomic Bomb

by Richard Rhodes

  Gresham’s Law operated with air raid shelters as it operates with good and bad money: the basements of better department stores like Dickens and Jones, where clerks carried around refreshments—chocolates and ice cream—filled up first. Because the bombing followed regularly, night after night, Londoners had time to get used to it, but adjustment could go either way, the confident beginner slowly unraveling, the frightened beginner moving beyond fear.

  More Londoners by far lived out the dangerous raids in their homes than in shelters: 27 percent fled to corrugated-iron Anderson shelters in back gardens, 9 percent to street shelters, only 4 percent into the Tube. By mid-November 13,700 tons of high explosives had fallen and 12,600 tons of incendiary canisters, an average of 201 tons per night; for the entire Blitz, September to May, the total tonnage reached 18,800—18.8 kilotons by modern measure, spread across nine months.1353 London civilian deaths in 1940 and 1941 totaled 20,083, civilian deaths elsewhere in Britain 23,602, for a total death by Blitz in the second and third year of the war (about which the United States was still officially neutral) of 43,685.1354 After that the bombing went the other way. Only twenty-seven Londoners lost their lives to bombs in 1942.

  At Oxford in December 1940, Franz Simon, now officially working for the MAUD Committee, produced a report nearly as crucial to the future of uranium-bomb development as the original Frisch-Peierls memoranda had been.1355 It was titled “Estimate of the size of an actual separation plant.” Its aim, Simon wrote, was “to provide data for the size and costs of a plant which separates 1 kg per day of 235U from the natural product.”1356 He estimated such a plant would cost about £5,000,000 and outlined its necessities in careful detail.

  Simon had never trusted the mails. He trusted them even less at the height of the Blitz. He duplicated some forty copies of his report, accumulated enough rationed gasoline for a round trip and shortly before Christmas drove from Oxford into bomb-threatened London to deliver the fruit of half a year’s hard work, his whole force in the struggle for his country, to G. P. Thomson.

  * * *

  The Germans may have been collecting radium, as Cockcroft thought MAUD RAY KENT signaled. They were certainly laying in industrial stocks of uranium. In June 1940, about the time Simon was hammering out his kitchen strainer, Auer ordered sixty tons of refined uranium oxide from the Union Miniére in occupied Belgium.1357 Paul Harteck in Hamburg tried that month to measure neutron multiplication in an ingenious arrangement of uranium oxide and dry ice—frozen carbon dioxide, a source of carbon free from any impurity other than oxygen—but was unable to convince Heisenberg to lend him enough uranium to guarantee unambiguous results. Heisenberg had larger plans. He had allied himself with von Weizsácker at the KWI. In July they began designing a wooden laboratory building to be constructed on the grounds of the Kaiser Wilhelm Institute for Biology and Virus Research, next to the physics institute. To discourage the curious they named the building the Virus House. They intended to build a subcritical uranium burner there.

  Germany had access to the world’s only heavy-water factory and to thousands of tons of uranium ore in Belgium and the Belgian Congo. It had chemical plants second to none and competent physicists, chemists and engineers . It lacked only a cyclotron for measuring nuclear constants. The Fall of France—Paris was occupied June 14, an armistice signed June 22—filled that need. Kurt Diebner, the War Office’s resident nuclear physics expert, rushed to Paris. Perrin, von Halban and Kowarski, he found, had escaped to England and taken Allier’s twenty-six cans of heavy water with them, but Joliot had chosen to remain in France.1358 (The French laureate would become president of the Directing Committee of the National Front, the largest Resistance organization of the war.)

  German officers interrogated Joliot at length when he returned to his laboratory after the occupation began. Their interpreter, sent along from Heidelberg, turned out to be Wolfgang Gentner, the former Radium Institute student who had confirmed that Joliot’s Geiger counter was working properly when Joliot discovered artificial radioactivity in 1933. Gentner arranged a secret meeting one evening at a student café and warned Joliot that the cyclotron he was building might be seized and shipped to Germany. Rather than allow that outrage Joliot negotiated a compromise: the cyclotron would stay but German physicists could use it for purely scientific experiments; Joliot would be allowed in turn to continue as laboratory director.

