IFR FaD 7 – Q&A on Integral Fast Reactors – safe, abundant, non-polluting power

Back in 2001, Dr George Stanford conducted an interview with the National Policy Analysis Center on the IFR. Nearly ten years later, in September 2010, George has updated this material, which is pitched squarely a layman audience. I post the updated version below (7-page printable PDF version here), since it fits like a glove with the IFR Facts & Discussion series that I am gradually developing. It’s probably the single best introduction to the IFR for the totally uninitiated — but there’s also real value here for the nuclear veteran, so everyone, read on!

Q&A on Integral Fast Reactors – safe, abundant, non-polluting power

By Dr George S. Stanford. George is is a nuclear reactor physicist, part of the team that developed the Integral Fast Reactor. He is now retired from Argonne National

Laboratory after a career of experimental work pertaining to power-reactor safety. He is the co-author of Nuclear Shadowboxing: Contemporary Threats from Cold War Weaponry.

What is the IFR?

IFR stands for Integral Fast Reactor. It was a power-reactor development program, built around a revolutionary concept for generating nuclear power—not only a new type of reactor, but an entire new nuclear fuel cycle. The reactor part of that fuel cycle was called the ALMR—Advanced Liquid Metal Reactor. In what many see as an ill-conceived move, proof-of-concept research on the IFR/ALMR was discontinued by the U.S. government in 1994, only three years before completion.

You might also see references to the AFR, which stands for “Advanced Fast Reactor.” It’s a concept very similar to the IFR, with some improvements thrown in. GE-Hitachi has the plans for a commercial version they call PRISM.

How was the IFR idea different from the concepts underlying traditional nuclear-power fuel cycles?

All of those fuel cycles were derived from technologies developed to meet special military needs: naval propulsion, uranium enrichment, weapons-plutonium production, and plutonium separation. Waste disposal has been approached as “someone else’s problem.” The IFR concept is directed strictly to meeting the needs of civilian power generation. It is an integrated, weapons-incompatible, proliferation-resistant cycle that is “closed”—it encompasses the entire fuel cycle, including fuel production and fabrication, power generation, reprocessing and waste management.

Do we need a new kind of reactor? What’s wrong with what we have now?

IFRs could reduce or eliminate significant difficulties that beset thermal-reactor fuel cycles—problems or concerns with:

* production and build-up of plutonium

* short-term management of plutonium

* disposition and long-term management of plutonium

* plutonium in national and international commerce

* other proliferation concerns

* long-term waste management

* environmental effects

* resource conservation

* long-term energy supply

* safety

What is a fuel cycle? What fuel cycles are there?

“Fuel cycle” refers to all the steps involving nuclear fuel that are needed to generate electricity: mining, milling, enrichment, fuel fabrication, reactor operation, reprocessing and waste management. Depending on the fuel cycle, some of those steps might not be needed. The three major fuel cycles of current interest are: thermal without reprocessing (“once-through,” or “throw-away”), thermal with reprocessing and IFR. The IFR will eliminate the need for mining (for centuries), and milling and enrichment (forever).

Who was working on the IFR? How far along was it?

The idea of the IFR originated at Argonne National Laboratory, which was about three years from finishing a study that was expected to establish firmly the technical and economic practicality of the concept. Progress had been spectacular. Design of the ALMR was being done at General Electric in San Jose, California. However, construction of a full-size prototype was not to be approved until Argonne’s research study had been completed and a current need was demonstrated.

What sort of reactor was the ALMR?

The ALMR was to be a “fast” reactor (one in which the chain reaction is maintained by high-energy neutrons)—so called because the energy spectrum of the neutrons is said to be fast.

Is there a slow reactor?

Yes, in concept, but it’s not called “slow,” it’s called thermal. Almost exclusively, current reactors are of the thermal variety: their chain reaction relies on thermal (slow) neutrons. In most of the thermal-spectrum reactors, the neutrons are moderated (slowed) by light water. Such reactors are called LWRs.

What is the most important difference in capabilities?

Probably this one: Inherently, thermal reactors are copious producers of plutonium, while IFRs can consume plutonium. In fact, two IFRs could consume the plutonium output of five LWRs of the same size, while generating electricity and bringing in revenue.

What’s so important about plutonium?

High-quality plutonium is the preferred bomb material for a sophisticated nuclear weapons program. It is even possible to make a nuclear explosive with low-quality plutonium, such as is found in power reactors.

What else can IFRs use for fuel, besides plutonium?

Their fast spectrum permits IFRs to burn any and all actinides from thorium on up. This is because in a fast neutron spectrum there are enough extra neutrons to convert the actinide isotopes that don’t fission easily into ones that do. The most important actinide elements are uranium (atomic number 92), plutonium (94) and, to a lesser extent, thorium (90). Since currently there is a growing glut of plutonium continuing to pile up from nuclear weapons and from thermal-reactor operations worldwide, the first IFRs will undoubtedly be fueled primarily with some of that plutonium.

How is that different from thermal reactors?

In a thermal neutron spectrum, many of the fission products and actinide isotopes absorb neutrons readily without undergoing fission (they have a high “capture cross section”), and the chain reaction is “poisoned” if too much of such material is present. Thus a thermal reactor cannot be a net burner of transuranic actinides. The main starting fuel for thermal reactors is a mixture of the fissile isotope U-235 (Pu- 239 can also be used), along with the fertile isotope U-238.

What in the world are “transuranic actinides?”

They are the elements beyond uranium—that is, their atomic number is 93 or greater: neptunium, plutonium, americium, curium and more. All of them are man-made elements, since they are so radioactive that the naturally created ones have long since decayed away in our little bit of the universe. They are also called higher actinides.

And what do you mean by “fissile” and “fertile?”

An isotope is called “fertile” when the addition of a neutron changes it into a fissile isotope—one that, like U-235, has a very high probability of undergoing fission when exposed to thermal neutrons. Both fissile and fertile isotopes are fissionable—it’s just that fertile ones require a high-energy neutron to make them split.

Burning and Breeding

What is a “breeder?”

A breeder is a reactor that is configured so as to produce more fissile material than it consumes. A fast reactor can be designed and operated to be either a net breeder or a net burner. A thermal reactor is a net burner of nuclear fuel, but—and this is very important—all thermal reactors are prolific breeders of plutonium.

What do you mean?

A thermal reactor starts out with no plutonium at all, and soon has a lot of it. In the process, though, it burns more fuel (mainly uranium) than it gives back as plutonium, and therefore is not called a breeder.

If IFRs can be either breeders or burners, why do some people insist on calling them breeders?

Partly for historical reasons (originally, fast reactors were investigated because of their potential to breed), partly because of genuine confusion, and partly for the emotional impact, since “breeder” carries the subliminal connotation of runaway plutonium production. The central fact that those people are missing is that with IFRs you can choose not to breed plutonium, whereas with thermal reactors you make plutonium whether you want it or not.

Then it is today’s reactors that are runaway producers of plutonium, and IFRs could put a stop to it.

Exactly.

What about the high-grade plutonium from dismantled nuclear weapons? Can we get rid of it?

Depends on what you mean by “get rid of it.” I suppose you could say we would have gotten rid of it if it were degraded to the point where it is as hard to deal with as the poor-quality plutonium in the used fuel from thermal reactors. That’s called the “spent fuel standard” for disposition of weapons plutonium.

How do you do that?

One straightforward way is to incorporate it in fuel for today’s thermal reactors. The fuel would consist of a mixture of the oxides of uranium and plutonium, called MOX. That process is now being started in the U.S. and Russia.

How long will it take?

There’s maybe 200 metric tonnes (1 metric ton equalling 1.1 standard tons) of weapons-grade plutonium in the world, most of which, we hope, will gradually become available for disposal over the next two or three decades. To process that much plutonium in 30 MOX-burning thermal-reactor plants (1000 MWe) would take approximately 20 years. Thus the MOX approach should be able to deal satisfactorily with the weapons plutonium.

Can IFRs help with this?

Eventually, when we get IFRs. When they do start up, they will of course begin consuming the accumulated plutonium— weapons-grade first, probably, if there’s any left, and then reactor-grade.

Commerce in Plutonium

You explained why the IFR is “fast.” Now, why is it “integral?”

“Integral” refers to the fact that the fuel processing facility can be an integral part of the IFR plant.

Is that important?

Very, if you are concerned about shipments of plutonium and spent fuel, or if you want to minimize national and international commerce in plutonium.

I think it is U.S. policy to discourage commerce in plutonium.

Yes, it is. And these days there certainly is commerce in plutonium—witness the controversy over shipments of plutonium from France to Japan a few years ago, and the recent controversy in England over a reprocessing plant at Sellafield. For the foreseeable future, and beyond, there will be no plutonium shipped out of IFR plants. The only shipments will be into them, from dismantled weapons and thermal reactors. Those are not extra shipments, but ones that otherwise would be to repositories. Thus the IFR all but eliminates commerce in plutonium.

How can that be?

An IFR plant will be a “sink” for plutonium: plutonium to be disposed of is shipped in, and there it is consumed, with onsite recycling as needed. Only trace amounts ever come out.

Safety

How safe are IFRs?

While the safety record of commercial reactors of Western design is superb, Three Mile Island notwithstanding, it would be desirable to have reactors that rely more on inherent safety features and less on engineered ones. ALMRs do that.

What is an “inherent safety feature?”

A safety mechanism that does not depend on human or mechanical intervention. For instance, ALMRs use metallic fuel rods, whereas LWRs use oxide fuel (as the Clinch River Breeder Reactor [CRBR] would have done).

Why are metallic fuel rods an inherent safety feature?

Metal is a good heat conductor, while oxide is a poor one. That means the interiors of the metal rods stay much cooler, which means that there is far less heat stored in an operating ALMR, which means that if there were a loss of coolant flow there would be much less heat present to raise the temperature of the fuel, which means that the consequences of a hypothetical accident would be much less severe.

Why is that?

Briefly, there’s a phenomenon called the “resonance Doppler effect,” which causes the reactivity to change somewhat with temperature. Because in an ALMR the temperature does not change much in a hypothetical accident, the reactor is much more stable.

O.K. What else?

ALMRs use liquid sodium for cooling and heat transfer, which makes the system intrinsically safer than one that uses water. That is because the molten sodium runs at atmospheric pressure, which means that there is no internal pressure to cause the type of accident that has to be carefully designed against in an LWR: a massive pipe rupture followed by “blowdown” of the coolant.

Also, sodium is not corrosive to steel like water is.

But doesn’t sodium burn in air and react violently with water?

Yes it does, and this of course requires prudent design, involving inert atmospheres and multiple barriers.

Not so fast! Seems to me there was a serious sodium leak and fire at a Japanese fast reactor.

You’re right. In December, 1995, at the Monju reactor, a temperature sensor broke and sodium leaked from a secondary sodium loop and caught fire. The plant was shut down, and has not yet been restarted.

How many people were hurt?

None.

Was radioactivity released?

No.

Was the reactor damaged?

No.

Was there any damage at all?

Yes. Some minor damage was caused by the burning sodium, and combustion products were spread through a portion of the building; cleaning them up took almost a year. The accident was classified as Category 1 on the international scale of 0 to 7 (with 0 being the least serious) by a committee of independent specialists.

So the sodium isn’t so safe after all.

When you think about it, it is pretty safe. There have been sodium fires, and undoubtedly there will be more. The Monju fire was a public-relations disaster, but did not even come close to being a public health threat. There is a great deal of industrial experience with liquid sodium, and there have been very few problems.

Well, I suppose that’s a risk we can tolerate, since we need electricity.

I think so.

What’s become of the Monju reactor?

Last I heard, it had commenced initial operations in May 2010, after a string of administrative and other non-technical snafus. But it’s not scheduled to go on the grid until 2013.

We were talking about inherent safety features. Are there any others?

The ALMR core sits in a pool of liquid sodium. In combination with the low heat content of the metal fuel rods, this means that, if there were to be loss of control power, the core would be cooled passively by convection.

Is this different than for other liquid-metal-cooled reactors?

Almost all the earlier fast reactors were of the “loop type”—relying more heavily on forced coolant flow—and also had oxide fuel, making passive cooling more problematic.

Wasn’t passive cooling tested in a prototype ALMR?

Yes, it was. All control power for the operating reactor was cut off. Coolant pumps stopped, control rods did not move, and the operators did nothing. The core temperature rose slightly, causing the reactor to go subcritical and shut itself down without incident. Unassisted convective cooling then prevented overheating.

Conservation and the Environment

What are the environmental considerations?

We already mentioned waste management. In addition, it can be argued that the major environmental problems with nuclear power are the consequences of the mining and milling operations. Because IFRs can use not only the surplus plutonium, but also the uranium (including U-238) that has already been mined and milled, they can eliminate for centuries any further need for mining or milling.

And of course, in common with all nuclear reactors, operating IFRs emit no carbon dioxide.

Do they put out any atmospheric pollutants?

None worth mentioning.

Then there some that aren’t worth mentioning?

Extremely small amounts of radioactive gas.

How small?

So small that there’s a lot more radioactivity from coal-burning plants.

You’re pulling my leg.

No I’m not. In coal there are trace amounts of radium and uranium, for instance, that come out of the smokestacks.

Then there’s dangerous radiation from coal plants?

No, there isn’t. It’s far below natural background levels. But nuclear plants put out even less.

Then I won’t worry. How do IFRs help conserve natural resources?

Thermal reactors are incredibly profligate with the earth’s endowment of potential nuclear fuel. The once-through, “throw-away” cycle in favor in the U.S. uses less than a hundredth of the energy potential of the mined uranium. Even with recycle, probably less than 1% can be extracted. IFRs can use almost all of it.

Wait a minute—less than 2% with recycle? I thought you could get nearly all of the energy that way.

Sorry, but you can’t. After two or three passes through a reactor, the fuel has gotten so contaminated with isotopes heavier than Pu-239 that various technical and operational problems arise. The only way to consume all of it is in a flux of fast neutrons.

I’ll be darned! Well anyway, with uranium so cheap, why do we care about conservation?

For the same reason we care (or should) about conserving petroleum, even though oil is now cheap. The current worldwide glut of reactor fuel is strictly temporary. Particularly with the U.S. throw-away cycle, the economically available U-235 is not predicted to last much longer than the petroleum reserves—a few decades.

Reprocessing & Proliferation

Thermal reactors with reprocessing would do at least a little better.

Recycling (it would be with the PUREX process, or an equivalent) could stretch the U-235 supply another few decades—but remember the consequences: growing stockpiles of plutonium, pure plutonium streams in the PUREX plants, and the creation of 100,000-year plutonium mines.

If you’re going to talk about “PUREX” and “plutonium mines” you should say what they are. First, what’s PUREX?

It’s a chemical process developed for the nuclear weapons program, to separate plutonium from everything else that comes out of a reactor. Weapons require very pure plutonium, and that’s what PUREX delivers. The pyroprocess used in the IFR is very different. It not only does not, it cannot, produce plutonium with the chemical purity needed for weapons.

Why do you keep referring to “chemical” purity?

Because chemical and isotopic quality are two different things. Plutonium for a weapon has to be pure chemically. Weapons designers also want good isotopic quality—that is, they want at least 93% of their plutonium to consist of the isotope Pu- 239. A chemical process does not separate isotopes.

I see. Now, what about the “plutonium mines?”

When spent fuel or vitrified reprocessing waste from thermal reactors is buried, the result is a concentrated geological deposit of plutonium. As its radioactivity decays, those deposits are sources of raw material for weapons, becoming increasingly attractive over the next 100,000 years and more (the half-life of Pu-239 being 24,000 years).

