Nuclear reactors and their hucksters
Apr. 30th, 2025 04:03 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
glaurungEvery so often, I see another media story about how a new kind of nuclear reactor that avoids all the horrible problems of the ones we have now is being studied, or a startup is looking to build one, or whatever. Don't be afraid of nuclear power, these articles proclaim, there is new, better technology that will resolve all the problems with the bad old reactors. The nuclear technologies I've seen mentioned like this make up a venn diagram of fast neutron reactors, molten salt reactors, and thorium reactors.
And these kinds of stories come from a mix of hucksterism and religion. Just as we have a religion of space enthusiasm (where solar power stations or helium 3 or whatever bullshit technology becomes the rallying cry for the religion's real goal, of having a city on the moon/Mars/in orbit, and having people permanently *Living in Space*), and the religion helps sustain huckster snake oil plutocrats who don't really care about space at all except as a way of siphoning off government subsidies and pumping up stock prices for their space technology companies -- so too we have a religion of Nuclear Power (it's the Future!), and nuclear power technology companies whose plutocrats have snake oil to sell. (PS: my data-free impression is that space enthusiasts vastly outnumber space hucksters, but for nuclear power, the ratio is much more even or perhaps reversed).
Some nuclear power advocates have at least half a leg to stand on (solar and wind are not 24/7/365 power sources, and nuclear power plants *are* a carbon-free way to provide round the clock power regardless of weather). But solar has become SO much cheaper than fossil fuel, let alone nuclear, that it leaves budgetary headroom for adding some kind of power storage to a solar farm and still being less expensive than the alternatives - and solar powered storage (like a lake that you pump full during the day and drain through hydroelectric generators at night) completely avoids all the regulatory and PR hassles of nuclear power.
Other nuclear advocates seem to Want to Believe in nuclear because they are right wing and regard solar as tainted by the leftist eco green conspiracy, or something? IDK.
But after encountering another "the new generation of (insert technical descriptor) nuclear power plants will completely avoid all the problems you've come to expect from nuclear power" article, I decided to try and figure out just how much truth there is to such articles. After picking away at the question for a while, I'm finally typing everything up so my time will not have been completely wasted. The rest of this post comes from reading far, far too many web pages (mostly on wikipedia but also elsewhere) devoted to nuclear power.
***
Nearly all US-based nuclear power and the vast majority of the world's nuclear power comes from uranium fueled, water-cooled, slow neutron reactors. That's because of two tiers of technical lock in.
First, from 1942 through the mid-50's, reactors were tools for transforming uranium-238 into plutonium 239, for bombs. Fission always produces fast neutrons, but slow neutrons are best for transmuting U-238 into plutonium. Water is a neutron moderator (it slows them down), so it made sense to submerge the reactor core in water, keeping the reactor cool and slowing down the neutrons produced at the same time. To minimize the formation of unwanted plutonium isotopes (plutonium 238, plutonium 240 and 241 are all undesirable in bombs - from some combination of emitting dangerous gamma radiation, emitting tons of heat, and/or being eager to spontaneously fission, causing premature detonation), reactors for making bomb plutonium would change out their fuel frequently, and maintain a rate of reaction to maximize production of the right kind of plutonium, with no regard for making use of the heat the reactor produced. While lots of experimental reactors were built and tested during that decade, including various fast neutron types, the focus was on producing lots of plutonium for the cold war arms buildup. By the mid 50's, the world's reactor engineers and technicians had lots of experience with water cooled uranium reactors, and little experience with other kinds.
Second, when, in the early 50's, Admiral Rickover started to work on his dream of a nuclear powered submarine (a sub that could stay underwater indefinitely, drastically transforming its strategic value), the teams he directed drew on the engineering expertise that existed, which meant they designed a water cooled uranium reactor. They made other tweaks to keep the reactor small and compact without reducing its energy output too much, and to minimize the volume that had to have radiation shielding. The result was a pressurized water-cooled reactor, PWR. The reactor core is bathed in an inner loop of water kept under pressure to keep it from boiling. That water, in turn, heats an outer loop of water into steam, which spins turbines to produce power to run the submarine. The water in the inner loop has three jobs: it transfers heat from the reactor to the outer loop, it slows the neutrons in the reactor, and calibrated amounts of neutron absorbing substances (boron in the form of boric acid, for example) are dissolved in it to keep the rate of fission constant. Graphite control rods can be inserted into the reactor to shut it down, but while it's running, boric acid is used to keep the neutron flux stable.
Tech/glossary note: anything that absorbs neutrons is called a *neutron poison*. As a reactor runs, fission products build up in the fuel (and decay, since fission products are all highly radioactive - more about this later). Some of the longer-lived fission products are neutron poisons, so as a reactor burns through its fuel supply, the rate of fission will go down as more and more neutrons are absorbed by the fission products. To keep the rate of fission steady, you need to have some poison present with new fuel, which you reduce gradually as the percentage of poisoning fission products in the fuel goes up, to keep the total energy output even. Having some boric acid in the coolant water, and diluting it over time by adding pure water, is one approach.
By the mid-50's, as part of Eisenhower's "atoms for peace" initiative, the first power reactors started being built... and they used the designs that already existed for powering naval vessels, and hired ex-navy nuclear techs to operate them. The first commercial atomic power plant in the US, Shippingport Atomic Power Station (commissioned 1958), started out using a reactor core originally built to power a (never built) aircraft carrier.
And that technical lock-in, of engineers who knew how to build water cooled uranium reactors and of naval technicians who knew how to run them, has resulted in a nuclear power industry that is overwhelmingly composed of PWR reactors, from the 50's through to today.
The good about PWRs: they're the nearest thing to a mature technology in the reactor industry. They can operate with only mildly enriched uranium (2-5% U-235), or (in the case of the Canada's CANDU reactors) natural, unenriched uranium. And, water is extremely well understood. It's cheap, stable, non-reactive, remains liquid at room temperature, and is transparent (allowing inspection of the reactor core via cameras).
But PWR reactors have numerous disadvantages.
First: problems with using water. It has a very low boiling point. Steam is not dense enough to slow neutrons, so you have to keep the water in the reactor from boiling, which requires keeping it under pressure. That in turn requires encasing the reactor in a thick steel pressure vessel, having your piping be heavy duty thick steel, and having pumps and other gear to maintain high pressure. All of that adds substantially to the cost. Plus, the neutrons which escape from the reactor core are absorbed by the steel pressure vessel, altering its composition and making it more brittle over time ("neutron embrittlement"), which shortens its life (in practical terms, there's no replacing a pressure vessel that's reached end of life - once the steel has degraded too much, the entire reactor has to be scrapped and replaced with a new one). Water is corrosive to steel, especially with boric acid added to it. The containment vessel and pipes need to be made of stainless steel (expensive) and even then, corrosion works to shorten the life of containment vessel and piping.
The water inside the pressure vessel is contaminated. If there's ever a failure in some of the reactor's plumbing, radioactive steam will leak out. The many safety precautions needed to prevent ruptures in the pipes and to contain steam leaks should they occur add even more to the reactor's cost.
