Yea!
Can we recycle a few nukes along the way?
Increasingly, datacenter operators are putting their faith in the promise of miniaturized nuclear power plants – better known as small modular reactors (SMRs) – to fuel their ever-growing energy demands. Sam Altman-backed SMR startup Oklo says it's obtained letters of intent from two major – but unnamed – datacenter providers …
Since I first heard of them back in 2015, I have always thought the Accelerator-Driven Subcritical Reactor to be an ideal way of recycling nuclear waste (including plutonium and old warheads) and producing safe controllable power, but it seems few people are working on them. Why?
Ok accelerators are expensive, but not THAT expensive, and we know how to build them. See ESS in Sweden: esss.se
Why?
They've constructed probably half a dozen reprocessing flowsheets already Including the one built for EBR II when it built in the first place. Naturally they all have flaws, starting with the issues of gas bubbles in the metal fuel of EBR II itself, although that seems to have been solved by improved metal alloys sometime in the 70's.
All reactors breed fuel and in fact modern PWR's couldn't run as long as they do between fuel reloads on the U235 in their fuel. Part of the time they run on Pu bred in the pellets in effectively an in-situ breed/burn cycle.
No doubt this will hoover up some R&D funds. Wheather it gets built, who knows?
France already recycle nuclear fuel, and other countries do as well .It's called MOX (mixed-oxide plutonium-uranium) fuel, and they use it extensively. This is reactor grade plutonium, which is a different isotope mixture than bomb grade plutonium (which has to be specially made for bombs in specialized reactors).
The reason it isn't done more widely is that currently uranium is so cheap that by most estimates is cheaper to use a once-through fuel cycle and just store the spent fuel until such time as uranium prices rise far enough to make it profitable to recycle the fuel. The French recycle fuel for reasons of energy security, so they don't have to import as much fresh uranium. They estimate that the costs of recycling are about the same as a once-through cycle plus long term storage. Most of the long lived radioactive elements in spent fuel are isotopes of plutonium which of course the French recycle back into fuel and get burned up in the reactor.
Canada has used MOX fuel made from surplus ex-Soviet nuclear weapons. However, this was more expensive to make than standard fuel (Canada uses natural non-enriched uranium fuel, which is cheaper than the enriched uranium which many countries use). It was only done in this case as part of an international agreement to dispose of surplus nuclear weapons left over from the collapse of the Soviet Union.
However, research has been done on recycling spent fuel from PWR reactors (a very common reactor style) and using it in Canadian designed CANDU reactors and derivatives (CANDUs are used in a number of countries). One method developed in South Korea involves simply chopping the spent fuel rods from a PWR into CANDU compatible lengths and welding the ends shut and feeding them into CANDU reactors. The fuel may be spent (used up) from the perspective of a PWR, but to a CANDU which runs on non-enriched uranium, this is high grade fuel.
The other method (developed by Canada) involves crushing the spent oxide fuel pellets from a PWR and blending them and mixing in fresh uranium to get a more consistent fuel mixture before reforming them into pellets and fuel rods and bundles. This is a dry process which is simpler and produces less waste than the conventional chemical reprocessing system used by France to produce their MOX fuel.
However, neither of these processes has been commercialized, again because uranium is currently so cheap and abundant that it's not economic to do so.
There are other fuel recycling methods as well, but they all run into the same issue of there being no market for it.
At this time the only thing that might make it worthwhile, aside from the national security reasons used by France, is if PWR reactor operators paid companies to recycle the fuel in order to take the waste off their hands. Using up the left over plutonium by recycling it gets rid of most of the long lived waste which would otherwise have to be stored.
Fun fact. Oklo is also the name of a place in Africa, where about 15 perfectly natural nuclear reactors have been found, dating from about 2 billion years ago.
It's "natural", so it's good for you, innit?
Hitachi (they go by various names) already have a customer building one of their SMRs in Canada, just east of Toronto. This is a 300MW reactor, and the plan is to build 3 more units alongside it for a total output of 1,200 MW.
The first reactor is already under construction, and it is scheduled to be connected to the grid in 2027, and delivering electric power within 2 years after that. This is being built next door to an existing nuclear power plant (Darlington) and is being operated by OPG (owned by the province of Ontario), who run a number of other nuclear power plants as well.
The actual fabrication of the reactor is being done by a company in Canada which has many years of experience building major components for nuclear power plants.
Hitachi are one of the shortlisted companies by the UK for SMR deployment in the UK. I understand their proposal is to basically clone the Canadian plant for the UK.
At 300MW this reactor is at the large end of the scale for SMRs, but the economics of very small SMRs is questionable. There are civil works (roads, cooling water handing, grid connections, etc.) which have inherent economies of scale, and these costs can be quite significant for very small reactors. At 300 MW, these SMRs are not drastically bigger than Pickering's 500MW reactors (also just east of Toronto), which date from the early days of nuclear power in Ontario. The "modular" (ability to build standard reactors in a factory and ship already assembled to the site) aspect is far more important to the SMR concept than the "small" aspect.
As for data centre companies building their own nuclear reactor, it would make far more sense for them to do what other companies that need lost of cheap electricity (e.g. aluminum smelters) do, which is to location operations in countries which have an abundant supply of electricity.