You listed, back at the beginning, some problems that the IFR would ameliorate. A lot of those problems are obviously related to proliferation of nuclear weapons.

Definitely. For instance, although thermal reactors consume more fuel than they produce, and thus are not called “breeders,” they inescapably are prolific breeders of plutonium, as I said. And that poses serious concerns about nuclear proliferation. And proliferation concerns are even greater when fuel from thermal reactors is recycled, since the PUREX method is used. IFRs have neither of those drawbacks.

Why does it seem that there is more proliferation-related concern about plutonium than about uranium? Can’t you make bombs from either?

Yes. The best isotopes for nuclear explosives are U-235, Pu- 239, and U-233. Only the first two of those, however, have been widely used. All the other actinide isotopes, if present in appreciable quantity, in one way or another complicate the design and construction of bombs and degrade their performance. Adequate isotopic purity is therefore important, and isotopic separation is much more difficult than chemical separation. Even so, with plutonium of almost any isotopic composition it is technically possible to make an explosive (although designers of military weapons demand plutonium that is at least 93% Pu-239), whereas if U-235 is sufficiently diluted with U-238 (which is easy to do and hard to undo), the mixture cannot be used for a bomb.

High-quality plutonium is the material of choice for a large and sophisticated nuclear arsenal, while highly enriched uranium would be one of the easier routes to a few crude nuclear explosives.

So why the emphasis on plutonium?

You’re asking me to read people’s minds, and I’m not good at that. Both uranium and plutonium are of proliferation concern.

Where is the best place for plutonium?

Where better than in a reactor plant—particularly an IFR facility, where there is never pure plutonium (except some, briefly, when it comes in from dismantled weapons), where the radioactivity levels are lethal, and where the operations are done remotely under an inert, smothering atmosphere? Once enough IFRs are deployed, there never will need to be plutonium outside a reactor plant—except for the then diminishing supply of plutonium left over from decades of thermal-reactor operation.

How does the IFR square with U.S. policy of discouraging plutonium production, reprocessing and use?

It is entirely consistent with the intent of that policy—to render plutonium as inaccessible for weapons use as possible. The wording of the policy, however, is now obsolete.

How so?

It was formulated before the IFR’s pyroprocessing and electrorefining technology was known—when “reprocessing” was synonymous with PUREX, which creates plutonium of the chemical purity needed for weapons. Since now there is a fuel cycle that promises to provide far-superior management of plutonium, the policy has been overtaken by events.

Why is the IFR better than PUREX? Doesn’t “recycling” mean separation of plutonium, regardless of the method?

No, not in the IFR—and that misunderstanding accounts for some of the opposition. The IFR’s pyroprocessing and electrorefining method is not capable of making plutonium that is pure enough for weapons. If a proliferator were to start with IFR material, he or she would have to employ an extra chemical separation step.

But there is plutonium in IFRs, along with other fissionable isotopes. Seems to me that a proliferator could take some of that and make a bomb.

Some people do say that, but they’re wrong, according to expert bomb designers at Livermore National Laboratory. They looked at the problem in detail, and concluded that plutonium-bearing material taken from anywhere in the IFR cycle was so ornery, because of inherent heat, radioactivity and spontaneous neutrons, that making a bomb with it without chemical separation of the plutonium would be essentially impossible—far, far harder than using today’s reactor-grade plutonium.

So? Why wouldn’t they use chemical separation?

First of all, they would need a PUREX-type plant—something that does not exist in the IFR cycle.

Second, the input material is so fiendishly radioactive that the processing facility would have to be more elaborate than any PUREX plant now in existence. The operations would have to be done entirely by remote control, behind heavy shielding, or the operators would die before getting the job done. The installation would cost millions, and would be very hard to conceal.

Third, a routine safeguards regime would readily spot any such modification to an IFR plant, or diversion of highly radioactive material beyond the plant.

Fourth, of all the ways there are to get plutonium—of any isotopic quality—this is probably the all-time, hands-down hardest.

The Long Term

Does the plutonium now existing and being produced by thermal reactors raise any proliferation concerns for the long term?

It certainly does. As I said earlier, burying the spent fuel from today’s thermal reactors creates geological deposits of plutonium whose desirability for weapons use is continually improving. Some 30 countries now have thermal-reactor programs, and the number will grow. To conceive of that many custodial programs being maintained effectively for that long is a challenge to the imagination. Since the IFR can consume plutonium, it can completely eliminate this long-term concern.

Are there other waste-disposal problems that could be lessened?

Yes. Some constituents of the waste from thermal reactors remain appreciably radioactive for thousands of years, leading to 10,000-year stability criteria for disposal sites. Waste disposal would be simpler if that time frame could be shortened. With IFR waste, the time of concern is less than 500 years.

What about a 1994 report by the National Academy of Sciences? The Washington Post said that the NAS report “denounces the idea of building new reactors to consume plutonium.”

That characterization of the report is a little strong, but it is true that the members of the NAS committee seem not to have been familiar with the plutonium-management potential of the IFR. They did, however, recognize the “plutonium mine” problem. They say (Executive Summary, p.3):

Because plutonium in spent fuel or glass logs incorporating high-level wastes still entails a risk of weapons use, and because the barrier to such use diminishes with time as the radioactivity decays, consideration of further steps to reduce the long-term proliferation risks of such materials is required, regardless of what option is chosen for [near-term] disposition of weapons plutonium. This global effort should include continued consideration of more proliferation-resistant nuclear fuel cycles, including concepts that might offer a long-term option for nearly complete elimination of the world’s plutonium stocks. [Emphasis added.]

The IFR, obviously, is just such a fuel cycle—a prime candidate for “continued consideration.”

Safeguards

You mentioned safeguards a while ago. Are you saying that IFRs need to be safeguarded?

Of course. Any kind of nuclear fuel cycle needs safeguards procedures, the most important job being to make sure that a power reactor is not operated so as to produce high-quality plutonium. The IFR is no exception, although it might be more easily safeguarded than other cycles.

Are there now any reactors that are not safeguarded?

Unfortunately, yes. A number of countries, such as India, Pakistan, and Israel, have not yet signed the Nuclear Non- Proliferation Treaty and do not permit all of their reactors to be inspected by the IAEA (International Atomic Energy Agency).

Why should there be IFRs if a country could expel the inspectors and make bombs?

Because a country could do that with any kind of reactor, and overall the IFR is by far the most proliferation-resistant nuclear fuel cycle.

Better than the thermal-reactor throw-away cycle?

Near-term, it’s comparable—quite possibly better. But when you factor in the long term (no plutonium mines), it’s the clear winner. And the IFR is far better than anything involving PUREX—which will inevitably be used increasingly in at least some countries, unless they go to IFR-type reactors.

Has a bomb ever been made that used reactor plutonium?

That is unclear. A U.S. weapons lab conducted an explosion in 1962 that made use of what was stated to be “reactor-grade plutonium,” the definition of which was different in 1962. While the details of that test and its results are still classified, some of the difficulties that would be encountered in making a weapon composed of reactor-grade plutonium are formidable and predictable. Dr. Peter Jones, who was Director of the Aldermaston weapons research establishment in England when they too tested some low-grade material, says the job is extremely difficult. We do know that, even with sophisticated weapons design, the explosive yield would be seriously degraded.

But suppose a country wanted to make some plutonium bombs, and had nothing but an IFR?

If their IFR plants were safeguarded, the material in the processing stream would be highly undesirable, as I explained earlier, and their chances of diverting it undetected would be slim indeed. If not safeguarded, they could do what they could do with any other reactor—operate it on a special cycle to produce good quality weapons material. But in either case, most likely they would do what everyone else has done: construct a special production facility, including a PUREX plant. Detecting such a clandestine facility is probably the main, immediate challenge facing international safeguards, and has nothing to do with whether a country has IFRs or LWRs.

But the uranium route to a simple bomb is so much easier that that’s probably the way a wannabe nuclear power would start out.

The Downside

There must be a downside. What is the single best technical argument against the IFR?

Isotopic composition of the plutonium. This is not at all a forceful argument, but it is probably the only one with any technical validity. (There are arguments, largely nontechnical, about whether the IFR is needed or whether it would be economical.)

How does that isotopic argument go?

In outline, like this: Designers of military weapons demand plutonium that is at least 93% Pu-239, although it is technically possible, with difficulty, to make an explosive with plutonium of almost any isotopic composition. Plutonium from LWR spent fuel runs around 60% or less Pu-239, while that from IFRs tends to be in the 70-80% range, and thus is somewhat closer to what weapons designers want.

Why isn’t that a forceful argument?

First, isotopic contamination is only one of many obstacles between a proliferator and a weapon from IFR fuel. I mentioned some of them a while back.

Second, having material that is 80% Pu-239 instead of 60% does not greatly lessen the difficulty of designing and building a bomb.

Third, and most important, remember that there are far easier ways to get fissile material for weapons—high quality material, at that—than from spent reactor fuel. Iraq, for instance, chose uranium enrichment. No country has ever used reactor-grade plutonium to make weapons.

Strongest Point

You mentioned the best argument against the IFR. What is the best argument for it?

Proliferation prevention. Near-term, the IFR makes PUREX illegitimate and plutonium inaccessible. Long term, it relieves future generations of the responsibility to guard the plutonium mines, and of the risks of not guarding them adequately.

There’s another huge benefit, of course. If nothing better comes along, the IFR can supply the world with pollution-free energy for as long as civilization lasts. Uranium becomes just as inexhaustible as wind or solar energy.

Since the IFR has so much going for it, research should be steaming full speed ahead, right?

Wouldn’t you think so? Nevertheless, at the Clinton administration’s urging, Congress terminated the research on October 1, 1994. The Senate voted to continue it, but the House prevailed in conference.

Well, I suppose at least we saved some of the taxpayers’ money.

Wrong again. Termination cost as much over the ensuing four years as finishing the research would have done, especially since the Japanese were all set to chip in $60 million.

You’re kidding. Why would our government do what it did?

Combination of factors, but the main one is plain misunderstanding of the facts I have just explained to you. Well-meaning but ill-informed people claiming to be experts confused pyroprocessing with PUREX, and convinced so many administrators and legislators that the IFR was a proliferation threat that the project was killed.

Isn’t it time to revive it?

I think it is. Other countries are working on their own fast rector designs—inferior, in my opinion, but this country abdicated its technological lead in 1994.

What should we do to get that lead back?

Needed first is one or two commercial-scale demonstration plants to finalize the details of the pyroprocessing techniques and to permit a better estimate of the probable cost of production-model IFR plants.

Add to FacebookAdd to NewsvineAdd to DiggAdd to Del.icio.usAdd to StumbleuponAdd to RedditAdd to BlinklistAdd to TwitterAdd to TechnoratiAdd to Furl

About these ads

152 Comments

  1. This all sounds good to a scientist or engineer, however, the IFR is still plutonium making machine. Plutonium is very ugly word in anti nuclear crowd and they cannot make any distinction between bomb grade plutonium and commercial reactor grade plutonium. They are convinced the bomb can be made from any plutonium and they effectively use that propaganda in their doomsday tactic. This article acknowledges such ridiculous possibility. It is like comparing mud and cement (both building materials) to construct skyscraper. You can use mud to do the job, however, the building will collapse before you are finished or it will be unsafe for occupancy, hence only total idiot would use mud for such job. The same goes for plutonium, nevertheless, it does not stop brain dead anti nuclear fanatics from preaching nonsense.
    In my opinion, the Thorium fuel cycle in Molten Salt Reactor does not produce the feared plutonium, therefore, it has better chance of public acceptance. Anyway, I believe LFTR it is better and safer technology. Uranium 233 was not yet linked to mass killing as is the case with U235 and Pu239

  2. Dr George Stanford has unfortunately left us with such a large number of problematic, and in some cases untrue statements that i could not list them all in a single comment. i will point out a couple of examples.

    in one shocking example, Stanford states, “thermal reactors consume more fuel than they produce, and thus are not called “breeders”. This is untrue for graphite or heavy water moderated thorium fuel cycle reactors, both of which can breed. The Shippingport reactor thorium experiment demonstrated that even Light Water Reactors can breed thorium.

    Stanford claims, “a thermal reactor cannot be a net burner of transuranic actinides.” Not true, thorium cycle thermal Molten Salt Reactors can be effective burners of transuranic actinides.

    I may post a detailed list of Stanford’s errors and problematic statements on Nuclear Green in a week or so.

  3. http://bravenewclimate.com/2009/11/29/ifr-fad-1/ Paragraph 4. Pretty much everyone here is on-board with both the LFTR and the IFR.

    As for the “shocking examples”, it’s obvious that Stanford is talking about existing thermal reactors. This is similar to some Thorium articles I’ve read where misleading claims are made about Thorium reactors being 200 times more efficient than Uranium reactors. Perhaps Barry could add an editors note adding that the LFTR & AHWR are exceptions?

  4. All human knowledge can be used for good or evil. As our energy storage and generation technologies become more potent dangers increase as well as the benefits.

    For example, dynamite is a truly dangerous material for storing energy, yet there is little irrational fear of it and “Nobel” is not a word used to frighten children.

    Plutonium and U235 deliver many orders of magnitude more energy per kilogram than dynamite does. It takes much more ingenuity to maintain sufficient control to the point that the benefits exceed the risks.

    There is widespread irrational fear of nuclear power owing to the association with nuclear weapons and radiation sickness. This blog can make a difference by educating the general public.

  5. I consider that Dr Stanford’s piece is harmful to the IFR cause for the following reasons:

    1) He is disingenuous about there being no alternative to breeding other than the use of fast reactors, contadicting the claims of the Gen IV Forum, of which he must be aware even if he doesn’t agree with them. He could, thus, be accused of deliberately misleading a reader who is totally unversed in the subject, though I doubt that was his intention.
    2) He has not considered whether breeding is necessarily essential for long term sustainability. Instead, he has assumed it withou justifying the assumption.t
    3) He has ducked any consideration of economics. While an accurate handle on these will have to await the construction of demonstration reactors and pyroprocessing units, there must be upper and lower bounds obtainable from modelling. How do these compare with equivalent modelling studies of MSCRs or LFTR thermal breeders and actual results LWRs, for example? We heard from Professor Peterson only recently that power from IFRs would be too expensive to compete with that from LWRs. Meanwhile, LFTR proponents claim that their technology could produce power for less than half the price. If nothing else, they appear to be winning the propaganda war.

    I want to support the rapid deployment of IFRs, but this article hasn’t helped. Are there better and more compelling reasons to do so than are set out by Dr Stanford?

  6. The claim about thorium reactors being 200 times more efficient is not too far from true statement. Present day reactors utilize only slightly over 0.5% of natural uranium without reprocessing, hence the reactor that would utilize 100% of fertile thorium or uranium to energy would be nearly 200 times more efficient as far as fuel utilization goes. Of course this can be done in Molten Salt Reactors or in Integral Fast Reactors.
    Charles Barton is right. The educated professionals should not make general statements to portray the other thermal reactor technology in negative limelight.

  7. Scott, even if we confine ourselves to existing reactor experiments, there is the Shippingport experiment, which did demonstrate thorium breeding in a thermal reactor. And if we confide ourselves to existing reactors, then we have to recognize that there has never been a IFR breeder. The notion that the IFR can breed, is based on theory, not established fact.

  8. If nothing else, they appear to be winning the propaganda war.

    With (or against) whom?

