IFR FaD Nuclear

IFR FaD 10 – metal fuel and plutonium

Over the next month or two, I will publish four extracts from the book Plentiful Energy — The story of the Integral Fast Reactor by Chuck Till and Yoon Chang.

Reproduced with permission of the authors, these sections describe and justify some of the key design choices that went into the making the IFR a different — and highly successful — approach to fast neutron reactor technology and its associated fuel recycling.

These excerpts not only provide a fascinating insight into a truly sustainable form nuclear power; they also provide excellent reference material for refuting many of the spurious claims on the internet about IFR by people who don’t understand (or choose to wilfully misrepresent) this critically important technology.

The first extract, on Fuel Choice, comes from pages 104-108 of Plentiful Energy. To buy the book ($18 US) and get the full story, go to Amazon or CreateSpace.


Metal Fuel

The IFR metal alloy fuel was the single most important development decision. More flows from this than from any other of the choices. It was a controversial choice, as metal fuel had been discarded worldwide in the early sixties and forgotten. Long irradiation times in the reactor are essential, particularly if reprocessing of the fuel is expensive, yet the metal fuel of the 1960s would not withstand any more than moderate irradiation. Ceramic fuel, on the other hand, would. Oxide, a ceramic fuel developed for commercial water-cooled reactors, had been adopted for breeder reactors in every breeder program in the world. It is fully developed and it remains today the de facto reference fuel type for fast reactors elsewhere in the world. It is known. Its advantages and disadvantages in a sodium-cooled fast reactor are well established. Why then was metallic fuel the choice for the IFR?

The Integral Fast Reactor (IFR) system

In reactor operation, reactor safety, fuel recycling, and waste product—indeed, in every important element of a complete fast reactor system—it seemed to us that metallic fuel allowed tangible improvement. Such improvements would lead to cost reduction and to improved economics. Apprehension that the fast reactor and its associated fuel cycle would not be economic had always clouded fast reactor development. Sharp improvements in the economics might be possible if a metal fuel could be made to behave under the temperature and radiation conditions in a fast reactor. Not just any metal fuel, but one that contained the amounts of plutonium needed for reactor operation on recycled fuel. Discoveries at Argonne suggested it might be possible.

Metal fuel allows the highest breeding of any possible fuel. High breeding means fuel supplies can be expanded easily, maintained at a constant level, or decreased at will. Metal fuel and liquid sodium, the coolant, also a metal, do not react at all. Breaches or holes in the fuel cladding, important in oxide, don’t matter greatly with metal fuel; operation can in fact continue with impunity. The mechanisms for fuel cladding failure were now understood too, and very long irradiations had become possible. Heat transfers easily too. Very little heat is stored in the fuel. (Stored heat exacerbates accidents.) Metal couldn’t be easier to fabricate: it’s simple to cast and it’s cheap. The care that must be taken and the many steps needed in oxide fuel fabrication are replaced by a very few simple steps, all amenable to robotic equipment. And spent metal fuel can be processed with much cheaper techniques. Finally, the product fuel remains highly radioactive, a poor choice for weapons in any case, and dangerous to handle except remotely.

Important questions remained—whether uranium alloys that included plutonium could be developed that had a high enough melting point and didn’t harm the fuel cladding, while at the same time retaining the long irradiations now possible for the uranium EBR-II fuel. Early metal fuel had swelled when irradiated—the reason it had been discarded. But the swelling problem had been solved for all-uranium fuel. EBR-II had been operating with fairly long burnup uranium metal fuel for over a decade. Long-lived metal fuel resulted from metal slugs sized smaller in diameter than the cladding that allowed the metal to swell within the cladding. If properly sized, the metal swelled out to the cladding in the first few months of irradiation, and when it did, it exerted very little stress on it. After that, the fuel would continue to operate without any obvious burnup limit nor any further swelling.

Before the metal swelled sufficiently to give a good thermal bond with the cladding the necessary thermal bond was provided by introducing liquid sodium inside the cladding. The compatibility of liquid sodium with uranium metal allows this. As the fuel swells, sodium is displaced into the empty space at the top of the fuel pin, provided to collect fission product gasses. The bond sodium is important. It provides the high conductivity necessary to limit the temperature rise at the fuel surface and therefore the temperature of the fuel itself. The swelling itself, it was found, is caused by the growing pressure of gaseous fission products accumulating in pores which grow in size in the fuel as operation continues. But as swelling goes on, the pores interconnect and release the gasses to the space above. At less than 2 percent burnup the point of maximum swelling is reached, and the interconnections become large enough that sodium enters the pores. This, in turn, has the effect of restoring heat conductivity, which then acts to minimize the fuel temperature rise in the fuel.

The Fuel Conditioning Facility at EBR-II, the prototype of the IFR. Near the centre of the picture (with checkered cardigan) is Dr. Charles Till, who directed the IFR programme from 1984-1994 and is one of the authors of “Plentiful Energy”.

The soundness of the basic uranium design had been established by thousands of uranium fuel pins of this design that had been irradiated without failure in EBR-II. But now, metallic uranium-plutonium would need to be designed to accommodate swelling. Would the plutonium content cause swelling behavior different from uranium alloy? And, more worrying, plutonium forms a low-melting-point eutectic (mixture) with iron, below the temperature required for operation. A new alloying element would be necessary to raise the eutectic melting point. Zirconium was known to be helpful in that. Zirconium also suppressed the diffusion of the cladding elements, iron and nickel, into the fuel. Iron and nickel form a lower melting point fuel alloy; worse, they form those alloys in the fuel next to the cladding. Zirconium solves these problems. Ten percent zirconium was chosen as optimal, because higher amounts gave fuel melting points too high for the techniques we intended to use to fabricate the fuel. Ten percent gave fuel with adequate compatibility with the cladding, and a high enough melting point to satisfy operating requirements, and could be fabricated with simple injection-casting techniques.

Thus the fuel would be a U-Pu-10Zr alloy. But would it work? Ten percent burnup, about three years in the reactor, was our criterion for success. We would have one set of tests initially, and everything depended on its success. In the event, the fuel passed 10% with no difficulty. It got close to 20% before it was finally removed from the reactor. There were no failures (such as burst cladding). The very first IFR fuel assemblies ever built exceeded the burnup then possible for oxide fuel in the large programs on oxide development of the previous two decades. Metal fuel which included plutonium had passed the test. All the benefits from its use were indeed possible. The program could then turn to a thorough sequence of experiments and analysis to establish, in detail, its possibilities and limitations.

Plutonium (Pu), a heavy metal (element 94)


The IFR fuel cycle is the uranium-plutonium cycle. In this, non-fissile uranium-238 is converted slowly and inexorably to fissile plutonium-239 over the life of the fuel. If there is a net gain in usable fuel material, the reactor is a breeder; if not, the reactor is a called a converter (of uranium to plutonium), as are all present commercial reactors. But all reactors convert their uranium fuel to plutonium to some degree. Water reactors convert enough that about half the power the fuel eventually produces comes from the plutonium they have produced and burned in place. A significant amount of the plutonium so created also stays in the spent fuel.

A large and lasting nuclear-powered economy depends on the use of plutonium as the main fuel. The truth about this valuable material is that it is a vitally important asset. Its highly controversial reputation has been built up purposefully from the activists, with little countervailing public awareness of its “whats and whys.” Its very existence is said to be unacceptable. In this way, breeder reactor development was stopped in the U.S. and today continues only fitfully around the world. The fact that present reactors fueled with uranium convert uranium to plutonium very efficiently indeed, creating new plutonium in yearly amounts comparable to the best breeders possible, is lost in the rhetoric. But facts are facts. The principal plutonium-related difference between breeders and converters is that breeders recycle their plutonium fuel, using it up, cycle after cycle, so the amount need not grow. Present reactors leave most of the plutonium they create behind as waste. For efficiency in uranium usage, there is little incentive to recycle it; perhaps a twenty percent increase in uranium utilization is achievable, at a considerable cost to the fuel cycle. (Other reasons, such as waste disposal, may make reprocessing of thermal reactor fuel attractive, but not the cost benefits of plutonium recycling.)

