This IAEA report [1] has more details about this design, and the dozens if not hundreds of other types of molten salt reactor designs. The relevant section for the Kairos Hermes design is 4.5 (pages 41-44).
[1] https://www-pub.iaea.org/MTCD/Publications/PDF/STI-DOC-010-4...
MSRs are a costly distraction. They are not viable without literally hundreds of billions in research and development money. That's why all the MSRs startups are failing long before they even start the licensing process.
> For example one quite unexpected effect in the MSRE was the enbrittlement of the reactor vessel caused by tellurium, which shows up as a fission product when U-235 burns.
It was not unexpected. The _main_ issue with MSRs is that they have to contain fluoride salts that release elemental fluorine radicals as a result of radiolysis. So the reactor vessel walls will be eaten up by them, rather rapidly. Especially when reactors are scaled up to a level that makes them practical. And then you have all the fission byproducts that literally include almost all the Periodic Table.
But yes, this design might be actually feasible for small reactors. But I bet that it won't be cheaper and it'll be impossible to scale to levels approaching PWRs.
> Both Hermes and ETU 3.0 will be built using modular construction techniques, with reactor modules fabricated in Kairos Power's facility in Albuquerque, New Mexico, which will be shipped to Oak Ridge for assembly.
https://www.world-nuclear-news.org/articles/work-begins-on-f...
Activation is also low. The two concerns would be traces of tritium from the two-step activation of beryllium (formation of 6He by (n,alpha) reaction, decay of 6He to 6Li, then (n,t) on 6Li), and also formation of 10Be (half life, 1.4 million years, but the thermal neutron capture cross section of 9Be is only 8.5 mb). The chemical toxicity of beryllium would considerably exceed its radiotoxicity, I imagine.
Also from the 6Li absorption of 1 neutron. The Lithium used is almost pure 7Li, at 99.995%. But there is still 50 ppm 6Li in it.
Yep.
But I'm personally more worried about the pelletized fuel. Pellets are not a great form-factor, and they don't have cladding. Fluoride salts are also far more aggressive than water (mechanically and chemically), so this will limit the maximum specific power of the reactor. So I don't think that something like 1GW molten salt reactor is even possible.
It might be OK if they want to continue using SMRs, though.
> The slightly reduced salt is also preferred to limit corrosion
Sidenote: that's actually not always a great idea. Steel is stainless because it's covered in a film of oxides, and without oxygen it might not be able to form.
This is a problem for the Russian BREST-300 reactor that is cooled by molten lead, they had to do almost 10 years of research to perfect a system that controls the amount of dissolved oxygen in the molten lead. And it's still not clear if they succeeded until the full-scale reactor is built.
In this case, though, I think that they can tune the reducers to react with fluorine preferentially, while still leaving enough oxygen.
https://www.sciencedirect.com/science/article/abs/pii/S00109...
The US produces 40000 billion kWh every decade, so that doesn't really seem that bad to me.
Modern PWRs are just as passively safe, if they lose cooling and melt down, the molten fuel will be safely contained by the core catcher.
Uranium is also not scarce, and if we want to get into breeding reactors, existing fuel reprocessing is an established industry, on which we _already_ have spent several hundred billion dollars.
Or maybe thorium reactors ON A ZEPPELIN! With microwave power transmission. After all, if the objective to spend money on research, then what can be cooler than a nuclear reactor on a blimp?
Hydrogen can be so hawt!1!!
CATL’s lithium phosphate batteries are more expensive for now at >$100kWh.
Sodium batteries are definitely cheaper in practice, but much lower energy density.
The question is how far CATL can push energy density, and the second generation is claimed to be >200Wh/kh.
https://www.autoevolution.com/news/catl-and-byd-to-start-pro...
https://www.westchestercleanenergy.com/post/record-low-lithi...
Here's one: https://energy-charts.info/charts/power/chart.htm?l=de&c=DE&... - look at the dates between 2019-01-16 to 2019-01-25.