    Douglas, your other points:
    1) He means significant rates of breeding, that don’t imply multi-decadal doubling times.
    2) He has, but this is done elsewhere.
    3) No, he hasn’t. It’s just not clearly the first priority for Gen IV. The cost advantages of IFR or LFTR are, at present, only theoretical and based on some logical inferences. It is not currently a strong suit for either technology, despite what the LFTR advocates would claim.

    In short, I strongly disagree with you on all your points, Douglas.

    Charles, George is quite clearly talking about currently deployed commercial thermal reactors, i.e. LWRs and HWRs.

    The notion that the IFR can breed, is based on theory, not established fact.

    That is not correct.

  9. Douglas Wise and Charles Barton,

    Excellent as this blog is, it cannot hope to pick a “Winner” in the contest between IFRs and LFTRs. That is something that governments have presumed to do without much success.

    Today, both technologies have another chance and I would like to see the “Winner” picked by the market place rather than another Jimmy Carter (no to PUREX) , Bill Clinton (no to the IFR) or some bureaucrat like Milton Shaw (no to LFTR). My guess is that both technologies will have commercial success when some government provides encouragements rather than impediments.

    Right now Russia seems to be leading in the application of fast neutron NPPs and India in Thorium reactors. OK, I have probably overlooked a bunch of interesting projects but the US seems to be shying away from any pretense to the leadership position it once held in NPPs.

  10. gallopingcamel,

    I don’t really disagree. Picking winners is fraught with danger. I think the UK Government’s choice of Magnox reactors exemplifies this.

    However, while it would be good to see all possibilities competing in a free market, I don’t think any one nation has the time or financial resources to do so. By the same logic, you could and might argue that the simultaneous pursuit of wind, solar, hot rock , CCS and nuclear technologies is the ideal approach to the avoidance of the dangers associated with picking winners.

    One could argue that a global/international pooling of resources might allow the above approach, but that is unlikely to happen in the context of a world of competing democratic and autocratic states and multinational companies. I accept, of course, that there already exist strategic alliances which fall short of being totally global.

    As a biologist, I tend to believe in survival of the fittest (being a winner and not a loser) which, I know, is a view that is repugnant to many on the left. You cite Russia and India as exemplars of leaders in two different nuclear approaches and the United States as a laggard. Do you not think that the last stance is one which is likely to make American citizens big time losers? Might it not be better to enter the race, even if one happens to choose an inappropriate mount? If one could arrange a team competition to share the burden, well and good.

    All I am suggesting is that certain Western nations, given the emergency situation, may have no choice but to try and pick winners. In any event, it is a race that could well have more than one winner. What I do believe is that any populous, industrialised state that fails to rollout nuclear power very quickly, regardless of the technology chosen, will rapidly become a failed state or, at least, a state with a relatively impoverished and discontented population. A state or collection therof that chooses a correct winner (defined as technolgy capable of producing a source of plentiful clean energy that is cheaper than that which can be derived from coal) is likely to aspire to world leadership, to the benefit of its citizens.

  11. Barry, George is a scientist, and scientists should be expected to use logically precise language, Heavy Water reactors are capable of breeding thorium, thus even your justification of George’s error, attributes an incorrect view to George. As i understand it, the EBR-ii, which to date has been the only iFR prototype, operated in a converter range. i have no doubt that the IFR will breed, but that statement has not been verified by an IFR prototype operating in a breeding range. Since Scott’s criterion required a prototype operating in a breeding range, i pointed out that his justification of George’s error was not logically sound. Thus George did not follow Scott’s criterion in the cases of the iFR or thermal reactors, and I am not inclined to excuse George on the grounds that his mistakes were due to linguistic or logical sloppiness, and thus he should be cut some slack. As I already stated, George is a scientist, thus he should be held accountable for mistakes that arise from linguistic or logical sloppiness.

  12. Barry,

    The propaganda war to which I referred was that being fought for the hearts and minds of readers of BNC. As one such, who may or may not be typical, I read your site for my own education (which may be of little or no consequence, given my limited influence) and it is greatly to the credit of you and some of your other contributors that I consider that I have learnt a great deal.

    I think it would be fair to say that not only have you a strongly pro nuclear stance, but that you are also a strong supporter of the IFR as being the best contender for efficient breeding. This is not to imply that you don’t accept the need for Gen 2/3 technologies in the interim period, nor that you have any hostility to LFTRs, except insofar as you regard them as less efficient breeders and further from potential commercial deployment and thus a potential distraction.

    Just as I was coming to the conclusion that the IFR was, indeed, the Gen 4 reactor of choice, several recent posts have given me pause for thought:

    1) The need for nuclear electric to be as cheap or cheaper than coal electric has been made frequently and convincingly (Steve Hirsch and Peter Lang, for example). Yet the post describing the work of Per Peterson, described as a highly influential nuclear scientist, made the point that the common perception (in the States) was that electricity emanating from IFRs was likely to be significantly more costly than that from LWRs.
    2) Per Peterson himself was described as working on a molten salt converter (rather than breeder) reactor. He felt that he could get it to deployment much more quickly than any type of breeder reactor, that it would produce electricity significantly more cheaply than that from LWRs and that it be would more economic in its use of fuel. Furthermore, it was argued that his technology would expedite the subsequent development of full blown breeder reactors.
    3) We also had a post that gave the views of Dr Koch, which I found a little bit daunting to the extent that I was left with the impression that many more years of very expensive research would be needed before the IFR concept could be shown to be a goer and live up to the potential claimed for it by its advocates, as exmplified by the recently posted proselytising views George Stanford. (I accept that Koch’s essay was out of date and that there may have been a lot of development since he wrote it). While I wanted to believe Stanford and have no real reason to doubt what he said, I felt I was being leaned on too hard and thus somewhat resisted the sales pitch.
    4) I then started pondering the question of peak uranium and the need for breeder reactors at all (mindful of a post by Finrod). Do we even need break even reactors in the mid term? It seems to me that what is needed for next generation rapid rollout is a reactor that will: a) produce power more cheaply than current LWRs; b) use fuel significantly more efficiently than LWRs; c) has a very high prospect of rapid deployment because of less R&D costs and time.
    5) The above led me to the conclusion that, were I to be picking the next winner (which I doubt would greatly appeal to gallopingcamel), it would be some sort of molten salt converter. This is not to imply that IFRs or LFTRs or both shouldn’t follow along behind.
    6) George Stanford likens the LFTR to being analogous to keeping one’s money under the bed, the LWR to burning it while the IFR is more akin to depositing it in the bank and earning interest. However, this disregards the relative merits the low start charge/long doubling or even breakeven approach to that of the large start charge/quicker doubling. It is not obvious to me which would be the more economic in the next couple of centuries.

    I hope I have given you insights into my thought processes which appear logical to me (but maybe not to others). However, logic based on lack of technical expertise is what judges and politicians use when coming to decisions and the decisions are not invariably correct. Barry, you say you disagree with all my points. That is fine by me, but it would be helpful to have detailed reasons for your disagreements. You know a hell of a lot more about the subject than I. I am more than willing to be shown the errors I have made. Nevertheless, I still can’t help wondering why things aren’t moving a lot faster for nuclear breeder development than they appear to be, certainly in Western nations.

  13. Douglas, we’ve been over most of this already via email, and further, I’ve stated my position several times, in several ways, on BNC. The ‘answer’, in caricature, is that we need lots of new LWR for the near term and IFR (and maybe LFTR) for the longer term. What is impeding nuclear deployment of any flavour in the near term, most especially in West, is the waste issue (however much it may be imagined vs real), and the IFR is one of (probably THE) most economical and environmentally sound approach to resolve this issue. That is why folks like Ernie Moniz and others at MIT (e.g. http://web.mit.edu/mitei/news/spotlights/nuclear-cycle.html) are both quite right (on the need for ALWRs) and yet also dead wrong (on the need for getting moving on fuel recycling), and I’m frankly astounded they haven’t got the wisdom to differentiate between the technical and socio-political motivations on the latter point. Ergo, in the short-term, we need a 100 t/yr LWR spent fuel conversion facility and a PRISM demo to demonstrate resolution of the waste issue. Market forces and government policy will determine when commercial IFRs will be deployed but only after we build a lot more LMRs (e.g. at least 100 in the US alone and many more in China and Europe) and the aforementioned demo facilities.

  14. Scott, you state, “it’s obvious that Stanford is talking about existing thermal reactors. ” The minimal requirement for meeting the existential criterion would be a prototype. If a prototype has never existed, and there are no commercial examples, then the reactor would seem t fail your existential test. But Stanford makes many statements about the IFR, We can assume that to the extent that Stanford describes the outcome of ERR-II experiments he is describing things that meet the existential criterion, even if the EBR-Ii no longer exists. If Stanford makes statements that cannot be confirmed by reference to prototype tests, then Stanford is not talking about an existing reactor. If we throw out evidence deduced from the scrapped EBR-ii prototype, then little of what Stanford says about the IFR would be about “an existing” reactor.

  15. Barry, I am not here arguing against the IFR. I am simply pointing out that George Stanford says things about thermal reactors that are literally untrue. What i have pointed out is that inferring that Stanford was only talking about a limited case does not help, because the limited case argument logically should then be applied to everything Stanford says as a test of its validity. We all have to eat crow every now and then. Even though you like and respect Stanford, you do not help yourself or the IFR case, by defending statements which are on their face incorrect.

  16. Not many people (none?) on this website are nuclear engineers, thus I find the notion of some propaganda war between the LFTR and IFR extremely disingenuous if not counterproductive simply because nobody here actually has the authority to state which design is better. I think all we need to know is that GEN-IV nuclear shared characteristics between each design. This website is mostly about the IFR, but that doesn’t make it anti-LFTR or any nonsense like that. No such claims about the IFR being better than the LFTR or vice versa have been made. Honestly I find bickering back and forth about semantics rather annoying (only about 3 members are actually engaged in this too – the readerbase doesn’t care).

    Charles, I said “it’s obvious that Stanford is talking about existing thermal reactors”. I am unfamiliar with any ‘existing’ thermal reactor that is a breeder. Shippingport does not exist in 2010, neither does the AHWR (perhaps you definition of ‘exist’ is different to mine). Stanford never claimed the IFR eixsts. The statement that “A thermal reactor is a net burner of nuclear fuel” seems like sloppyness to me because it is of course true that thermal reactors can be designed to breed, and done this in the past. Like I said, I support an editors note adding ‘existing’ to the statement. I abologize in advance but I regard the rest of what you wrote as nonsensical jibberish.

    If you really get down to it, could you please correct the similar claims here:
    http://www.telegraph.co.uk/finance/comment/7970619/Obama-could-kill-fossil-fuels-overnight-with-a-nuclear-dash-for-thorium.html

    Thanks.

  17. Charles, I’ve asked George to come and post here to defend/clarify his statement. I’m not in a position to edit or add clarifications to his post.

    However, on a broader point, why are you concerned about whether MSRs are argued to be a 1:1 converter, or a slight breeder? The way your arguments logically run, surely a breeding ratio >1 is unnecessary for an MSR. Trying to up the BR to much above a full CR would only increase the fissile inventory required (without doing anything much to produce meaningfully short doubling times), and FI is surely what you’re trying to minimise. Please clarify.

    Scott – quite right, thanks.

  18. Douglas, your assessment “that many more years of very expensive research would be needed before the IFR concept could be shown to be a goer and live up to the potential claimed for it by its advocates” is probably correct. The Indians have launched a second generation FBR program,. That program would appear to develop at least some IFR features (ie. metallic fuel) through a multiple prototype stage that will last at least 15 years. Argonne IFR research has focused on the IFR as a burner not a breeder. The highest breeding ratio for an IFR design i have found coming out of Argonne is 1 to 1.07, which is comparable to the ORNL MSBR. Argonne can probably develop a high breeding ratio design, but that will take considerable time, and probably the most economic approach would be to throw in with the Indians.

    There is research from both Oak Ridge and Russia, which suggests that the MSRs can be competitive with IFRs as actinide burners. Part of the LFTR sales pitch will be that you can dump nuclear waste into a bunch of LFTRs and get a heck of a lot of electricity in return.

    As for development time, I would be willing to bet you a sushi dinner that were LFTR and IFR R&D given equal financing, the LFTR would come to market first,

  19. Barry,

    Thanks for your prompt and instructive reply. I was particularly interested in your assertion that it was primarily the waste issue that was holding back nuclear development in the West. Apparently, the inside view, therefore, is that proliferation, safety, cost and sustainability concerns are all down the pecking order in the socio-political stakes. I had rather gained the impression from some Pugwash members and Dr Stanford that proliferation concerns would be higher ranked.

    That said, I can follow the logic of ALWR deployment being eventually followed by IFR (or LFTR breeder) rollout at a later stage. What I find difficult to understand is your apparent opposition (if I have interpreted you correctly) to an interim phase as exemplified by David LeBlanc’s proposed DMSR. The claims for it are based on the possibility of more rapid deployment than that possible for breeders (because reprocessing could be delayed for 30 years) and cheaper power – similar, in fact to those of Per Peterson. Given that reprocessing is purported to be simple and could take place, as I mention, after 30 years, the waste issue is addressed as well by this route as by your preferred one.

    I must say that I have become preoccupied more by costs than by any other factor. I am not sure that, even were safety concerns to be addressed, nuclear rollout would rapidly follow in the West due to private equity investor uncertainty, high discount rates and very slow rate of return. Governments will need to give strong and unconditional support before much progress is made.

    It seems that you yourself have a vision of LWRs being joined by several hundred LMRs – to be up and running before one can move on to the pyroprocessing side and close the fuel cycle. To the extent that there are already several LMRs up and running now, would it be possible for you to comment on their relative economic merits vis avis LWRs? If they are likely to produce power more cheaply when construction costs settle down, I can see the merits. However, if they aren’t, I can’t help making the anology to the building of new CCS ready coal plants, probably an outrageous statement, but one I couldn’t resist making.

    Do you believe that, in order to overcome the socio-political problems associated with nuclear rollout in the West, its generated power will have to become more expensive than that from levy-free coal? Closing the fuel cycle clearly has great attractions, but at what cost? What of Peter Lang’s oft stated views? I can’t believe that it would be a socio-political masterstroke to set out deliberately to make nuclear power less safe as Peter demands, seeing it, as he apparently does, as an essential means of making nuclear deployable. I had been desperately hoping that new nuclear technology would result in the strong possibility of generating power more cheaply than from LWRs while, at the same time, providing the other benefits discussed above. Are my hopes groundless?

  20. Cost is important to investors, so will ultimately be a strong determinant for governments, utilities and financial institutions. Proliferation concerns are something that must be dealt with politically, via the IAEA etc., and my view is that most of the public do not consider this to be a major factor, even though it is still the #1 point used by the anti-nuclear protest organisations (well, perhaps equal #1, along with radiation scare mongering, which is related, somewhat tangentially, to the waste issue). Safety has been satisfied to most people’s satisfaction, but again, radiation scaremongering is used to raise this as a potential spectre. Waste management and the related issue of perceived sustainability of supply are, to my mind at least, the “big ones” for the public. But perhaps social and political researchers can shed more light on this matter than I. My opinion is just that – mine. I will answer your other questions at some point later.

  21. Douglas, Of the issues that turned me into a Molten Salt Reactor/LFTR advocate in 2007 were nuclear costs, and scalability, although by the time I completed my initial analysis, I was convinced that the MSR solution would offer superior safety, and a solution to the problem of nuclear waste. Later analysis was to reveal that the MSR offered solutions to the problems of industrial heat, intermediate and peak power, while offering in costal areas desalinization. i found multiple paths to nuclear cost saving, for example recycling old coal fired power plants as LFTR sites.

