The Making of the Atomic Bomb

MeV threshold. What Bohr had drawn was thus a visualization of thorium's changing response to bombardment by neutrons of increasing energy.

Bohr moved to the next section of blackboard and drew a second graph. He labeled it with the mass number of the isotope most plentiful in natural uranium. “He wrote the mass number 238 with very large figures,” Rosenfeld says; “he broke several pieces of chalk in the process.” Bohr's urgency marked the point of his insight. The second graph looked exactly like the first:

But a third graph was coming.

Francis Aston had found only U238 when he first passed uranium through his mass spectrograph at the Cavendish. In 1935, using a more powerful instrument, physicist Arthur Jeffrey Dempster of the University of Chicago detected a second, lighter isotope. “It was found,” Dempster announced in a lecture, “that a few seconds' exposure was sufficient for the main component at 238 reported by Dr. Aston, but on long exposures a faint companion of mass number 235 was also present.” Three years later a gifted Harvard postdoctoral fellow named Alfred Otto Carl Nier, the son of working-class German emigrants to Minnesota, measured the ratio of U235 to U238 in natural uranium as 1:139, which meant that U235 was present to the extent of about 0.7 percent. By contrast, thorium in its natural form is essentially all one isotope, Th232. And that natural difference in the composition of the two elements was the clue that set Bohr off. He drew his third graph. It depicted one cross section, not two:

Having made a hard copy of his abrupt vision, Bohr was finally ready to explain himself.

Both thorium and U238 could be expected on theoretical grounds to behave similarly, he pointed out to Rosenfeld: to fission only with fast neutrons above 1 MeV. And it seemed that they did. That left U235. It followed as a matter of logic, Bohr said triumphantly, that U235 must be responsible for slow-neutron fission. Such was his essential insight.

He went on to explore the subtle energetics of the several reactions. Thorium was lighter than U235, U238 heavier, but the middle isotope differed more significantly in another important regard. When Th232 absorbed a neutron it became a nucleus of odd mass number, Th233. When U238 absorbed a neutron it also became a nucleus of odd mass number, U239. But when U235 absorbed a neutron it became a nucleus of even mass number, U236. And the vicissitudes of nuclear rearrangement are such, as Fermi would explain one day in a lecture, that “changing from an odd number of neutrons to an even number of neutrons released one or two MeV.” Which meant that U235 had an inherent energetic advantage over its two competitors: it accrued energy toward fission simply by virtue of its change of mass; they did not.

Lise Meitner and Otto Frisch had realized in Kungalv that a certain amount of energy was necessary to agitate the nucleus to fission, but they had not considered in detail the energetics of that input. They were distracted by the enormous 200 MeV output. In fact, the uranium nucleus required an input of about 6 MeV to fission. That much energy was necessary to roil the nucleus to the point where it elongated and broke apart. The absorption of any neutron, regardless of its velocity, made available a binding energy of about 5.3 MeV. But that left U238 about 1 MeV short, which is why it needed fast neutrons of at least that threshold energy before it could fission.

U235 also earned 5.3 MeV when it absorbed a neutron. But it won Fermi's “one or two MeV” in addition simply by adjusting from an odd to an even mass. That put its total above 6 MeV. So any neutron at all would fission U235 — slow, fast or in between. Which was what Bohr's third graph demonstrated: the probably continuous fission cross section of U235. From slow neutrons on the left only a fraction of an electron volt above zero energy, to fast neutrons on the right above 1 MeV that would also fission U238, any neutron an atom of U235 encountered would agitate it to fission. Natural uranium masked U235's continuous fissibility; the more abundant U238 captured most of the neutrons. Only by slowing the neutrons with paraffin below the U238 capture resonance at 25 eV had experimenters like Hahn, Strassmann and Frisch been able to coax the highly fissionable U235 out of hiding. In a burst of insight Bohr had answered Placzek's objections and replenished his liquid drop.