The hotter you make the steam in the outer loop of a reactor, the more efficiently you can produce electricity. But there's a sharp limit to how hot you can let the inner loop of the reactor get before the pressure required to keep the water from becoming steam becomes impossibly difficult to maintain. Thus PWR reactors are less efficient than reactors that use a coolant with a higher boiling point. (ideally, a coolant with a boiling point higher than the temperature at which the reactor normally gets, would let you have an unpressurized coolant system, vastly reducing cost, complexity, and removing one major source of potential accidents).
Second: problems with neutron absorption. Water moderated uranium reactors produce plutonium and other transuranic isotopes, which remain in the spent fuel and are a major source of long term radioactive waste.
Tech/glossary note:
All elements after radium in the periodic table can fission, but most of them do so reluctantly. They only split sometimes, only when hit by a fast neutron, and when they do split, the neutrons they release are not quite fast enough to cause further fissions. In contrast, *fissile* isotopes are eager to split (and even have a tendency to spontaneously split). Fissile isotopes can split from both fast and slow neutrons (but slow ones work better), and the neutrons they release *can* cause further fissions, even in non-fissile isotopes. *Only* fissile isotopes can support a chain reaction. Bombs must be made from a nearly pure fissile isotope, and reactors must contain at least some fissile material. The fissile isotopes we need to care about are uranium 233, uranium 235, plutonium 239, and plutonium 241.
More tech note:
Neutrons can be thought of as an electron and a proton squished together. When a fissionable atom is hit by a neutron, three things can happen - it can absorb the neutron and carry on (possibly transforming one of its neutrons into a proton by giving off a beta particle, aka a fast electron, thus becoming a different element), it can spit two neutrons right back out, or it can fission. The absorbing one and spitting out two is called an (n, 2n) reaction, just to be difficult, and only applies to certain isotopes some of the time. Fissionable isotopes that, on absorbing a neutron, become fissile, are referred to as *fertile* isotopes. Thorium 232, U-238, and plutonium-240 are all fertile. Thorium gives off two beta particles in succession to become U-233 (two steps up the periodic table), and uranium does the same to become Pu-239. Plutonium 240 just absorbs the neutron without any fuss, becoming plutonium 241.
Yet more tech note:
Fissile isotopes are more likely to fission, but there's still a chance that they will absorb the neutron instead. For uranium 235, about 1 in 6 neutron impacts result in absorption rather than fission, producing U-326 (not fertile or fissile). For both plutonium 239 and 241, about 1 in 4 impacts are absorbed, producing pu-240 (fertile) and 242 (not fertile or fissile). Those isotopes can in turn absorb neutrons themselves, producing yet still heavier isotopes, some of which give off beta particles to become transuranic elements.
In short, reactors can produce modest amounts of all sorts of odball isotopes in addition to the main nuclear reactions you read about in high school physics. Some neutron absorption is good - over 1/3 the energy produced in a PWR reactor comes from fission of plutonium 239 and 241.
But Uranium-236 and plutonium-242, as well as other more exotic transuranic elements (neptunium, americium, curium, etc), just build up in the reactor core, have little use, and make the spent fuel significantly more radioactive than fresh fuel. You'll also see these waste elements referred to as actinides (after the series in the periodic table that includes thorium, uranium, and everything after them). Extracting useful exotic isotopes from the hodge podge of elements in spent fuel requires reprocessing the fuel - which a) involves technologies useful for extracting and purifying plutonium for bombs (thus heavily regulated/restricted), b) requires mucking around with highly radioactive materials, and c) is generally more difficult/expensive than producing just the desired isotope by inserting a precursor material into a reactor designed to bombard stuff with neutrons.
Coming back around to the question of those pesky fission products. In addition to a build up of neutron poisons, the accumulation of fission products in nuclear fuel is problematic for another reason - they are highly radioactive. About 7 percent of the heat from fission is released gradually by fission products as they decay. By the time you have to change out the fuel because of poison buildup, the fission products will be generating tons of heat, and the spent fuel will need to be kept cool for a few years while the worst of the residual radioactivity tapers off. Fortunately the fission products mostly have short half lives. After a few years, the spent fuel stops being too hot to handle, and after about 3 centuries, the radioactivity from fission products will be mostly gone. (as opposed to the radioactivity from plutonium and other actinide waste, which will continue for millennia).
What seldom gets mentioned is that in "spent" fuel from water cooled reactors, only about 5 to 10 percent of the original fuel has fissioned - the remaining 95 percent is still there, contaminated with actinide waste and fission products. In a few countries, including France, spent fuel gets reprocessed, with the waste separated out and the rest being turned into new fuel rods and returned to the reactor. But in the US and many other countries, there is no reprocessing of spent fuel - it's cheaper to just restock the reactor with new uranium and treat used the fuel rods as waste.
Some alternative reactor designs have a somewhat higher "burn" rate, but all of them suffer from this problem - the ash from fission eats neutrons, and that plays havoc with the reactor's ability to sustain a reaction long before you've used up the fuel. Spent fuel will always still contain plenty of perfectly good fissile and fertile material, which you must either throw away or reprocess.
In the 50's and early 60's, before environmental impacts were given any consideration, at a time when America's plutonium production facilities just buried nuclear waste in their backyards in barrels, it was assumed that nuclear power, in the long term, would involve reprocessing of spent fuel. Then several things happened:
- the promise ^H^H^H blatant lie that nuclear power would lead to electricity being "too cheap to meter" was proven wrong. The rate of new power plant construction slowed once the true cost of nuclear power in comparison to other power plants became apparent (60's).
- environmentalism became a thing, leading to government regulation of waste disposal (early 70's). A souring of the attitude towards all things nuclear - bombs, fallout, waste, etc, began.
- between new mines coming online and lowered demand due to the end of new reactor construction, the cost of uranium went down, while the cost of reprocessing (now with actual attention paid to safe waste disposal) went up. In other countries, governments subsidized the building of reprocessing facilities. In America, power companies were expected to pay the full cost, and said no thanks, treating their used fuel (still with 95% of the original fuel present) as waste. (not sure of the date range for this)
- power companies discovered that it was cheaper to encourage their customers to be more efficient in their use of electricity, than to build new power plants. The predicted rate of increase in electrical demand never materialized, and with it, the predicted demand for new nuclear power plants. (probably 70's and later? not sure)
- the three mile island accident turned people completely against nuclear power. New nuclear plant construction essentially ceased (late 70's).
- India made a bomb out of reprocessed fuel from one of their civilian reactors, prompting a ban in the US on reprocessing. Reprocessing became a weapons proliferation concern. Countries that did not want to join the nuclear club, and that had not already built reprocessing facilities, had a new reason to not build them (late 70's).