    From my perspective, you have focused on the right questions. Until recently i was in agreement with Barry that the short term energy solution solution, but after attending David Le Blanc’s presentation at ORNL in May, and discussing the uranium fueled MSR option with Kirk Sorensen, i now think that non-breeder uranium fueled MSRs can be developed quickly, and that they would be a low cost alternative to LWRs.

    In addition DMSR type reactors could serve as a bridge to the LFTR.

  22. Charles,

    Thank you for your replies. I find your case persuasive, but, as I have repeatedly stated, I’m not sure that my opinion is worth a great deal, lacking as i am in any relevant technical expertise.

    You suggest that you became an MS reactor advocate in 2007. However, given your father’s background in Oak Ridge developments, would it not be fair to suggest that your advocacy owed a certain loyalty to his beliefsand possibly predated 2007? This is not a criticism, merely an indication that i am attempting to remain impartial. My background as an academic research worker, albeit in a different field, has inculcated into me a deep sense of cynicism. Furthermore, I have always found that the best way for me to learn is to play devil’s advocate, provided, of course, that the subject i’m interrogating is prepared not to bre offended by what, at first sight, may appear to be an offensive line of questioning.

  23. Barry,

    I shall look forward to your future more detailed replies. i really do want to believe that some solution or other will be available for the crisis that the younger generations face. I’m satisfied that it has to be a nuclear one, but remain open minded beyond that.

  24. Douglas, since my father played a significant role in the development of the light water reactor, and indeed regarded this as the professional accomplishment that gave him the most pride. What lead me to the molten Salt reactor was first a question about big a LWR build out could be. It is 3 years later and no one has come up with a doable large scale LWR buildout that would offer even a significant part of the 2050 solution.

    I thus decided that that alternatives were needed. I looked at 2. Robert Hargraves was advocating the Pebble Bed Reactor in 2007, and I seriously considered it, before I decided that the LFTR was a far better candidate for a rapid buildout than the PBMR.

    I knew about MSR technology because i was still living in Oak Ridge during the MSRE, and we read about it in the news paper.. i really talked to my father more about MSRs in 2007-2008, than i had during the 1950’s and 60’s. What drew me to the MSR was its simple design, compact size, high operating temperature, safety, and potential for solving the nuclear waste problem.

    Actually i began to appreciate my father’s contribution to MSR development after i got better acquainted with its design features and the history of its development.

    There are undoubtedly some brand loyalty issues, but since my father contributed to the the development of several reactors, i have some choice regarding which brand to be most loyal too. Brand loyalty would seem to preclude criticizing the LWR, but that has not stopped me.

  25. One response given by Dr. Stanford that strikes me as particularly disingenuous is in regards to the question “But doesn’t sodium burn in air and react violently with water?” ….to which he answers “Yes it does, and this of course requires prudent design, involving inert atmospheres and multiple barriers.”
    At best, this addresses only the “burn in air” concern: How do you put “inert atmospheres and multiple barriers” between the water and sodium in a steam generator ?
    While its true that the secondary sodium circuit is non-radioactive, so was the sodium fire at Monju, that put the reactor out of commission for fifteen years.
    Such an outage would be TERMINAL to a commercial operation.
    Given the choice, the inert salt in MSRs is by far the prefered option — not just because such accidents are totally ruled out, but also because expensive safety features involving “inert atmospheres and multiple barriers” are simply avoided.
    In the latter case (MSRs) there is a chance for “energy cheaper than coal”. In the former ( IFR/ALMR ) there isn’t such a chance.

  26. The Candu breeds thorium if you want it too. It also burns PWR waste with very little processing.

    I still don’t get why we even need to talk about breeding. The worlds supply of nuclear waste has so much energy in it, it could supply all the world’s energy needs for centuries.

    David LeBlanc’s DMSR is so simple. it seems obvious that such a simple reactor can be developed and in operation in a couple of years in some skunkworks operation. I’m sure Charles know who is up to something in that regard.

    Wouldn’t it be great though if the equally unfunded no money Focus Fusion, or the Polywell folks breakthroughs occurred as hoped, and the nuclear fission went the way of the buggywhip. We’ll know a lot more about this time next year.

    Meanwhile the attorneys that rule us are dumping all our development money into thousand year old solar and wind tech, hoping for what ?

  27. It would be useful to add these questions and answers:

    When is it reasonable to expect to see the first IFR’s start commercial production of electricity?

    Which countries are likely to be the early adopters?

    Once in production, at what rate p.a. could we reasonably expect to see IFR’s added to the international production inventory?

  28. For NPPs to make a real contribution to reducing CO2 emissions we need to build them in huge numbers (a nuke per day). Other than on this blog I don’t see much interest in building NPPs at such a scale.

    For a dramatic build up to have any chance of success the economics have to be right. If China can build an NPP for $1/Watt compared to $5/Watt in the USA, don’t expect to see the USA in a lead position.

    Likewise, Leblanc points out that a rapid build can deplete the fissile material stocks needed for “Start Charges”. If one picks the wrong technologies it is simply impossible to have a really rapid build up of NPPs owing to this constraint alone.

    I have confidence that engineers and scientists will make good choices on the above two issues but will they get the chance?

    The main problem at least in the “West” is public opinion. Right now public opinion is against NPPs, so I hope that blogs like this can make a difference.

  29. Likewise, Leblanc points out that a rapid build can deplete the fissile material stocks needed for “Start Charges”. If one picks the wrong technologies it is simply impossible to have a really rapid build up of NPPs owing to this constraint alone.

    How much fissile material is required for the start charge, per reactor?

    There is quite a lot of surplus plutonium from weapons stockpiles in the world.

    There’s also HEU from weapons stocks, and the small U-233 stockpile in the U.S., for LFTRs.

    There is also quite a lot of used LWR fuel in the world, which can be reprocessed for its transuranic content.

  30. Luke, for a 1000 GWe LFTR starting with a U-235 start charge, 450 kgs. For a 1000 GWe LFTR starting with RGP, a start charge of 800 kgs. For a LMFBR with a RGP start charge of from 10,000 to 15,000 kgs. For an Indian combined thorium-uranium fuel cycle LMFBR 4000 kg RGP.

  31. I would like to see an abbreviated comparison table of IFR versus MSR that lists the major issues. Evidently that includes mass of start charges, possible composition of start charges, operating temperature and pressure of primary and secondary fluids, containment issues, expected long run operating cost compared to pulverised black coal, unresolved metallurgical problems, onsite processing method, waste disposal volumes/storage time and weaponisable isotopes production.

    It would be good if both sides could agree to the contents of such a table without throwing in numerous afterthoughts.

  32. Luke, for a 1000 GWe LFTR starting with a U-235 start charge, 450 kgs. For a 1000 GWe LFTR starting with RGP, a start charge of 800 kgs.

    Okay, a couple of questions:

    1. Why the difference between U-235 and RGP? If RPG was 20% Pu-240, then that doesn’t imply a near doubling.

    2. Where are you getting this information? When has a LFTR ever been run on RPG? What was the size of the experimental MSR run by ORNL in the 1960s and what was it’s fissile loading?

    Sorry Charles, but my bulls!it detector has gone off a number of times in relation to your comments, as you seem to be unable to separate speculation of POSSIBLE performance of some future MSR vs actuality.

  33. Also, for the theoretical LFTR ‘breeder’ as opposed to burner or converter, what are the start charges? Barry Brook asked this above and was ignored by you. If you go to a BR of 1.05, what is the start charge then? What about LeBlanc’s dream of 1.13? (which still implies a very long doubling time).

  34. Wow Charles, I didn’t expect that. You haven’t answered ANY of the questions I posed.

    That makes me thing you either don’t know, or don’t like the answer. This is really contrary to the way you harp constantly on Nuclear Green about being open, and the way you like to try and flay Barry over here on IFR matters.

    Quite disappointing.

  35. Rapid Rabbit, you are mighty free with accusations of Bullshitting. If you are going to slander me, you need to back it up.

    If you want information feel free to ask, but without being nasty. I do not take gratuitous insults well. People who rely on insults are intellectually lazy, and are seldom interested in serious responses. You got the answer you deserved.

    Bye the way my source start up charges for LFTRs and LMFRBs is David Le Blanc. I found the Indian FBR start up charge in Indian press accounts. There are numerous papers and reports devoted to MSR breeding, For example, NUCLEAR CHARACTERISTICS OF A lOOO-MW(e) MOLTEN-SALT BREEDER REACTOR By J. R. Eogel, H. T. Kerr, E. J. Allen, or ORNL-4812, the Developmental Status of Molten Salt Breeder Reactors, which calculates the breeding ratio of a single fluid MSBR to be, 1 to 1.071, with an uncertainty of + or – .016.

    I also suggest that you read WASH-1097 for a better understanding of the thorium fuel cycle and of thorium breeding. Finally I suggest you loose your attitude and learn how to treat people with respect. You might get a little respect in return.

  36. Charles, I apologise for the use of that word and for the offense so taken. I actually was alluding to the ‘smell test’ turn of phrase of De Volphi, who is quoted extensively on Nuclear Green and talks about this when referring to Lovins’ rubbish. I thought your claims also smelled and so I asked those questions and made reference to male cow pies.

    Beyond that unfortunate word, I don’t think I was offensive. I was surprise by the way you evaded the question and your ongoing sensitivity. I thought you had thick skin.

    I am not disputing the possibility of MSR breeding. I am asking what the relationship is between the BR and the start charge. I think, from my reading, that it is significant, but you don’t explicitly acknowledge this when on the one hand touting LFTR as a breeder and on the other hand waxing lyrical about its low fissile inventory.

    As to your response above, where does David Le Blanc get his LFTR start charge estimate from, given that he has never run a LFTR or any other MSR as far as I’m aware?

  37. when I first read about IFRs (see stanford comment below my comments), I was excited about the waste reduction (inseparable from efficient use of fuel) on the one hand, and the reduction of “disposal time” from 10,000 years etc. to under 500.

    but when I read that, I didn’t really understand half lives and thought that the long lived waste was more dangerous. GIven that the long lived waste is much less radioactive, what precisely is there to be worried about with the long lived waste?

    under what circumstances is the long lived waste “appreciably radioactive”? How much long lived waste would you need and in what proximity to human beings to give them an “appreciable dose”?

    On this list elsewhere, we have dismissed the thousands of years fear. Yet here it is being renewed. Is there really a difference of opinion here? It seems to me we should not cater to radiation fear to sell the IFR.

    Perhaps no one is doing that, but then I’d like to know more about the radiological dangers of this longer lived waste.

    g (see stanford below)

    Are there other waste-disposal problems that could be lessened?

    Yes. Some constituents of the waste from thermal reactors remain appreciably radioactive for thousands of years, leading to 10,000-year stability criteria for disposal sites. Waste disposal would be simpler if that time frame could be shortened. With IFR waste, the time of concern is less than 500 years.

  38. as a follow up, why talk of “100,000 year plutonium mines,” unless the intent is to engage in radiation fear?

    plutonium can be made into atomic bombs so such material would need safeguarding, always. not just for 100,000 years.

    which means the 100,000 year comment suggests that plutonium really is a 100,000 year radiological hazard.

    I don’t think we should cite this material and bernard cohen’s realism about plutonium (often referenced on BNC) fears side by side without comment or criticism. they can’t both be right. or if they can, some mediation is required between the two views.

    In short, what my two posts really point to is perhaps Doug Wise’s sense that IFRs are sold at the expense of current nuclear power–I think this is a disabling dichotomy we should avoid.

    How make the case for a gen three transition to gen four if we’re helping to spread the usual fears about waste, radiation and proliferation?

    On proliferation, primarily a political problem, Dr. S almost looks like he’s talking out of both sides of his mouth (I don’t think this is intentional)–treating reactor grade plutonium as a proliferation worry on the one hand and then in other answers, showing it to be a minimal worry.

    Look at the following:

    Then it is today’s reactors that are runaway producers of plutonium, and IFRs could put a stop to it.

    Exactly.

    What about the high-grade plutonium from dismantled nuclear weapons? Can we get rid of it?

    Depends on what you mean by “get rid of it.” I suppose you could say we would have gotten rid of it if it were degraded to the point where it is as hard to deal with as the poor-quality plutonium in the used fuel from thermal reactors. That’s called the “spent fuel standard” for disposition of weapons plutonium.

    Me again:

    In nearly the same breath, Dr. S talks of runaway plutonium dangers, only to reveal in the next comment that this plutonium danger is virtually already not a problem due to “poor quality.”

    am I missing something, or does this indeed appear to be either double talk or serious lack of rhetorical care?

  39. The abbreviation RGP (RPG?) is being thrown around. Pardon my ignorance, but what does it mean? It’s presumably a kind of fuel. A quick google suggests Royal Gibraltar Police, Role Playing Game, Rocket Propelled Grenade and such. Maybe the last one could be converted into fuel …

  40. If our unburned fuel is vitrified within storage casks, once we have IFRs or LFTRs can this fuel be extracted from the cask to burn in these reactors? Seems to me like it would be very difficult to de-glass?
    Howard

  41. Having said that, I think George Stanford’s Q&A is a great achievement in communicating the technical aspects of the nuclear fuel cycles to a lay audience.

    I think complicated ideas are actually rare in science, but complicated explanations all too common. This is a great piece of technical writing.

  42. Rapid Rabbit, Apology accepted. I discuss start charges here:

    http://nucleargreen.blogspot.com/2010/01/ifr-breeding-lftr-breeding.html

    David Le Blanc briefly discussed IFR “PRISM” start charges here:
    http://memagazine.asme.org/Articles/2010/May/Too_Good_Leave_Shelf.cfm
    David states:
    “a sodium-cooled fast breeder such as the PRISM design requires as much as 18 tons of either U-235 or spent fuel plutonium. ”

    The figure i attributed to David came out of his Mar 2010 ORNL lecture. My link to the video of that lecture does not produce any sound, but I think that the start up charge range, found in these two links is sufficient to establish that I was not bullshitting.

  43. am I missing something, or does this indeed appear to be either double talk or serious lack of rhetorical care?

    You are missing something, but perhaps in this case George is too close to the subject and so didn’t provide a full explanation. But that is modestly technical, so you can appreciate is reticence in presenting this to a lay audience after reading my explanation below.

    Reactor Grade Plutonium contains a high proportion of Pu-240 and Pu-241, via neutron capture of Pu-239 (itself formed from U-238 neutron capture) during the reactor’s operation. Bomb makers do not like this isotopic mix (“poor quality”), because the spontaneous decay of isotopes higher than Pu-239 make a device ‘hot’ and unstable (e.g., Pu-240 releases 910 spontaneous neutrons per gram per second, whereas Pu-239 releases 0.022, i.e. ‘bomb quality’ Pu is 41,300 times less active).

    However, Pu-240 has a half-life of 6,560 years, and its decay product, Am-241, has half-life of 430 years. Pu-241 is a mere 14 years. By comparison, the useful fissile Pu-239 is 24,100 years (and it decays to U-235).

    So, let’s say you have the RPG with an initial composition of 70% Pu-239, 28% Pu-240 and 2% Pu-241. After it has sat in a repository for 30,000 years, the new transmuted composition is U-235 = 40%, Pu-239 = 30%, Pu-240 = 4% and Pu-241 = 0%. Separate the U-235 chemically and you have Pu with perfect isotopic quality for bomb makers. You also have U-235, another useful bomb material.

    That’s why George refers to long-term repositories eventually being “Plutonium mines”.

  44. Although ORNL’s Molten Salt Reactor (MSR) was shut down circa 1980 their estimates for start charges for LFTRs are widely accepted. What we need now are full scale advanced reactor programs.