> It's clearly visible in the charts of the last years
I have just provided you an example. Want more? Here's another: https://energy-charts.info/charts/power/chart.htm?l=de&c=DE&... - the period between 2023-02-04 and 2024-02-08, then followed by 2023-02-12 to 2023-02-15.
But hey, it's all fake news. When the next Dunkelflaute happens, the citizens are supposed to just sit in their cold homes and think how great renewable generation is during the other times.
https://www.researchgate.net/figure/Correlation-coefficients...
Like today we will need peaking capacity in the future, likely either based on hydrogen, synfuels or biofuels.
What we don't need is nuclear power plants which sit idle at all times unless there is a winter dunkeflaute across half of Europe.
The problem is, this is average. You _have_ to plan the grid for the worst case scenario. For Germany, the worst case is 1 month of straight Dunkelflaute. It's estimated to happen once in 100 years.
> Like today we will need peaking capacity in the future, likely either based on hydrogen, synfuels or biofuels.
It's not peaking capacity. It has to be more than 100% of the current capacity (once Germany switches to electricity instead of gas for heating). And it'll be mostly sitting idle.
> What we don't need is nuclear power plants which sit idle at all times unless there is a winter dunkeflaute across half of Europe.
Build nukes, remove wind generators. Problem solved.
Consider an even worse issue and expect 80% to be offline for the 100 year catastrophe.
https://www.nytimes.com/2022/11/15/business/nuclear-power-fr...
In reality we work with statistics. In for example Sweden the “reliability guarantee” is at most one hour of demand exceeding production per year.
That’s a reliability of 99.8%.
Building nuclear power plants would mean locking in enormously more expensive energy costs for everyone, stalling the green revolution in terms of electrification of industry.
This is the problem with nukebros, there’s no logic to the suggestions. Only a complete fixation on nuclear power as the solution to everything.
Which means widespread continued fossil use and energy crisis power cost for the general public.
But that’s a price you’re willing to pay as long as we spend trillions of dollars on subsidies of nuclear power.
Utterly wrong. Grids _are_ sized for that.
> Then we need to make the same calculation for when half the French nuclear fleet was offline.
French shutdowns were _planned_. Nobody died as a result, and it only cost more money than planned as a result of unlucky confluence of events.
> In reality we work with statistics. In for example Sweden the “reliability guarantee” is at most one hour of demand exceeding production per year.
You have it backwards. Sweden is ready to accept one hour of outage per year. Not more than that.
The failure scenario for Germany is not an hour of outage, but a month without energy. Leading to millions of people dead and a total economic collapse.
> This is the problem with nukebros, there’s no logic to the suggestions. Only a complete fixation on nuclear power as the solution to everything.
There's no fucking shame and zero self-reflection with you greenies.
Germany _wasted_ close to $500B on useless Energiewende and is still directly _subsidizing_ _new_ _gas_ _powerplants_. And it'll need to spend even more than that to have even a _hope_ of carbon neutrality, contingent on multiple speculative bets coming true.
Instead they could have used the same amount of money to build a completely carbon-neutral nuclear-powered grid. That would have been available by now. Using technology that was tried and tested in 2000-s.
So maybe stop with the hyperboles?
Then a ton of post-fact reasoning for why it was actually fine that half the French nuclear supply was offline during the largest energy crisis in a generation.
Please get back to reality.
Wasted? For the first time since the Industrial Revolution we have found a new cheapest near infinitely scalable energy source: renewables.
Nuclear power was the last attempt, it never delivered on the promises.
I love how you say that it would be “available now” when Flamaville 3 started at the same time currently is 6x over budget and 12 year late on a 5 year planned construction timeline.
Germany would have had massive cumulative emissions when still waiting for nuclear power to come online.
But that’s the problem with you guys. You don’t care about emissions. It’s completely fine locking in fossil fuels for 20 years as we wait for nuclear power to come online.
