During World War II the scientists and technicians in Britain and the US who were building an atom bomb feared that Nazi Germany might produce such a weapon first. Yet at the end of the war it became clear that there had been no such threat. Not only was no German bomb under construction, but no extensive effort was being made to develop one. Why was that so?

The first answer was given by the physicist Sam Goudsmit, who was sent to Europe in 1947 on what was called the “Alsos” mission to discover the fact about the German project. In his book Alsos1 he attributed the failure to gross incompetence on the part of the scientists. In particular he claimed that Heisenberg did not know the difference between a bomb and a reactor. This last statement he later withdrew, but he remained convinced that Heisenberg, the most famous German physicist, did his best to develop a bomb, but his best was not good enough.

Another explanation was put forward in 1956 by Robert Jungk in his book Brighter Than a Thousand Suns,2 and supported by Walter Kaempfert of The New York Times. According to them, the German scientists refrained from making a bomb for moral reasons. I know of no evidence that Heisenberg made this claim himself in public, and it is not clear whether he really took that position in his interview with Jungk. But he also never repudiated it publicly.

Heisenberg later gave his own explanation in articles in scientific journals and in his autobiographical book Physics and Beyond.3 While he and his colleagues knew a bomb to be possible in principle, he wrote, they saw that the effort to develop it was far too great to be undertaken by Germany in wartime. (In retrospect this seems reasonable, because even the Americans, who had greater technological resources and were not subject to air raids, did not complete the bomb before V-E day.) As a result the German scientists did not ask for a crash program, and took as their goal the construction of an experimental reactor, in order to prove the feasibility of a chain reaction and to prepare the way for a future nuclear power program. They felt relieved that they did not have to make a decision on whether to work on a weapon. It is anybody’s guess whether they would have worked on a bomb if they had regarded this as a practical possibility.

I find this explanation convincing; it seems consistent with the evidence I have seen. It is true that Heisenberg was known to allow his memory to be colored by what he would like to be true, but I have learned from experience that one should never discount the possibility that a person might actually mean what he is saying.

The recently released “Farm Hall Transcripts,” of recorded conversations of German atomic scientists who were interned in England shortly after the war,4 firmly ruled out the idea of their moral superiority. The transcripts record the reaction of the scientists to the announcement that a bomb was dropped on Hiroshima. Apart from a tentative remark by von Weizsäcker, none of them mentioned moral objections as reasons for not having made a bomb, although they argued about the reasons for their own failure and the consequences of the bomb for their own future in Germany. Yet the controversy has continued, and in Heisenberg’s War Thomas Powers produces a new theory.

Before I consider his view, I should recall the history of the development of the bomb both in Germany and in the US and Britain. Both sides started with rather tentative basic research to ascertain the possibilities for larger projects. These studies got underway in 1939 in Germany, and in the spring of 1940 in Britain and the US. In the summer of 1941 the British “Maud” committee reported that a bomb making use of the separated uranium isotope 235 was feasible and worth pursuing, and the report persuaded the British government to step up the pace of research; a little later the US project was reorganized and the “Manhattan Project” started as a crash program. By this time it was known that bombardment by neutrons would produce a reaction in the abundant uranium isotope U-238, which would eventually lead to the new element plutonium, expected to be an alternative fuel for a weapon. In 1942 Enrico Fermi succeeded in building an experimental reactor, thus demonstrating the first man-made neutron chain reaction.

To accomplish this, and go on to produce a bomb, many obstacles had to be overcome. To a physicist, the idea of separating substantial quantities of the isotopes of such a heavy element as uranium first appeared utopian. This had been done only for light elements, for which the percentage differences between the masses of isotopes were much bigger than for the heavy uranium. For the separation, moreover, it would be necessary to use a gaseous compound of uranium; but the only such compound is the very corrosive hexafluoride, an awkward substance to work with. The work on light elements had produced only milligrams of isotopes, so to produce the kilograms needed for a bomb the process previously used would have to be scaled up a million times.


For these reasons the idea of a bomb using uranium 235 was at first not taken seriously, and the question of how much U-235 would be required remained academic, although most scientists who might have addressed the problem would intuitively have guessed that tons were needed. Once one took the possibility seriously and thought about the “critical mass” of U-235 needed to produce a chain reaction one could see it was a matter of kilograms rather than of tons; to separate the isotopes would still need a tremendous effort, but not a prohibitively difficult or expensive one.