    Both the IFR and the LFTR appear to have the potential to allow a rapid “roll out” without running into “start charge” constraints.

    We need to see many commercial size advanced reactors operating for several years before a massive roll out can be planned with real confidence.

    One of my fears for advanced reactor design is the danger of attempting too much. For example the F-111 project suffered from introducing the swing wing and a titanium skin on the same aircraft.

    I get the same uneasiness when I read about high temperature LFTRs that use the Brayton cycle to increase thermodynamic efficiency. It would make more sense to take a more modest first step by building advanced reactors that can simply replace or augment the heat sources in existing coal fired power plants.

  45. greg myerson,

    As Barry points out, the RGP mines will mature like a fine wine so that after 30,000 years they will be a wonderful source of bomb grade fissiles like U235 and weapons grade Plutonium (>93% Pu239).

    Personally I don’t think that any nuclear waste rich in Actinides will last 100 years let alone 30,000. As they say in Yorkshire:
    “Where there’s muck there’s money”.

    As pointed out by Dr. Stanford, the IFR (and the LFTR) can “burn” Actinides as fuel. Once we adopt more advanced reactor technologies, yesterday’s problem (nuclear waste) becomes tomorrow’s opportunity (cheap fuel).

    Incidentally, LFTRs produce only tiny quantities of Plutonium compared to Uranium reactors but I don’t see this as a major selling point for LFTRs.

  46. thanks barry:

    I (understandably) misunderstood the plutonium mines point. learned something too. didn’t know pu 239 decayed to u 235.

    Your explanation clears that up nicely.

    but I still think the piece needs some clearing up on the waste/radiation issue: the issue of what constitutes “appreciable” radioactivity that needs 10,000 year sequestering.

  47. thanks gc:

    btw, I think the plutonium mines point can be clarified, as you both do here, in the original.

    I don’t think the explanation for how RGP becomes potential weapons grade plutonium is that out of sync with the rest.

    couldn’t you use the u235 for bombs as well–while we’re at it?

  48. Greg, you ask,

    am I missing something, or does this indeed appear to be either double talk or serious lack of rhetorical care?

    Yes, there is something you have missed. You’re correct that the longer the half-life, the less dangerous the substance is (per pound). The “plutonium mine” reference is not to radiological hazard, but to the usefulness of the plutonium for making bombs. While the reactor Pu is not attractive to bomb-makers now, its isotopics will make it much less radioactive and generally better for bombs in 100,000 years (if there is anyone around then who cares).

    There’s no conflict whatever with Bernard Cohen — he’s a very careful investigator, and right on in what he says. The concern about the hazards of carefully handled nuclear waste is way overblown.

  49. Greg, you wrote,

    but I still think the piece needs some clearing up on the waste/radiation issue: the issue of what constitutes “appreciable” radioactivity that needs 10,000 year sequestering.

    As Cohen has pointed out, the activity of spent LWR fuel in the repository in 10,000 years is no realistic danger whatsoever. But still, there’s a court decision that declares that the depository must be secure for a million years. Go figure.

    With IFRs, the long-term activity (per GWe of energy generated) in the repository would be lower by a factor of ~100.

    I’ll see if I want to revise my piece to make that clearer.
    — George

  50. I’ll have to learn to be less hasty in hitting the Submit button. I said, “BTW, Pu-239 is fissile in its own right. Its slow decay to U235 is irrelevant.” Wrong. Barry is correct, in that there will be perfectly good U235 down there too.
    — George

  51. Howard asked:

    If our unburned fuel is vitrified within storage casks, once we have IFRs or LFTRs can this fuel be extracted from the cask to burn in these reactors? Seems to me like it would be very difficult to de-glass?

    I don’t think anyone is proposing to vitrify spent fuel from U.S. reactors. The fuel pins would be deposited in the repository intact, so they would be recoverable, at least for a while. Separated fission products, however, might be vitrified for permanent disposal.
    — George

  52. George S.Stanford
    Thank you again for this excellent piece, which I find to be pitched exactly at the right level for lay people like myself and friends and acquaintances who have now read it. I would suggest that, for your target audience, you don’t need to change anything. John Morgan (above) – I’m glad to see you agree with me about the worth of this brilliantly written article. Complicated scientific writing is not needed, nor warranted for a IFR Q & A aimed at the general public and these are the people we need to convert.

  53. George:

    Your post was an excellent and informative introduction to the benefits of the IFR for the uninformed layman.

    As a slightly less uninformed layman, I wonder whether you could explain your reasons for believing the IFR to be the Generation 4 reactor design of choice.

    I suspect that this may have a great deal to do with its potentially superior breeding potential. If such should be the case, what do you make of the recent MIT report which claims that sustainability of nuclear fuels should not be considered a major issue except in the very long term?

    If, alternatively, your preference for the IFR stems from the fact that you believe it to be the Generation 4 design than can be soonest deployed on a commercial scale, could you give your best estimate of when one might expect to see a commercial demonstration prototype (reactor and reprocessor) up and running should R&D funding be made available immediately? Also, could you give a guesstimate of the scale of the funding necessary. Finally, if one disregards the R&D funding, FOAK costs etc, do you anticipate that the LCoE from commercial IFRs will be higher or lower than from ALWRs and by what extent?

    I ask these questions in view of the alternative options available which your post did not discuss and would be very grateful if you could expand a bit on your original contribution.

  54. gallopingcamel, I now favor quick roll outs of transition designs. in the case of the IFR, the ARC-100, a reactor that is closely based on the EBR-II prototype. The ARC-100 converts in something like the 0,87 range. A Molten Salt Reactor that operates on a uranium fuel cycle and is closely based on the MSRE, with a goal of producing 600 C or 650 C, would be practical in the short hall and would cost half the cost of a Light Water Reactor to build. If a full court press approached to cost lowering were applied the cost of the UMSR could be lowered to as little as 1/4th the cost of a LWR. it would be practical in the near term. The full court press method would involve recycling the sites and part of the facilities of old coal burning power plants, the adoption of rapid manufacturing/construction technology, underground reactor housing, and other cost lowering measures.

  55. Charles,

    To the extent that my opinion is worth a damn, I ,too, have been working my way towards believing in the need for “quick rollouts of transition designs”.

    My judgement is based on the fact that they appear to have greater potential to produce power significantly more cheaply than ALWRs and to become deployable more quickly than IFRs or LFTRs. They also address waste and proliferation issues.

    Do you have views on the likely relative costs and times to deployability of the ARC-100 and UMSR? On the face of it, both would seem highly desirable and there would be no necessity to push for only one type unless the relative differences in cost or deployment time are going to be very great.

  56. gallopingcamel said:
    “Once we adopt more advanced reactor technologies, yesterday’s problem (nuclear waste) becomes tomorrow’s opportunity (cheap fuel).”

    Spot on. A few years ago I saw nuclear energy as perhaps being just a portion of the short to medium to term mitigation options for climate change (i.e. by helping to displace fossil-fuels). Learning of Gen IV reactors was a major turning point in my thinking on this, well before I came to realise just how central energy is to virtually everything civilisation relies upon.

    I think Barry is quite correct in stating that perceived waste and fuel availability issues are the biggest constraints to public acceptance of nuclear power currently. The latter was my particular concern. I strongly believe using the “yesterday’s problem, tomorrow’s solution” type marketing pitch is a strong point for nuclear advocacy. And I know from talking to people (friends, colleagues etc.) about nuclear power, this is always the point of conversation that’s a bit of a “game-changer”.

    I should also add, I think this is a fantastic introductory piece of work by Professor Stanford, one which I have shared with people already, and intend to share with others. Thank you for reproducing it here for us Barry, and to Prof Stanford creating it!

  57. Douglas, If the ARC-100 design is modified to produce power through a steam turbine, it could be on the market within 10 years. Ditto for a UMSR. Both will need production facilities and very good construction engineers in order to creatr the reality.

  58. Douglas: You say,

    I wonder whether you could explain your reasons for believing the IFR to be the Generation 4 reactor design of choice.

    I’m trying to look at our probable energy needs farther into the future than most people do, and see efficient breeding as absolutely essential by the latter part of this century. And to get there, we need to start deploying breeders as soon as possible. Right now, the only efficient breeder close to commercialization is the IFR. If in the future something like the LFTR can demonstrate comparable breeding capability and is otherwise competitive, then we should start deploying those also.
    In more detail, here are comments on some things I see that you wrote earlier:

    1) [Stanford] is disingenuous about there being no alternative to breeding other than the use of fast reactors, contradicting the claims of the Gen IV Forum, of which he must be aware even if he doesn’t agree with them. He could, thus, be accused of deliberately misleading a reader who is totally unversed in the subject, though I doubt that was his intention.

    Of the six Gen IV candidate technologies, three are fast reactors with a closed fuel cycle, one uses a thermal, once-through cycle, one (the SCWR) claims to have a fast-spectrum version with a closed cycle. So the “claims of the Gen IV Forum” that you mention must refer only to the MSR, mentioned above. Maybe there is an MSR concept with an appreciable, reasonably assured breeding gain, and if there is, I’d like to hear about it. (Charles? But please don’t cite Shippingport, which did achieve break-even operation, but its breeding gain was so marginal that it barely covered the losses in fuel reprocessing and refabrication.)

    2) He has not considered whether breeding is necessarily essential for long term sustainability. Instead, he has assumed it without justifying the assumption.

    You can do your own calculation. Start with the rules-of-thumb that (a) 1 GWe-year of LWR energy requires 150 – 200 tons of elemental uranium to be mined, and (b) that 1 GWe-year of LWR energy discharges ~20 tons of spent fuel containing ~200 kg of transuranics. Then pick a growth rate for LWR-powered nuclear energy for the next century (starting with ~375 GWe this year, globally), and figure the resources needed and the “waste” generated.
    In my scenario, with nuclear providing most of the energy for a stabilized global population of some 10 – 11 billion by 2110, problems are really piling up by 2060 or so without breeding.

    BTW, the self-sustained growth rate GR of a reactor type is given approximately by

    GR = [HC x (BR-1)]/FI,

    where HC is the amount of HM consumed per GWe-yr (~1 tonne),
    BR is the breeding ratio (number of fissile atoms created per heavy-metal atom fissioned),
    and FI is the fissile inventory in the reactor plant.

    Relevant here is something Chris Sanderson wrote earlier:

    It would be useful to add these questions and answers:
    a. When is it reasonable to expect to see the first IFR’s start commercial production of electricity?
    b. Which countries are likely to be the early adopters?
    c. Once in production, at what rate p.a. could we reasonably expect to see IFR’s added to the international production inventory?

    All very good questions.
    a. When? Before deployment begins, there has to be a commercial-scale demonstration, mainly to iron out some residual uncertainties in scaling up the pyroprocess, but also to demonstrate (to Charles Barton’s satisfaction) the operating characteristics of the reactor itself. Also to permit a better estimate of the eventual economics of the IFR system. That *could* be done in 5or 10 years, but when it will be done is still an open question. Seems to me that we’ll be lucky to have a few operating IFRs by 2030, and LFTRs (if they pan out) by 2040 or 2050. But then, my accuracy with the crystal ball not all that great.
    b. Where? Could be India, China, Russia, South Korea, or the United States — all of which except the last have a fast reactor program or are seriously considering one, and have expressed an interest specifically in the IFR technology. Unfortunately the U.S. lead in the technology of nuclear energy was abdicated in 1994, and there’s scant indication of any determination to try to recover it.
    c. Rate of increase? Ramping up a new industry is a gradual process. Once about 10 IFRs are operating, the fleet will grow at a rate determined by energy demand, resource limitations, and competing energy sources. We can expect construction of LWRs to taper off as IFRs (and/or LFTRs) come on line. IFRs can grow rapidly until the supply of fissile plutonium from LWRs dries up, and thereafter at a rate of about 5% per year (determined by the breeding gain).

    First, what we urgently need is a demonstration of the commercial viability of the IFR system. Current cost predictions range all over the map. From where I sit, I see no reason why the IFR should not be very competitive, given its waste handling advantages (repository savings); the lack of need for mining, milling, and enrichment of uranium; it’s sequestration of plutonium; and the fact that pyroprocessing is far easier and cheaper than the PUREX-type processing needed for oxide fuel.

  59. Apologies for this somewhat off-topic question, but since Dr. Stanford has been kind enough to respond here:-

    I’ve seen a number of statements on the web to the effect that IFR advocates make unsupported assertions about IFR capabilities, and never publish the underlying data. Information on MSRs, in contrast, is publicly available in the original ORNL official reports.

    Elsewhere, I have seen claims that all information on the IFR was subjected to the US ‘born secret’ classification for nuclear-related information, and has been declassified only to the extent of giving some details to GE for developing PRISM. Publishing details would therefore be a violation of GE’s intellectual property rights, if you’re lucky, or of US national security laws if you’re not.

    Does anyone – particularly Dr. Stanford, of course – know what the actual facts are here?

    Thanks

  60. Luke-UK:
    There’s nothing classified about the IFR technology. Some of the (perhaps important) tricks-of-the-trade exist only in the minds of the aging people who did the work. GE might see some of the information they developed as proprietary. From Argonne’s point of view, the status of the technology as of 1994 is described in detail in

    Hannum, W.H., ed. “The Technology of the Integral Fast Reactor and its Associated Fuel Cycle.” Progress in Nuclear Energy, Volume 31, Number 1–2 (1997).

    What do you want to know?

    — George

  61. George,

    I’ extremely grateful for the lengths to which you went to my to respond to what might have been deemed to be an impertinent line of questioning.

    It does appear that you regard the IFR to be vitally important from the viewpont of sustainability and consider that its potentially superior breeding performance relative to that of the LFTR, along with its probable advantage in terms of nearer market readiness, give it the edge.

    I am in no position to disagree with your evaluation, lacking, as i do, any experience of the subject. However, could I prevail upon you to give your views on the pros and cons of transitional technologies (say ARC-100 or UMSR) that would render ALWRs redundant before breeders could be put in place? My reading suggests that these may be deployable more quickly than breeders and push back your presumed sustainability date beyond 2060 (which you think would obtain if we stuck with LWRs in the interim), It is also claimed that they could produce electricity more cheaply than LWR while addressing waste and proliferation concerns (for what they’re worth). In short, do you agree or, alternatively, consider them to be a distraction?.

  62. Douglas:
    What I regard as vitally important is efficient breeding. And it’s urgent to start deploying breeders as soon as possible. Meanwhile, let’s let the market decide what reactor types are deployed.. The ARC-100 is a small reactor based on IFR technology, and deploying that would be a valuable boost for the technology, IMHO. ( I don’t know what the “UMSR” is, and Google doesn’t help me.)
    Most of the reactors built in the near future will probably be LWRs of some sort (plus some HWRs), presumably Gen-III. And that’s fine. The more of them there are, the more fissile material we’ll have to start up the Gen-IV reactors when they mature.
    I don’t see building any kind of reactor as a “distraction.” The major mistake would be to hold off on demonstrating the IFR technology while waiting for R&D to come up with something better.

    BTW, I don’t know why the format of my earlier post seems to be screwed up. I thought I was doing everything right.
    — George

    Ed: I fixed the earlier comments

  63. George Please, ORNL Researchers did a very systematic investigation of the MSR breeding potential. You should know this. They did it repeatedly, from the late 1950’s till the late 1970’s. They looked at one fluid cores and two fluid cores. They looked at different carrier salts and different core geometries, thorium content, and fissile inventories. There work was more than credible, after all their boss Alvin Weinberg was looking at their results, and Weinberg understood core nuclear processes as well as any scientist of his generation. Since then the ORNL findings have been recalculated by the Reactor Physics Group of the University of Grenoble. The RPG findings fully supported the earlier ORNL findings.

    i believe that the ORNL and RpG scientists used methods similar to those employed by Argonne scientists to determine IFR breeding ratios.