And then in the next sentence you turn around denigrating even a single percentage of fossil fuels as we transition into a renewable grid, before said nuclear would come online.
It’s simply completely senseless. You’re making ridicule of yourself.
For the _second_ time. The first time was with the nuclear power. Keep forgetting that, yes?
> Nuclear power was the last attempt, it never delivered on the promises.
France has an 8 _times_ less carbon-intensive grid than Germany. RIGHT NOW. Keep forgetting that, yes? They can go to carbon-neutral with fairly minimal changes.
> I love how you say that it would be “available now” when Flamaville 3 started at the same time currently is 6x over budget and 12 year late on a 5 year planned construction timeline.
Russia is finishing the Bangladesh nuclear power plant. On time and within budget. 2 reactors completely built within 10 years. If Russia can do that, other countries can certainly replicate that.
> Germany would have had massive cumulative emissions when still waiting for nuclear power to come online.
Germany is right now already _locked_ into future emissions for more than 20 years. Even freaking _coal_ (!) is not being phased out until 2038: https://www.cleanenergywire.org/news/german-government-says-...
> And then in the next sentence you turn around denigrating even a single percentage of fossil fuels as we transition into a renewable grid, before said nuclear would come online.
Again, zero reflection from your side. Zero contrition, zero knowledge, and endless excuses.
France is literally the next door. Their carbon intensity for energy production is just 12% of Germany's. Right now. Germany in the best possible case won't be able to match that until 2040-s. In reality, it won't happen unless something magical occurs.
Hydro power did, and is famously called "geographically limited" because we in short order exploited near every single river globally.
France made the right choice in the 70s in the name of energy independence and nuclear weapons. They did not care the slightest about emissions.
The equivalent choice in 2024 to nuclear power in the 1970s is renewables.
"Hurr durr my cherry picked reactor!!!"
While completely ignoring all western projects. The facts are: Flamanville 3 still haven't entered commercial operation and the projected was started at about the same time as energiewende.
Flamanville, HPC, Olkiluoto 3 and Vogtle are the successful western projects. The unsuccessful get stuck in financing limbo like Sizewell C because the needed subsidies are truly stupid.
https://www.ft.com/content/2a5d9462-b921-4577-82c1-4eb508775...
You are proposing that Germany in 2024 should have emissions closer to Poland because you value building nuclear power above curbing emissions.
Lets do a thought experiment in which renewables somehow end up being wholly incapable of solving the last 20% of carbon emissions.
Scenario one: We push renewables hard, start phasing down fossil fuels linearly 4 years from now, a high estimate on project length, and reach 80% by 2045.
The remaining 20%, we can't economically phase out (remnant peaker plants).
Scenario two: We push nuclear power hard, start phasing down fossil fuels linearly in 10 years time, a low estimate on project length and reach 100% fossil free in 2060.
Do you know what this entails in terms of cumulative emissions?
Here's the graph: https://imgur.com/wKQnVGt
The nuclear option will overtake the renewable one in 2094. It means we have 60 years to solve the last 20 percent of renewables while having emitted less.
Do you still care about our cumulative emissions when any dollar spent on nuclear power increases them?
Even better, if you look at energy flow in the EU grid for e.g. the period from 2023-02-12 to 2023-02-15, you'll see that we actually exported energy to nuclear-powered France, and imported from renewables-powered Denmark.
That's actually a proof that renewables saved the day there, because lack of wind is always only a regional phenomenon.
Also, both li-on and sodium are cheap and getting cheaper, so the latter being more expensive currently is kind of moot.
For that you need to have 100% of idle capacity just sitting there.
> The priority for climate change mitigation is to get to 97% carbon free soon
No, it's not. The priority is to shake down consumers to get as much money to Greenie cronies as possible. That's why Germany is straightforwardly _subsidizig_ new natural gas power plants. With a pinkie-promise to be "hydrogen ready", maybe sometime in 2030-s.