However, it is plutonium that brings the potential for unlimited amounts of electrical power. Plutonium no longer exists in nature except in trace amounts. Its half-life is too short: 24,900 years. The earth’s original endowment decayed away in the far distant past. It has to be created from uranium in the way we just described. Plutonium is a metal. It’s heavy, like uranium or lead. It is chemically toxic, as are all heavy metals if sufficient quantities are ingested, but no more so than the arsenic, say, common in use for many years. It is naturally radioactive, but no more so than radium, an element widely distributed over the earth’s surface. Its principal isotope, Pu-239, emits low-energy radiation easily blocked by a few thousands of an inch of steel, for example, and it is routinely handled in the laboratory jacketed in this way. It is chemically active, so in fine particles it reacts quickly with the oxygen in the air to form plutonium oxide, a very stable ceramic. If this is ingested, either through the lungs or the digestive system, as a rule the ceramic passes on through and the body rids itself of it. A popular slogan by the anti- nuclear organizers is that “a little speck will kill you.” Nonsense—a little speck of the ceramic plutonium oxide will not react further, and will generally pass through the body with little harm.

Plutonium has been routinely handled, in small quantities and large, in laboratories, chemical refineries, and manufacturing facilities around the world for decades. There have been no deaths recorded from its handling in all this time. A study of the wartime Hanford plutonium workers gave the unexpected result that these people on average lived longer than their non-plutonium-exposed cohort group. This was explainable as the likely result of better and more frequent checkups because they were involved in the study, but at the very least there was certainly no shortening of lifespan.

The last point is plutonium’s use for weapons. The very fact that Pu-239 is fissile makes this a possibility, as it does also for the two fissionable isotopes of uranium, U-233 and U-235. But plutonium for a time was exceptional because it could be chemically separated from the uranium that it was bred from, and it did not require the large, expensive diffusion plants necessary for the separation of the fissile U-235 isotope of uranium. But this ease of acquisition argument changed with the development of centrifuges. Now the fissile element U-235 can be separated from bulk uranium with machines. And instead of a stock of irradiated fuel, a chemical process, and facilities for handling, machining, and assembling a delicate implosion device, as one must have for plutonium, for uranium one has a nearly non-radioactive natural uranium feed, centrifuges that can be duplicated to give the number needed, and a nearly non-radioactive product, easily machined and handled, which allows a more simply constructed weapon. Plutonium can no longer be singled out as more susceptible to proliferation of nuclear weapons than uranium. The fact is that uranium now is probably the preferred route to a simple weapon in many of the most worrying national circumstances. Iran’s current actions are a case in point.

Plutonium-238 (plutonium oxide). Nuclear fission reactions release enough energy to increase the temperature of this sample of plutonium to red-heat. The heat produced by plutonium has been used as an energy source on spacecraft. Photo: Department of Energy.

Weapon-making is complicated by the presence of radioactivity. Plutonium processed by an IFR-type process remains very radioactive; it must be handled remotely, and delicate fabrication procedures are correspondingly difficult. Uranium is so much easier. This is not to imply that the large and sophisticated weapons laboratories like Los Alamos or Livermore could not use such isotopically impure reactor-grade plutonium; it is sufficient to say they would not choose to do so with much more malleable material available. And the beginner would certainly avoid the remote techniques mandatory for IFR plutonium.

This is the situation: Plutonium, as used in the IFR, cannot be simply demonized and forgotten. It is the means to unlimited electricity. The magnitude of the needs and estimates of the sources that might be able to fill those needs lead to one simple point: Fast reactors only, taking advantage of the breeding properties of plutonium in a fast spectrum, much improved over any uranium isotope, can change in a fundamental way the outlook for energy on the necessary massive scale. Their resource extension properties multiply the amount of usable fuel by a factor of a hundred or so, fully two orders of magnitude. Fine calculations are unnecessary. Demand can be met for many centuries, by a technology that is known today, and whose properties are largely established.

This technology is not speculative, as are fusion, new breakthroughs in solar, or other suggested alternatives. It can be counted on.

By Barry Brook

Barry Brook is an ARC Laureate Fellow and Chair of Environmental Sustainability at the University of Tasmania. He researches global change, ecology and energy.

86 replies on “IFR FaD 10 – metal fuel and plutonium”

I’m also part-way through my first reading of the book, nearly finished in fact.
Mr. Holland, If you want technical details about pyroprocessing and the IFR design only I suggest you skip the first couple of chapters, as they’re mainly history (and therefore political) chapters. I did find those chapters helpful for getting a grip on the history of the technology though, but each to his own.

I’ve found that Till and Chang tend to repeat what they’re saying about the basic IFR characteristics quite a bit in the chapter introductions, which gets a bit tedious.
I would concur that they don’t mention many downsides that are inherent to the reactor technology.

I would also add that they say nothing (or I have missed in my first pass of reading so far) about the power characteristics such as theoretical electrical power ramp-up/ramp-down rates in comparison to existing reactors. Are we therefore safe to assume that as they use steam turbines similar to PWRs (minus the large sodium-water heat exchangers) that they would have similar ramp-up/ramp-down rates? Or are there IFR-specific features perhaps such as faster reactivity controls that would lead to better control of power output?


Does this design reduce the following risks of nuclear power plants, you still require pump power to cool do you not.
The biggest risk now with nuclear power is almost anyone can cause a meltdown. WE know from TMI that a switch out of position can cause a meltdown; it is easy for a person with basic electronics to get hired on at a plant and screw with switches and relays. A person spending about $5,000.00 could wire in wireless remote controls to change readings at will. Also I have read about sun surge in 1860’s that was so massive that the electrical surge was so high that it will take out all electrical power including battery, generators and main power grid. This was many times greater than the surge that hit the Quebec power grid several years ago that toke it out. This could mean that all the plants in the world facing the sun when the surge hits will go into a meltdown all at the same time could be over 50 plants involved. These plants are vulnerable to cosmic events including comet showers. But this does not stop the industry I do not know what will other then demanding the leaders of each country shut them down no courts involved.


Richard Perry, the IFR has passive cooling of its fuel. If there is a total loss of external power, like what happened at Fukushima, the reactor can self-cool indefinitely with an external heat exchanger. There is no need for an active core-cooling system like in water-moderated reactors, and the type of meltdown that has occurred at TMI and Fukushima (due to loss of cooling, conversion of RPV water to steam etc.) does not occur. Indeed, part of the PRISM design is to cool the recently used fuel (pre-recycling) within the reactor sodium pool, rather than in cooling ponds.


(Comment deleted – off topic.)

Please re-post this critique of the book in the “Plentiful; Energy” thread. Once done I will delete this comment.This thread is for comment on the IFR Fad 10. Time limit for re-post has expired.


One question that came up for me as I started to understand the importance of the choice of metal for the fuel is, what about GE’s PRISM? Can anyone point me to material I can study that will clarify what the difference is between GE’s PRISM and the IFR?


David Lewis, there’s no real difference. The GEH S-SPRISM is a 311 MWe commercial implementation of the IFR, although the initial proposal in the UK is not immediately integral because the fuel recycling would not be on-site. To quote a recent post from Dan Yurman:

Fuel for the PRISM reactor is created by converting the plutonium from powder form mixing it with uranium and zirconium to make a metal fuel. The resulting spent fuel contains plutonium in a form that cannot be used to make nuclear weapons.

Eric Loewen, chief engineer on the project (and president of the American Nuclear Society), said that the waste form is much the same as what comes out of light water reactors. Once the plutonium has been in the PRISM reactor for five years, it is mixed with other nuclear materials that make it nearly impossible to retrieve the metal for the purpose of making a weapon.

The PRISM reactor is a so-called “fast reactor” because it uses liquid metal sodium rather than water to cool the system. The sodium allows the neutrons to maintain higher energies and to cause fission in elements such as plutonium more efficiently than light water reactors. (large image)

Based on the design of the Integral Fast Reactor (EBR-II) developed at the Argonne National Laboratory in Idaho, the PRISM reactor uses passive safety features that cause it to shut down automatically. In the event of a complete loss of electrical power, it simply stops working and passively dissipates residual heat. EBR-II was canceled in 1994, but not before a safety analysis showed that there were no technical barriers to getting a license and safely operating one.