The reactor project also faced problems; it required a “moderator,” a light material to slow down the neutrons, since slow neutrons are more effective in producing nuclear fission. Of the obvious light elements, carbon, in the form of graphite, was the favorite for use as a moderator, but experiments showed that even pure graphite absorbed too many neutrons to make it suitable. The vital step was the intuition by Leo Szilard and Fermi that commercially “pure” graphite contained impurities in amounts that were too small to show up in chemical analysis but still contributed to a very high absorption of neutrons. By insisting on further purification of graphite, they solved this problem. The primary reason for building the experimental reactor was to establish the reality of the chain reaction. The plan to produce plutonium for weapons came later.

In Germany the pessimism about separating isotopes on a large scale was never overcome. There was some small-scale research on such separation, but the whole matter was regarded as a very long-term prospect. Since it was so academic, nobody, not even Heisenberg, made a serious effort to estimate the critical mass needed to produce a chain reaction. Heisenberg used different arguments at different times, giving answers ranging from “as big as a pineapple”—about 15 kilograms of metallic uranium—to many tons. This vagueness seems surprising, but Heisenberg, though a brilliant theoretician was always very casual about numbers. When I was his student in the late 1920s the first assignment he gave me was to check whether a recent observation in a spectroscopic experiment could be explained as an example of his uncertainty principle. A simple back-of-an-envelope estimate would have shown that the effect was 100 or even 1,000 times greater than could be explained by his hypothesis. When he first heard, at Farm Hall, of the reality of the Hiroshima bomb, he at last thought through the question of critical mass seriously and came up with a reasonable answer.

In attempting to make reactors, the Germans, too, found that graphite absorbed excessive amounts of neutrons, but did not insist on further purification, and turned instead to heavy water—i.e., water containing the isotope of hydrogen called deuterium—as a moderator that would slow down neutrons. This is more efficient and makes it possible to use a much smaller reactor, but it is hard to obtain. The world’s only facility for producing heavy water at that time, in Norway, was destroyed by sabotage and in Allied bombings, so that there was a continuing shortage of it in Germany.

The large-scale production of plutonium seemed to the Germans also a very distant goal. It is true that Heisenberg said after the war: “We saw an open road to the atomic bomb,” implying that, since they knew a reactor moderated by heavy water would work, they could see how to produce plutonium (actually neptunium, the intermediate element that decays into plutonium, and was not yet known to them) and hence weapons. But this open road was still a rocky one: it would have required the design of large-scale reactors and methods for cooling them as well as chemical technology for extracting the plutonium from the highly radioactive reactor fuel; finally, the method of detonating a plutonium bomb was the hardest of the problems solved at Los Alamos. Heisenberg’s statement was a little naive.

The evidence supports the contention that since he and the other scientists regarded the bomb as a long-term problem with no relevance to the war, they did not ask the Nazi government to support a crash program. Their aim, like the initial aim of Fermi, was to prove the existence of a chain reaction. They pursued this with a very small team, and without great urgency. They still hoped they might be the first to achieve a chain reaction and thought this would be good for the prestige of German science. It seems never to have entered their thoughts that others might be ahead of them. After they first heard the news of Hiroshima, some of the Farm Hall group comforted themselves by saying that, while the Americans had made the bomb, history would recognize that the Germans had made the reactor!


Meanwhile the Allies were afraid of a German bomb. Their intelligence services made an intensive effort to discover what was going on. While working on the British Atomic Energy Project early in the war I was able to convince myself there was no German crash program by checking the list of physics lecturers in German universities published each semester by the Physikalische Zeitschrift. This showed that nearly all the physicists were in their usual positions, teaching their usual subjects, a very different picture from that which would have been shown by a similar list for the UK or the US. In addition, many messages arrived from German opponents to the regime, a few warning of a German bomb program, but most saying that only a small reactor project was actually underway. Some, notably the publisher Paul Rosbaud and the physicist Fritz Houtermans, claimed that Heisenberg and other German scientists were trying to delay work on a bomb. Yet the Allied authorities continued to worry. When Strasbourg was captured by the Allied forces, documents were found showing the small scale and slow progress of the work, but American concern remained intense. At the end of the war the leading German atomic scientists were rounded up and interned in Farm Hall until early 1946, presumably to prevent their being hired or kidnapped by the Russians.

Thomas Powers’s book covers these events in enormous detail. He has collected an impressive amount of information, though much of it is from secondary sources, including such controversial writers as Robert Jungk and David Irving. He describes many episodes that are not widely known. Many of these concern the relations, and the squabbles, among the different US intelligence services, including their travels to and within Europe, with precise details of their departure and arrival dates. Readers who are not fascinated by intelligence operations may not find these accounts very interesting; but many of the stories are surprising and intriguing. Unfortunately it is not easy to decide which of these stories to believe, in spite of their extensive documentation, partly because of the book’s unfortunate errors, of which I shall give examples later.