If we hope to go 100% renewable, storage is a key piece of that puzzle.
Much of the pumped hydro that exists today was built to handle excess nuclear.
Furthermore, too much energy is a far easier problem to solve than too little energy. People can desalinate water, or do any other energy intensive things.
Economically? Load following with nuclear power means an even worse business case than running at 100% 24/7. And nuclear power is already a laughably bad business case when running at 100%.
Economically? Load following with nuclear power means an even worse business case than running at 100% 24/7. And nuclear power is already a laughably bad business case when running at 100%.
Nuclear is a bad business case compared to a fossil fuel grid. Solar and wind backed by fossil fuels are a better business choice, too. But when it comes to a fossil-fuel free grid, it's the only viable option if you don't have a big source of hydropower nearby. Batteries can't deliver the required storage capacity. Remember, the world uses 60,000 GWh of electricity per day. And as transportation and industrial uses of fossil fuels are electrified, that'll increase.
Hydroelectric storage is the only grid-scale energy storage system available to us, and it's geographically dependent. And the places that are suitable for hydroelectric storage usually don't need it because they can just generate electricity via hydropower anyway. Until your hypothetical breakthrough in power-to-gas or giant flywheels actually happens, this is the state of grid storage.
You have a baseload. That's basically the minimum load that always exists. You don't have to worry about selling the power, it's already sold. And then you have peak. Both of these are time / weather dependent but we're still quite able to plan days, if not weeks and months, in advance for what those two values will be. As an example, let's get back to the original topic of the article.
If you're Google and you have a particular datacenter at a location, you know what that baseload is. You know what the peak is. It's a pretty simple calculation to figure out what is most cost-effective here. It's probably even easier for you than the local power company, as base and peak loads likely don't fluctuate much for you outside of HVAC keeping up with weather. It might be to use nuclear for just base load and use the local grid for the rest. It might be to over-produce occasionally because that's still cheaper than buying from the local system operator. Hell, you can probably sell any access, but that's more of a problem than most people think.
But the point is nuclear is king when it comes to baseload power supply. And a datacenter, which consumes a lot of power consistently, is almost entirely baseload.
And power plants having individual finances in the first place is a policy choice, not a law of nature.
If you find cryptocurrency mining objectionable for some reason, you could apply the same basic principle to things like aluminum refining or desalination.
https://reneweconomy.com.au/much-storage-needed-solar-wind-p...
Sadly, there's far, far, far too much FUD floating around about storage (understandably, coz wind+solar threatens the nuclear+carbon lobbies), and not enough thorough and realistic studies like this one.
I've heard people say "oh you cant pay attention to this study because it's in Australia which must be discounted because [reasons], what about [ other country ]?", and I'd welcome seeing an alternative study making appropriate assumptions, but none of these comments so far come attached to anything other than FUD.
I've also seen far, far too many people build or cite a "naive" models that make inappropriate assumptions (e.g. that zero power is generated at night by wind).
So a study for Australia should be applicable everywhere else? Like in Germany or Norway?
The issue is that electrical grids have very high reliability requirements (upwards of 99.95%). Plenty of models claiming that small amounts of storage required neglect to mention how frequently they will encounter insufficient generation. Remember, even fulfilling demand 99% of the time is a 20x increase in blackouts.
Also, even just 12 hours of storage globally would be 30,000 GWh of storage. That's still about 50 times the amount of batteries produced annually. The reality is that hydropower is the only feasible form of grid storage.
I am, as I said, still waiting for models which demonstrate that it is wildly different...
Here's a good overview article: https://energytransition.org/2017/07/germanys-worse-case-sce...
> Germany will always need dispatchability roughly at the level of its annual peak demand
The solution they're proposing is power-to-gas. Which so far has been way too expensive to matter.
Power to gas for the "worst case scenarios" would also be the last thing you'd do to transition a grid from 97% carbon free to 100% carbon free.