    I will site you some examples, In ORNL-4541 on Page 33, a table reports the breeding ratios for three different single fluid MSBR designs. Breeding ratios of 1.051.1.055, and 1.060 were reported. Did the ORNL scientists just make this up? Or did they make errors in their calculations? If they made it up, you should write a professional paper exposing the fraud and, If they made calculation errors you should also write a professional paper exposing them. Be sure to get your paper on MSR breeding in a reputable Nuclear Physics Journal, and not just on BNC.

    Unless and until you can write a definitive paper exposing ORNL and RPG MSR breeding calculations as errors or frauds, your claims lack credibility.

  64. George:
    Thank-you for your reply, and the reference. My immediate interest is in how clean the fission product / actinide split can be made. Much of the public hostility to nuclear power seems to relate to the long lived waste. Greenpeace never fail to raise this issue. ‘Clean’ fission products will decay to radiotoxicity no worse than the uranium you started with in about 300 years, which is a much easier sell than 100,000. However, leaking even 1% actinides into the waste stream invalidates this claim.
    I think Barry intends to post about this (and other aspects of IFR technology, like fissile inventory in the reactor and reprocessing loop, breeding ratio, deployment rate…) later, so that would be a better place for further discussion. Meanwhile, I’ll see if I can find your paper.

  65. George, There are numerous technical documents from ORNL and the University of Grenoble Reactor p
    Physics Group, that address the MSR breeding issue. i have no intention to plow through as large stack of documents, so one example should be sufficient. ORNL – TM – 1467 Table 10 compares breeding ratios for a MSBR design with and without protactinium removal. Without the removal the ratio is 1.0491, with removal the ratio is 1.0713, about the same as the maximum breeding ratio for recently published Argonne IFR designs. If the ORNL calculations were not reasonably assured, why should we consider the Argonne calculations to be better assured?

    Based on the indian proposed development program for a metal fueled FBR, it is going to take more than one IFR prototype to get all of the issues involved in a high breeding ratio IFR sorted out. The IFR to date has been designed to be a converted and not a breeder. Given the amount of work that has gone into the design of the IFR burner, it is quite obvious that a lot more will be required to make the IFR breeder work satisfactorily. . The IFR breeder will require a multi-billion dollar development program that will take years to reach fruition. You hint at what I have heard, that an unacceptably high amount of plutonium is lost to waste in the IFR fuel processing stream. How much is that going to cost to fix? What if the problem cannot be fixed?

    I am not opposed to IFR development, and I am supportive of the ARC-100 commercial reactor project, i
    just happen to think that IFR advocates are acting more like snake oil salesmen, than serious advocates for a particular nuclear technology. The IFR is far from developed. You know that. It may well turn out to be far more expensive to develop the iFR breeder than the LFTR. And commercial LFTRs may well sell for half the price of a commercial iFRs, or even less. You need to honestly face the issues and let the chips fall where they may.

  66. Charles Barton, on 21 September 2010 at 19.24 Said:
    “gallopingcamel, I now favor quick roll outs of transition designs. in the case of the IFR, the ARC-100, a reactor that is closely based on the EBR-II prototype.”

    Sounds like an approach that could be implemented quickly using materials that are well understood. I think you support my point about the F-111; introducing more than one major innovation in a development project often leads to serious problems and delays. Solving the problems of high temperature materials for driving an efficient Brayton cycle heat engine should not be part of the early LFTR design objectives.

    George Sanford said that LFTRs would roll out in 2040 or 2050, more than ten years behind IFRs. Probably LeBlanc and the ORNL folks would say the exact opposite. Personally, I would not like to see the proponents of these reactors wasting much energy fighting amongst themselves when the real opposition is coming from the general public and the entrenched nuclear bureaucracies.

  67. Galloping Camel wrote:

    Personally, I would not like to see the proponents of these reactors wasting much energy fighting amongst themselves when the real opposition is coming from the general public and the entrenched nuclear bureaucracies.

    Amen. There’s no reason that I can see to object to development work on the MSR. What does bother me is to see some people arguing that the MSR’s potential to be a competitor of the IFR should delay the overdue demonstration of the latter’s commercial viability. The market for nuclear power will be big enough for a number of designs to compete in and find their niche.

  68. I was just reading through some of the comments further up this thread, and I noted that dialogue between Rapid Rabbit and Charles.

    Was there a final resolution on the relationship between fissile inventory and BR in the MSR? That is, if you have a BR of 1.05, or 1.13, as opposed to a CR = 0.95 or 1.0, what does that do to the required FI? I would guess that the relationship is strongly non-linear.

  69. “commercial viability” are indeed the magic words ! ….thanks for mentioning them, George.
    Of course as we all know, commercial interests are risk averse : they will not bet investors’ money on things that can put them out of business in a hurry — like the 15-year shutdown of the Monju reactor sodium fire.
    I still haven’t seen anything explaining how “inert atmospheres and multiple barriers” are to be installed between the water and sodium in a steam generator.
    Non-combustible fluoride salts are a far safer bet, from a commercial viability point of view — its a total waste of time to even try convincing anybody otherwise !

  70. Jaro, as you know, the 15 year shutdown of Monju was not on technical grounds. The Russians were famous for restarting their BN-600 reactor the same day as a sodium leak was detected and repaired – multiple times. So let’s use the commercial viability word here – the BN-600 has among the highest availability factors of any reactor in the Russian fleet, the Russians are now building a BN-800, and the Chinese just bought 2 x BN-800s off the Russians, starting construction in 2012.

    The multiple barriers between the non-radioactive sodium and water in the heat exchanger are, most simply, double-walled steel piping with high quality welds, as was installed at EBR-II. The heat exchanger is in an argon-filled chamber. There was never any contact as a result.

  71. Jaro:
    I go along pretty much with what Barry said in his response to your sodium concerns. I’m sorry I didn’t respond until now to your earlier comment along the same lines. You said:

    One response given by Dr. Stanford that strikes me as particularly disingenuous is in regards to the question “But doesn’t sodium burn in air and react violently with water?” ….to which he answers “Yes it does, and this of course requires prudent design, involving inert atmospheres and multiple barriers.”
    At best, this addresses only the “burn in air” concern: How do you put “inert atmospheres and multiple barriers” between the water and sodium in a steam generator ?

    Why ascribe nefarious motives when simple incompetence might be the explanation? Perhaps I should have mentioned steam-generator leaks, but they didn’t occur to me at the time. Their possibility is well recognized, and the matter of single-walled versus double-walled piping is being debated. As you say, there’s an economic trade-off. However, such steam generator leaks are relatively trivial, as industrial accidents go. As Barry said, the Russian BN-600 had a number of such accidents, which took out one of the three heat-transfer loops, and the reactor was running while the damaged steam generator was repaired. Sloppy engineering gets the blame for the accidents, and industry learns from such events. In any other industrial field, the occasional accident much more serious than those is regarded as a normal part of doing business, and attracts little attention when it happens. The occasional steam-generator leak, if it occurs, is no public hazard, and will be a small part of the operating cost, not a make-or-break consideration.
    You also said,

    While it’s true that the secondary sodium circuit is non-radioactive, so was the sodium fire at Monju, that put the reactor out of commission for fifteen years.

    That is indeed seriously is indeed seriously misleading, since the fire itself put the reactor plant out of service for only a year or less. What kept it shut down until just recently was a combination of management snafus and political factors.

  72. Barry Brook wrote:

    Was there a final resolution on the relationship between fissile inventory and BR in the MSR? That is, if you have a BR of 1.05, or 1.13, as opposed to a CR = 0.95 or 1.0, what does that do to the required FI? I would guess that the relationship is strongly non-linear.

    . Barry, the required FI (fissile inventory) is determined by factors unrelated to the BR (breeding ratio). The BR depends on the reactor design. Charles is correct in observing that the MSR’s low fissile inventory is a point in its favor — about ten MSRs could be started up for every IFR. And if the MSR were to have a breeding ratio of 1.05, that would put it in the same growth rate category (5% per year) as the IFR with a BR of 1.5.
    . Both types of reactor — and others — should have a chance to show their mettle. Right now, the fact that GE has an operational design makes the IFR much closer to demonstration of commercial viability, and therefore we ought to proceed, while expecting competiton to develop down the road.
    . Note, BTW, that MSR (Molten Salt Reactor) is a generic category, not a specific reactor type. Its a category that includes the LFTR and a number of others. The MSR community seems to be divided as to which concept is the most promising. Even the LFTR seems to have variants that differ significantly. The Hargraves/Moir variant has no net breeding capability, whereas I gather that Charles Barton knows of LFTRs with BR’s of 1.05 and up.

  73. Technologies that at first sight seem to be competing often turn out to be synergistic. For example the introduction of intercontinental telephone links via satellites seemed to threaten the dominance of submarine cables, the established technology.

    Instead of eating into the market for submarine cables, the satellites caused an explosion in demand that would not have occurred without the “competition” that satellites provided. The cable business blossomed as never before.

    Similar synergies may apply to the IFR/LFTR “competition”. Bear with me, I may not be as crazy as you think.

    IFRs need rather large “Start Charges” and that will be a real advantage when there are not enough Gen IV reactors to make much impression on the growing inventories of nuclear waste. Eventually, the price of U235 and Plutonium will rise and then the start up economics of LFTRs with their small “Start Charges” will become attractive.

    If LeBlanc is right the core of a two fluid LFTR will be a cylinder ~2 meters in diameter, scalable according to rather simple rules. It is easy to imagine such reactors being built in small sizes in factories and delivered to site on trucks.

    Maybe 30 years from now most of the large reactors being delivered will be IFRs and the small ones will be LFTRs. By using a mix of large and small reactors the complexity and capacity of distribution networks will be reduced.

    During WWII the US built 2,700 “Liberty Ships” each weighing almost 10,000 tons. The peak production rate was ~3 ships per day. Building a nuke per day should be a much easier task especially as the USA would not be doing it alone.

  74. gallopingcamel, it’s funny that you mention those IFR/LFTR synergies. There is an interesting French study, done a few years ago, that discussed exactly that:

    “Scenarios for worldwide deployment of nuclear power” (2005)
    http://hal.in2p3.fr/docs/00/04/33/43/PDF/scenarios_full.pdf

    There is also this:
    “Synergy between Fast Reactors and Thermal Breeders for Safe, Clean and Sustainable Nuclear Power”
    http://www.worldenergy.org/documents/p001515.pdf

  75. There appears to be growing consensus among contributors here that a mix of future different reactor technologies will best serve global interests and this might well be the case. Eventually, it is anticipated that a few of the best will survive and that the chaff will be removed from the wheat – all good free market competition, guaranteeing the optimum outcome.

    While I don’t discount this approach, I think it might be worth considering matters from a different perspective, given the vast R&D expenditure required to develop any one design and the long delays before any return is going to become possible.

    Suppose one puts oneself in the position of a fictitious CEO of, say, GHE. The company makes and sells ALWRs and fuel assemblies for them. Its boffins have also developed a breeder reacor design, and are pressing for more research funding. Would one be serving one’s investors better by giving the boffins the money or shutting them up? I don’t think the answer is obvious. What may be obvious is that the public interest might benefit from a decision to fund whilst the company might not. The ALWR income stream, requiring much less further investment, could be prejudiced and the eventual breeder, if licensed, might never get sold in any quantity if a cheaper and equivalent alternative emerged. All these are awkward decisions that the CEO must contemplate and explains why most in dominant companies would prefer cartels to true competition.

    Our fictitious CEO might decide that only one other company has the potential to develop an alternative breeder. He decides that his ALWR income stream might eventually suffer even if he blocks his own breeder development if the competitor goes ahead and succeeds. Possibly a merger or partnership then becomes attractive as a means of insuring against risk. Thus, with double R&D spending, a high chance of one of the designs succeeding and the ability to recoup both lots of spending by building them into the price of the succeeding product, one might get the best of both worlds.

    However, this cosy scenario is too good to be true. A few nations will decide to get involved and use their taxpayers’ money to develop their nuclear industry. Various designs are explored and, because picking winners isn’t guaranteed, some will fail to the detriment of their taxpayers. However, the nation that gets it right will put itself into an extremely powerful position because it will have spent less on R&D, can control its regulators to get domestic startups accelerated and then move on to export at competitive prices once FOAK problems are ironed out.

    Perhaps our CEO decides that he should forget about his private competitor and attempt to forge alliances with any nation that is thinking of backing his company’s- type design, thus giving them a head start and saving them time and money.

    In the real world, all these combinations seem to be taking place in a fairly leisurely fashion and might lead to design optimisation. However, those who worry about peak oil and AGW may consider the approach to be too slow to achieve the optimisation in time to do any good. Hence, some are calling for a “war footing” or even an integrated global approach to expedite matters. Possibly, while global agreement on emissions control has failed, agreement on joint R&D funding on clean energy solutions might have more chance of succeeding and even of being effective.

    I have no answers but would like to hear the reasoned views of others. It does seem clear, however, that technological answers are certainly available, but that it is all the problems associated with financing that are constraining nuclear rollout in liberal democracies and that this will be to their long term economic detriment.

  76. @Douglas Wise

    I think your point about international cooperation to finance development of advanced reactors is very valid. If it can be done for ITER, then why not for fission? There does not seem to be any inherent reason why not – other than the usual political anti nuclear stuff.

    It’s not really an issue of money. The US military budget is something like $600 billion. Say 2% of that – $12 billion per annum – would surely be sufficient to see both sodium cooled fast reactors and molten salt reactors off to a flying start. The US could easily do it alone – but spread around some of the G20 countries it would be more than affordable.

  77. George, given the lower required fissile inventory for MSRs, it would be possible to start MSRs capable of producing the entire united States electrical supply, using the current stock of RGP held at United States Reactors. Current United States Government stockpiles of U-235 and Pu-239, would be sufficient to start enough LFTRs to power the entire United States energy economy. Thus there would be no need for higher breeding ratios,, and indeed breeding ratios higher than 1.01 might create proliferation issues.

    LFTRs can fill roles that are not envisioned for IFRs, for example powering ships, or providing heat for many industrial processes including hydrogen and ammonia production.

    I have suggested on a number of occasions that the best path forward for American fast reactor development would be a partnership with India which recently embarked on a metal fueled fast reactor program. The Indian plan focuses not only on the breeding of plutonium, but also on the breeding of thorium. The U-233 produced in Indian fast breeders will go into low fissile inventory thermal reactors. Since the Indians lack a large RGP inventory, it works better for them to use FBRs to create start up charges for their thermal breeders. Given the history of government support of FBRs inthe United States, the indian path is more likely to yield commercializable results.

    Finally, I do not envision large US government LFTR program. it was revealed at the FHR workshop at ORNL on monday that uranium fueled Molten Salt cooled reactors have the potential of having the cost of NPP produced electricity. MSRs can at least equal this using uranium fuel. The technology for these reactors is already largely in the can. Thus it will not require a large R&D program to bring these reactors to market. Once they reach the market, fluoride cooled, and liquid fluoride fueled reactors will quickly make the LWR obsolete.