Perhaps you would like to revise that to "drastically smaller"?
The fact is all water moderated reactors (BWR, PWR, PHWR which is what CANDU is) are going to produce poor quality steam due to their low operating temperatures while being quite slow and expensive to construct. Those pressure vessels (excluding the CANDU again) and the large containment buildings to cope with worst case water (and subsequent flash boiling to steam) need a lot of concrete (about 2x that of a similar capacity FPP according to an MIT report).
CANDU is the best of this bunch but the GE-Hitachi is a mini-BWR, As such it has the same combination of high pressure(but lower than a PWR)/low temperature architecture that make it both difficult to build and thermally inefficient in operation, which has a knock-on effect of needing special steam turbines that have been described as being up to 10x the cost of normal FPP turbines that operate in the supercritical or the ultra supercritical temperature range. US FPP's run mostly at about 530-540c due to the availability of the US bad high sulphur (but very cheap) coal, while the rest of the world (including China) moved up to 600c+ decades ago.
Oklo has the right combo (High temp/low pressure) but the massively increased neutron damage (no moderator --> no reduction of neutron speed --> greatly increased wall damage) of the fast spectrum, and the high enrichment level (enrichment is 1/4-1/2 the cost of the fuel for normal PWR/BWR) will make fuel sourcing a PITA as most suppliers are only certified < 5% due to criticality (IE heaps of it forming a low yield nuclear radiation burst. not an actual explosion, but quite lethal to anyone standing nearby, or more likely running for their very lives).
Yes thank you, it should have been the new Hitachi SMRs are not drastically smaller than the reactors at Pickering.
As for build time, they did site prep work for all 4 units this past autumn. Construction of the nuclear works will start in early 2025, and the first unit is expected to be in commercial operation by 2029. This fairly quick, so the concrete work doesn't appear to be a serious bottleneck.
As for the type of steam turbines used, I don't think anyone really cares. Ontario gets the majority of its electric power from nuclear energy and has for many years, so the steam burbines seem to work just fine as is.
The UK AGR (Advanced Gas-cooler Reactors) were designed around the idea of being able to use high temperature steam turbines and higher thermal efficiencies, but the practical benefits of this were much less than the problems associated with higher temperatures. The AGR design proved to be a technological dead end, as has every other high temperature reactor.
Oklo's reactor is a liquid metal cooled fast neutron reactor. A number of countries have built these types of reactors over the decades and found them too complicated, expensive, and impractical. I haven't seen anything which would lead me to believe that Oklo's design will be any better.
The SMRs being built at Darlington are based on well proven technology. However, Ontario are also planning to build large nuclear reactors as well, again based on existing technology.
Yes, I read the original promotional Oklo press release, and now your article based on the press release.
It's all airball.
Like everything from Oklo, it is full of nightmarish Orwellian language and misinfomation, probably AI generated and smoothed - note that the press release had a disclaimer nearly >twice as long< as the release itself. Clearly legal is weighing jn heavily to sanitize everything since they are courting so much liability through their misrepresentations.
This has the stink of earlier Altman propaganda, such as his announcement of an Air Force contract last year, which was quickly rebuked by the Defense Logistucs Agency as an illegal sweetheart deal, currently under DLA investigation.
Altman promoted that fake deal because Oklo refused to divulge safety info to the NRC, so they refused his reactor license. Without a license, his IPO tanked, forcing him to take it back into his VC firm for a modified reverse IPO. With that boost, his last ditch private placement round squeaked through based on the false claims of the Air Force deal. Because the DLA investigation was suppressed until the week before his firing from OpenAI, and he succeeded in suppressing the Board's findings and then brought in NSA leadership and others to run OpenAI, we'll never know what role his nuclear dealings may have played in the ethics charges against him.
But he made sure to move with the rest of the industry to lobby, successfully, to slash NRC safety regulations and insurance and licensing requirements of the Price Anderson nuclear insurance act, eliminating NRC licensing of "foreign" reactors, along with the move to sell power directly as B2B in an effort to evade utility regulation.
Now with Trump, enforcement and regulation will be greatly reduced and the tech bros will get their shot at the goal of 10,000 reactors in everyone's back yard.
With both proliferation and safety protections slashed, the Atomic Veterans programs and their civilian public health components eliminated (and ignored by a complacent press), every ignorant wanne be is falling in ljne to support our glorious nuclear future, despite proven better, faster, and cheaper renewables plans (that don't benefit the centralization of power in big tech and big capital).
Since no absurdity is merely absurd for long these days, I'll nominate Sam Altman for Secretary of Misinfirmation and Energy - who's laying odds?
FWIW the issue with the NRC is true, and happened in 2022, though the upstart is getting along at least a little better with the government lately.
That said, as much as we like nuclear power, we are maintaining a healthy skepticism of this SMR project, and in future coverage about Oklo, we'll include some more history of the startup. I've added the above links to the piece.
C.
Yaaay, just what I need--a small nuclear reactor disaster in my back yard that forces my family out of our home never to return for a hundred years due to the fallout and contamination. Thanks guys, I guess they wipped the history books about Chernobyl, Three mile island, or Fukushima already.