Tech/glossary note: As discussed above, plutonium-239 absorbs neutrons about 1/4 of the time, leading to formation of Pu-240, which has a high rate of spontaneous fission, giving off a high level of gamma and neutron radiation. Too much spontaneous fission does not lead to a lump of plutonium blowing up spontaneously, but it does mean when you try to detonate a bomb, the plutonium core blows apart before the chain reaction can build fully. Instead of making a big explosion, it "fizzles" with a much smaller explosion, which, from a bomb maker's perspective, is bad. Because the two isotopes differ by just a single neutron, instead of trying to purify plutonium-239 by removing the Pu-240, bomb makers prevent the formation Pu-240 as much as they can, and settle for "bomb grade" plutonium containing up to 7% Pu-240 (for bombs on planes and ICBMs) and the more expensive "supergrade" plutonium with no more than 3% (for bombs on submarines and ships, where the crew has to work in close proximity to the bombs, and thus low radiation is important for safety).
More tech: Removing heavier isotopes from the Pu-239 using the same enrichment process as uranium, while theoretically possible, would be significantly more difficult since the isotopes differ by a single neutron, aka less than 0.5% in weight, whereas uranium 235 is 1.2% lighter than U-238. Enrichment technology is not done on the pure elemental metal, but on uranium hexafluoride, which is a gas at modest temperatures, so the weight difference during enrichment is even less (plutonium hexafluoride is also a gas at modestly elevated temperatures).
Last tech: the enrichment process of choice in the US (and I think Europe?) throughout the 20th century was gaseous diffusion, which uses tons of electricity. The dramatically cheaper gas centrifuge alternative did not see wide use outside of the USSR (which kept it secret) and Pakistan (which developed it on their own in the 70's) until this century. By the time enrichment became cheaper and easier, the cold war had ended and bomb-having nations had largely stopped making plutonium.
Now on to other kinds of reactors.
***
Two variations on the PWR reactor exist.
PWR were originally designed to be small and compact, for submarines, and to confine their radiation inside the reactor's pressure system. Boiling water reactors (BWR) give up both those aspects in exchange for a much less complicated system. They allow the water inside the reactor to boil, and feed the steam directly into the turbines (no inner loop/outer loop). The turbines have to be shielded and access restricted while the reactor is running, but the gains are many: the reactor operates at half the pressure of a PWR, allowing the steel pressure vessel and plumbing to be much less robust and reducing the risk of a leak. BWRs operate at lower temperatures and are slightly less efficient than PWRs. But, allowing steam to form inside the reactor core means if the reactor overheats, more steam bubbles form, allowing neutrons to speed up, which in turn reduces the rate of fission, lowering the temperature - this feedback loop increases the safety of the system.
The other common variant are CANDU reactors. The hydrogen in ordinary water is a weak neutron poison. About 10% of the neutrons in a PWR reactor get absorbed by hydrogen, turning regular H-1 into H-2 (deuterium). CANDU reactors use heavy water (in which most of the hydrogen is already deuterium) instead of regular water, which increases their neutron supply by 10%. Those extra neutrons are enough of a boost to enable the reactor to operate with natural uranium instead of enriched uranium (the higher cost of heavy water is more than offset by the lower cost of natural uranium). Since uranium enrichment is a necessary first step to making bombs, being able to use natural uranium makes nuclear non-proliferation regulators happy. Alternatively, CANDU reactors can burn used fuel from PWR reactors (their higher neutron flux lets them use a few more percent of the fuel before the neutron poisoning of fission products gets too much).
***
Getting away from PWR reactors, fast neutron reactors eschew water completely. Usually they use liquid metal (lead or sodium) as coolant instead. Obviously this makes it impossible to use cameras to monitor the status of the reactor visually, requires you to drain out the coolant before performing maintenance on the reactor, and means you have to melt your coolant before you can pump it back into the reactor and turn it back on. Sodium is more commonly used than lead, which brings its own problems (unlike water, sodium is not inert - it smoulders on contact with air, and burns explosively on contact with water). It's harder to achieve a chain reaction with fast neutrons - you need at least 20% fissile material in your fuel, instead of 5% for a PWR. Fast neutron reactors are more expensive to build and operate than PWRs, which has hindered their adoption.
Fast neutron reactor advantages include: Fast reactors can burn about 20% of their fuel before needing to change it out, four times more than PWR reactors. Fast neutrons produce a completely different balance between neutron absorption and fission inside the reactor, which means that their spent fuel contains significantly less transuranic waste than a PWR's fuel. Fast reactors operate at a higher temperature without pressurization of the coolant, making them more efficient than PWRs, without any risk of a pressurized coolant leak. Fast reactors can be configured as "breeder" reactors - in which, thanks to Pu-239 production, the spent fuel contains more fissile material than you started with (obviously you have to reprocess the spent fuel to take advantage of this).
In theory, a fast neutron reactor can also be configured as an "actinide incinerator," burning up transuranic waste from spent fuel, transforming nuclear waste from a problem for many future millennia into a problem for the next few centuries.
Fast reactors are largely a solution to a problem (shortage of nuclear fuel) that no longer exists. Wikipedia lists 23 which have shut down, and eight with are currently operating or are temporarily down for repairs. Only two of those are power plants (both ex-soviet reactors), the rest are experimental. Four more are under construction, two for power production (in India and China) and two for research (both in Russia).
Fast reactors cost more than PWR, and while I have seen at least one glowing writeup about their future potential, I don't think it's happening.
***
So far, we've been talking about uranium fueled reactors. An alternative would be thorium reactors. Natural thorium (thorium 232) contains no fissile material. When Thorium absorbs a neutron, though, it gives off two beta particles to become U-233, which is fissile. So before you can start making thorium reactors, you must have a working reactor that can be used to irradiate your thorium fuel to enrich it with U-233.
U-233 is special because its ratio of absorption to fission is much more favorable than any other fissile material - instead of a 1 in 6 chance of absorbing a neutron (like uranium-235) or a 1 in 4 chance (like plutonium), it has a 1 in 12 chance. The rest of the time, it fissions. The few U-233 atoms that do absorb neutrons become U-234, which on capturing another neutron in turn becomes fissile U-235. Since a thorium reactor contains no U-238, the only way the reactor can produce transuranic waste is by multiple neutron captures. Thus, spent fuel from a thorium reactor contains dramatically less long term actinide waste than any kind of uranium reactor, fast or slow.
Thorium can be used in either a slow (PWR) or fast neutron reactor. Uniquely, it's possible to run a PWR thorium reactor as a breeder reactor (where the spent fuel contains more U-233 than you started with). Uranium based breeding is possible only with a fast neutron reactor.
Another frequently cited plus turns out to be bogus: U-233 is often said to be unsuited to bomb making due to those pesky (n, 2n) reactions producing small amounts of U-232, which is also fissile, but which gives off gamma radiation as it decays. The plutonium for bombs can be safely handled in a glove box with minimal shielding, and (it is alleged) this is not possible with uranium-233 due to U-232 contamination. For safe handling, the amount of u-232 contamination in U-233 needs to be less than 50 parts per million, compared to 30,000-70,000ppm Pu-240 contamination being acceptable in bomb grade plutonium.