  The Virus House was finished in October. Besides a laboratory the structure contained a special brick-lined pit, six feet deep, a variant of Fermi’s water tank for neutron-multiplication studies. By December Heisenberg and von Weizsäcker had prepared the first of several such experiments. With water in the pit to serve as both reflector and radiation shield they lowered down a large aluminum canister packed with alternating layers of uranium oxide and paraffin. A radium-beryllium source in the center of the canister supplied neutrons, but the German physicists were able to measure no neutron multiplication at all. The experiment confirmed what Fermi and Szilard had already demonstrated: that ordinary hydrogen, whether in the form of water or paraffin, would not work with natural uranium to sustain a chain reaction.

  That understanding left the German project with two possible moderator materials: graphite and heavy water.1359 In January a misleading measurement reduced that number to one. At Heidelberg Walther Bothe, an exceptional experimentalist who would eventually share a Nobel Prize with Max Born, measured the absorption cross section of carbon using a 3.6-foot sphere of high-quality graphite submerged in a tank of water. He found a cross section of 6.4 × 10−27 cm2, more than twice Fermi’s value, and concluded that graphite, like ordinary water, would absorb too many neutrons to sustain a chain reaction in natural uranium. Von Halban and Kowarski, now at Cambridge and in contact with the MAUD Committee, similarly overestimated the carbon cross section—the graphite in both experiments was probably contaminated with neutron-absorbing impurities such as boron—but their work was eventually checked against Fermi’s. Bothe could make no such check. The previous fall Szilard had assaulted Fermi with another secrecy appeal:

  When [Fermi] finished his [carbon absorption] measurement the question of secrecy again came up. I went to his office and said that now that we had this value perhaps the value ought not to be made public. And this time Fermi really lost his temper; he really thought this was absurd. There was nothing much more I could say, but next time when I dropped in his office he told me that Pegram had come to see him, and Pegram thought that this value should not be published. From that point the secrecy was on.1360

  It was on just in time to prevent German researchers from pursuing a cheap, effective moderator. Bothe’s measurement ended German experiments on graphite. Nothing in the record indicates the overestimate was deliberate, but it is worth noting that Walther Bothe, a protege of Max Planck, had been hounded from the directorship of the physics institute of the University of Heidelberg in 1933 because he was anti-Nazi. “These galling fights so affected my health,” he wrote later in a brief unpublished memoir, “that I had to spend a long period in a Badenweiler sanitorium.” When Bothe was well again Planck appointed him to the Kaiser Wilhelm Society’s Heidelberg physics institute, but “the Nazis continued to harass me, even to the accusation of scientific fraud.”1361

  At nearly the same time—early 1941—Harteck learned at Hamburg what Otto Frisch had recently learned at Liverpool. Frisch had moved to the industrial port city in the northwest of England to work with Chadwick and Chadwick’s cyclotron. He built a Clusius tube there with a student assistant Chadwick assigned him—they moved in such energetic coordination through the laboratory that they won the nickname “Frisch and Chips”—and discovered, says Frisch, that “uranium hexafluoride is one of the gases for which the Clusius method does not work.”1362 The discovery set the British program back not at all, since Simon was already hard at work on gaseous barrier diffusion. But the German researchers had placed such faith in thermal diffusion that they had not bothered to develop a
lternatives. They quickly began doing so and identified several promising methods; oddly enough, barrier diffusion was not among them. Restudying the separation problem made it even clearer that U235 and U238 could only be separated by brute-force methods and at great expense.

  When Harteck reported to the War Office in March 1941, following a conference with his colleagues, he stressed their consensus that isotope separation would be feasible “only for special applications in which cheapness is but a secondary consideration.”1363 Only for a bomb, he meant—so he told the historian David Irving after the war. The German physicists gave “special applications” second place on their list; they recommended urgent work first of all on the production of heavy water. Like Fermi and Szilard, they opted initially for a slow-neutron chain reaction in natural uranium. Make that reaction work and “special applications” might follow. Knowing no more than they knew, they hardly had a choice.