    Profits from the sale of transitional uranium fueled reactor may well fund LFTR research. The primary rational for the LFTR is not its breeding but how rapidly it can be deployed. The LFTR offers a technological route to the2050 goal of reducing global carbon emissions by 80%, without destroying the energy basis of industrial society.. The IFR can play a role in reaching that goal by converting enough thorium to U-233 to start LFTRs in countries that lack large RGP stockpiles, or large stockpiles of weapons grade U-235 and Pu-239.

  78. @Quokka re:: “usual political anti-nuclear stuff”

    With no exception, BNC writers (if not perhaps lurkers, but who knows?) blank out proliferation as a concern for civilian nuclear, as one can see from the ferocity with which Jim Green of FOE in Aust. was attacked here some months ago. The arguments for this blanking are well-known, as also the tacit acceptance (because this blog is in English) of US nuclear weapons threat/posture AKA umbrella since 1945.

    However, the new-style cyberwar worm Stuxnet, of possible Israeli origin and planted via an apparently innocent Russian NPP supplier:

    http://www.csmonitor.com/USA/2010/0921/Stuxnet-malware-is-weapon-out-to-destroy-Iran-s-Bushehr-nuclear-plant

    attacks SCADA software, thereby implicitly compromising precisely the software category on which civilian NPPs run. Or does BNC see this as yet another argument to have passive-safety NPPs that could survive sabotage of their SCADA software?

    As neither nuclear engineers nor Australian diplomats ever write “in the clear” to BNC, is it too much to ask BNC men to check their collective spectacles for myopia?

  79. Peter Lalor, Jim Green was criticized not attacked. Criticism is part of the quest for truth, and anyone who argues for something being true, has to endure and respond to criticism. Your suggestion that poor Jim Green was somehow martyred by his critics, is simply not true. it is more likely that Green did not know hot to respond to his critics, and quit the field.

    As for proliferation concerns, believe it or not some of us, at least, are concerned about proliferation. My concerns go beyond proliferation to the reduction of the stockpiles of fissionable materials currently set aside for nuclear weapons, and the fissionable materials currently tied up in nuclear weapons.

    i am also concerned about well the charted paths to nuclear proliferation followed by countries like South Africa, Pakistan, and North Korea. I am critical of people who focus on imaginary paths to proliferation while not offering realistic and effective methods of preventing proliferation by the well charted paths.

    American NPPs were designed and built before the days of computerized control, and are quite old fashion in this regard. Hacking the computers of old nuks is not going to cause harm.

    Finally I support the development of Molten Salt Reactors which do not require external control.

  80. @Peter Lalor

    If you believed every rumour, supposition and piece of disinformation that is circulated about Iran’s nuclear program you would come to the conclusion that Armageddon is almost upon us. It isn’t. Without any real evidence, the stuxor stuff is pure speculation. Not impossible, but then a lot of things are not impossible.

    I could say a lot more about in inadvisability of using Windows in these environments, but I don’t want to derail this thread.

  81. @Barry, thanks for pointing out the synergy papers again. In addition to the synergy of using high-breeding-gain IFRs to provide start-up U-233 charges to low fissile inventory LFTRs, there may be an operational synergy, due to the likelihood that simpler and cheaper LFTR designs, and particularly reprocessing systems, are possible if the breeding requirement is dropped down to 1.0 – 1.02. The simplified reprocessing is leaky, though, so of the order of 1% actinides – U-233, Np-237 and Pu-238 in a thorium fuelled system – goes into the ‘waste’. If this material is allowed to cool for a few years, then shipped to a site with an IFR, the residual actinides can be pulled out and burned, and the waste prepared for final disposal. Meanwhile, the IFRs supply start up charges, and fuel for small converter reactors like the salt-cooled AHTR that might be the most economic choice for process heat, or even ship propulsion.

    I suspect (but have no way to prove) that putting a full pyroprocessing / waste preparation plant at every reactor site will be more expensive than centralising some of the more difficult work, particularly if we end up wanting a lot of smallish plants to supply power + district heating, as is common using fossil fuel plants in much of Northern Europe today. This is much easier to do with LFTRs in the mix.

    This ‘nuclear ecosystem’ suggestion splits the fissile inventory roughly equally between IFRs and LFTRs, with 80% – 90% of the energy output from thorium cycle reactors, but nearly all the breeding gain from U/Pu cycle IFRs, some of them run with thorium blankets to make U-233. Whether it is economic, rather than just looking neat from a neutronics/waste reduction perspective is an open question, as far as I can tell at the moment.

  82. Peter, it seems that the Stuxnet worm uses several zero-day Windows vulnerabilities to attack certain models of Siemens PLCs, if they’re connected to a vulnerable Windows PC running vulnerable software, with a mechanism by which the worm can be inserted into that PC. (i.e. it can only get onto the PC either through removable media or over the Internet.)

    SCADA is not a specific piece of software, it is a general family of systems. There are lots of different PLCs on the market from different vendors, and lots of different software to interface to them on a PC.

    Personally, I wouldn’t touch Windows with a ten foot pole in this situation, I’d be far more comfortable with something such as EPICS running on top of Linux.

    It’s plausible that a dedicated, well-resourced, well-planned digital attack on a NPP could shut down the plant for some time, *if* the plant uses vulnerable PLCs connected to vulnerable software on a vulnerable OS which is connected to threats in the outside world without an appropriate firewall or air gap. Is this what Stuxnet is designed to do? I don’t know, but it’s plausible.

    Knocking the plant offline is plausible. Causing an actual accident such as core damage, LOCA or any accident with any health physics significance is, however, not plausible, even if you can mount such an attack.

  83. Pingback: TerraPower’s Travelling Wave Reactor – why not use an IFR? « BraveNewClimate

  84. Luke Weston,

    Thanks for some clarity on the PLC hackability issue. I suspect that few of Barry’s fans are familiar with EPICS.

    EPICS stands for “Experimental Physics and Industrial Control System”. Its prime developer was Los Alamos National Laboratory with a bunch of major laboratories and universities including my employer as co-developers.

    This is a “state of the art” control system with much better security than the people who use Windows can enjoy. Our EPICS based system manages over 3,000 devices in real time and it runs best on powerful Unix boxes (OK, I like Linux too). It can easily be isolated from exterior intervention.

    For PLCs we have tried all kinds but for some years have bought almost exclusively from one manufacturer. I will not mention my recommended supplier by name other than to say it is not Siemens.

  85. Barry Brooks,

    After plowing through those papers on synergy it appears that a rapid build up of NPP capacity cannot be achieved without at least one type of Gen IV reactor. I was beginning to get the feeling that “you can’t get there from here” but that French paper shows that even an eight fold build up in NPP capacity by 2050 is possible.

  86. Douglas Wise,

    Your speculation about companies being conflicted when it comes to introducing game changing innovations is well founded. In some cases there is deliberate suppression of new technologies as for example with General Motor’s EV1 electric car.

    Even when deliberate suppression does not take place, the dominant companies often have difficulty allocating sufficient resources to emerging technologies when their current products are “cash cows”. This explains why an upstart company like Cisco Systems can compete successfully with the huge companies that previously dominated the telecommunications business.

  87. Barry,

    I read with interest the links you provided upthread relating to synergy between different reactor technologies. The conclusions, based on the starting premises relating to extent of fuel reserves and the technologies selected, seemed convincing and both studies showed that, for sustainability, we would need to end up with both fast and thermal breeders with the latter probably eventually predominating.

    However, the Wider et al paper assumed total uranium reserves at 14 million tonnes and the Merle-Lucotte et al one picked a figure of 24 million tonnes were a price of $400/kgU to be acceptable.

    The latter, when defining sustainability to 2100 was only allowing for 28% of total power to come from nuclear (an eightfold increase from present), an equal amount from renewables and the rest from fossil fuels. While the authors did acknowledge that even more nuclear could be rolled out than this by using their recommended technology combination, I was nevertheless somewhat alarmed by their conclusions and apparent lack of ambition because their solution to nuclear sustainability doesn’t appear to provide a solution to AGW.

    Thus, in order to cheer myself up, I started wondering whether they were correct about the extent of affordable nuclear fuel reserves. Given the recent reference to the MIT study in which Moniz appeared to be saying that there would be so much uranium around for such a long time that there was no reason to rush to advanced fuel cycles, I wondered whether the MIT researchers had come up with new information indicating that affordable reserves had been vastly understated. Having seen only summaries of the study, I received no further enlightenment.

    Next, I wondered about the claims of IFR proponents to the effect that a simple ALWR/IFR combination would provide all the energy we need sustainably and in time without even the necessity of mining any more uranium at all. These claims appear totally inconsistent with the findings of the two papers that you cited.

    The most cheering statement that I have read was that of David LeBlanc to the effect that uranium reserves were inexhaustible if one considers sea water extraction (David McKay also mentions this) . LeBlanc claims that U from seawater currently costs $5000/kg (with the potential for reduction). More importantly, he asserts that this would still be affordable for the type of DMSR he is advocating (but not for ALWRs because of their low burn up). I hasten to add that LeBlanc is not suggesting that breeders should not be developed, only that they may not be a necessity ever to do so. The decision would rest on economics rather than the amount of finite resources. Similarly, Rubbia’s spallation approach to U233 production is one that was not considered in the papers you cited, but clearly seems to have the potential to allow fast thermal breeder startups if economic considerations were set aside. Presumably, spallation would only be an interim cost until one had enough self sustaining units.

    Anyway, I would be very interested to know whether you found the papers you cited to be as “spirit dampening ” as I did (initially, at any rate).

  88. I think 4th generation nuclear will have to be a key power source by 2050 not 2100. Oil will be minor, only one or two countries will have plentiful gas and coal will be heavily taxed. Some (Patzek et al) already think world coal will peak on a net energy not tonnage basis in the next year or two. I’d put renewables at only say 20% of current stationary energy levels by mid century. Nuclear will also have to significantly power transport via electricity and synfuels.

    Without knowing the numbers perhaps the breeding requirement can be postponed til after 2050. It may be possible to work out the minimum mix of Gen 3 and IFRs needed by 2050 to do that assuming very little help from fossil and renewable sources. The next question is whether that mix can actually be achieved.

  89. @Douglas
    If you need more cheering up, remember that the French paper assumed a U/Pu fast reactor requiring 6 Te of plutonium per load, with a full 2nd load in the reprocessing cycle, and a breeding rate of 300 kg Pu / year. This gives a doubling time of 40 years. All these assumptions are highly conservative – they may refer to Superphenix, which used oxide fuel and PUREX recycling. Unlike an IFR, Superphenix was actually built, so there is no denying that it can be done, but IFRs are better on every count, with only 4 Te / GW(e) in core + 4 Te out, 400 kg / year (or more?) of Pu produced and a doubling time of 20 years (or less?). Doubling the rate of exponential growth makes a huge difference, as it compounds over time. They are also assuming 4.6 Te / Gw(e) fissile inventory for the MSRs This is fine if you want a reasonable breeding gain and a conservative reprocessing schedule of 300 days to turn the inventory over, but their own earlier work shows you can make do with only 1.5 Te, at the expense of breeding gain, even without adding graphite moderator.

  90. John Newlands, the achievement of that very goal has been the major focus of Nuclear Green since its inception. My inspiration has been the Manhattan Project, in the shadows of which i grew up. If you ever see what is left of K-25 in Oak Ridge, you could have little reason to doubt what we are capable of doing, once we accomplish unity of purpose. The path to a nuclear future, involves a simple, low cost, yet safe and efficient nuclear design that can be built in very large numbers in factories, and transported to where ever power is needed. The problems would include creating a reactor that will handle peak electrical generation in a cost effective fashion. i have discussed possible solutions on Nuclear Green. Industrial process heat can be supplied from Liquid Fluoride reactors. Surface transportation can be electrified, while ships can be powered by small MSRs. Aircraft power is still an unresolved issue. Short and intermediate range travel can be handled by high speed electric trains. Long range aircraft travel is an unresolved problem.

    i and others have come to the conclusion that Molten Salt technology reactors will cost about half of what LWRs cost. Further economies are possible. Of course we cannot be certain about this, but the argument seems compelling.

    i have also come to the conclusion that direct and indirect savings from the nuclear conversion will actually pay for the conversion. Electrifying transportation will also create significant savings. There are other economic benefits of nuclear generation of electricity. For example waste heat which now sometimes creates water supply problems, can be used to desalinate sea water, creating new water supplies in areas where fresh water is in short supply.

  91. Luke_UK

    Just the sort of information that does cheer me up. Thank you. Your insights persuade me that rollout can be faster than the authors of the French paper intimated, given a different selection of reactor technologies and given that the proposed new designs live up to their promise. Can I therefore assume that it should be technically possible to generate, say, 3 times the power that the paper contemplated by mid century without hitting constraints relating to nuclear fuel (a level necessary to put a serious dent in AGW)? I am assuming that fuel, as defined by raw uranium and/or thorium stocks is only likely to be a constraint until such time as a requisite number of full or iso breeding plants are up and running. In other words, sustainability becomes a non issue thereafter except in the multi century sense.

    I remain interested in whether, from an economic perspective, the David LeBlanc approach of a transition technology (either prior or even instead of a full breeding approach) would be superior to sticking with ALWRs until such time as breeders were ready. It seems to have additional merit as it appears that George Stanford suggests that rapid ARC-100 deployment would, in fact, accelerate rather than hinder that of the IFR just as David LeBlanc suggests that his DMSR would do the same for the LFTR iso-breeder.

  92. Charles, you say,

    Long range aircraft travel is an unresolved problem.

    But here’s one nuclear-powered resolution:
    – Get hydrogen from water by electrolysis
    – Get carbon from CO2 (perhaps as emitted from coal plants — this would be better than sequestration)
    – Create synthetic liquid fuels.
    – Distribute them with the existing infrastructure.
    . We know how to do all that. And it’s carbon-neutral.
    — George

  93. Seeing as we’re more or less talking about the future of nuclear power here, I’ll post this link to a new study released by the Massachusetts Institute of Technology (MIT), entitled “Future of the Nuclear Fuel Cycle”.

    http://web.mit.edu/mitei/docs/spotlights/nuclear-fuel-cycle.pdf

    Some of their statements/conclusions:

    There is no shortage of uranium resources that might constrain future commitments to build
    new nuclear plants for much of this century at least.

    For the next several decades, light water reactors using the once-through fuel cycle are the preferred option for the U.S.

    For the longer term, there are multiple viable fuel cycle options with different economic, waste management, environmental, resource utilization, safety and security, and non- proliferation benefits and challenges. A significant research agenda is needed to explore, develop and demonstrate the advanced technologies to the point of allowing informed future market place and policy choices.

    Also of interest, it states

    Since the 1970s major decisions on development of sustainable closed fuel cycles have been
    based on the assumptions that uranium resources are limited and that consequently a reactor with as high a conversion ratio as feasible (which turns out to be 1.2 to 1.3) is required. These assumptions drove fuel cycle decisions. Our assessment is that both assumptions are
    incorrect—uranium resources are large and a conversion ratio of unity is preferred.

    I’d be interested to know what Dr Stanford thinks of this assessment.

  94. Dougla Wise, you said:

    “Similarly, Rubbia’s spallation approach to U233 production is one that was not considered in the papers you cited, but clearly seems to have the potential to allow fast thermal breeder startups if economic considerations were set aside.”

    Rubbia’s idea is criticized because it creates fissile material which currently is so abundant that we have concerns about weapons of mass destruction. However, we could run out of fissiles if a rapid expansion of NPP capacity was attempted.

    Has anyone calculated what it would cost to produce fissile Uranium using ADRs compared to more conventional means such as gas centrifuges?