The problem being, the US produced two tons of bomb grade U-233 (with suitably low levels of U-232 contamination) during the cold war. It's thought that at least some US and Soviet nuclear tests used U-233 or a combination of plutonium and U-233 in their design. India (which is all-in on a thorium based nuclear industry) has tested one U-233 bomb (out of seven or so total tests).
Furthermore, in 1966, a one page internal memo (document DUN-677) written by an employee of Douglas United Nuclear, Inc. (a contractor for the Hanford plutonium production site in Washington state) summarized a conversation between the author and researchers at the Lawrence Radiation Laboratory about U-233. The LRL team felt that U-233 was "highly satisfactory as a weapons material." It was, however, not any better than plutonium - there was no reason to switch away from plutonium for bombs. On the other hand, if they had been using U-233 for bombs all along, there would be no reason to switch to plutonium either - the two were equal. The LRL people said they wanted to have a modest annual supply of U-233 for experimentation, but that was all.
Wikipedia, being what it is, has a paragraph in the article on U-233 talking about how thorium is good because it is non-proliferative due to the U-232 thing, followed by a paragraph summarizing the DUN-677 memo along with the nuclear tests involving U-233.
There are various technical downsides to thorium reactors, but the main issue (other than the need to create U-233 with which to enrich thorium) is that the founder effect has geared all nuclear infrastructure outside of India to use uranium, and thorium adoption has had to struggle against that.
There have been numerous thorium reactors over the years, but most of them have shut down - India currently has about ten PWR thorium reactors in operation (India's nuclear roadmap is to continue to build more as they can be designed and built entirely by Indian companies and fueled with Indian thorium), but outside of India, there are none.
***
All the reactor types discussed above have multiple examples existing today producing electricity at commercial scale. Molten salt reactors (MSRs) exist only on paper and in three experimental reactors, two of them from the mid-20th century. The US experimented with one of these in the 50's as part of their nuclear powered supersonic bomber program (a bondoogle which never went anywhere), and then the same facility (Oak Ridge National Laboratory) built another one a decade later, specifically to gather more data on this type of reactor. In this century, there is currently a scaled down prototype MSR reactor operating in China (started operating in 2023). There are plans to begin building bigger versions for power production in the next decade.
MSRs are nowhere near a production ready technology, which might be why the hucksters touting them claim that they are the solution to every possible problem with nuclear power.
In a normal reactor, the fuel exists as pellets of uranium oxide, thorium oxide, and/or plutonium oxide, which are in turn loaded into fuel assemblies (usually metal rods) that hold the pellets in place and keep the fuel separated from the coolant. A meltdown, in which the fuel and/or the rods get hot enough to liquefy, would be a Very Bad Thing indeed, destroying most of the reactor's control mechanisms and spewing radioactive fuel into the coolant, or worse, melting through the reactor vessel and spewing it all over the place.
As the name implies, a molten salt reactor would rely first, on combining the reactor fuel with halogens to create salts - fluorine salts for a slow neutron reactor (fluorine is a neutron moderator) and chlorine salts for a fast neutron reactor. The two 20th century experimental reactors mixed UF4 with LiF and BeF2 - lithium and beryllium are also neutron moderators, and the mixture had a lower melting point than pure UF4. The boiling point of the salt mixture was much higher than the temperature at which the reactors operated, allowing them to run unpressurized.
The two experimental reactors did not attempt to do anything useful with the heat they generated (they were small enough to be cooled by blowers). Designs for MSR power plants call for either piping water/some other thing to spin turbines directly through the reactor, or using an inner loop of liquid metal/some non-radioactive salt to pull heat out of the reactor and an outer loop to put that heat into the turbines. Because all of this is on paper, discussions of MSR power plants often veer off into using helium or nitrogen gas to spin turbines for super high efficiency ratings, despite gas based electrical turbines being itself a not-entirely-there-yet, immature technology.
In an overheating situation, or simply when the reactor needs to be shutdown, a valve would allow the liquid fuel to drain out of the reactor into multiple containers, each too small to sustain a chain reaction, instantly shutting the reactor down. Also, heat affects the rate of nuclear fission. The hotter the fuel gets, the lower the rate of fission due to thermal expansion. At the same time, the hotter the non-fuel parts of a reactor get, the more avidly those elements absorb neutrons, also lowering the rate of fission. Having the fuel be liquid (which can expand and contract more than a solid) would make the slowing effect of the reactor heating up more pronounced.
Since the fuel is liquid in an MSR, fission products that are gaseous (like Xenon-135, one of the worst neutron poisons) can bubble out of the fuel, thus removing their effect from the reactor, which would enable a higher burn percentage. I found a claim that 50% fuel utilization could be achieved, but all of this is speculative until someone actually builds an MSR reactor and sees how long they can run it before the fuel becomes too poisoned (neither of the 20th century experimental reactors operated long enough to find out).
The other thing that liquid fuel enables, in theory, is on-site reprocessing, separating out the chemically distinct fission products from the fuel as the reactor runs, adding new fuel as need be, removing plutonium from the reactor as it is bred, etc. Obviously this is a huge problem for proliferation concerns, one which the glowing writeups about the glorious future of MSRs completely ignore, focusing instead on the amazingly high fuel utilization levels that such reprocessing enables.
The 1950's aircraft reactor experiment only ran for a few days. The 60's MSR experiment ran, on and off, for four years, first with uranium and then with thorium, but it spent much of that time turned off due to problem after problem. Some of those problems were due to poor, slapdash reactor design and cavalier operation (ie, the cooling system they built was inadequate and even after an upgrade, they could not run the reactor at the designed power level. Unsurprisingly, after the experiment ended, they did not bother unloading or decommissioning the reactor, leading to many expensive issues decades later when it began to leak fluorine gas and radiation from its location in a basement into the building above) (both reactor experiments happened in the era of burying nuclear waste in barrels in the backyard, remember).
Other problems were more fundamental, mostly revolving around embrittlement. The reactor vessel and its piping was built with a custom designed corrosion resistant alloy, which was immune from highly corrosive fluoride salts in the lab, but in the reactor, tellurium (a fission product) bonded to the metal and made it extremely brittle. Time after time, the reactor experienced leaks from one place or where the corrosion resistant alloy cracked. The wikipedia article on MSR claims that the embrittlement problem has since been solved, but again, that's on paper, not ever tested in an actual running reactor.
Someday, there may be a MSR industry and we will be able to know just how much of the benefits they present are real instead of snake oil, but that remains in the future (the news of the Chinese MSR includes that they achieved ten days of operating at full design power in October 2024, this is definitely not very long at all).
***
Having read far too much and gone down far too many rabbit holes, I think I can say confidently that 90% of the claims in articles touting the bright future of new! improved! no longer dangerous or scary! nuclear power are hogwash. Things those articles tout nearly always involve making proliferation-enabling technologies routine (no one other than members of the Church of Nukes want this), assume that technologies still on the drawing board will work out as advertised (they never do), and/or gloss over many, many hard to solve problems. Meanwhile, right now, we already have the ability to just use renewables paired with energy storage. We soon won't need nuclear power anymore, yet somehow there are still scads of acolytes of the Nuclear Church who refuse to accept that their God has become irrelevant.
