Dark Sun: The Making Of The Hydrogen Bomb

Serber. But it was also soon obvious from work by Niels Bohr that a formidable obstacle stood in the way of making bombs: only one isotope of uranium, U235, would sustain a chain reaction, and U235 constituted only 0.7 percent of natural uranium; the other 99.3 percent, chemically identical, was U238, which captured secondary neutrons and effectively poisoned the reaction.[3] There were then two difficult technical questions that needed to be resolved by any nation that proposed to explore building an atomic bomb: whether it might be possible to achieve a controlled chain reaction — to build a nuclear reactor — using natural uranium in combination with some suitable moderator, or whether the U235 content of the uranium would have to be laboriously enriched; and how to separate U235 from U238 on an industrial scale for bomb fuel when the only exploitable distinction between the two isotopes was a slight difference in mass. Enrichment and separation were essentially identical processes (“separated” bomb-grade uranium is natural uranium enriched to above 80 percent U235) and would use the same massive, expensive machinery that no one yet knew how to build; while a reactor fueled with natural uranium, if such would work, might be a straightforward enterprise.

Khariton and Zeldovich approached these questions from first principles, as it were, carefully calculating what was not possible as well as what might be. In the first of three pioneering papers they published in the Russian Journal of Experimental and Theoretical Physics in 1939 and 1940 (papers that went unnoticed outside the Soviet Union) they demonstrated that a fast-neutron chain reaction was not possible in natural uranium. Isotope separation would therefore be necessary to build a uranium bomb.

A second, longer paper, delivered a few weeks later on October 22,1939, developed important basic principles of reactor physics. Khariton and Zeldovich correctly identified the crucial bottleneck that experimenters would have to bypass to build a natural-uranium reactor that worked. Visualize a stray neutron in a mass of natural uranium finding a U235 nucleus, entering it and causing it to fission. The two resulting fission fragments fly apart; a fraction of a second later they eject two or three secondary neutrons. If these fast secondary neutrons encounter other U235 nuclei they will continue and enlarge the chain of fissions. But there is much more U238 than U235 in the mass of natural uranium, making an encounter with a U238 nucleus more likely, and U238 tends to capture fast neutrons. It is particularly sensitive to neutrons moving at a critical energy, twenty-five electron volts (eV), a sensitivity which physicists call a “resonance.” On the other hand, U238 is opaque to slow neutrons. To make a reactor, then, Khariton and Zeldovich realized, it would be necessary to slow the fast secondary neutrons from U235 fission quickly below U238's twenty-five eV resonance. The way to do that, they proposed, was to make the neutrons give up some of their energy by bouncing them off the nuclei of light atoms such as hydrogen. “In order to accomplish [a chain] reaction [in natural uranium],” they wrote, “strong slowing of the neutrons is necessary, which may be practically accomplished by the addition of a significant amount of hydrogen.”

The simplest way to mix uranium with hydrogen would be to make a slurry — a homogeneous mixture — of natural uranium and ordinary water. But Khariton and Zeldovich demonstrated in this second paper that such a mixture would not sustain a chain reaction, because hydrogen and oxygen also capture slow neutrons, and in a reactor fueled with natural uranium such capture would subtract too many neutrons from the mix. Important consequences followed from this conclusion. One was that instead of hydrogen in ordinary water it would apparently be necessary to use heavy hydrogen — deuterium, H² or D, an isotope of hydrogen with a smaller appetite for neutrons than ordinary hydrogen — perhaps in the form of rare and expensive heavy water. (In a review article published in 1940, Khariton and Zeldovich proposed carbon and helium as other possible moderators, both materials that later proved to work.) Alternatively, wrote the two Soviet physicists, “another possibility lies in the enrichment of uranium with the isotope 235.” They calculated that natural uranium enriched from 0.7 percent U235 to 1.3 percent U235 would work in a homogeneous solution with ordinary water.

