Why Hitler never got the bomb: Nazi Germany's last reactor experiment, recalculated after 80 years

One of the driving fears behind the United States' push to develop the atomic bomb during World War II was the possibility that Nazi Germany might get there first. Germany had, in fact, begun nuclear research in April 1939 — four months before Albert Einstein wrote to President Franklin D. Roosevelt urging the United States to pursue its own nuclear program.

A new analysis finds that Germany's final reactor experiment would have needed roughly twice as much uranium and about 2.2 times as much heavy water as it actually used to achieve a self-sustaining chain reaction — more than the country could have assembled even by pooling all available materials.

That conclusion stands in sharp contrast to what German physicist Werner Heisenberg, who led the April 1945 nuclear experiment, wrote after the war. In a 1947 paper in Nature, Heisenberg claimed that "a comparatively small additional quantity of uranium would in all probability have been sufficient."

The reassessment appears in Volume 5, Issue 9 of PNAS Nexus. A research team led by Patrick Park of Columbia University and Timothy Koeth of the University of Maryland recalculated Heisenberg's final reactor experiment using surviving records and physical uranium cubes, reconstructing the device computationally.

In 1944, after Allied bombing destroyed much of Berlin, Heisenberg's team relocated to Haigerloch, a village in southwestern Germany, where they built an experimental reactor device known as B8. The apparatus suspended 664 natural uranium cubes — each measuring 5 centimeters on a side — on chains submerged in 1,400 liters of heavy water, with the assembly surrounded by graphite and water.

In a nuclear reactor, neutrons released by fissioning uranium split other uranium atoms, sustaining a chain reaction. Heavy water slows the fast-moving neutrons, making them more effective at triggering further fission.

Of the roughly 1,100 uranium cubes used in the German experiment, only 14 have been accounted for today.

The research team obtained two of those cubes and measured the uranium's actual density at 18.53 grams per cubic centimeter — lower than the theoretical value due to microscopic internal voids. The team also incorporated a 1947 measurement by the US National Bureau of Standards showing the heavy water used in B8 had a purity of 96.8 percent, along with surviving records from graphite analysis.

For a reactor to reach criticality — the point at which it sustains a chain reaction on its own — each fission event must produce, on average, at least one neutron that goes on to trigger another fission in the next generation.

When the team simulated the 1944 experiment, B8 yielded a value of 0.94, meaning that for every 100 neutrons produced, only 94 survived to the next generation. At that level, the chain reaction would inevitably die out.

Running 1,222 combinations of varying uranium and heavy water quantities, the team calculated that B8 would have needed 1,343 uranium cubes — weighing 3,111 kilograms — and 3,121 liters of heavy water to reach criticality.

That is roughly twice the uranium and 2.2 times the heavy water actually used in the experiment. The team said Heisenberg had significantly underestimated the amount of material required.

Even combining all of Germany's available materials would not have been enough. Of the 2,840 liters of heavy water Germany produced during the war, losses from French diversion, an experimental accident, and a Norwegian special forces operation that sank a shipment left Germany with a maximum of 1,836 liters.

Even assuming all of Germany's uranium and heavy water had been loaded into B8, the simulation returned a value of only 0.96.

The team also calculated whether Germany could have succeeded using graphite instead of heavy water as a neutron moderator. The United States had operated the world's first nuclear reactor using graphite in 1942.

The key variable with a graphite reactor is boron. Even trace amounts of boron in graphite absorb neutrons efficiently enough to halt a chain reaction.

When the team estimated the boron content of Germany's graphite and ran the calculations, they found that no configuration using German uranium and graphite could have achieved criticality. The team said it could not be certain whether the outcome would have been better even if Germany had produced high-purity graphite comparable to American standards.

The team said it hopes the calculations will serve as a benchmark for reassessing Germany's nuclear program and can be applied to the analysis of other early reactor experiments.

DOI: 10.1093/pnasnexus/pgag282

Patrick J Park, Brittany Robertson, Miriam E Hiebert, Jason Zhao, Ciara B Sivels, Timothy W Koeth, "Nuclear archaeology reassesses Heisenberg's last reactor experiment," PNAS Nexus, Volume 5, Issue 9, September 2026, pgag282.