And So It Begins
by David Archibald
17 May 2025
Diamondback Energy is one of the largest tight oil producers in the United States, with a production rate of about 500,000 barrels per day from the Permian Basin. In a letter to stockholders dated 5th May, the CEO of the company stated that: “Today, geologic headwinds outweigh the tailwinds provided by improvements in technology and operational efficiency. Therefore, we believe we are at a tipping point for U.S. oil production at current commodity prices” and “it is likely that U.S. onshore oil production has peaked and will begin to decline this quarter.”
World oil production had peaked in 2018. Now a competent operator has called the top in U.S. oil production. The consequence will be that the perception of a permanent energy cornucopia will flip to one of continually tightening supply. US natural gas production looks like it is also peaking. Longer term, things that can substitute for oil will go to the oil price in energy content terms less the cost of conversion. So, using coal and natural gas will be as expensive as if you were burning oil to get the energy you need. This starts from an oil price about twice what it is now.
When all the rocks we can dig up and burn have been burnt, our choice in energy supply will be nuclear power or horse-drawn carts. The next question will be how much nuclear energy will be required to replace the energy currently supplied by carbon fuels from the ground? It would be about 2,000 GWe which, using the current dominant nuclear technology of U235-burning light water reactors, would require mined production of 1.5 million tonnes of uranium each year. This compares to current uranium demand of about 60,000 tonnes per annum. World uranium resources are 8 million tonnes up to a production cost of US$98 per pound of yellowcake, equating to US$260 per kilo of uranium metal.
Figure 1: World mined uranium 1950 to 2022. Blue is mined uranium, red is reactor requirements.
At full tilt, the world’s identified uranium resources would last just over five years using light water reactors. These only use 0.4% of the contained uranium though. If nuclear technology switched to breeder reactors, currently identified uranium resources would last 1,300 years.
The other major fissile element is thorium. The current estimate for the world’s thorium resource is seven million tonnes. There is currently no commercial use for thorium, but doubtless more will be found once people start paying for it. It is no less useful as a nuclear fuel than the main uranium isotope U238, so eventually thorium will trade at the uranium price. Thorium is three times as abundant in the Earth’s crust as uranium.
Figure 2: Map of the world distribution of uranium resources
The bulk of Australia’s uranium resource is in the Olypmic Dam deposit in South Australia. Currently copper production from Olympic Dam is worth five times its uranium production.
The way fission works is that odd-numbered isotopes fission, and thus produce energy, while even-numbered isotopes accept a neutron to become the next element up, which can then be fissioned. So, U238 can be bred to Pu239 and Th232 can be bred to U233. The technologies to do this haven’t been commercialised yet. But we know enough to understand what they need to be. For uranium, the future needs to be lead-cooled reactors in the fast neutron spectrum. For thorium, the future needs to be molten salt reactors operating in the thermal neutron spectrum. It is all about breeding margins and safety. Thermal neutrons are fast neutrons that have been slowed down by a moderator, normally graphite, or ordinary water in light water reactors.
Figure 3: Uranium Production by country 2013 – 2022
The potential to breed plutonium from U238, and thus use the 99.6% of mined uranium that is currently wasted, was realised back in the 1950s. The technology to do this needs to produce more plutonium than it consumes. This is called the breeding margin. The first Soviet experiment to that end didn’t use a coolant and had a breeding margin of 80%. For a reactor design to produce power, you need a coolant to circulate through the reactor core and then be used to heat water to steam. Molten sodium has the best characteristics possible for breeding in the fast neutron spectrum, so ‘fast breeder’ research has concentrated on using sodium. The first US nuclear submarine, the USS Nautilus, was powered by a light water reactor. But the second one, the USS Seawolf, was powered initially by a sodium-cooled reactor. It proved to be problematic. Admiral Rickover, who developed the US nuclear submarine fleet, made the observation in 1956 that sodium-cooled reactors are “expensive to build, complex to operate, susceptible to prolonged shutdown as a result of even minor malfunctions, and difficult and time-consuming to repair.”
Figure 4: Uranium reserve distribution by country and operating cost in US$ per kilo
Nevertheless, many countries persevered with sodium-cooled reactor research projects because of its breeding margin of 20%. This means that, ideally, the number of reactors could double every 3.6 years. France was able to develop a successful breeder reactor at a rated capacity of 1.2 GWe, after a great deal of expense and trouble, but shut it down in a deal with that country’s Green party. Russia has been able to run a couple of sodium-cooled reactors for decades but is now building a lead-cooled breeder reactor. This is the Brest-OD-300 reactor in the city of Seversk in the Tomsk region of Russia, with a rated capacity of 300 MWe. The current stage of the project is building the reactor vessel.