And these kinds of stories come from a mix of hucksterism and religion. Just as we have a religion of space enthusiasm (where solar power stations or helium 3 or whatever bullshit technology becomes the rallying cry for the religion's real goal, of having a city on the moon/Mars/in orbit, and having people permanently *Living in Space*), and the religion helps sustain huckster snake oil plutocrats who don't really care about space at all except as a way of siphoning off government subsidies and pumping up stock prices for their space technology companies -- so too we have a religion of Nuclear Power (it's the Future!), and nuclear power technology companies whose plutocrats have snake oil to sell. (PS: my data-free impression is that space enthusiasts vastly outnumber space hucksters, but for nuclear power, the ratio is much more even or perhaps reversed).
Some nuclear power advocates have at least half a leg to stand on (solar and wind are not 24/7/365 power sources, and nuclear power plants *are* a carbon-free way to provide round the clock power regardless of weather). But solar has become SO much cheaper than fossil fuel, let alone nuclear, that it leaves budgetary headroom for adding some kind of power storage to a solar farm and still being less expensive than the alternatives - and solar powered storage (like a lake that you pump full during the day and drain through hydroelectric generators at night) completely avoids all the regulatory and PR hassles of nuclear power.
Other nuclear advocates seem to Want to Believe in nuclear because they are right wing and regard solar as tainted by the leftist eco green conspiracy, or something? IDK.
But after encountering another "the new generation of (insert technical descriptor) nuclear power plants will completely avoid all the problems you've come to expect from nuclear power" article, I decided to try and figure out just how much truth there is to such articles. After picking away at the question for a while, I'm finally typing everything up so my time will not have been completely wasted. The rest of this post comes from reading far, far too many web pages (mostly on wikipedia but also elsewhere) devoted to nuclear power.
***
Nearly all US-based nuclear power and the vast majority of the world's nuclear power comes from uranium fueled, water-cooled, slow neutron reactors. That's because of two tiers of technical lock in.
First, from 1942 through the mid-50's, reactors were tools for transforming uranium-238 into plutonium 239, for bombs. Fission always produces fast neutrons, but slow neutrons are best for transmuting U-238 into plutonium. Water is a neutron moderator (it slows them down), so it made sense to submerge the reactor core in water, keeping the reactor cool and slowing down the neutrons produced at the same time. To minimize the formation of unwanted plutonium isotopes (plutonium 238, plutonium 240 and 241 are all undesirable in bombs - from some combination of emitting dangerous gamma radiation, emitting tons of heat, and/or being eager to spontaneously fission, causing premature detonation), reactors for making bomb plutonium would change out their fuel frequently, and maintain a rate of reaction to maximize production of the right kind of plutonium, with no regard for making use of the heat the reactor produced. While lots of experimental reactors were built and tested during that decade, including various fast neutron types, the focus was on producing lots of plutonium for the cold war arms buildup. By the mid 50's, the world's reactor engineers and technicians had lots of experience with water cooled uranium reactors, and little experience with other kinds.
Second, when, in the early 50's, Admiral Rickover started to work on his dream of a nuclear powered submarine (a sub that could stay underwater indefinitely, drastically transforming its strategic value), the teams he directed drew on the engineering expertise that existed, which meant they designed a water cooled uranium reactor. They made other tweaks to keep the reactor small and compact without reducing its energy output too much, and to minimize the volume that had to have radiation shielding. The result was a pressurized water-cooled reactor, PWR. The reactor core is bathed in an inner loop of water kept under pressure to keep it from boiling. That water, in turn, heats an outer loop of water into steam, which spins turbines to produce power to run the submarine. The water in the inner loop has three jobs: it transfers heat from the reactor to the outer loop, it slows the neutrons in the reactor, and calibrated amounts of neutron absorbing substances (boron in the form of boric acid, for example) are dissolved in it to keep the rate of fission constant. Graphite control rods can be inserted into the reactor to shut it down, but while it's running, boric acid is used to keep the neutron flux stable.
Tech/glossary note: anything that absorbs neutrons is called a *neutron poison*. As a reactor runs, fission products build up in the fuel (and decay, since fission products are all highly radioactive - more about this later). Some of the longer-lived fission products are neutron poisons, so as a reactor burns through its fuel supply, the rate of fission will go down as more and more neutrons are absorbed by the fission products. To keep the rate of fission steady, you need to have some poison present with new fuel, which you reduce gradually as the percentage of poisoning fission products in the fuel goes up, to keep the total energy output even. Having some boric acid in the coolant water, and diluting it over time by adding pure water, is one approach.
By the mid-50's, as part of Eisenhower's "atoms for peace" initiative, the first power reactors started being built... and they used the designs that already existed for powering naval vessels, and hired ex-navy nuclear techs to operate them. The first commercial atomic power plant in the US, Shippingport Atomic Power Station (commissioned 1958), started out using a reactor core originally built to power a (never built) aircraft carrier.
And that technical lock-in, of engineers who knew how to build water cooled uranium reactors and of naval technicians who knew how to run them, has resulted in a nuclear power industry that is overwhelmingly composed of PWR reactors, from the 50's through to today.
The good about PWRs: they're the nearest thing to a mature technology in the reactor industry. They can operate with only mildly enriched uranium (2-5% U-235), or (in the case of the Canada's CANDU reactors) natural, unenriched uranium. And, water is extremely well understood. It's cheap, stable, non-reactive, remains liquid at room temperature, and is transparent (allowing inspection of the reactor core via cameras).
But PWR reactors have numerous disadvantages.
First: problems with using water. It has a very low boiling point. Steam is not dense enough to slow neutrons, so you have to keep the water in the reactor from boiling, which requires keeping it under pressure. That in turn requires encasing the reactor in a thick steel pressure vessel, having your piping be heavy duty thick steel, and having pumps and other gear to maintain high pressure. All of that adds substantially to the cost. Plus, the neutrons which escape from the reactor core are absorbed by the steel pressure vessel, altering its composition and making it more brittle over time ("neutron embrittlement"), which shortens its life (in practical terms, there's no replacing a pressure vessel that's reached end of life - once the steel has degraded too much, the entire reactor has to be scrapped and replaced with a new one). Water is corrosive to steel, especially with boric acid added to it. The containment vessel and pipes need to be made of stainless steel (expensive) and even then, corrosion works to shorten the life of containment vessel and piping.
The water inside the pressure vessel is contaminated. If there's ever a failure in some of the reactor's plumbing, radioactive steam will leak out. The many safety precautions needed to prevent ruptures in the pipes and to contain steam leaks should they occur add even more to the reactor's cost.