In a third paper submitted in March 1940, Khariton and Zeldovich identified two natural processes that would make it easy and “completely safe” to initiate and control a chain reaction in a nuclear reactor. The fissioning process would heat the mass of uranium and cause it to expand, which in turn would increase the distance the neutrons would have to travel to cause additional fissioning and would therefore slow down the chain reaction, allowing the mass of uranium to cool and the chain reaction to accelerate. This natural oscillation could be controlled by increasing or decreasing the volume of uranium. Another natural process — delayed neutrons released in fission which would “significantly increase” the oscillation period — subsequently proved more significant for reactor control. (Apparently critics within the Soviet scientific community had made safety a point of attack; in this third paper Khariton and Zeldovich vigorously disputed what they called “hasty conclusions… on the extreme danger of experiments with large masses of uranium and the catastrophic consequences of such experiments.” Because of the natural processes they had identified, they scoffed, such conclusions “do not correspond to reality.”)

Khariton and Zeldovich summarized these early and remarkable insights in the introduction to their third paper:

It would appear (the lack of experimental data precludes any categorical assertions) that by applying some technique, creating a large mass of metallic uranium either by mixing uranium with substances possessing a small capture cross-section (e.g., with heavy water) or by enriching the uranium with the U²³⁵ isotope… it will be possible to establish conditions for the chain decay of uranium by branching chains in which an arbitrarily weak radiation by neutrons will lead to powerful development of a nuclear reaction and macroscopic effects. Such a process would be of much interest since the molar heat of the nuclear fission reaction of uranium exceeds by 5 · 10⁷ [i.e., 5,000,000] times the heating capacity of coal. The abundance and cost of uranium would certainly allow the realization of some applications of uranium.

Therefore, despite the difficulties and unreliability of the directions indicated, we may expect in the near future attempts to realize the process.

At the annual All-Union Conference on Nuclear Physics, held in 1939 in November at Kharkov in the Ukraine, Khariton and Zeldovich reported their conclusion that carbon (graphite) and heavy water were possible neutron moderators. They also reported that a controlled chain reaction even with heavy water would be possible in a homogeneous reactor only with uranium enriched in U235. Since uranium enrichment was notoriously difficult, and would require the development of an entirely new industry, their conclusion made the possibility of building a working nuclear reactor within a reasonable period of time and for a reasonable amount of money appear remote. But there are other possible arrangements of natural uranium and graphite or heavy water that they overlooked, even though their second 1939 paper had offered an important clue. Why two such outstanding theoreticians should have overlooked more promising alternative arrangements is a question worth exploring.

The effectiveness of a moderator such as graphite or heavy water is limited crucially by its probability of capturing rather than reflecting neutrons. That probability, called a “cross section,” can only be determined by experiment. Physicists quantify capture cross sections (and other such probabilities) in extremely small fractions of a square centimeter, as if a cross section were the surface area of a target the incoming neutron might hit. The two theoreticians had calculated that to achieve a chain reaction in a mixture of ordinary uranium and heavy water, the cross section of deuterium for neutron capture must not be larger than 3 · 10^–27 cm². They lacked the laboratory equipment they needed — a powerful cyclotron and a large quantity of heavy water — to measure the actual capture cross section of deuterium (the entire Soviet supply of heavy water at that time amounted to no more than two to three kilograms). For the 1939 All-Union Conference they must have offered an approximation drawn from the international physics literature.

Apparently they continued to search the literature to see if someone had determined a more accurate value for the deuterium capture cross section. They found an estimate in a letter to the editor of the American journal Physical Review published in April 1940. In that letter, University of Chicago physicists L. B. Borst and William D. Harkins noted a “quantitative estimate” of 3 · 10^–26 cm², a full order of magnitude too large (^–26 rather than ^–27). “Thus,” Igor Kurchatov would explain in 1943 in a top secret report, “we came to the conclusion that it is impossible to achieve a chain reaction in a mixture of [ordinary] uranium and heavy water.” And if not in heavy water without investing expensively in isotope enrichment, then also not in carbon, where tolerances were even closer. “Contrary to the opinion of a small group of enthusiasts,” Khariton would comment late in life, “the dominant opinion in our country was that a technical solution to the uranium problem was a matter for the remote future, and that success would require fifteen to twenty years.” Khariton and Zeldovich's disappointing conclusion must certainly have contributed to that conservative assessment. But the “small group of enthusiasts,” which included Khariton, Zeldovich, Kurchatov and Flerov, was not deterred. “In the case of a homogeneous reactor, the enterprise looked doomed,” Khariton would