  * * *

  Lieutenant Colonel Suzuki reported back to Lieutenant General Yasuda in October 1940.1364 He confined his report to a basic issue: the availability to Japan of uranium deposits. He looked beyond Japan to Korea and Burma and concluded that his country had access to sufficient uranium. A bomb was therefore possible.

  Yasuda turned then to the director of Japan’s Physical and Chemical Research Institute, who passed the problem on to his country’s leading physicist, Yoshio Nishina. Nishina, born late in the Meiji era and fifty years old in 1940, known for theoretical work on the Compton Effect, had studied with Niels Bohr in Copenhagen, where he was remembered as a cosmopolitan and exceptional man. He had built a small cyclotron at his Tokyo laboratory, the Riken, and with help from an assistant who had trained at Berkeley was building in 1940 a 60-inch successor with a 250-ton magnet, the plans for which had been donated by Ernest Lawrence. More than one hundred young Japanese scientists, the cream of the crop, worked under Nishina at the Riken; to them he was Oyabun, “the old man,” and he ran his laboratory Western-style with warmth and informality.

  The Riken began measuring cross sections in December. In April 1941 the official order came through: the Imperial Army Air Force authorized research toward the development of an atomic bomb.

  * * *

  Leo Szilard was known by now throughout the American physics community as the leading apostle of secrecy in fission matters. To his mailbox, late in May 1940, came a puzzled note from a Princeton physicist, Louis A. Turner. Turner had written a Letter to the Editor of the Physical Review, a copy of which he enclosed.1365 It was entitled “Atomic energy from U238” and he wondered if it should be withheld from publication. “It seems as if it was wild enough speculation so that it could do no possible harm,” Turner told Szilard, “but that is for someone else to say.”1366

  Turner had published a masterly twenty-nine-page review article on nuclear fission in the January Reviews of Modern Physics citing nearly one hundred papers that had appeared since Hahn and Strassmann reported their discovery twelve months earlier; the number of papers indicates the impact of the discovery on physics and the rush of physicists to explore it.1367 Turner had also noted the recent Nier/Columbia report confirming the attribution of slow-neutron fission to U235. (He could hardly have missed it; the New York Times and other newspapers publicized the story widely. He wrote Szilard irritably or ingenuously that he found it “a little difficult to figure out the guiding principle [of keeping fission research secret] in view of the recent ample publicity given to the separation of isotopes.”1368) His reading for the review article and the new Columbia measurements had stimulated him to further thought; the result was his Physical Review letter.

  Since U235 is responsible for slow-neutron fission, the letter pointed out, and ordinary uranium contains only one part in 140 of that isotope, “it is natural to conclude that only 1/140 of any quantity of U can be considered as a possible source of atomic energy if slow neutrons are to be used.”1369 But the truth may be otherwise, Turner went on. The fission energy of most of the U238, if it could not be used directly, might yet find indirect release.

  Turner was referring to the possibility that bombarding uranium with neutrons converted some of the uranium to transuranic elements, the transuranics that Bohr had hoped might have been banished by the discovery of fission. When an atom of U238 captured a neutron it became the isotope U239. That substance itself might fission, Turner suggested. But whether or not U239 did so, it was energetically unstable and would probably decay by beta emission to new elements heavier than uranium. And one or more of those new elements might be fissionable by slow neutrons—which would thereby indirectly put U238 to work.

  The next element up the periodic table from uranium would be element 93. Turner selected as the likeliest candidate for fission not 93 X 239, however, but the element next along, the element that 93 would probably decay to, 94 X 239, which he called “eka-osmium.”1 And 94 EkaOs 239, Turner proposed, changing from an odd to an even number of neutrons when it absorbed a neutron preparatory to fissioning (239 nucleons—94 protons = 145 neutrons + 1 = 146) just as U235 changed to U236, ought to be even more fissionable than the lighter uranium isotope: “In 94EkaOs240 . . . the excess energy would be even larger than in 92 U 236 and a large cross section for fission would be expected.”1370, 1371