The hotter you make the steam in the outer loop of a reactor, the more efficiently you can produce electricity. But there's a sharp limit to how hot you can let the inner loop of the reactor get before the pressure required to keep the water from becoming steam becomes impossibly difficult to maintain. Thus PWR reactors are less efficient than reactors that use a coolant with a higher boiling point. (ideally, a coolant with a boiling point higher than the temperature at which the reactor normally gets, would let you have an unpressurized coolant system, vastly reducing cost, complexity, and removing one major source of potential accidents).
Second: problems with neutron absorption. Water moderated uranium reactors produce plutonium and other transuranic isotopes, which remain in the spent fuel and are a major source of long term radioactive waste.
Tech/glossary note:
All elements after radium in the periodic table can fission, but most of them do so reluctantly. They only split sometimes, only when hit by a fast neutron, and when they do split, the neutrons they release are not quite fast enough to cause further fissions. In contrast, *fissile* isotopes are eager to split (and even have a tendency to spontaneously split). Fissile isotopes can split from both fast and slow neutrons (but slow ones work better), and the neutrons they release *can* cause further fissions, even in non-fissile isotopes. *Only* fissile isotopes can support a chain reaction. Bombs must be made from a nearly pure fissile isotope, and reactors must contain at least some fissile material. The fissile isotopes we need to care about are uranium 233, uranium 235, plutonium 239, and plutonium 241.
More tech note:
Neutrons can be thought of as an electron and a proton squished together. When a fissionable atom is hit by a neutron, three things can happen - it can absorb the neutron and carry on (possibly transforming one of its neutrons into a proton by giving off a beta particle, aka a fast electron, thus becoming a different element), it can spit two neutrons right back out, or it can fission. The absorbing one and spitting out two is called an (n, 2n) reaction, just to be difficult, and only applies to certain isotopes some of the time. Fissionable isotopes that, on absorbing a neutron, become fissile, are referred to as *fertile* isotopes. Thorium 232, U-238, and plutonium-240 are all fertile. Thorium gives off two beta particles in succession to become U-233 (two steps up the periodic table), and uranium does the same to become Pu-239. Plutonium 240 just absorbs the neutron without any fuss, becoming plutonium 241.
Yet more tech note:
Fissile isotopes are more likely to fission, but there's still a chance that they will absorb the neutron instead. For uranium 235, about 1 in 6 neutron impacts result in absorption rather than fission, producing U-326 (not fertile or fissile). For both plutonium 239 and 241, about 1 in 4 impacts are absorbed, producing pu-240 (fertile) and 242 (not fertile or fissile). Those isotopes can in turn absorb neutrons themselves, producing yet still heavier isotopes, some of which give off beta particles to become transuranic elements.
In short, reactors can produce modest amounts of all sorts of odball isotopes in addition to the main nuclear reactions you read about in high school physics. Some neutron absorption is good - over 1/3 the energy produced in a PWR reactor comes from fission of plutonium 239 and 241.
But Uranium-236 and plutonium-242, as well as other more exotic transuranic elements (neptunium, americium, curium, etc), just build up in the reactor core, have little use, and make the spent fuel significantly more radioactive than fresh fuel. You'll also see these waste elements referred to as actinides (after the series in the periodic table that includes thorium, uranium, and everything after them). Extracting useful exotic isotopes from the hodge podge of elements in spent fuel requires reprocessing the fuel - which a) involves technologies useful for extracting and purifying plutonium for bombs (thus heavily regulated/restricted), b) requires mucking around with highly radioactive materials, and c) is generally more difficult/expensive than producing just the desired isotope by inserting a precursor material into a reactor designed to bombard stuff with neutrons.
Coming back around to the question of those pesky fission products. In addition to a build up of neutron poisons, the accumulation of fission products in nuclear fuel is problematic for another reason - they are highly radioactive. About 7 percent of the heat from fission is released gradually by fission products as they decay. By the time you have to change out the fuel because of poison buildup, the fission products will be generating tons of heat, and the spent fuel will need to be kept cool for a few years while the worst of the residual radioactivity tapers off. Fortunately the fission products mostly have short half lives. After a few years, the spent fuel stops being too hot to handle, and after about 3 centuries, the radioactivity from fission products will be mostly gone. (as opposed to the radioactivity from plutonium and other actinide waste, which will continue for millennia).
What seldom gets mentioned is that in "spent" fuel from water cooled reactors, only about 5 to 10 percent of the original fuel has fissioned - the remaining 95 percent is still there, contaminated with actinide waste and fission products. In a few countries, including France, spent fuel gets reprocessed, with the waste separated out and the rest being turned into new fuel rods and returned to the reactor. But in the US and many other countries, there is no reprocessing of spent fuel - it's cheaper to just restock the reactor with new uranium and treat used the fuel rods as waste.
Some alternative reactor designs have a somewhat higher "burn" rate, but all of them suffer from this problem - the ash from fission eats neutrons, and that plays havoc with the reactor's ability to sustain a reaction long before you've used up the fuel. Spent fuel will always still contain plenty of perfectly good fissile and fertile material, which you must either throw away or reprocess.
In the 50's and early 60's, before environmental impacts were given any consideration, at a time when America's plutonium production facilities just buried nuclear waste in their backyards in barrels, it was assumed that nuclear power, in the long term, would involve reprocessing of spent fuel. Then several things happened:
- the promise ^H^H^H blatant lie that nuclear power would lead to electricity being "too cheap to meter" was proven wrong. The rate of new power plant construction slowed once the true cost of nuclear power in comparison to other power plants became apparent (60's).
- environmentalism became a thing, leading to government regulation of waste disposal (early 70's). A souring of the attitude towards all things nuclear - bombs, fallout, waste, etc, began.
- between new mines coming online and lowered demand due to the end of new reactor construction, the cost of uranium went down, while the cost of reprocessing (now with actual attention paid to safe waste disposal) went up. In other countries, governments subsidized the building of reprocessing facilities. In America, power companies were expected to pay the full cost, and said no thanks, treating their used fuel (still with 95% of the original fuel present) as waste. (not sure of the date range for this)
- power companies discovered that it was cheaper to encourage their customers to be more efficient in their use of electricity, than to build new power plants. The predicted rate of increase in electrical demand never materialized, and with it, the predicted demand for new nuclear power plants. (probably 70's and later? not sure)
- the three mile island accident turned people completely against nuclear power. New nuclear plant construction essentially ceased (late 70's).
- India made a bomb out of reprocessed fuel from one of their civilian reactors, prompting a ban in the US on reprocessing. Reprocessing became a weapons proliferation concern. Countries that did not want to join the nuclear club, and that had not already built reprocessing facilities, had a new reason to not build them (late 70's).
Tech/glossary note: As discussed above, plutonium-239 absorbs neutrons about 1/4 of the time, leading to formation of Pu-240, which has a high rate of spontaneous fission, giving off a high level of gamma and neutron radiation. Too much spontaneous fission does not lead to a lump of plutonium blowing up spontaneously, but it does mean when you try to detonate a bomb, the plutonium core blows apart before the chain reaction can build fully. Instead of making a big explosion, it "fizzles" with a much smaller explosion, which, from a bomb maker's perspective, is bad. Because the two isotopes differ by just a single neutron, instead of trying to purify plutonium-239 by removing the Pu-240, bomb makers prevent the formation Pu-240 as much as they can, and settle for "bomb grade" plutonium containing up to 7% Pu-240 (for bombs on planes and ICBMs) and the more expensive "supergrade" plutonium with no more than 3% (for bombs on submarines and ships, where the crew has to work in close proximity to the bombs, and thus low radiation is important for safety).