According to a fact sheet from GE Hitachi, the PRISM reactor’s relatively small size and simpler design would allow it to be built in modules and transported for assembly on site. Another benefit of the reactor is that while it is disposing of weapons materials, it is also generating electricity.


Would commenters please note Barry’s directions that any critiques on the book should be posted in the “Plentiful Energy” thread.
Please reserve this IFR Fad 10 thread for comments relating to that subject only.
Failure to do so will result in a messy and confusing thread.


One point on plutonium covered elsewhere in the book concerns how breeders and plutonium got their present “bad” reputation.

“The methods of reprocessing commercial nuclear fuel in current use in several nations (but not in the US) were actually developed originally to provide very pure plutonium for use in nuclear weapons. The commercial plants have that same capability….”. It is this point the antis have exploited to the point people who should know better repeat mindlessly that if civilization turned to breeders in order to phase out fossil fuels quick there will be weapons grade plutonium transported around all over the place as a result.

I.e. Dr. Kevin Anderson, former Director of the Tyndall Centre in the UK explained exactly that, i.e. breeders can’t save us, because the threat from the huge quantities of “weapons grade” plutonium that will inevitably be produced is too great, recently. His line is civilization, because of the size of the present fossil infrastructure and the unprecedented and therefore plausibly impossible rates of emission phaseout that would be required to keep warming below 2 degrees, faces an “impossible” future, i.e. warming of such extent it will prove to be “incompatible with an organized global community”. Etc. voice of doom. But he is not acquainted with the IFR design.

The book explains the pyroprocessing in great detail. What Till and Chang want people to understand is this: “the process cannot purify plutonium from the IFR spent fuel; it is scientifically impossible for it to do so. (page 42). The IFR technology should not contribute to weapons proliferation. On the contrary, by replacing present methods it should substantially reduce such risks”.


DL, quite right. Indeed the only way to get weapons-grade Pu out of a pyroprocess is to put weapons-grade Pu in. I wish more people would understand this, especially the Kevin Andersons of this world, who advocate for substantive decarbonsiation but have not taken the (relatively small amount of) time to familiarise themselves with the crucial technology solutions that they try and critique. Very frustrating.


This link says the UK has 108 tonnes of plutonium. My reading is that is mostly as already separated oxide (via PUREX type processes) and mostly held at Sellafield. It also explains why the MOX option is not preferred.

Reduction to metal must be a costly and difficult extra step. I wonder how many people inhale lead compounds which have that distinct odour when you clean the positive terminal on a car battery.


JN, I doubt that reduction from LWR oxide fuel to metal will be particularly expensive, but this step needs to be proven at a commercial scale (which is what the UK implementation would do). There is a whole chapter in Plentiful Energy on the LWR oxide –> metal fuel process.


John, PE describes a pyroprocess (electroylsis in molten salt) for reduction of spent oxide fuel to metal. The oxide forms the cathode, and a voltage applied through molten LiCl / 1% Li2O at 650 C to deposit the metal on a Pt anode. Nothing’s simple, but it doesn’t appear more complex than the pyroprocessing loop for the main IFR fuel cycle.


I had seen documents, such as this one, where GE envisions at least some installations where fuel “recycling” (they prefer this word to reprocessing) will be on site.

On the other hand, GE offers documents like this which leave out any mention of the fuel recycling feature.

So metal was rejected by the Rickovers of this world who thought they knew all about how to commercialize new designs. But some at Argonne couldn’t forget about metal’s advantages.

Because they kept on experimenting with metal reactors by running one they overcame the problem. The result was the sodium filled rod that allowed the metal fuel pins to expand somewhat before hitting the cladding design. By solving metal’s great problem they discovered a great advantage. The onsite reprocessing they were forced to resort to in order to make their metal reactor experiment economical, when it became “good enough”, turned this bug into a feature.

So now they’ve got a system where all plutonium produced is recycled out of the waste stream and sent back into the reactor because they’ve got onsite el cheapo recycling facilities. The way Till and Chang write about the evolution of the design it almost sounds like the famous “Puffed Grass” commercial where, we’re told, “science works for YOU here at Eating Corporation of America”, when the gardener dumped the lawn clippings into the Puffed Wheat machine by mistake and created the first Puffed Grass.

However they came up with it, it is a breakthrough advance. They do caution in the book that the process needs further development. It was disappointing to read that. Blees had sort of convinced me the IFR was a design that was ready to deploy at full scale right now. I’d be interested in more detail on what a civilization that was on a war time footing now that it had realized how badly it had blown its climate problem could do.


I think the fuel fabrication could be done more economically. The uranium could be converted to metal, which has a reasonably high melting point of 1132C. Reduction of PuO2 to a metal with a melting point of 639C could be avoided. Uranium could hold the ceramic PuO2 in a Ceramic-Metal form. With nearly three-fourth of lighter oxygen nuclei gone, the breeding ratio advantage will be proportionally obtained.
If we use in cermet fuel higher melting and better heat conducting thorium metal, we sacrifice the fast fission of U238 but go on to create better fissile U233. Thorium could be used with ceramic PuO2 or with metallic U233. Metallurgical problems of metallic Pu are better avoided.


How about what happened at Monju Nuclear Power Plant in Japan a few years back. This looks bad as there are few of these type of reactors and have problems. The problem was caused from a break in the piping for the cooling system. It is bad when all it takes is a pipe to break and the plant goes into critical mode.


Interesting comments thread on Depleted Cranium, thanks Seamus. Richard Perry, trolling is not allowed on BNC, and you are skirting perilously close to that by the tone/nature of your comments. Be on notice.

Monju had a sodium leak in its secondary piping, which was quickly cleaned up, but also covered up, and when someone found out about the cover up, it created all of the legal/bureau/admin problems. It didn’t send the plant into ‘critical mode’, and to say this indicates that you have no idea what you are talking about on this issue.

Putting that aside, the IFR is not a Monju-type reactor anyway. The IFR uses metal fuel instead of oxide, it has a pool design instead of loop, it uses double-walled piping instead of single-walled like Monju (so a leak in an IFR internal pipe can be detected and repaired without any sodium leaking), etc. As a result of these innovations, the EBR-II ran perfectly for many decades, without leaks.


Interesting that the metal fuel solution was to just give the fuel a volume in which to expand. Very simple….and undoubtedly thought up by someone with a mechanical engineering background :)

Loved the book. Too bad homo sapiens is too backward and stuck in their ways to implement the IFR.

Perhaps by 2100 or so the human race will wake up and realize there is not much choice.


I’ve just ordered the book!

Can someone please flesh out one detail for me. The “remote handling” point about IFR plutonium. If you had a work force
willing to die, could you handle plutonium other than remotely. How quickly will it sicken or kill? I’m not actually sure that such a work force is possible … suicide bombers are not the same thing as suicide engineers … but I’d still like the detail.


Geoff, GRL Cowan constructed an estimate of suicide engineer survival time for spent IFR fuel in this comment 3 years ago:

.. a conservative estimate of how quick the supposed thief, having neglected to bring a 50-tonne self-propelled shielding flask, will decide to sit down for a little rest, and never get up again: two seconds.


I think, to put it in plain language, there is simply no way to remove this material from the reactor site without an engineered system in place to do it, with all the forethought, design, planning, construction, testing and capital expenditure that implies.

It is not the sort of thing that could be achieved by either surreptitious material subversion, or a raid, no matter how well manned and armed, to address two popular fantasies.


Barry Brook, with the amount of money put into this plant and time not running I would call that critical. I have read now they have shut it down. Barry, I am very glad to hear that there are safe guards put in place since this plant, this is good to see. Zirconium be careful at different temperatures and repeated changes in temperature. Looks good but I believe with all the new standards the cost will be in the way. Sorry if I came across as aggressive, but I believe nuclear is the way to go but only if it is safe but the owners want it built and run cheap to compete in the market is a problem.


Richard, the IFR design doesn’t use zirconium in its cladding like in water-moderated reactors – it uses stainless steel. There is also no water present in the reactor vessel, so no Zr-catalysed hydrogen can be evolved.

Zr is alloyed in the metal fuel pins, to control the melting point (U-Zr was used in EBR-II, the final IFR fuel is envisaged as U-Pu-Zr10 alloy). I’d encourage you to read the Plentiful Energy book to find out more.