  While Turner was thinking these theories through, two Berkeley men, Edwin M. McMillan and Philip M. Abelson, were moving independently toward demonstrating them. McMillan, a slim, freckled, California-born experimentalist, had been one of the men most responsible in the 1930s for improving Ernest Lawrence’s cyclotrons to the point where they worked steadily and produced reliable results. Soon after the news of the discovery of fission reached Berkeley in late January 1939 he had devised an elegantly simple experiment to explore the phenomenon. “When a nucleus of uranium absorbs a neutron and fission takes place,” McMillan told an audience later, “the two resulting fragments fly apart with great violence, sufficient to propel them through the air, or other matter, for some distance. This distance, called the ‘range,’ is a quantity of some interest, and I undertook to measure it.” He did so first with thin sheets of aluminum foil “like the pages of a book” stacked on a layer of uranium oxide backed with filter paper.1372 He bombarded the uranium with slow neutrons. Some of the fission fragments recoiled up into the stack of foils; each fragment embedded itself in a single sheet of foil at the end of its range, which depended on its mass; McMillan could then simply check successive sheets of foil in an ionization chamber, look for the characteristic half-lives of various fission products and read out the range (the uranium nucleus splits in many different ways, producing many different lighter-element nuclei).

  But aluminum itself is activated by neutron bombardment, which made half-life measurements difficult. So McMillan replaced the foils with a stack of cigarette papers previously treated with acid to remove any trace of minerals that might develop radioactivity under bombardment. “Nothing very interesting about the fission fragments came out of this,” he comments. The uranium coating on the filter paper under the stack of cigarette papers, on the other hand, “showed something very interesting.”1373 It showed two half-life activities different from those of the fission products that had recoiled away. And since whatever had remained in the uranium layer had not recoiled, the two different activities were probably not fission products. They were probably radioactivities induced in the uranium by captured neutrons. McMillan suspected that one of the two activities, the one with a half-life of 23 minutes, was one that Hahn, Meitner and Strassmann had identified in the 1930s as U239, “a uranium isotope produced by resonance neutron capture.”1374 The other activity left behind in the uranium layer had a longer half-life, about 2 days. In his report on his foil and cigarette-paper experiments McMillan chose not to speculate on what that second activity might be, but privately, he remembers, he thought “the two-day period could . . . be the product of the beta-decay of U-239, and therefore an isotope of [transuranic] element 93; in fact, this was the mo
st reasonable explanation.”1375

  To check that explanation McMillan needed some hint of the substance’s chemical identity. He expected that element 93 would behave chemically like the metal rhenium, element 75, next to osmium on the periodic table—would be “eka-rhenium” in the old terminology. He bombarded a larger uranium sample and enlisted the aid of Emilio Segré, who was now working as a research associate at Berkeley. “Segré was very familiar with the chemistry of [rhenium], since he and his co-workers [studying rhenium] had discovered [a similar element], now called technetium, in 1937.” Segrè began a chemical analysis of the irradiated uranium; in the meantime McMillan sharpened his half-life measurement to 2.3 days. Segrè, says McMillan, “showed that the 2.3-day material had none of the properties of rhenium, and indeed acted like a rare earth instead.” The rare earths, elements 57 (lanthanum) to 71 (lutetium), form a chemically closely related and odd series between barium and hafnium. Because of their middle-table atomic weights near barium, they often turn up as fission products. When Segrè found the 2.3-day activity acting not like rhenium, as expected, but like a rare earth, McMillan assumed that was what it was: “Since rare earths are prominent among the fission products, this discovery seemed at the time to end the story.” Segrè even published a paper on his work titled “An unsuccessful search for transuranic elements.”1376

  McMillan might have left it there, but the fact that the 2.3-day substance did not recoil away from the uranium layer nagged at him. “As time went on and the fission process became better understood, I found it increasingly difficult to believe that one fission product should behave in a way so different from the rest, and early in 1940 I returned to the problem.”1377 The 60-inch cyclotron, with a massive rectangular-framed magnet spacious enough to shelter Lawrence’s entire crew between its poles for a photograph—twenty-seven men, two rows seated on the lower jaw of the beast, Lawrence prominent at center, and a third row standing inside its maw—was up and running by then; McMillan used it to study the 2.3-day activity in more detail. He studied the activity chemically as well and managed the significant observation that it did not always fractionally crystallize out of solution as a rare earth would.