> However, the share of solar generation increases less, or even decreases, in higher-latitude countries like Russia, Canada, and Germany (Fig. 2b). These trends continue as more storage is added, so that with 12 h of energy storage and no excess annual generation, 83–94% (average 90%) of electricity demand is met with mixes of 10–70% solar power (49% on average; Fig. 2c).
Even with 12 hours of storage, Germany would be seeing blackouts weekly.
To put this in perspective:
> reliability standards in industrialized countries are typically very high (e.g., targeting <2–3 h of unplanned outages per year, or ~99.97%). Resource adequacy planning standards for “1-in-10” are also high: in North America (BAL-502-RF-03), generating resources must be adequate to provide no more than 1 day of unmet electricity demand—or in some cases 1 loss of load event—in 10 years (i.e., 99.97% or 99.99%, respectively)
So no, even with 12 hours of storage and 50% overcapacity, we'd have an unacceptably unreliable grid.
Also, your linked article is not modelling a carbon-free grid:
> Graham says that the CSIRO modelling showed that at very high levels of wind and solar, a maximum of half a day’s average demand was needed for storage. In some areas of the grid, only around three hours might be needed.
What are these "very high levels of wind and solar"? How much of the remaining demand is satisfied by fossil fuels? The article doesn't say.
This isn't really accurate. The majority of pressure cookers you can buy operates at an absolute pressure of only about twice the atmospheric pressure.
I'm glad they are, actually! Personally, I'm not really convinced by the "small modular reactor" concept. Ok, so building a big nuclear power plant is expensive. But is it really cheaper to build 10 smaller nuclear power plants (which all need to conform to the same safety regulations, need maintenance, personnel etc.) instead?
I also wonder how much of the cost is tied up in the upfront construction and engineering compared to operations.
Everyone in this thread is assuming this, but you get economy of scale by building lots of the same thing, not by constantly changing things.
The US has more hope to learn by iterating on smaller projects.
About your comparison to SpaceX: the approach of building a rocket, launching it, letting it explode and then using the gathered data to make the next one explode later (or not at all) is fine for rockets, but I wouldn't want to see it applied to nuclear reactors.
I think small modular reactors are the way out of the similar cycle we've got for nuclear reactors. And I think that building a small number of large ones is going to be a lot better if we're also building a large number of small ones and learning.
It's like not building a y houses for 30 years and then building a massive skyscraper. If that's all we do then we'll only ever have one way of building a skyscraper because there's no room to experiment on other construction materials and techniques.
I'm pretty sure that was what NASA was going to do all the way back in the 70s before their funding got slashed. It was a novel idea but not a novel idea that SpaceX was particularly responsible for, just one they threw capital at because the government stopped caring after the space race.
Nuclear power research, by contrast, never really suffered from a lack of available funds. They were throwing money at mini reactors back in the 90s, saying all the same stuff about how mass production would bring down the price.
So, while in theory it may have had the money, in practice it did not.
It's not entirely bad though, I'm sure lots of those folks are doing good and important stuff. But I don't think the balance between employment and building things is quite right. At least to my tastes.
The route that Nuscale and Kairos are taking is both cheaper and safer than large scale reactors. And we are going to need something to fill the gaps when there is no wind or solar generation available if we want to get off fossil fuels.
Large physical scale is everything with old school power generation technologies due to various scaling laws.
https://spectrum.ieee.org/the-forgotten-history-of-small-nuc...
> a 400-MW reactor requires less than twice the quantity of concrete and steel to construct as a 200-MW reactor, and it can be operated with fewer than twice as many people. Writing in Science in 1961, a senior member of the AEC worried that “competition [from fossil fuel plants] is indeed formidable” and suggested that “with current pressurized-water reactor technology, lower nuclear power costs can be achieved most readily with large plants.”
Almost all the SMRs we've built have been water cooled reactors, which require a large amount of concrete and steel. Newer designs, such as molten salt reactors, can use a lot less, because they operate near atmospheric pressure, don't have to leave room for a high-pressure pipe break introducing a large volume of steam, and have nothing that can cause a chemical explosion. They're also inherently stable and adaptive to load, with little need for active control. A small modular MSR could well work out.