More tech: Removing heavier isotopes from the Pu-239 using the same enrichment process as uranium, while theoretically possible, would be significantly more difficult since the isotopes differ by a single neutron, aka less than 0.5% in weight, whereas uranium 235 is 1.2% lighter than U-238. Enrichment technology is not done on the pure elemental metal, but on uranium hexafluoride, which is a gas at modest temperatures, so the weight difference during enrichment is even less (plutonium hexafluoride is also a gas at modestly elevated temperatures).
Last tech: the enrichment process of choice in the US (and I think Europe?) throughout the 20th century was gaseous diffusion, which uses tons of electricity. The dramatically cheaper gas centrifuge alternative did not see wide use outside of the USSR (which kept it secret) and Pakistan (which developed it on their own in the 70's) until this century. By the time enrichment became cheaper and easier, the cold war had ended and bomb-having nations had largely stopped making plutonium.
Now on to other kinds of reactors.
***
Two variations on the PWR reactor exist.
PWR were originally designed to be small and compact, for submarines, and to confine their radiation inside the reactor's pressure system. Boiling water reactors (BWR) give up both those aspects in exchange for a much less complicated system. They allow the water inside the reactor to boil, and feed the steam directly into the turbines (no inner loop/outer loop). The turbines have to be shielded and access restricted while the reactor is running, but the gains are many: the reactor operates at half the pressure of a PWR, allowing the steel pressure vessel and plumbing to be much less robust and reducing the risk of a leak. BWRs operate at lower temperatures and are slightly less efficient than PWRs. But, allowing steam to form inside the reactor core means if the reactor overheats, more steam bubbles form, allowing neutrons to speed up, which in turn reduces the rate of fission, lowering the temperature - this feedback loop increases the safety of the system.
The other common variant are CANDU reactors. The hydrogen in ordinary water is a weak neutron poison. About 10% of the neutrons in a PWR reactor get absorbed by hydrogen, turning regular H-1 into H-2 (deuterium). CANDU reactors use heavy water (in which most of the hydrogen is already deuterium) instead of regular water, which increases their neutron supply by 10%. Those extra neutrons are enough of a boost to enable the reactor to operate with natural uranium instead of enriched uranium (the higher cost of heavy water is more than offset by the lower cost of natural uranium). Since uranium enrichment is a necessary first step to making bombs, being able to use natural uranium makes nuclear non-proliferation regulators happy. Alternatively, CANDU reactors can burn used fuel from PWR reactors (their higher neutron flux lets them use a few more percent of the fuel before the neutron poisoning of fission products gets too much).
***
Getting away from PWR reactors, fast neutron reactors eschew water completely. Usually they use liquid metal (lead or sodium) as coolant instead. Obviously this makes it impossible to use cameras to monitor the status of the reactor visually, requires you to drain out the coolant before performing maintenance on the reactor, and means you have to melt your coolant before you can pump it back into the reactor and turn it back on. Sodium is more commonly used than lead, which brings its own problems (unlike water, sodium is not inert - it smoulders on contact with air, and burns explosively on contact with water). It's harder to achieve a chain reaction with fast neutrons - you need at least 20% fissile material in your fuel, instead of 5% for a PWR. Fast neutron reactors are more expensive to build and operate than PWRs, which has hindered their adoption.
Fast neutron reactor advantages include: Fast reactors can burn about 20% of their fuel before needing to change it out, four times more than PWR reactors. Fast neutrons produce a completely different balance between neutron absorption and fission inside the reactor, which means that their spent fuel contains significantly less transuranic waste than a PWR's fuel. Fast reactors operate at a higher temperature without pressurization of the coolant, making them more efficient than PWRs, without any risk of a pressurized coolant leak. Fast reactors can be configured as "breeder" reactors - in which, thanks to Pu-239 production, the spent fuel contains more fissile material than you started with (obviously you have to reprocess the spent fuel to take advantage of this).
In theory, a fast neutron reactor can also be configured as an "actinide incinerator," burning up transuranic waste from spent fuel, transforming nuclear waste from a problem for many future millennia into a problem for the next few centuries.
Fast reactors are largely a solution to a problem (shortage of nuclear fuel) that no longer exists. Wikipedia lists 23 which have shut down, and eight with are currently operating or are temporarily down for repairs. Only two of those are power plants (both ex-soviet reactors), the rest are experimental. Four more are under construction, two for power production (in India and China) and two for research (both in Russia).
Fast reactors cost more than PWR, and while I have seen at least one glowing writeup about their future potential, I don't think it's happening.
***
So far, we've been talking about uranium fueled reactors. An alternative would be thorium reactors. Natural thorium (thorium 232) contains no fissile material. When Thorium absorbs a neutron, though, it gives off two beta particles to become U-233, which is fissile. So before you can start making thorium reactors, you must have a working reactor that can be used to irradiate your thorium fuel to enrich it with U-233.
U-233 is special because its ratio of absorption to fission is much more favorable than any other fissile material - instead of a 1 in 6 chance of absorbing a neutron (like uranium-235) or a 1 in 4 chance (like plutonium), it has a 1 in 12 chance. The rest of the time, it fissions. The few U-233 atoms that do absorb neutrons become U-234, which on capturing another neutron in turn becomes fissile U-235. Since a thorium reactor contains no U-238, the only way the reactor can produce transuranic waste is by multiple neutron captures. Thus, spent fuel from a thorium reactor contains dramatically less long term actinide waste than any kind of uranium reactor, fast or slow.
Thorium can be used in either a slow (PWR) or fast neutron reactor. Uniquely, it's possible to run a PWR thorium reactor as a breeder reactor (where the spent fuel contains more U-233 than you started with). Uranium based breeding is possible only with a fast neutron reactor.
Another frequently cited plus turns out to be bogus: U-233 is often said to be unsuited to bomb making due to those pesky (n, 2n) reactions producing small amounts of U-232, which is also fissile, but which gives off gamma radiation as it decays. The plutonium for bombs can be safely handled in a glove box with minimal shielding, and (it is alleged) this is not possible with uranium-233 due to U-232 contamination. For safe handling, the amount of u-232 contamination in U-233 needs to be less than 50 parts per million, compared to 30,000-70,000ppm Pu-240 contamination being acceptable in bomb grade plutonium.
The problem being, the US produced two tons of bomb grade U-233 (with suitably low levels of U-232 contamination) during the cold war. It's thought that at least some US and Soviet nuclear tests used U-233 or a combination of plutonium and U-233 in their design. India (which is all-in on a thorium based nuclear industry) has tested one U-233 bomb (out of seven or so total tests).
