The year is 1972, at the Tricastin uranium enrichment site, in France. A sample of processed uranium coming from the Oklo mine, in Gabon, is examined as part of a routine inspection. Surprise: it contains only 0.60% of U-235, instead of the 0.72% expected in natural uranium. The difference might appear negligible to the uninformed eye, but it does matter gravely. The bulk of natural uranium is U-238, which is not fissile—it cannot sustain a nuclear reaction. U-235 and U-238 are two isotopes of the atom of uranium. They contain the same amount of protons and electrons, 92 of each, but the U-235 isotope contains three less neutrons. This small difference makes it fissile. All in all, around 200 kg of fissile material was missing, which, if verified, could have been a serious security breach.

Further investigation showed that the ore also contained lighter elements that result from the splitting of uranium during the fission reaction. They eventually understood that this sample was natural and had gone through fission on its own. At some point in the past, the conditions were right in the Oklo uranium deposit to form not one, but seventeen natural fission reactors. From the missing U-235, they calculated that the natural reactor had produced 100 kW, roughly the same as an electric car fast charging station, for hundreds of thousands of years. [3]

At the core of nuclear reactors is a chain reaction. An atom of U-235 is hit by a neutron and splits into two smaller atoms and emits a handful of neutrons. From here there are three possibilities. If all these neutrons are lost, either because they escape the core or because they are captured by non-fissile elements such as U-238, then the chain reaction dies out. If two or more neutrons are captured by other atoms of U-235, then these atoms will also split, producing more neutrons and splitting exponentially more atoms. That is the very principle of a runaway chain reaction which, rest assured, is extremely complex to achieve. If exactly one neutron is captured by another atom of U-235, then there is a controlled chain reaction.

This number—the ratio between the number of neutrons in one generation over the preceding one—is called the multiplication factor. Achieving a controlled chain reaction requires thus a multiplication factor of exactly one. In other words, the neutron population has to remain constant.

This mostly depends on three factors. First, the lump of uranium has to be larger than the average length that neutrons travel. If it wasn’t, too many neutrons would escape the deposit and be lost to the surroundings before having had a chance to meet another U-235. Second, the concentration of fissile U-235 has to be large enough within the lump of uranium. As it turns out, it is actually an unstable isotope: it decays naturally, with a half-life of slightly over 700 million years. While this is far too long to be noticeable over the span of human civilization, it matters when considering geological time scale. Rewind to 1.7 billion years ago. Back then, the concentration of U-235 in the Oklo deposit was around 3%. Last but not least, neutrons have to be slowed down by a moderator. Their energy, directly after fissioning, is actually too large to be captured by U-235 atoms. In the Oklo deposit, this function was fulfilled by groundwater.

That’s the essentials of what’s needed to run a nuclear reactor: enough uranium, with a large enough concentration in fissile U-235, and a moderator. Controlled critical fission is an old and crude technique. Mother nature built and operated at least one on her own billions of years ago—and perhaps more, waiting to be discovered. We have done so ourselves since the 2nd of December 1942, the day a team led by Enrico Fermi ignited the Chicago Pile-1 (CP-1), the first artificial self-sustaining nuclear reactor.

Today’s nuclear reactors are much more efficient than the Oklo natural reactor and the CP-1 experiment. Schematically, however, the basic principle remains the same. The vast majority of nuclear reactors are pressurised water reactors (PWR).

The basics of a nuclear power plant. From left to right: water from the primary circuit is heated up while circulating through the hot core in a pressure vessel, and cooled down in a steam generator where it exchanges heat with water from the secondary circuit. This all takes place in a containment structure. From there on, it works like any thermal energy production unit. Source: Encyclopaedia Britannica, Inc. 2013.

Their fuel typically consists of cylindrical pellets, roughly 1 cm in diameter and 1.5 in height, made out of a uranium oxide ceramic. Despite their small size, each of these contains as much energy as one ton of coal. Read that sentence again, and take a moment to think about this. The manufacturing of these fuel pellets is long and complex. One particularly critical step is enrichment, during which the proportion of U-235 is artificially increased to around 5%. This process is not only heavily regulated, it also requires a substantial industrial effort. Weapons grade uranium contains 85% or more of U-235, which requires an entirely different infrastructure, with yet another scale in terms of complexity and cost.

Modern reactors may contain up to 10 million pellets, arranged in fuel bundles. The amount and geometrical arrangement of these bundles is meant to achieve a critical chain reaction. These are inserted in the reactor vessel, which is in turn filled with pressurised water.

As they undergo fission, U-235 atoms release energy, which is transferred to water as thermal energy. Water heats up to a temperature of around 315°C. The high pressure, about 155 bar, ensures that it remains liquid. Hot water is pumped into a heat exchanger, where it is cooled down, and is then recirculated into the reactor vessel. This loop is called the primary circuit. The heat exchanger is placed in a separate vessel, called the steam generator. There, hot water coming from the reactor vessel transforms the cool water from the secondary circuits into hot steam. This is accomplished without mixing primary and secondary circuit water, thus avoiding any radioactive contamination. The hot steam is fed to a steam turbine which drives an electrical generator connected to the grid. The mixture of water and steam exiting the turbine is further cooled in a condenser, before being pumped back to the steam generator.

Pressurised water from the primary circuit also acts as a moderator, reducing the energy of neutrons to increase their chance of inducing fission in other U-235 atoms. Should the reactor core become supercritical, thus creating a runaway chain reaction, the fuel would quickly heat up causing the water around it to boil. The apparition of gaseous bubbles would in turn impede water’s ability to slow down neutrons, thus reducing the fission reactor rate and lowering the multiplication factor back to criticality. This negative feedback loop is called the negative void coefficient. It is one of the many safety features of modern nuclear reactors. One of the key reasons that led to the Chernobyl accident is that the design of the Soviet reactors was such that they had a positive void coefficient—the loss of coolant increased the fission reaction rate.

This is the core principle of nuclear reactors. All the surrounding technology is there to make sure this process takes place safely and sustainably. From that perspective, nuclear energy fits multiple criteria of low-tech innovation as defined by Arthur Keller and Emilien Bournigal: sobriety, efficiency, and durability. [4] These characteristics will be further developed throughout the rest of the book.

There are some variations on that basic design. Some reactors use boiling water instead of pressurised water. Others use uranium fuel that has been enriched just below the limit of 20% of U-235. Some are meant to operate using fast neutrons, removing the need for a moderator. In fact, the current nuclear renaissance is driving a whole new generation of reactor designs, each more creative than the next. We will meet some of them through the pages of this book. Nonetheless, the overarching principle is always the same.

As of fall 2023, humanity had accumulated over 19,600 years of operational experience with power reactors. There are 411 power reactors in operation around the world, 26 in suspended operation, and 58 more under construction. [5]

But that’s not the end of the story. There are also an estimated 200 nuclear reactors powering over 160 ships—submarines, aircraft carriers, icebreakers. More than 14,000 reactor-years of nuclear marine operation have been accumulated. The US Navy alone has in excess of 6,200 reactor-years of accident-free experience. [6]

There’s more. There are 224 operational research reactors, out of a staggering 819 ever built, and 22 more under construction or planned. They are used for research and training, materials testing, or even radioisotopes production for medical and industrial applications. More than a few of these are in populated areas such as university campuses. [7]

Last but not least, over 70 nuclear power systems have been launched in space!