Well that is more because the American grid is suffering from a mountain load of neglect-debt to a tune that paints the infamous German railways as saints - the Camp Fire for example was most likely caused by a C-hook breaking after many, many decades of neglect [1].
Here in Europe, we get by just fine with outages measured in maybe a dozen minutes (!) a year [2] - Germany got rid of all nuclear reactors, many old ones in other countries were retired as well (Fessenheim, the most infamous one, in 2020), and while there are new builds, they are often decades late and many billions over budget.
How do we do this? We have strict regulations across the board (mandating stuff such as resilience against bad weather, unlike Texas which IIRC refuses to tie in to the other US grids to avoid such regulation), a cross-continental spanning grid [3], and most especially... we just love to bury our cables belowground, so even in the case lightning or storms hit, the actual impact on the consumers is all but negligible. And as the infamous summer of 2023 shows, the power grid still didn't fail even as dozens of NPPs in France and Switzerland had to completely shut down or significantly reduce output because there was not enough cooling water [4]. Hell we even manage to supply an entire country at war with decent power, despite Russia continuously attacking the power grid.
[1] https://www.nbcbayarea.com/news/local/long-term-wear-found-o...
[2] https://de.statista.com/statistik/daten/studie/37960/umfrage...
[3] https://www.entsoe.eu/data/map/
[4] https://www.reuters.com/business/energy/high-river-temperatu...
https://en.wikipedia.org/wiki/List_of_Falcon_9_and_Falcon_He...
Comparisons to other rockets here:
https://arstechnica.com/science/2022/02/spacexs-falcon-9-roc...
https://www.guinnessworldrecords.com/world-records/most-succ...
At the moment, of course, two astronauts are stuck on ISS after Boeing's new spacecraft, developed along more traditional lines, experienced problems in its first crewed flight. They're awaiting rescue by SpaceX's Dragon, which has flown 16 times with crew and delivered astronauts to the ISS 10 times, all without a glitch. Both companies were awarded their crew contracts in 2014.
Actually proving this with a deployed design seems to be the challenge, though with strong headwinds against nuclear across the board, favoring proven designs over newer ones is at least a little faster. That's not a flaw of the design, but intrinsic to the social and political environment.
Hopefully it is just a stepping stone.
History is not really against them. Our current reactors (mainly pressurized water reactors) are the way they are because Admiral Rickover determined that PWRs are the best option for submarines. He was not wrong, but civilian power reactors are not the same as the reactors powering submarines.
PWRs are expensive mainly because of the huge pressure inside the reactor core, about 150 times higher than the atmospheric pressure. For comparison, a pressure cooker has an internal pressure about 5 times higher than the atmospheric pressure, and such a cooker can explode with a pretty loud bang.
The Kairos Hermes reactor design is based on a design that was tested in the '60s, the Molten-Salt Reactor Experiment [1]. While such a reactor can be used to burn thorium, Kairos decided to go with the far more conventional approach of burning U-235. The reactor operates at approximately regular atmospheric pressure. This should reduce considerably the construction costs.
Of course, there are unknowns. While the world has built thousands of pressurized water reactors, it has built maybe 10 molten salt reactors. For example one quite unexpected effect in the MSRE was the enbrittlement of the reactor vessel caused by tellurium, which shows up as a fission product when U-235 burns.
The Nuclear Regulatory Commission is a very conservative organization, and they don't have much experience with molten salt reactors because nobody has. It took them 6 years to give NuScale an approval for a pressurized water reactor, design that they knew in and out. My guess is that they will not give Kairos an approval without at least 15 years of testing. But Google's agreement with Kairos is quite crucial to keep this testing going.
[1] https://en.wikipedia.org/wiki/Molten-Salt_Reactor_Experiment