Furthermore, in 1966, a one page internal memo (document DUN-677) written by an employee of Douglas United Nuclear, Inc. (a contractor for the Hanford plutonium production site in Washington state) summarized a conversation between the author and researchers at the Lawrence Radiation Laboratory about U-233. The LRL team felt that U-233 was "highly satisfactory as a weapons material." It was, however, not any better than plutonium - there was no reason to switch away from plutonium for bombs. On the other hand, if they had been using U-233 for bombs all along, there would be no reason to switch to plutonium either - the two were equal. The LRL people said they wanted to have a modest annual supply of U-233 for experimentation, but that was all.
Wikipedia, being what it is, has a paragraph in the article on U-233 talking about how thorium is good because it is non-proliferative due to the U-232 thing, followed by a paragraph summarizing the DUN-677 memo along with the nuclear tests involving U-233.
There are various technical downsides to thorium reactors, but the main issue (other than the need to create U-233 with which to enrich thorium) is that the founder effect has geared all nuclear infrastructure outside of India to use uranium, and thorium adoption has had to struggle against that.
There have been numerous thorium reactors over the years, but most of them have shut down - India currently has about ten PWR thorium reactors in operation (India's nuclear roadmap is to continue to build more as they can be designed and built entirely by Indian companies and fueled with Indian thorium), but outside of India, there are none.
***
All the reactor types discussed above have multiple examples existing today producing electricity at commercial scale. Molten salt reactors (MSRs) exist only on paper and in three experimental reactors, two of them from the mid-20th century. The US experimented with one of these in the 50's as part of their nuclear powered supersonic bomber program (a bondoogle which never went anywhere), and then the same facility (Oak Ridge National Laboratory) built another one a decade later, specifically to gather more data on this type of reactor. In this century, there is currently a scaled down prototype MSR reactor operating in China (started operating in 2023). There are plans to begin building bigger versions for power production in the next decade.
MSRs are nowhere near a production ready technology, which might be why the hucksters touting them claim that they are the solution to every possible problem with nuclear power.
In a normal reactor, the fuel exists as pellets of uranium oxide, thorium oxide, and/or plutonium oxide, which are in turn loaded into fuel assemblies (usually metal rods) that hold the pellets in place and keep the fuel separated from the coolant. A meltdown, in which the fuel and/or the rods get hot enough to liquefy, would be a Very Bad Thing indeed, destroying most of the reactor's control mechanisms and spewing radioactive fuel into the coolant, or worse, melting through the reactor vessel and spewing it all over the place.
As the name implies, a molten salt reactor would rely first, on combining the reactor fuel with halogens to create salts - fluorine salts for a slow neutron reactor (fluorine is a neutron moderator) and chlorine salts for a fast neutron reactor. The two 20th century experimental reactors mixed UF4 with LiF and BeF2 - lithium and beryllium are also neutron moderators, and the mixture had a lower melting point than pure UF4. The boiling point of the salt mixture was much higher than the temperature at which the reactors operated, allowing them to run unpressurized.
The two experimental reactors did not attempt to do anything useful with the heat they generated (they were small enough to be cooled by blowers). Designs for MSR power plants call for either piping water/some other thing to spin turbines directly through the reactor, or using an inner loop of liquid metal/some non-radioactive salt to pull heat out of the reactor and an outer loop to put that heat into the turbines. Because all of this is on paper, discussions of MSR power plants often veer off into using helium or nitrogen gas to spin turbines for super high efficiency ratings, despite gas based electrical turbines being itself a not-entirely-there-yet, immature technology.
In an overheating situation, or simply when the reactor needs to be shutdown, a valve would allow the liquid fuel to drain out of the reactor into multiple containers, each too small to sustain a chain reaction, instantly shutting the reactor down. Also, heat affects the rate of nuclear fission. The hotter the fuel gets, the lower the rate of fission due to thermal expansion. At the same time, the hotter the non-fuel parts of a reactor get, the more avidly those elements absorb neutrons, also lowering the rate of fission. Having the fuel be liquid (which can expand and contract more than a solid) would make the slowing effect of the reactor heating up more pronounced.
Since the fuel is liquid in an MSR, fission products that are gaseous (like Xenon-135, one of the worst neutron poisons) can bubble out of the fuel, thus removing their effect from the reactor, which would enable a higher burn percentage. I found a claim that 50% fuel utilization could be achieved, but all of this is speculative until someone actually builds an MSR reactor and sees how long they can run it before the fuel becomes too poisoned (neither of the 20th century experimental reactors operated long enough to find out).
The other thing that liquid fuel enables, in theory, is on-site reprocessing, separating out the chemically distinct fission products from the fuel as the reactor runs, adding new fuel as need be, removing plutonium from the reactor as it is bred, etc. Obviously this is a huge problem for proliferation concerns, one which the glowing writeups about the glorious future of MSRs completely ignore, focusing instead on the amazingly high fuel utilization levels that such reprocessing enables.
The 1950's aircraft reactor experiment only ran for a few days. The 60's MSR experiment ran, on and off, for four years, first with uranium and then with thorium, but it spent much of that time turned off due to problem after problem. Some of those problems were due to poor, slapdash reactor design and cavalier operation (ie, the cooling system they built was inadequate and even after an upgrade, they could not run the reactor at the designed power level. Unsurprisingly, after the experiment ended, they did not bother unloading or decommissioning the reactor, leading to many expensive issues decades later when it began to leak fluorine gas and radiation from its location in a basement into the building above) (both reactor experiments happened in the era of burying nuclear waste in barrels in the backyard, remember).
Other problems were more fundamental, mostly revolving around embrittlement. The reactor vessel and its piping was built with a custom designed corrosion resistant alloy, which was immune from highly corrosive fluoride salts in the lab, but in the reactor, tellurium (a fission product) bonded to the metal and made it extremely brittle. Time after time, the reactor experienced leaks from one place or where the corrosion resistant alloy cracked. The wikipedia article on MSR claims that the embrittlement problem has since been solved, but again, that's on paper, not ever tested in an actual running reactor.
Someday, there may be a MSR industry and we will be able to know just how much of the benefits they present are real instead of snake oil, but that remains in the future (the news of the Chinese MSR includes that they achieved ten days of operating at full design power in October 2024, this is definitely not very long at all).
***
Having read far too much and gone down far too many rabbit holes, I think I can say confidently that 90% of the claims in articles touting the bright future of new! improved! no longer dangerous or scary! nuclear power are hogwash. Things those articles tout nearly always involve making proliferation-enabling technologies routine (no one other than members of the Church of Nukes want this), assume that technologies still on the drawing board will work out as advertised (they never do), and/or gloss over many, many hard to solve problems. Meanwhile, right now, we already have the ability to just use renewables paired with energy storage. We soon won't need nuclear power anymore, yet somehow there are still scads of acolytes of the Nuclear Church who refuse to accept that their God has become irrelevant.
(no subject)
Date: 2025-05-01 12:00 am (UTC)