One of the more exciting episodes was the plan by the US to kidnap or assassinate Heisenberg. This is reported with a dramatic effect not unlike that common in TV thrillers, where at dramatic moments the presentation is interrupted by a commercial. Powers, too, builds up tension, only to change the subject for many pages. We read how Colonel Carl Eifler was selected and briefed for the job and then, after long deliberation, was eased out because it was feared he might fall into German hands, and might under pressure reveal vital information about the American project. Instead the intelligence officials chose another man, Morris “Mo” Berg, a former baseball player and an accomplished linguist, who, at least officially, was ignorant of the Manhattan Project. Plans were made to get him, along with a support team, into Germany, or into Switzerland during a visit by Heisenberg to Zurich where he was to give a lecture. But Berg had a mind of his own, and went alone to Zurich while Heisenberg was there.

The author quotes the physicist Philip Morrison, who worked at Los Alamos, as saying that he was doubtful whether at this time, December 1944, there was any sense in putting Heisenberg out of action. If there was a German project that had any chance of success before the war ended, it would have progressed by this time to the stage where the technological problems would have been central and individual scientists would not have been vital.

Berg made contact with Swiss physicists sympathetic to the Allied cause and, with a gun in his pocket, he found a seat among the audience at Heisenberg’s very abstract theoretical seminar. So far, there is no reason to question Powers’s account. But we are told that Berg was authorized to make his own decision whether to shoot Heisenberg. He did not shoot, it is said, because Heisenberg did not mention anything about an atom bomb in his talk.

I find it wholly incredible that it should have entered anyone’s head that Heisenberg could, or would, talk about an atom bomb in an open academic seminar in Switzerland, or that a junior agent of the OSS would have been authorized to decide whether to kill him or not. Surprising things happen in war, especially within secret intelligence agencies, but I would need very strong additional evidence to make me accept this story.

Another episode that I find surprising concerns Niels Bohr, who arrived in America with a crude sketch of a reactor, which Heisenberg had given him during his visit to Copenhagen in 1941. American physicists considered whether the reactor might have been intended as a weapon. The author claims that Bohr himself regarded it as possible that a weapon could be made using slow neutrons. This seems quite unlikely, since Bohr had been the first to point out the impossibility of doing so. Hans Bethe is cited as the author’s principal source for this story, but Bethe tells me he does not think Bohr ever held this view.

In the main thesis of his book, stated explicitly in the last two chapters but present in the background throughout, Powers advances his own, new explanation of the failure of the German project. Heisenberg, he argues, deliberately withheld information from the authorities and from his colleagues in order to stop the construction of a German atom bomb. The most important aspect of this deception was that Heisenberg kept stressing that the development of a bomb was too difficult and involved too great an effort to be completed in wartime Germany. No doubt Heisenberg said this on every possible occasion; the questions are, did he believe it, and was it true?

It was indeed true, as I have pointed out earlier. Powers, however, disagrees. He argues that the German work started in 1939, whereas the Manhattan Project began only in the summer of 1942, so German scientists could have been testing their first bomb two years before the Americans did, in 1943. This is a completely mistaken comparison. One cannot launch an enormous technological program based on a new discovery without preliminary laboratory studies of its feasibility and its basic principles. Such studies went on in Britain and the US from about early 1940, and Germany could not do without similar studies either, so there was, in fact, at most a six months’ difference in the respective starting dates. In this early phase Western scientists, unlike the Germans, had the use of a number of cyclotrons and other accelerators to determine the behavior of nuclei and they also had mass spectrographs to produce small samples of the relevant isotopes. During the later, technological phase of manufacturing a bomb the hugely expensive technology, assembled, for example, at Oak Ridge, Tennessee, and Hanford, Washington, could not have been matched by the Germans.

From the evidence I have seen, Heisenberg certainly believed that the atom bomb was something for the distant future. In fact he was sure that even the Americans could not make one during the war. This is quite clearly demonstrated by his reaction at Farm Hall to the news that a bomb had been dropped on Hiroshima, when he at first firmly believed that the announcement was a bluff, and that the US did not really have an atom bomb. Of all the scientists at Farm Hall he seems to have been the last to accept the reality of the event.

A contrary impression is given by the report, quoted by Powers, that in 1942 Heisenberg, in answer to a question from the Luftwaffe general Erhard Milch, estimated that the Americans could not complete their first reactor before the end of 1942, and the first bomb would take another two years. What is the source for this surprising report? It turns out to be a 1947 interview by Heisenberg with Der Spiegel. News magazines are not the most dependable secondary sources, and we know of many instances when Heisenberg’s memory deceived him.