In January Bohr had produced a 700-word paper in three days to protect his European colleagues' priorities. Now, in his eagerness to spread the news of U235's special role in fission, he produced an 1,800-word paper in two days, mailing it to the Physical Review on February 7. “Resonance in uranium and thorium disintegrations and the phenomenon of nuclear fission” was nevertheless written with care, more care than it received in the reading. Everyone understood its basic hypothesis — that U235, not U238, is responsible for slow- neutron fission in uranium — though not everyone concurred without the confirmation of experiment. But probably because, as Fermi recalled, isotopes at that time “were considered almost magically inseparable,” everyone overlooked its further implications. Szilard explained to Lewis Strauss that month that “slow neutrons seem to split a uranium isotope which is present in an abundance of about 1 % in uranium.” Richard Roberts at the DTM, in a 1940 draft report of considerable significance, asserted that “Bohr… ascribed the [slow] neutron reaction to U235 and the fast neutron reaction to U238.” Roberts' misstatement was probably no more than a rough first approximation that he would have corrected in a polished report. Szilard's and Roberts' comments illustrate, however, that the slow-neutron fission of U235 preoccupied the physicists at first to the exclusion of a more ominous potentiality.

Bohr acknowledged it indirectly in his paper for the Physical Review. The slow- neutron fission of U235 occupied the foreground of his discussion because it explained the puzzling difference between uranium and thorium. But Bohr also considered U235's behavior under fast-neutron bombardment. “For fast neutrons,” he wrote near the end of the paper, “…because of the scarcity of the isotope concerned, the fission yields will be much smaller than those obtained from neutron impacts on the abundant isotope.” The statement implies but does not ask a pregnant question: what would the yields be for fast neutrons if U235 could be separated from U238?

The latest incarnation of Orso Corbino's garden fish pond in Rome was a tank of water three feet wide and three feet deep that Fermi and Anderson set up that winter in the basement of Pupin Hall. They planned to insert a radon-beryllium neutron source into the center of a five-inch spherical bulb and suspend the bulb in the middle of the tank. Neutrons from the beryllium would then diffuse through the surrounding water, which would slow them down. The neutrons would induce a characteristic 44-second half-life in strips of rhodium foil, Fermi's favorite neutron detector, set at various distances away from the bulb. Once he estabhshed a baseline of neutron activity using the Rn + Be source alone, Fermi intended to pack uranium oxide into the bulb around the source and make a second series of measurements. If more neutrons turned up in the water tank with uranium than without, he could deduce that uranium produced secondary neutrons when it fissioned and could roughly estimate their number. One neutron out for each neutron in was not enough to sustain a chain reaction, since inevitably some would be captured and others drift away: it needed something more than one secondary for each primary, preferably at least two.

Upstairs on the seventh floor Szilard discovered a different experiment in progress. Walter Zinn, a tall, blond Canadian postdoctoral research associate who taught at City College, was bombarding uranium with 2.5 MeV neutrons from a small accelerator. He had reasoned in terms of neutron energy rather than quantity; he was trying to demonstrate secondary neutron production by looking for neutrons faster than the 2.5 MeV's he supplied. So far he had managed only inconclusive results.

“Szilard watched my experiment with great interest,” Zinn recalls, “and then suggested that perhaps it would be more successful if lower energy neutrons were available. I said, ‘That's fine, but where do you get them?’ Leo said, ‘Just leave it to me, I'll get them.’”

Szilard meant to help Zinn, but he also coveted Zinn's ionization chamber. “All we needed to do,” he said later, “was to get a gram of radium, get a block of beryllium, expose a piece of uranium to the neutrons which come from the beryllium, and then see by means of the ionization chamber which Zinn had built whether fast neutrons were emitted in the process. Such an experiment need not take more than an hour or two to perform, once the equipment has been built and if you have the neutron source. But of course we had no radium.”

The problem was still money. The Radium Chemical Company of New York and Chicago, a subsidiary of the Union Miniere du Haut-Ka-tanga of Belgium, the dominant source of world radium supplies, was willing to rent a gram of radium for a minimum of three months for $125 a month. Szilard wrote Lewis Strauss at his Virginia farm on February 13 “to see whether you could sanction the expenditures” and presciently briefed the financier on the meaning of the latest developments. The letter's crucial paragraph addresses Bohr's new hypothesis that U235 is