Two surviving uranium cubes show Heisenberg’s last reactor was nowhere near working

Germany’s last wartime reactor experiment, the B8 pile built in April 1945 at Haigerloch, held 664 cubes of natural uranium and roughly 1,400 litres of heavy water, and a new reassessment led by Timothy Koeth of the University of Maryland concludes it would have needed about twice as much uranium and more than twice as much heavy water to sustain a chain reaction. The calculation rests on measurements of two surviving cubes from the German project. Werner Heisenberg’s team never achieved criticality.

Tech news worth reading: Get Techpresso, the daily email with the most interesting tech news and insights, read by 400,000+ professionals from Google, Apple and OpenAI.

Get Techpresso →

We may earn a commission if you sign up.

The authors go further than a near miss. Their reading is that the whole effort was short of the basic means to succeed.

Two cubes and what the measurement showed

The study, published in PNAS Nexus as Nuclear archaeology reassesses Heisenberg’s last reactor experiment, was written by Patrick Park, Brittany Robertson, Miriam Hiebert, Jason Zhao, Ciara Sivels and Koeth. Park is at Columbia University and Reed College, Robertson at Pacific Northwest National Laboratory, Hiebert and Koeth at the University of Maryland, Zhao at Texas A&M University and Sivels at the Johns Hopkins University Applied Physics Laboratory.

The team measured the density of the two cubes with a helium pycnometer and found 18.5349 grams per cubic centimetre, give or take 0.1412, against a nominal 19.01 for solid uranium. A lower density means less fissile material packed into the same lattice, and it feeds directly into the reactor model, which the authors checked against 1945 measurements. The published abstract does not say which institutions hold the two cubes. It notes only that the locations of 14 of roughly 1,100 cubes are known, spread across personal, university and museum collections.

Earlier Maryland reporting on one such cube said the Smithsonian Institution and Harvard University each hold one, and that Koeth owns another, according to a 2019 university feature. That earlier story did not say which cubes were used in the new measurement, so the custody of the two examined here is not established in the sources reviewed.

The shortfall in density is about 2.5 percent, a small number that still matters. Each cube carried slightly less uranium per unit of volume than an ideal block would, and across 664 cubes that adds up to a lattice that was a little poorer in fuel than the planners assumed. The authors treat density, together with the purity of the heavy water, as an input whose uncertainty they carry through the simulations, and they show how the multiplication factor shifts as each varies.

A multiplication factor of 0.94

A reactor sustains a chain reaction when its effective multiplication factor, written k-effective, reaches 1. The authors’ neutronics model puts the B8 as built at about 0.94, a value at which the reaction fades away instead of growing. The ScienceDaily summary of the paper says the pile had never achieved criticality and that its findings indicate the project was not simply missing a few materials.

How big was the gap? The paper’s own abstract gives roughly twice the uranium, which is about 1,300 cubes against the 664 actually loaded. One detailed account of the study puts the figure at about 1,343 cubes weighing 3,111 kilograms and about 3,121 litres of heavy water, roughly 2.2 times the amount used. That account adds that Germany’s maximum available heavy water was about 1,836 litres, and that even with every cube and every litre loaded, the simulation reached only 0.96.

The result does not rest on a single lucky measurement. The optimisation behind it searched layouts of between 500 and 2,100 cubes to find configurations that reach criticality, so the answer does not hinge on the particular lattice Heisenberg’s crew happened to build. Even under favourable assumptions, the shortage remained.

The heavy water and graphite problem

Heavy water was the second wall. According to ScienceDaily, the Vemork plant in Norway, raided on November 16, 1943, was the only European facility capable of supplying it. The Debrief’s reading of the paper is that a heavy-water reactor would have demanded more of it than Germany produced during the entire war, which it calls far outside logistical possibilities. Heisenberg had claimed only a small amount of uranium stood between his pile and criticality, a claim that these figures contradict.

Graphite was no way out. The B8 reused graphite from an earlier experiment, and the authors found that German graphite made from metallurgical coke carried more boron than American graphite made from petroleum coke. Boron absorbs neutrons, so the material could not have supported a reactor, and the purer grade was beyond wartime industrial capacity. The authors conclude that Germany could not have built either a heavy-water or a graphite reactor during the war, as the journal’s abstract states, and the article itself carries the full simulations.

Park, Koeth and their co-authors call the approach nuclear archaeology: rebuilding a historical machine from surviving objects, documents and modern simulation. The Debrief describes the method as combining measurements of the two cubes, the known purity of the heavy water available to German researchers and a reactor model, then running simulations.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

Tech news worth reading: Get Techpresso →

We may earn a commission if you sign up.