The Nazi bomb that never was: what two uranium cubes tell us

How do you lose a race and then spend decades implying you were a nose behind? Werner Heisenberg's postwar account of the German reactor effort is a masterclass in that genre, and a new study has just put numbers on how generous it was.

In April 1945, in a cellar at Haigerloch, Heisenberg's team lowered 664 natural uranium cubes into heavy water in a pile called B8. It never started a chain reaction. Heisenberg later argued that a bit more uranium would have done it, a version published in Nature in 1947. A study in PNAS Nexus, led by Princeton's Patrick Park and Timothy Koeth of the University of Maryland, rebuilt the reactor digitally from two surviving cubes plus archival data. The model matches the 1945 measurements and gives a neutron multiplication factor of 0.94. Anything below 1 fizzles out.

The shortfall was structural, not a rounding error

The headline number is 0.94, but the more telling one is the bill to fix it. According to Orange's coverage, closing the gap would have taken about 3.1 tonnes of uranium and 3.5 tonnes of heavy water, against roughly 1.5 tonnes of each in B8. ScienceDaily sums it up as about twice the uranium and more than twice the heavy water the reactor actually held, and Park told Live Science that Heisenberg was off by a factor of about two.

That matters because "a little more uranium" is a story about willpower. "Twice as much of everything" is a story about industry. Heavy water came from a single plant in occupied Norway, and graphite, the American route, was out because German graphite carried boron, which soaks up neutrons. Remove both escape hatches and the reactor was not almost working. It was a different project from the one Heisenberg described.

A digital twin is evidence, not a verdict on a man

We should be careful here. The twin tells us about neutrons, not about intentions. Historian Mark Walker called it interesting engineering analysis while noting its counterfactual side, and he is right to. The study cannot say whether Heisenberg slowed the program on purpose, and it does not claim to. Whether his postwar claims were self-serving is our reading, not a finding. But it is a reasonable reading: a physicist who lost the war had every incentive to present the failure as either a near miss or a moral choice, and the numbers fit neither.

There is also a methodological point worth applauding. The code is open on GitHub, so anyone can rerun the calculation. A claim that stood since 1947 is finally being checked the way it should have been: by someone else, with the inputs on the table. The team also ran 1,222 reactor variants, and even the best one needed more cubes and more heavy water than B8 held.

The myth survived because it was comfortable

Why did "almost" last so long? Because it flattered everyone. For Germans, it made the scientists look competent and possibly principled. For the Allies, it justified the fear behind the Manhattan Project. A tale of a program starved of materials, and led by people who misjudged what they needed, is less dramatic than either.

One caveat keeps us honest: reactor is not bomb. Even a critical B8 would have been only a first step toward plutonium, as the reporting stresses. The study does not overturn the whole history of the Uranprojekt, and the sources we have are press coverage of a single paper, so independent replication would help.

Our verdict: stop treating "how close were the Nazis?" as an open question. The physics now says far, and the interesting question is why the answer was ever disputed. The cubes are a good reminder that memory writes the first draft of history, and measurements, eventually, edit it.