I may have missed a point in my reading of PE. The book explains that the alloy of U-Pu-10Zr is soft enough to allow the gaseous fission products to form pockets, so that the fuel foams and expands to fit the the width of the stainless steel canister. With enough gas in the pockets and the plenum at the top of the canister, any expansion of the fuel will be taken up without splitting the container. Well, almost any.

Elsewhere (on BNC if not in the book itself) I have gathered that in the event of a nuclear excursion, the fuel would expand so that its increased surface area would bleed excess neutrons and limit the excursion. I can imagine the entire core expanding over a few minutes into a giant popcorn. But that is an end-of-reactor event rather than occasional control mechanism. Is there a gentler self-limiting response that doesn’t require the fuel to break out of the canisters?


Roger, the scenario you describe is not damaging to the IFR and would not result in burst cladding. We know this, because during the two 1986 ATWS safety tests (LOF and LOHS), the EBR-II remained completely undamaged during the induced power excursion. The fuel pins expanded, criticality was reduced, and the reactor shut down to low power levels and self cooled thereafter. This is all described in the PE book.

Indeed, one test was run in the morning, then the reactor was powered up, and the second test was run in the afternoon. The expansion required to reduce criticality is far below that which might cause irreversible fuel warping. No popcorn event, we know that for sure.


Roger Clifton, I think whats going on here is that gaseous fission product builds up slowly, so the fuel expands over a period of presumably months, before dimensions stabilize when the pore structure percolates. Reactivity changes are compensated by control rod operation to keep the criticality factor k~1.

But the fuel density will respond immediately to a change in temperature, so any power excursion immediately has a negative feedback on reactivity (time constant of a few ms). The difference between prompt and delayed criticality is less than 1% of the neutrons (for a LWR at least), so it doesn’t take much neutron loss to turn the wick down. In one of the EBRII pull-the-plug experiments the temperature shot up a couple of hundred degrees in not many seconds before negative feedback kicked in. Metal density will change a fair bit in that range. I also think there is a negative feedback from high temperature increasing the relative speed of neutrons to nuclei, so the cross section is reduced.

In other words, I don’t think its an irreversible popcorn event, rather a reversible expansion/contraction.


Problems of sodium coolant have been brought up in many posts. There are two lines to a solution of the fire risk, but a related issue is compatibility of coolant with containers and piping.The two lines are:-
1. Salt coolant. A combination of NaF, PbF2, SnF2 and ZrF4 in various combinations will meet the requirement. More is given in document

Click to access Pub29596.pdf

2. (a) Metal coolants tried out by Russians like Lead or Lead-Bismuth. It is more damaging to metals than sodium.
(b) A Al-Mg eutectic expected to be benign to metals. It burns at a higher temperature bringing in a higher margin of safety.


Another choice of fuel in a fast reactor but generally discussed as a different class of reactors, is liquid fuel. Chlorides of uranium and Plutonium have low melting points and can easily be used as liquid. One advantage of liquid fuel is that fission gases, Xe and Kr can be constantly removed.This would, however require isotopic separation of chlorine and Cl37 used.
Another advantage of liquid chloride fuel is the ease of processing to partition the irradiated fuel into fission products, uranium and Transuranics.


Please do not post off topic comments. The Open Thread is for this purpose. BNC does not have the facility to move posts for commenters so please keep a record of your comments to enable you to easily re-post elsewhere, when directed.
Off topic comments tend to distract, confuse and lead the thread off-track. When I am not on the blog these comments build up and cause people replying to them to be mis-led as to the proper place to respond. Thankyou.


I agree with the metal fuel and metal bond philosophy outlined here.

Apparently the biggest scale-up/commercialization issue is the fuel processor. So why not just build the reactor on low enriched uranium metal fuel? Operate in once-through mode and a bit higher burnup. Reserve some space for a future processor unit. Get started right now with the reactor. When the processor is available, add it on and start processing spent fuel, and switch to a bit lower burnup mode.


Isn’t this what GE are proposing for the UK reactor, except for using depleted U + plutonium out of the UK stockpile rather than paying for enrichment? There is no plan to build the fuel processing plant at the moment, but nothing to stop it being done later if policy changes. They get to demonstrate the reactor, and establish real world costs and schedule for FOAK plant. It will not be economic on fuel costs because there will be far too much fissile left in the discharged fuel. You have to reprocess to reuse this, even if you aren’t trying to breed, or you end up throwing away more fissile than an equivalent-power conventional reactor would need in total.


Enriched uranium for fast reactors may be feasible in source countries like Australia, Canada, Russia or Kazakhstan but reuse of materials is the best policy for others. Costs are relative and vary with the place, time and circumstances. The China currently has lowest costs and highest number of reactors under construction.
If the UK do not want any further reprocessing and only want to use up the plutonium stocks, they should include it in thorium pellets as the fissile feed and burn it in their AGRs and the future EPRs. They will avoid capital expenses of a FOAK reactor and get return on the money invested in reprocessing in the past.


Luke, Pu startup without the online processor will certainly be uneconomic. But low enriched uranium startup with a high burnup should be pretty decent – you can get well over 10 years fuel life which mostly makes up for the higher fissile charge. One of the advantages of the fast spectrum is that fission products aren’t such a big deal, and the bred plutonium is better fissile than the low enriched uranium startup. I guess this is getting nearer to a travelling wave reactor than an IFR.


Regarding the cladding, I’ve always wondered why IFRs have always suggested stainless steels and such for cladding. This seems like a case of reinventing the wheel. Why not use the zirconium based PWR cladding? That’s what is being used right now, it works fine, no low melting plutonium eutectics. All the nuclear giants are still investing in better zirconium based claddings, recently we’ve seen the oxygen dispersal strengthened zirconium claddings from the French and Russians. With no water in the core, zirconium’s primary safety issues (hydriding, oxidising, hydrogen evolution) disappear.


Found the ref. on (un)economics of IFRs in once-through mode

Chang & Till wrote

The LMR simply cannot compete with the LWR in a once-through mode. The fissile inventory is much higher in fast spectrum and the spent fuel has too high fissile value. Fuel cycle closure is mandatory and its economics then determines the economic viability of the LMR.

LMR = liquid metal (cooled) reactor


Luke, I don’t see where the evidence is. I just see some very low fuel cycle costs in general with some bold assertions by IFR people begging for an online processor unit. Even if I don’t try to read between the lines.


In India, the per MW costs of the PFBR and the Russian design VVER (both in advanced construction stage) are similar. It must be the same in Russia. Some would call the RG Plutonium costlier. Others would call it recycling of waste. Without recycling in fast reactors, the uranium peak may be close. There are news predicting it and studied statements against it. With fast reactors, you are safe for centuries.


I wonder if the 10% zirconium is to increase the softening range of the fuel, rather than to lift the melting point as stated in the book. As a viscous material against the steel cladding, it would have less power to dissolve the stainless steel. Similarly, a viscous alloy would be more able to accommodate the bubbles in the metallic foam.

As temperature increased during some extreme event, the hotter and thus softer foam would be more able to expand and provide the proven self-limiting safety response of the IFR.

I note that the EBR2 fuel also contained zirconium, perhaps providing the proof of concept.


Roger, the melting point itself doesn’t appear the basic problem. Wikipedia has a cool graph of the different phases of the actinides:

Uranium has a low temperature transition. Adding zirconium would increase that transition, causing the crystalline phase change related swelling to occur at higher temperatures, ie above normal operating temperature, which is what you want.

It also looks as if the crystalline change absorbs a lot of heat, just like actual phase change absorbs latent heat. That’s cool. I mean, it keep things cooler during a transient.

The metal fuel and metal bond are truely genial inventions. I wonder why it is not used more often. PWRs could use lead as a metal bond in the oxide fuel. There was some work on this but it doesn’t appear to have made it to commercial plants. It is really worrying that even modest innovations in existing reactors have a hard (and long) time to get into commercial operation,.


CyrilR gave an explanation of the role of zirconium, suggesting that it would raise the temperature of a phase change in otherwise pure uranium. This would have been apropos EBR2’s mainly-uranium fuel. However, I would also suggest that zirconium may be adding glassy bonds that impede transitions between crystalline phases.