Lead has a breeding margin of up to 10% under ideal conditions; the Seversk reactor is designed to breed with a 5% margin. Russia had previously operated lead-cooled reactors in its Alfa Class submarines from 1971 to 1996. These were able to maintain a submerged top speed of 41 knots. The problems of lead-cooled reactors include erosion and corrosion of the steel components of the reactor. A Swedish company, Blykalla AB, was formed in 2013 to commercialise lead-cooled reactors and solved these problems by developing an aluminium oxide forming steel for protecting cladding tubes, an austenitic steel for protecting reactor vessels, and a martensitic steel suited for lead pump impellers. The founders of Blykalla had been motivated by a desire to mitigate global warming. From their understanding of the science, they realised that using lead-cooled reactors for power generation would be the best way to mitigate that perceived problem. They are now offering a small modular reactor of 55 MWe rated capacity.
Figure 5: Blykalla’s 55 MWe Sealer reactor
On the subject of small modular reactors, they are not suited for grid-scale application. They are approaching the problem from the wrong end. Nuclear energy in the long term, at the grandchildren scale of time, needs to be approached with breeding as the first criterion. As a civilisation, U235 is the match that Nature endowed us with to start the nuclear fire that will sustain our civilisation for thousands of years. We are still burning through that match instead of starting the main fire. When humanity’s energy use has settled down to a steady state, at a 10% breeding margin, 90.9% of nuclear energy will be from breeder reactors and the remaining 9.1% will use the excess plutonium for site-specific applications. This might be for remote mines off the grid or for ships. At one stage the US Navy calculated that nuclear-powered ships could be cost-competitive down to a tonnage of 8,000 tons. It all depends upon the cost of the competitor fuel which will be synthetic diesel made from biomass.
In the interim, running small modular reactors on enriched uranium is simply burning more of the match. There is also the problem of capital intensity with scale. With most industrial processes, doubling the size of a plant results in a 30% reduction of the capital cost per unit of production. This tendency may be even more so with nuclear power, and the capital cost reduction might be 40% with a doubling of size. This works in reverse if you try to make the reactor smaller for no good reason. Smaller reactors might be safer in that the ratio of surface area to volume is higher than in normal reactors so they can shed heat faster, but lead-cooled reactors are inherently safe in that if the reactor has a leak, the lead will freeze on the floor of the reactor building and nothing else will happen. A lead-cooled reactor operates at atmospheric pressure and can even be refuelled while it is producing power.
There has been a tendency in recent decades for full-scale reactors to have enormous cost overruns. These have been in erecting the containment buildings rather than the reactor vessels, and are ultimately due to a lack of continuous experience in building nuclear power plants. The Russians, Chinese and South Koreans don’t experience the same sort of cost overruns, because they have continuous experience in building reactors.
There is another technology that needs to be optimised for our ideal future to happen. This is chemical separation of spent fuel to produce the plutonium for the breeding cycle. A few countries around the world make a token effort at this to be able to separate out weapons-grade plutonium if they wanted to. Otherwise they would rather not because reprocessed plutonium won’t be competitive with mined uranium until the yellowcake price is US$200 per lb. It is currently US$72 per lb. Each tonne of reprocessed plutonium displaces 250 tonnes of mined uranium in light water reactors.
But to have a future that lasts more than just a few years, the nuclear industry needs to switch from enriching uranium to reprocessing plutonium from spent fuel. China has a research program in optimising the hydrometallurgical reprocessing spent fuel. Once it is optimised, they intend to build capacity at a scale of 800 tonnes of plutonium per annum. This would displace 200,000 tonnes of mined uranium if used in light water reactors, about three times current world consumption. Or they could operate 800 breeder reactors of 1 GWe each. China also has a research program in lead-cooled breeder reactors. As with Blykalla, it seems they have read the literature and have realised what is needed to have a future that works.
Figure 6: Currently identified Australian thorium deposits
Back to thorium. There is at least three times as much thorium as uranium. So, three quarters of the nuclear power plants at some point in the future will be molten salt reactors breeding Th232 to burn U233. They will use a two-fluid design because of the better breeding margin of this design, rather than trying to do it all in one container. Thorium, currently worthless, will become as valuable as uranium.
Australia has a lot of catching up to do in nuclear energy. Things we should be doing include:
- Current Australian uranium exports are sent overseas for processing. This starts with enrichment of the U235 component into 14% of the mined uranium that the process started with. The remaining 86% is called ‘depleted uranium’ in which the U235 has been reduced from 0.7% to 0.2%. The depleted uranium is currently considered worthless but one day, not too far off, it will be almost as valuable as mined uranium processed to uranium hexafluoride. The Australian Government should require that the depleted uranium produced from Australian-sourced yellowcake be sent back to it. We would then be selling something and then getting 86% of the inherent value back.
- Obtain spent reactor fuel in their steel storage casks and store that as a service to our uranium customers. One day the contents will be valuable and the 0.7% plutonium in the spent fuel will be needed to start breeder reactors operating.
- Start a research program in hydrometallurgical separation of plutonium from spent fuel. This technology is known as PUREX but we need experience in operating our own version of it.
- Design, build and operate a lead-cooled plutonium breeder reactor to optimise that technology.
- Design, build and operate a two-fluid molten salt thorium breeder reactor.
David Archibald is the author of The Anticancer Garden in Australia