Why did Heisenberg overestimate the difficulties? The explanation is connected with his failure to make an accurate estimate of the critical mass, on which I have already commented. Most of the time he relied on an argument that was conceptually wrong, though he quoted different estimates at different times. After Hiroshima, as I have said, he took time to think through the problem rationally, and a week later had a perfectly reasonable figure. This was not, as Powers says, because he had initially used a wrong figure for the “mean free path” of neutrons in uranium 235, for which he could not have discovered a better number at Farm Hall, but because his original argument was logically flawed. Knowing that many “generations” of neutrons were needed to make a chain reaction, he assumed, in estimating the critical size, that each neutron must, on the average, make a certain number of collisions with uranium nuclei, instead of enough collisions to generate one further neutron before escaping.

In his revised estimate he also took into account that the bomb core could be surrounded by a reflector, which would scatter back some of the escaping neutrons and thereby reduce the size of the critical mass needed. According to Powers he must have devoted much thought to bomb design for such a subtle idea to have occurred to him. Yet Powers ignores the fact that reflectors are also used for the design of reactors, on which Heisenberg had been working for a long time; so it would have been quite obvious to him to apply the idea of a reflector to a bomb.

Powers also stresses, as part of his argument that Heisenberg was deceiving his colleagues, that he never explained such points to them; but even if he had any interesting thoughts about bomb design, why should he have discussed them with others when nobody was planning to develop an atom bomb?

Powers also relies on the message from Fritz Houtermans carried to the US by Fritz Reiche, which said that many German physicists were working on an atomic-bomb project, but that Heisenberg was trying to delay the work. We must remember, however, that Houtermans was not a member of the project, and not necessarily familiar with what was going on. The first part of the message was certainly misleading because, as we know, there was no work on a bomb, so it is difficult to accept the second part. I therefore conclude that the new and ingenious explanation of the German failure is without substance.

I earlier mentioned wrong or misleading statements in Powers’s book. Many of the facts of which I have direct knowledge are in some way garbled, although some do not affect the substance of the argument. For example, Powers claims that I originally estimated the critical mass of U-235 as many tons; I never did so. The memorandum I wrote with Otto Frisch in 1940 for the British authorities estimating the critical mass is dated in one place in the book as 1940, in another as 1941. The German town of Duisburg (to which some of the uranium ores was traced) is identified in the book as “Duisburg in Belgium.” The famous physicist Max Born is said to be a mathematician on one page and a physicist on another, and the well-known nuclear physicist Walther Bothe is called a chemist.

Other errors are more serious. Powers has a persistent tendency to interpret any work on uranium fission as bomb research; he expresses surprise that Bohr allowed his “bomb lecture” to be published in 1941. In fact this is an academic exposition of fission physics which contains the remarks that no explosion is possible in natural uranium and that, with present technical means, it is impossible to purify enough of the rare uranium isotope to cause an explosion. A paper by Frédéric Joliot-Curie, written in 1939, showed that there were secondary neutrons emitted in the fission process, which pointed to the possibility of a chain reaction. This, the author says, “was just another way of saying bomb,” but a chain reaction also drives a reactor.

In 1939 the German physicist Siegfried Flügge wrote about the possibility of a nuclear reactor generating power and referred to the need to avoid a runaway chain reaction. This, the author again comments, was just another way of saying “bomb.” In fact Flügge evidently referred to the kind of nuclear disaster that occurred at Chernobyl.

In 1977 Goudsmit wrote to me as follows:

Armin Hermann, whose opinion I asked, claims that I am now “softer” on Heisenberg than I was in my book, written in 1947. I have had time to think. I believe, that we resent that this great physicist, our idol, wasn’t any better than we are, that he didn’t present the humane leadership we hoped for. Did anyone among our prominent colleagues? Perhaps Laue or Blackett or Franck?

Powers quotes only a phrase of this letter and gives the impression that Goudsmit’s attitude had not changed:

Goudsmit felt nothing but abiding cold fury for the possibility that Heisenberg and his friends might claim any iota of moral responsibility for the failure of Germany to build a bomb. Even after thirty years Goudsmit told Rudolf Peierls he resented the fact that “this great physicist, our idol, wasn’t any better than we are.”

Power’s prose is lively and readable, in spite of occasional clumsy passages: “Heisenberg was the single greatest physicist remaining in Germany” or “explain … what had to be done to the commander of the Eighth Air Force” (meaning “explain to the commander … what had to be done”). The narrative frequently jumps back and forth in time, so it is not always easy to see what period the author is talking about.

This Issue

April 22, 1993