IFR fuel is to contain 40 to 50% plutonium, which shows up as particularly complex in CyrilR’s link to the crystalline phases of the actinides, so the need to inhibit crystalline phase transitions increases.

In order to machine plutonium metal, bomb makers add about 5% gallium to ensure a workable material. In an article in Physics Today, we read that the plutonium alloy in weapons is thermodynamically unstable, in a state of suspended transition.

If Pu-Ga weapons-grade metal is to be turned into fuel in an IFR pyroprocess, the gallium would presumably accumulate at the anode along with iron from the cladding.


To avoid the sacrifice of good in pursuit of the best, you could have a CERMET (CERamic PuO2-METallic thorium) fuel. You will be sacrificing the fast fission of thorium as well accepting nearly 25% of O atoms. What you get is high melting (2115K) metal matrix and creation of U-233.
U-233 can be used as a metal with Uranium or thorium metals.


@Jagdish proposes a fuel composite material of PuO2 in a matrix of Th metal.

The high neutron flux in an IFR would excessively increase the probability of Pa233 absorbing a neutron instead of decaying to U233. It would however make sense in a low flux reactor like CANDU. In fact the IAEA proposed such a mix as a good way to burn plutonium.

If an IFR were to be used to burn plutonium rather than breed it, the plutonium still belongs in the core, whereas thorium could be used in the blanket. Then in the pyroprocessing, the cheap, disposable thorium could be dropped out at the anode (if my guess at its rank in the electrochemical series is correct) whereas the precious U233 would travel on into the cadmium cathode. However this would not serve the self-sustaining intention of an IFR.


Roger, CANDUs are quite high flux in terms of numbers of neutrons. But their excellent thermalized spectrum reduces the losses to Pa233. And their low fissile inventory reduces the cost of buying the reprocessed Pu (it is expensive).

However a very fast spectrum also reduces the losses to Pa233, and what is more, the U234 produces from captures has a good fast fission profile. So it is not so bad for a fast spectrum, though the IFR’s high power density does gobble up more Pa233.

It should be quite feasible to breakeven on breeding with an IFR running on Pu-Th fuel. You could also have a hybrid cycle with some U238 in it to increase breeding ratio & denature the bred U233 to low enriched uranium levels. That would be a very proliferation resistant cycle.

As for reprocessing, I think you can also boil out most of the fission products in metallic form using a single stage distillation unit. That leaves all the actinides together – they’re all very high boiling.


If the single unit uses chlorides, themost volatile elements will be thorium and uranium. Plutonium and other actinide chlorides are close to lanthanides in volatility. They are best separated electrolytically by Pyroprocessing associated with IFR.


@CyrilR — it is good news is that U234 fissions well. Already enriched in any LWR fuel used as top up, it would otherwise accumulate in the IFR, as its cycling burns its way through the trickle of input actinides.

What might accumulate is any (non-fissile) fission product, or component of cladding metal, with an electrode potential that gets it swept into the cadmium cathode. Jagdish, you spoke of electrolytic separations, would you be able to link us up with a table of electrode potentials?


Roger – this reference has what you’re looking for:

Click to access lecture5.pdf

It has a table of Gibbs free energy which tells you where the different atoms will tend to go into. Cd, Fe, Mo, Nb, Tc, Rh, Pd go into the cadmium pool (anode). Actinides and Zr are electrotransported for recovery. More stable stuff such as Ba, Cs, Rb, Sr stay in the salt as chloride salt.

I wonder how the processor is cooled – surely some passive cooling system?


@CyrilR, thank you for the link. These figures are those in the book, presumably from the same source. In the book, captions refer to the values as kilocalories per mass, whereas here it is more clearly in kilocalories per mole of electrons. Converting to joules per coulomb, ie volts, shows the range of voltages needed to re-collect the elements for the core fuel at 500°C :

CmCl3 -2.77
PuCl3 -2.71
AmCl3 -2.69
NpCl3 -2.52
UCl3 -2.39
ZrCl2 -2.02

However the table refers only to the elements in the study, which would not include all of the fission products, nor all of the elements in the alloys in the LWR fuel, stainless steel etc. I wonder that with continuing recycling, unwelcome elements might accumulate. Magnesium, aluminium and beryllium come to mind. Quite possibly, the activation products would add minimal radioactivity, however spallation of beryllium may bleed tritium for example.

On the other hand, it would be neat if subsequent processing of chlorides were able to separate cesium (-3.81 V is too high for the pyroprocessor to separate it) and somehow return it to the core fuel, so that the long-lived Cs135 could be transmuted.


Unfortunately the pyroprocessing was carried out only on one batch of fast reactor fuel. Let us hope that Russia, China or someone else continues the good work. Indians are reluctant to even think about salt fuel, which is the medium for extracting TRU’s by electro-refining.


Magnesium will only be present in minute quantities in the sodium coolant and sodium bind, from sodium neutron activation. But magnesium forms a stable chloride that should stay in the salt electrolyte. Aluminium I don’t think is present at all in the IFR system. Beryllium might be present as minor fuel constituent, but its presence is beneficial to neutronics and fuel integrity. It reduces swelling, rod surface oxidation, improves delayed neutron fraction, etc. Zirconium is convenient to recover since a goodly amount is needed in the fuel.

Cesium chloride could probably be recovered by vacuum distillation (ORNL developed this techology decades ago for the MSRE). This is probably the simplest way to go; the salt is already molten in the processo and need only be routed to a small single stage still.

I wouldn’t worry much about Cs-135, though. It’s a long lived beta emitter that doesn’t bioaccumulate.


Jagdish, pyroprocessing is being actively developed in Korea. From the Korean Atomic Energy Research Institute’s page on fuel cycle development we learn:

KAERI is focusing its efforts on the development of pyroprocess technology, which separates and refines various nuclear materials contained in spent fuels with an electrochemical method at a high temperature. Utilization of this technology can substantially reduce the volume and temperature of spent fuels discharged from nuclear power plants, and the separated and refined materials can be used in fast reactors. It is called a dream technology as it can increase the uranium utilization rate by up to 100 times. Furthermore, its proliferation resistance has been internationally recognized due to the impossibility to recover plutonium.

* Reducing the volume and heat of spent fuels up to 1/20 and 1/100 each
* Expanding the uranium utilization rate through the reuse of spent fuels
* Construction of an Advanced spent fuel Conditioning Process Facility (ACPF)
* Securing international transparency for the separation of nuclear materials
* Participating in 10 joint collaborative research programs with the USA

Given Korea’s commitment to nuclear technology development I think this is cause for optimism.


This is good news. Bulk of reactor construction is now in Asia. Advantage research should also logically be at the same places. Fast reactors and pyroprocessing, the main components of the IFR are the next big step in nuclear power development. Fast reactors in Russia, China and India and pyroprocessing in Korea could lead to IFR development. Let North America, Europe and even Japan have a long breather of a decade or two.


Reducing the volume and heat of spent fuels up to 1/20 and 1/100 each

One cannot simultaneously reduce the volume and the specific heat generation. Indeed, they are opposed. Concentrating fission products (which generate almost all the decay heat) by removing the 95% uranium bulk (which generates almost no decay heat) increases the specific heat generation.

Also, the fission products don’t go away – their total heat rate is fixed at a familiar declining rate, no matter what the volume.

The selling points for the hot processing “pyroprocessing” technologies are in the small physical equipment size & related cost, almost complete elimination of long lived wastes (waste-to-energy), high temperature operation allowing passive cooling (safe) and the use of shorter lived spent fuel, and lack of bulky secondary wastes creation.

It is likely that the fission products can be used to generate power, because they are sitting so hot in a molten salt bath. If the power or cooling fails, the higher temperature rise will remove heat passively.


I’m not very worried about long-lived fission products. The “worst” of these is Tc-99, which has about the same yield as Cs-137, but ~7000x longer halflife and ~4x lower decay energy, for a total activity ~28 000x lower. Cs-135 is ~10x lower still with ~11x longer halflife, a bit lower decay energy, but somewhat higher yield than Tc-99.

Sn-126 might represent a problem due to the 2,8 MeV gammas from its daughter Sb-126, but 230 ka halflife and low yield keeps the activity fairly low. Other LLFPs are even less of a worry.
Speedy – BNC Comments Policy requires scientific refs to support your comment. Please supply tthem so others, not as knowledgeable may check out the information in more depth. Thankyou.


Technetium-99 is a high value industrial catalyst and corrosion protector. It’s best use is for nuclear applications where the slight radiation is no problem, for example nuclear primary loop and steam generator components (adding some Tc-99 in the alloys protects the alloys from corrosion).


Sources for my previous post:

Wikipedia on LLFPs:

2,8 MeV gammas from Sb-126 comes from the ANL fact sheet on tin:

Click to access tin.pdf

I’ve compiled a spreadsheet of fission products showing nuclides with halflifes >1d, >10d, >1a, long lived and the final stable isotope for each atomic number along with thermal yields for U233, U235, Pu-239 and Pu-241:
Feel free to share it. There are two omission from this spreadsheet, meta-states (e.g. Sn-121m) and the effects of neutron capture (e.g. Cs-134 is not produced directly since it’s shadowed by stable Xe-134, but is produced with capture from Cs-133).

Halflifes vary somewhat between sources, I’ve used this chart:

Yields are from here:


According to a presumably authoritative report in the Korea Herald , the liquid cadmium cathode “collects most of the actinides and some of the rare earth elements”. Consequently fission product rare earths must accumulate to some degree in the recycling fuel.

As fission products tend to be neutron rich, they would also tend to have low neutron capture cross-section. For the same reason they would tend to beta decay to Z=Z+1, as an element which may escape the voltage window. However some species may simply accumulate neutrons until they contribute a delayed neutron instead. At first glance they seem to be of little problem, if anything adding a few gammas and maybe neutrons to the thief-proofing.

I imagine most of the thief-proofing would be due to neutrons arising (1.6 million /g/h ) from the spontaneous fission of Pu240. Surely, untouchable by human hands. Does anyone have an advance on this?


With plutonium in oversupply nowadays, there is no longer a need for a separate blanket. Presumably the IFR could be reconfigured without a blanket so that sufficient breeding occurs in the core. Then only a single stream of irradiated metal pieces goes from the reactor to the pyroprocessor, returning as a single stream of cleaned, but still hot metal back to the core. Considering how difficult remote handling must be, such simplification should be very welcome.

With cycling of core metal only, the ratio of 240/239 increases (before, after). That is, there is no rich Pu 239 to worry the proliferation people.

The reduced requirement for security and oversight may make such an IFR much more politically palatable.


Roger Clifton, yes all good points. One other advantage of a homogneous fuel element core (no blanket) is increased power production. That is because a blanket fuel element produces much less power than a seed fuel element. Since there are limits to the power density of the core, a non-blanket core produces much more power than a core with a blanket. This is an important economic improvement.

Also if there is a reduced requirement for breeding, say only breakeven breeding is targeted, then a much higher burnup is possible. This again makes the cycle much more economical.

Personally I prefer to add thorium to the fuel. Thorium stays with plutonium chemically, and makes Pu238 that produces heat and sponteneous fission neutrons. It also makes U232 whose daughters produce powerful gammas, hurting bomb electronics and bomb builders alike. At the same time the bred fissile from thorium, U233, is diluted in the U238 so that it is always present as low enriched uranium. Those are important proliferation resistance improvements. Thorium also has a higher melting point, and can breakeven on breeding in a slower spectrum where you need less fissile to start up the core and you can benefit from improved doppler reactivity safety. One way to achieve this is to add beryllium to the fuel. This moderates neutrons a bit. It also makes extra neutrons from the n,2n reaction. Finally, it also produces neutrons from alpha radiation, which further improves the proliferation resistance.


I imagine most of the thief-proofing would be due to neutrons arising (1.6 million /g/h ) from the spontaneous fission of Pu240. Surely, untouchable by human hands. Does anyone have an advance on this?

No Pu240 offers little hazardous radiation to stop bomb builders, with a half life of thousands of years, and it produces less than a 1000 spontaneous fission neutrons per g/s, not 1.6 million. It’s really mostly an alpha emitter. I would handle it, if I had standard issue gloves and a gas mask. But Pu240’s spontaneous fission does cause predetonation, causing the bomb to fizzle rather than fission. A fizzle bomb is quite powerful though not anywhere near a Hiroshima or Nagasaki bomb.

However, the Pu241 is a great proliferation deterrent. With a half life of 14 years it is very energetic, producing loads of heat and it also fissions faster than Pu239, further increasing proliferation resistance.

The pyroprocessing would add a number of fission products to the fuel mix. I’m not sure but I think you will get zirconium and a couple of rare earths. These are of course very energetic and further deter proliferation.

Further design proliferation resistance can be had by the addition of thorium and beryllium, as I’ve written above.


CyrilR – yes, if the IFR were required to burn up a plutonium-only stockpile, addition of thorium would provide the delayed neutrons needed for control without breeding higher actinides. Even the Pu238 that you mention is only present at 30 g (per each GWa generated by the thorium component ref). I don’t know about the metallurgy, though, especially the requirement for softness at operating temperatures.

If the un-bomb-worthy fuel were considered innocent by the proliferation people, it might allow relocation of the pyroprocessing plant. Further, a large community of breeders and burners trafficking their fuel through a large, presumably more efficient, pyroprocessor, might also ease the establishment of nukes in non-nuclear countries, while their pyroprocessing was conducted in an already-nuclear country.



Very many thanks for answering my questions on the “Off to Russia” thread. Given that you have been discussing some of the same issues with Roger Clifton on this thread, I have decided to hop across to join you.

It would seem that both you and Roger see advantages in abandoning a breeder blanket and going with a mixed metal homogeneous fuel in the core only. You both cite increased proliferation resistance and ease of reprocessing as advantages while accepting that reduced breeding will be a consequence. Nevertheless, isobreeding at worst will still be possible. In addition, Cyril cites increased power production and longer cycle times as giving additional economic advantages to blanket removal.

Cyril, you go on to suggest the addition of thorium will provide even more proliferation resistance, improved safety and isobreeding capability with lower start charges (slower spectrum being achieved by beryllium addition). Are you suggesting that one’s fuel should contain plutonium, uranium, thorium and beryllium or just plutonium, thorium and beryllium (or either option)?

I am wondering whether proliferation concerns are really particularly legitimate for a nation that already has nuclear weapons capability. Surely, one should be more concerned over safety and economics? In any event, it would seem that any potential proliferation worries are principally to be found in the reprocessing stage of the fuel cycle? Thus, onsite pyroprocessing seems attractive (assuming it to be cheap and doable). I had assumed (wrongly) that pyroprocessing was more or less synonymous with electrorefining. Cyril, however, would apparently prefer to see distillation of spent fuel as a means of separating mixed heavy metal from fission products if that were possible. Is it likely to be and to what extent would it delay rollout? It would still be pyroprocessing since the term merely relates to the temperature at which it is carried out (robotically in a hot cell). Could you elaborate on the technical and economic implications of distillation versus electrorefining?

Finally, the S-PRISM is a medium sized reactor that has the capability of being constructed in a modular manner, allowing economic advantages. However, neutron losses are higher than in larger reactors. Clearly, therefore, the potential to breed efficiently is somewhat compromised. Are there also economic implications? Are the advantages of lower construction costs of modularity likely to be partly or completely offset by increased power production costs? .



You have been promoting lead as a coolant. I haven’t the technical expertise to disagree. It is clear, however, that its principal advantage over sodium relates to fire risk. links to a paper which compares and contrasts the pros and cons of lead, lead-bismuth and sodium as coolants. The cons of lead are listed as follows
1) Damage of components and fuel elements caused by corrosivity of heavy metals. (Possibly an obsolete criticism in view of your earlier dismissal of it).
2) Freezing in steam generators if high pressure feed water heaters fail.
3) Difficulty of repair, maintenance and refuelling at temperatures high enough to prevent freezing.
4) Potential blockage of coolant flow cross section in the fuel sub assembly caused by water/steam interaction with coolant.
5) Problems with production by lead of long lived radioactive isotopes (?Pb 205).

The authors suggested that none of these issues need be show stoppers, but the conclusions they arrived at were sufficient to suggest to me that it would take considerably longer to licence a lead cooled fast reactor than a sodium cooled one. If you agree with this conclusion, might it not be better to target molten salt coolants as natural heirs to sodium?

On a slightly broader topic, it seems to me that, in the medium/short term, we need a nuclear reactor that can produce power as cheaply as possible while providing reasonable assurances over safety. Breeding or isobreeding would be bonuses, but necessary in the longer term. However, I have gained the impression that, technically, we are not far off being able to achieve much higher conversion than currently possible with Gen 3 reactors at power production costs that are significantly lower (eg DMSR [MSCR] a la Le Blanc or pebble bed AHTR a la Peterson). Delays are primarily attributable to regulators. It would seem that regulators would potentially be able to arrive at decisions far more quickly if they were funded more generously and it was made clear to them by their Governments that it was their role to facilitate rather than block the provision of safe nuclear power. I think, too, that publicy-owned nuclear operators would offer a faster route than those owned privately. I am somewhat worried that the reactors being ordered now will become obsolete before they are time expired because of their higher power production costs than those which will be achievable by different designs in the near future. i would value your thoughts on these deliberations.


Douglas Wise: yes the fuel would then contain plutonium, uranium, thorium and beryllium. Enough plutonium to startup and have good burnup, enough U238 (depleted uranium) to keep the bred U233 at low enrichment levels (which is 13% or less for U233), enough beryllium for moderation and some alpha-neutron creation (good for delayed neutrons and proliferation resistance).

On proliferation I’d like to quote Lars Jorgensen:

You WILL get proliferation arguments. Sometimes these arguments logically lead to the conclusion that we should ban the existence of uranium – so just because an argument is put forward doesn’t mean that there is a real issue there.

You should also keep in mind that one “could” take seawater, extract uranium from it, enrich it, and then make a bomb. So watch out for arguments that one “could” – otherwise logically we’d have to ban the oceans.

The key issue is whether LFTR is more resistant than other choices a proliferator might make not whether some genius “could” make a weapon from the stuff.

One form of LFTR, Denatured Molten Salt Reactor – is the most proliferation resistant reactor around. In this reactor, fissile is always isotropically protected – which is the strongest protection possible. In addition, we have 232U which makes the electronics difficult and the broadcasts its location so you can’t sneak such a weapon into a country.

In general, the largest proliferation risk I see is enrichment technology. Building an enriched uranium weapon is much easier than a plutonium one. It is a difficult task to keep enrichment technology from spreading to those who don’t have it but want it. If we succeed in designing and building LFTRs that are iso-breeders (make just enough fissile to keep themselves running no extra and no shortage) and lower cost than LWRs then we eliminate the need and market for enrichment services. That will make it reasonable to buy out the companies with enrichment technology and shut down the enrichment business and redirect the company skills elsewhere. This would be a significant step forward in making proliferation more difficult.

Second, while there is plenty of 238U to isotropically protect fissile uranium it is difficult to do so for plutonium. The only recognized isotropic protection for plutonium is 238Pu which is very rare. The good news about LFTR is that 238Pu is exactly the kind of plutonium we generate (and not that much of it).

Third, dealing with existing plutonium. The once-through plans of the US mean that the 250 kg or so of reactor grade plutonium is planned to be buried. Buried is not the same as destroyed so there are those on the proliferation side that argue you have only succeeded in creating a rich ore deposit for some future maniac plutonium miner. Using LFTR we could fission off the plutonium – destroying it forever (while actually doing good).

Could we design a LFTR to serve as a weapons fissile generator?
Yes, but it would take a lot of development and current methods work just fine so there really isn’t a purpose to doing this.

Could a LFTR designed and operating as a power plant clandestinely support weapons material generation?
Less so than a solid fuel reactor since your fuel is all mixed together. I would envision a requirement that the plant sensor data be provided real time to the IAEA. We could then model the plant and know how much fissile is supposed to be present. The plant operates at 600C, in an intensely radioactive environment so it needs to be a robotic facility. I see no reason why it should not also be sealed with an alarm to IAEA. Doesn’t mean a country couldn’t break into their reactor but it does mean that they could not do so clandestinely.

Could a terrorist organization take over a power plant and use it?
Not at all easily. Assuming they overwhelm the local security force everyone in the world will know about it immediately so they had better finish up their work quickly before a real force arrives. But the reactor salt is hot, heavy, radioactive, sealed, and broadcasts its location (through the 232U decay product gamma emissions). So I think transporting the salt is out. Since things are chemically stable I also can’t see how a threat to “explode” the reactor would be effective. In contrast, if they succeeded in taking over an LWR they could threaten to cut off the cooling and start a zirconium burn.

Could a country repurpose their LFTR for weapons production?
Probably but it would take more work than building a purpose-built weapons production reactor AND everyone will know you are doing it whereas you can keep a purpose-build reactor hidden for a while.

In summary, the proliferation argument doesn’t stand (especially since it is used in countries that have or could easily have weapons already). It takes walking through the various scenarios and that isn’t pleasant thinking. There are some things we have to do in designing the reactor to make it proliferation resistant (like continuous sensor reporting). There are some options we have to forgo in some countries (like equipment to separate plutonium from fission products). But assuming we do the reasonable things then proliferation from LFTRs won’t be a problem.

You should also know that of the 70,000 or so nuclear weapons built to date, one test was done using 233U and then that route was abandoned.

One other argument I’ve heard is that in allowing a country to have nuclear power they gain the nuclear knowledge and practical experience and that builds a base for them to build on to start a weapons program. Some truth to that but then you can take MIT nuclear engineering course on line so knowledge about nuclear power isn’t controlled and knowledge about nuclear power isn’t the same as knowledge about weapons.

David is right though, LFTR is not proliferation proof (but neither is sea water).


Cyril, however, would apparently prefer to see distillation of spent fuel as a means of separating mixed heavy metal from fission products if that were possible. Is it likely to be and to what extent would it delay rollout? It would still be pyroprocessing since the term merely relates to the temperature at which it is carried out (robotically in a hot cell). Could you elaborate on the technical and economic implications of distillation versus electrorefining?

Distillation I think attractive because the IFR already needs to handle molten uranium fuel, to cast it into new fuel rods. If you have a system that can handle that, it is not a big deal to increase the temperature a bit and boil off the fission products. This leaves the actinides – plutonium, uranium, thorium – all together, that can then be cast into new fuel rods normally.

For further reading I’ll refer you to this thread:


1) Damage of components and fuel elements caused by corrosivity of heavy metals. (Possibly an obsolete criticism in view of your earlier dismissal of it).
2) Freezing in steam generators if high pressure feed water heaters fail.
3) Difficulty of repair, maintenance and refuelling at temperatures high enough to prevent freezing.
4) Potential blockage of coolant flow cross section in the fuel sub assembly caused by water/steam interaction with coolant.
5) Problems with production by lead of long lived radioactive isotopes (?Pb 205)

1. As answered before, simple niobium liners are sufficient. Also today’s most commonly used claddings, zirconium claddings, are compatible with lead.
2. If all of the high pressure feed water heaters fail, the turbine will trip, which will trip the reactor. It’s not hard to design the steam generators with sufficient natural circulation flow to keep lead temperatures well above freezing. In fact that’s a great way to get passive cooling to deal with the initial decay heat. Personally I think it’s disingenuous to complain about lead freezing being an engineering problem in steam generators; after all, sodium steam generators need various engineered provisions for steam and sodium isolation, to stop sodium water reactions after any tube rupture… such systems are much more demanding (they need to act fast) than simple freeze protection. A simple steam feed regulation system for the lead reactor would prevent overcooling.
3. The ELSY lead cooled reactor system uses cover gas phase manipulation and fuel shuffeling, by extending the top of the fuel rods with steel bars to above the free lead liquid level, and all components are replaceable in full. Keeping the temperature up avoids thermal cycling stresses, and allows you to use all the passively cooled decay heat removal systems – which need high temperature to operate – if you have eg a station blackout during refuelling. There’s not much need of maintenance anyway if you use niobium liners on the equipment.
4. Like the IFR, use a separate cooling loop to avoid pressurization/steam ingress into the primary loop.
5. The reactivity level of high purity lead coolant is very low. Lead is also not volatile, reactive or mobile. That means you can bury it pretty much anywhere and realistically not worry about it. Compare it to sodium, which is quite terrible; the initial radioactivity in Na-24 activation is similar to the fission products (!) and stays very high for many days. In any case, realistically there’s no such thing as a 100% pure coolant so impurities will be present, and there will be some corrosion products like chromium and iron, these will then activate, and present a radiotoxicity problem. Also keep in mind that fission products sometimes leak from cladding. These are highly radioactive and so make the coolant radioactive. There’s a lot of old sodium coolant still being stored because of the residual activity. To store it very long term you’d have to convert to a salt or something and store in a salt deposit.

I do agree that sodium reactors are a step ahead in tech development. But I don’t care about 10 years more or less, because I’m very sceptical any Gen IV is going to matter in the short term. So my perspective is more on the long term.



Thanks again.

On distillation. I read the discussion you linked to. It was mostly dealing with heavy metal salts whereby one one boils off and condenses what one wants and retains unwanted fission products in the still. By May 10th, you came up with a reciprocal process appropriate for distillation of metals rather than their salts. In this, temperatures would need to be much higher (circa 2000 deg C) such that fission products boiled off and actinides remained in the still.

I can quite see that this could be a very elegant approach. However, you yourself raised still lining issues, suggesting tungsten and doubting the suitability of pyrolytic graphite which someone else suggested. Further, Iain worried that fissile accumulation in still bottoms might give rise to criticality concerns and I don’t think you addressed this.

Is the metal distillation idea receiving any attention from any recognised laboratories? How far into the future must one look before deployability becomes a possibility?

You say that you are only interested in the long term because you don’t think Gen IV matters in the short. However, you failed to address my comments relating to the potential near term economic advantages of DMSRs or AHTRs. It could even be that the Prism might prove to produce power more economically than currently licensed Gen IIIs while simultaneously providing all its other benefits. .


Douglas: the criticality concerns with stills are easy enough to deal with; use multiple smaller stills that don’t have a critical mass.

It may also be possible to use chloride salt distillation, as well. Use a first still to boil off volatile fission products. Then a second still to boil off uranium and thorium chloride, and recover it. Then a third still to boil off the more volatile fission products, leaving plutonium chloride and noble metals. Noble metals can be simply drained or scooped off (being almost certainly insoluble in plutonium chloride).

But if metal distillation works that would be simpler, and all the actinides always stay together. I don’t know if anyone is actually working on it right now.

I didn’t discuss AHTR and DMSR. You ask so many questions its easy to miss several. But I like them and have discussed much about their great near term potentials as have many on the Check it out, lots of detailed discussion for people like you who are very interested in specifics. Both AHTR and DMSR don’t require online fuel processing which puts them a step ahead of IFR on at least that front. AHTR I think is deployable around 2020 in GW quantity if everything works out, DMSR hopefully following a decade later if all works out. But it’s still not available right now. For deployment right now we need AP1000s and EPRs. That’s fine by me – these produce coal free electricity and produce valuable plutonium for later startup of Gen IV.


Chlorides are lowest boiling of all the metal halides in most cases, including thorium and uranium.That leaves a small residue including Pu.
PuF6 and any other volatile fluoride can then be removed by fluoridation.
The LWR’s available right now can best use 20% LEU with thorium.

Click to access viewer

Any other fissile available could be used but uranium rich Australia could get uranium enriched in France or Russia. 20% LEU could also be used as initial metallic fuel in fast reactors. Russians are using 16.5% LEU as oxide/nitride in SVBR-100.


Douglas Wise queried whether we should be calling it “pyroprocessing” or “electrorefining”. Quite right, our usage does confuse. “Pyroprocessing” more accurately describes distillation of metal vapour, as CyrilR proposes, or the distillation of chlorides. It does describe the original processing of EBR 2 fuel, which evolved into electrorefining, while the book keeps using the old term.

However, the electrolysis of hot chlorides planned for the IFR would be more accurately described as electrorefining, so “electrorefining” would be clear to new readers. I will try to conform to that!


I think pyroprocessing is a more general term referring to any high temperature, dry (non-aqueous) nuclear reprocessing method. The IFRs chloride electrolysis may be electrorefining, but it is also a high temperature, dry, non aqueous process. But Roger makes a good point that we shouldn’t confuse too much over terminology. Electrorefining for IFRs, pyroprocessing for LFTRs.



Thanks for your continuing patience. Actually, I am already a regular reader of energy from thorium discussions, but don’t feel myself to be adequately equipped in technical knowledge to become an active participant.

I think it important, if one wishes to lobby for nuclear, to have some sort of a grasp of detail over the possible reactor options and, given a fair wind, their likely time to deployability. Having satisfied myself that worries over safety,waste and proliferation, though important, are now more or less solved from the technological viewpoint (albeit not from the political), it is the issue of power production cost that I regard as of extreme importance in reactor design choice.

There would appear to be two schools of thought over the cost issue. The first and more prevalent takes the view that the cheapest practical way forward is to go with current Gen III plus designs while diddling along with a bit of R&D aimed at eventually developing more advanced reactors.The second (espoused,among others, by Per Peterson) is that, in developed democratic states, one can expect little in the way of a nuclear renaissance unless and until nuclear power becomes significantly cheaper than that which can be produced from current designs. I am attracted to the second line of thinking, particularly as the inefficient use of fuel by current reactors may become more important with a decline in easily accessible uranium stocks such that fuel costs become more significant.

The attraction, rightly or wrongly, towards the second school of thought is predicated upon the belief that, given sufficient will, it should not take too long before more efficient and potentially cheaper reactors can be brought to market. Lobbyists for an isobreeding LFTR, I think, greatly weaken their case by pretending that its undoubted potential benefits could be available soon. Even if the will were present,I get the impression that one would probably be a couple of decades away at least.

Against this background, Cyril, I was very interested in and encouraged by your guesstimate that the PBAHTR, designed specifically to produce power significantly more cheaply than is currently possible from existing designs, was likely to be commercially deployable within a decade. Should such, indeed, be the case, it would seem sensible for nations such as Australia to leapfrog the current technology which may be becoming obsolescent by the time a decision is taken to go nuclear at all.

As a UK citizen, I am greatly attracted to the S-Prism idea, given our large plutonium stockpile and need for energy security. I am unimpressed with current plans relating to the manner in which it will be used, but that could change if pyroprocessing is demonstrated to be economically possible. It appears that this fast reactor technology is available now and it seems to me that there exists the possibility of its producing power more cheaply than from current designs. However, I accept that there are plenty of dissenters to the last point. The UK Government is keen to deploy nuclear power, but the willingness of the private generating companies to purchase reactors is far from strong ( for well understood reasons). It will therefore take until 2018 at earliest before we can expect to any new plant ready to go on line. Thus, even in the UK, the case for waiting for PBAHTRs is tempting, particularly if we are also going ahead with IFRs.

I would accept that I am voicing opinions which I am not really qualified to make. I would, therefore, welcome the views of others.


I don’t really see a big competition between Gen III that we have now like EPR and AP1000, and the PB-AHTR, DMSR, LFTR, IFR etc.

In fact they’re highly synergistic. The Gen III plants will displace disgusting coal plants on the near term, which is very important (China alone is building one or two coal plants a week, do we want this to continue another 20-30 years?). In so doing they make valuable transuranic elements that can be used to start up PB-AHTR, DMSR, LFTR, IFR, etc.

My primary worry about the IFR remains that sodium is bad publicity and it is bad for protecting the capital investment. That’s my main reason to look at lead and fluoride coolants for such a reactor in stead (but there are other more factual reasons such as superiour heat capacity, gamma shielding etc.)

I don’t think we should wait and keep putting the bogeyman on nuclear’s shoulders. That’s the worst decision we could make. It’s what we’ve done for decades, and it has resulted in tripled global coal use. Even if we go full steam on all the nuclear tech that we can get, it will be extremely tough to reduce global coal use over the next 50 years. With a more luxurious selective attitude, it’s a lost cause in my opinion.

That’s my perspective.


Leave a Reply (Markdown is enabled)

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s