  95. Douglas Wise,
    As an after thought to my previous quote, the cost of neutrons is coming down. The SNS at ORNL produces a 1 MW proton beam and L.N. Chang (Virginia Tech) says he can get more than 30 neutrons for every proton.

    Charlie Bowman (ADNA) built a prototype ADR back in 2002.

  96. Tom Keene wrote:

    I’d be interested to know what Dr Stanford thinks of [the MIT] assessment.

    . Tom, I haven’t had time yet to read and digest it. But I’m struck by this excerpt:

    For the next several decades, light water reactors using the once-through fuel cycle are the preferred option for the U.S.

    . The meaning of “several” is not clear, but it’s all-important. If it means two, then that’s just a statement of reality — we’re won’t be ready to build much of anything other than LWRs for the next 20 years (although there will be some fast reactors going by then in forward-looking nations like China, Russia, and India).
    .
    . But if “several” means five, that’s a different kettle of fish, particularly from a global perspective. For a simplistic example, suppose every reactor that comes on line before 2060 is an LWR, and every subsequent reactor to come on line is an IFR. And suppose also that the global population ,stabilizes at perhaps 11 billion by 2110, and that our global goal is to have enough clean, carbon-free energy to provide everyone with 50-60% of the per-capita energy now enjoyed by Americans.
    .
    . Data: To fuel a 1-GWe LWR for one year, ore containing about 140 tonnes of natural uranium has to be mined, and each LWR can be expected to last at least 60 years.
    .
    . To meet our global goal, after allowing considerable growth in wind and solar power, nuclear capacity would have to grow at a compounded rate of about 5.4%, starting with the current capacity of about 300 GWe. This puts the nuclear capacity in 2060 (5 decades hence) at ~4000 GWe. (Most of the numbers here are very approximate, but in the ballpark.)
    .
    . That means that between now and 2060 there will be ~94,000 GWe-years of operations requiring 94,000 x 140 = 13.6 million tonnes of uranium to have been mined and fed to enrichment plants.
    .
    . And say that the entire 4000 GWe is provided by LWRs, which in 2060 will have an average life expectancy of something like 40 more years. Then the used-plus-committed natural uranium will amount to 13.6 million + (4000 x 60 x 140) = 47 million tonnes of uranium.
    .
    . Undoubtedly that much uranium exists, but not at today’s prices (~$140/kg). The European Nuclear Society says, “The global uranium resources with mining costs up to US $ 130 per kilogram amount to about 3.3 million tonnes.”

    .
    . But that’s not all. In 2060 the annual enrichment capacity needed would be 4000/300 = 13 times today’s enrichment capacity — not a situation to please those who worry about proliferation of uranium weapons. And the amount of spent fuel would be almost 20 times what we have today and are wondering what to do with.
    .
    . In this simplistic scenario, we are assuming that, starting in 2061, all new reactors coming on line are breeders that get their initial fissile charge existing supplies of plutonium and require no further increase in the rate of uranium mining or enrichment. Whether there’s enough plutonium to accomplish this is a topic for another discussion.
    .
    . For a really unrealistic exercise, consider the case where the breeders (or reactors such as LFTRs with unity conversion ratio) don’t come on line until nuclear expansion ceases in 2110. By then, there will have been about 1.4 million LWR-years accumulated and about 2.8 million more LWR years committed, for a net requirement of 4.2 million x 140 = 588 million tonnes of uranium. The enrichment capacity would be about 250 times what we have today.
    .
    . But if, on the other hand, we commit ourselves now to deploy breeders as soon as possible, the global future uranium requirements will total something like 16 million tonnes, an amount that is probably available at prices not many times today’s. Also, the need for any mining or enrichment capacity will have disappeared by 2050 or so. And the entire global inventory of plutonium could be safely sequestered in operating nuclear plants by the end of the century.
    .
    . Caveat: Some of the above numbers come from an algorithm I have been working on, and it has not been thoroughly tested. If anybody can independently confirm or refute any or all of the above, please do so.

    — George

  97. George Stanford:

    Your response to Tom Keene was very informative. It is good to hear from authorities on the subject. I would really appreciate answers to a few follow up questions.

    1) You suggest that 4000GW nuclear capacity will be needed by 2060 after allowing for considerable growth of wind and solar power. How much power have you allowed from the renewables? I would suspect that a more economic route might be to increase the proportion of nuclear should such be possible within the time frame. Is it?

    2) You suggest that we are stuck with LWRs for 2 to 3 decades. Does this have to be the case? Could we move quicker by using transitional technology?

    3) Is it correct to state that the ARC-100 is to the IFR what the uranium converting DMSR is to the LFTR and that both could be deployed more quickly than full breeders? If so, how much more quickly?

    4) Is it correct to think that the above will defer the requirement to demonstrate and hence licence the appropriate reprocessing technologies for two to three decades and that both are designed to be constructed as small, factory built units such that one might hope that they would be cheaper to build and capable of more rapid rollout? If so, are they more likely to facilitate or delay the deployment of IFRs and LFTR iso breeders?

    5) What, in essence differentiates the IFR from the ARC-100 anyway other than onsite reprocessing in the case of the latter? If small factory built modular units are going to be cheaper and quicker to build, wouldn’t it be more pragmatic to have regional reprocessing centres rather than one with each plant?

    I am acutely aware that a little learning is a dangerous thing and I therefore hope for some expert advice.

  98. Douglas, below are my responses to your questions. Be aware, however, that I don’t claim to be the horse’s mouth, and that this a fanciful scenario with no predictive value whatever.

    1) You suggest that 4000GW nuclear capacity will be needed by 2060 after allowing for considerable growth of wind and solar power. How much power have you allowed from the renewables? I would suspect that a more economic route might be to increase the proportion of nuclear should such be possible within the time frame. Is it?

    . My scenario is only partly baked at present. I’m assuming that the non-nuclear portion of the energy economy remains at its current value (in quads per year), leaving unspecified the split between fossil fuel and non-nuclear renewables. If the latter were to come anywhere near the current fossil-fuel contribution, that would be a considerable expansion.

    2) You suggest that we are stuck with LWRs for 2 to 3 decades. Does this have to be the case? Could we move quicker by using transitional technology?

    . I said *two* decades — we can hope that the pace of change will be picking up during the third. And I do expect that “transitional technology” will start to be deployed within the next two decades — but it will mainly be non-breeding fuel cycles for niche applications (e.g., non-breeding SMRs), and it will be make only a minor contribution to the overall energy supply by 2030. But yes, their deployment will provide valuable base-line info. Things will move faster in countries other than the U.S.

    . Incidentally, I’m using “breed” in its intuitive sense, namely, implying that the conversion ratio CR — the number of fissile atoms created per heavy-metal atom fissioned — is significantly greater than unity. The customary term for a reactor with CR of unity or less is “converter.”

    3) Is it correct to state that the ARC-100 is to the IFR what the uranium-converting DMSR is to the LFTR, and that both could be deployed more quickly than full breeders? If so, how much more quickly?

    . The ARC-100 is a SMR based on IFR technology, and its deployment should advance the state-of-the-art. Whether it is deployed before the PRISM is hard to predict. Regarding the DMSR & LFTR, that technology seems to be a decade or more behind the PRISM, in terms of readiness for commercialization.

    4) Is it correct to think that the above will defer the requirement to demonstrate and hence licence the appropriate reprocessing technologies for two to three decades and that both are designed to be constructed as small, factory built units such that one might hope that they would be cheaper to build and capable of more rapid rollout? If so, are they more likely to facilitate or delay the deployment of IFRs and LFTR iso breeders?

    . No. The urgent need now is to show how to bring pyroprocessing up to commercial scale, so that breeders can be coming on line by 2030 or so. It would be short-sighted to delay finalizing development of the IFR technology — which, after all, can do just about everything that needs to be done — in the hope that something marginally better will come along if only we do more R&D.

    5) What, in essence differentiates the IFR from the ARC-100 anyway other than onsite reprocessing in the case of the latter? If small factory built modular units are going to be cheaper and quicker to build, wouldn’t it be more pragmatic to have regional reprocessing centres rather than one with each plant?

    . The ARC is a small (100-MWe) modular reactor with a CR of unity. It is intended for long-life “nuclear battery” applications, for which breeding would not be appropriate and centralized reprocessing of the returned cores would be.

    . The PRISM version of the IFR is somewhat larger (~300-GWe modules), for use in multi-unit power plants. Keep in mind that pyroprocessing is done in batches, which means that economy of scale is less important than you might think. One advantage of in-plant processing is that it can be integrated with the reactors, and commerce in plutonium is all but eliminated. The annual input to a 1-GWe IFR plant operating in breeder mode consists of ~1.5 tonnes of uranium. The output is heat and electric power, one tonne of fission products (with trace amounts of actinides), and some very radioactive fuel containing 0.5 tonne of Pu, to be delivered to a new IFR that is starting up somewhere else.

    — George

  99. George, thanks for all that information in your response to Douglas, particularly on the ARC-100 reactor.

    From my reading of the ARC white paper and the EBR-II report, it looks like the ARC-100 is almost exactly the EBR-II. In other words, ARC is basically commercialising EBR-II, while GE is taking the scaled up version (PRISM) to market. Its the same reactor concept targeted at different markets, differentiated by power outputs.

    Is that a fair reading of the situation?

    If so, what differentiates the isobreeder (ARC) from the breeder reactor? Could the ARC reactor be run as a breeder by using a different fuel configuration? Or does breeding require changes to the reactor design?

  100. John:

    . The ARC-100, Toshiba 4S, PRISM, and TWR are all variants on the IFR theme, which Argonne developed using EBR-II as a test bed. IFRs can be designed to have a conversion ratio anywhere between approximately 0.5 and 1.5, depending on the loading configuration. I suspect, however, that the reactor vessels for the 4S and ARC-100 would too small to accommodate the somewhat larger core that would be needed for appreciable net breeding (although maybe it could be done by reducing the interval between core changes). And the TWR has other, unique constraints.

    . On the Web there’s considerable info on the ARC-100 — for example, see
    , and
    .

    . — George

  101. While the full LFTR iso-breeder definitely needs some more research, I don’t think a 10-year gap between IFR and DMSR is fair. There is a detailed design, based on extrapolating results of smaller scale experiments (MSRE). A prototype at 100 MW(e) scale, preferably bigger, is needed. That’s not very different from the IFR case. My main reservation about DMSR is that it is too good as a stopgap. By reducing uranium needs by a factor of five vs PWRs, it allows us to get well into the 2nd half of the century (on George’s scenario above) without running into severe uranium constraints. This ought to be a good option to have, but if it is simply seen as an excuse to do nothing about developing real solutions that need no fissile input and leave no long term (actinide) waste output, then it might not be. DMSR waste also makes terrible start-up fuel – lots of non-fissile Pu isotopes and heavy actinides. The DMSR design priority was proliferation resistance, not waste minimisation.

  102. Douglas Wise:

    Can I therefore assume that it should be technically possible to generate, say, 3 times the power that the paper contemplated by mid century without hitting constraints relating to nuclear fuel (a level necessary to put a serious dent in AGW)?

    Yes, at least. My next post on BNC, which I’ll put up in a few days, outlines some scenarios for this. It’s a work in progress, and I’d hoped to do more on it before posting, but I think it would be better to just put it up and invite comments/improvements.

    Short answer is I can see a pathway to 12 TWe of nuclear (a thermal/IFR/LFTR combo) by 2050 without any extreme assumptions regarding fuel supply or breeding rates.

  103. Luke_UK:

    Luke, I was interested in your DMSR comments. Could you elaborate?

    1) You thought George was being unfair to suggest that the DMSR would be a decade behind the IFR (while agreeing that this would be the case for the full LFTR iso-breeder). I don’t have the wherewithal with which to judge whose view is correct. However, you were not clear as to whether you thought the DMSR would be ready, say, a decade before the IFR or at about the same time. One of its potential attractions would disappear unless it could be brought on stream sooner than the IFR.

    2) It would appear that the term, DMSR, is generic. When you say that its waste makes terrible start up fuel, does this apply to all designs? I was thinking specifically about those that relied on uranium only and didn’t blend it with thorium. While this approach might be less efficient than a version using blended fuel, it would nevertheless seem to be considerably more efficient than LWRs in terms of fuel economy. If one then considers the potential relative economic advantages, it would seem to have virtue unless those same advantages were true for the ARC-100 or unless the DMSR would post date the IFR in its rollout , as George suggests.

  104. Douglas,
    Sorry to have been unclear. I should also point out that my opinion is in no way expert. Look up David LeBlanc’s articles on DMSR, and judge from them. To clarify my own opinion, it is that the gap in technical readiness between the IFR and DMSR is less than 10 years, but whether it’s 2 or 5 I’m not qualified to determine. However, the major problem for IFR deployment is not technology but politics. IFR development was shut down for political reasons due to misplaced fears about proliferation. That political opposition still exists. The DMSR was designed specifically to allay these concerns, at the expense of other features. These design decisions – like removing most of the reprocessing system – also make it simpler. Which is the better course, to try to defeat political opposition, so we can build the technically best reactor , or to accept that we aren’t going to be allowed to do that and build what we can? That might be DMSRs, or ARC100s.

    There are lots of possible variations on the DMSR theme, but the only one with real design work done is the graphite moderated, thorium + uranium mixed fuel version. No version produces much Pu-239 for starting up IFRs or LFTRs. Any that is formed is left to stew until it fissions or transmutes in to something else.

  105. Luke,

    You make an important distinction between whether it is better to plot the way forward with the technically best or politically most acceptable nuclear designs. This, I think, is a decision that is only of real concern in liberal democracies, but is nonetheless of supreme importance to their denizens. I think that, maybe, one should add in economic concerns which are not necessarily related to political ones in the short term.

    Having tried to digest and integrate all the answers to the questions I have been asking on this and other recent threads, I am coming round to the view that the ARC-100 approach might be the best way forward in the short term for the following reasons:

    1) It should be capable of deployment sooner than the IFR.
    2) Its development is more likely to expedite than delay deployment of the IFR.
    3) It is, to all intents and purposes, an iso-breeder.
    4) It can be deemed to be more proliferation resistant than the IFR.
    5) It has the capability of being factory built and rapidly rolled out.

    The potential disadvantage might relate to economy of scale. Will the advantages of modular construction outweigh each module’s relatively small electrical output? Would 10 modules/GW be competitive with PRISM’s 3 modules/GW?

    I can fully appreciate why it is somewhat pointless for an ignoramus such as myself to worry about such matters when there are others with a great deal more knowledge. However, if it is one’s intention to try and exert influence to bring about nuclear solutions to the peak oil and AGW problems we face, there would seem to be advantages in being well briefed. It seems to me that many of the experts working in the nuclear industry are not driven by a sense of urgency. For example, is GE maximising its propaganda effort towards rapid deployment of the PRISM or is it happy to go along the ALWR route or any other route that will generate the best returns in the short term?

    While we are stuck with ALWRs to produce all our nuclear power, nuclear critics will probably continue to flourish until such time as our economic state is so parlous that we can’t move fast enough to prevent catastrophe. The sooner we can demonstrate the means to make nuclear power sustainable and more proliferation resistant and show that the waste problem is of no practical consequence, the more quickly opposition will wilt.

    I think it possible/probable that molten salt technolgy will eventually provide a more economic way forward, but I think the priority now is to go with the fastest demonstration we can get of new generation nuclear in the liberal democracies.

  106. George Stanford and I have sorted out our disagreement on LFTR breeding, but other issues remain to be settled. For example the notion that the IFR is ahead of MSRs in the race to deploy. However, a review of IFR development related documents on the information bridge suggests that all IFR development to date was directed to low IFRs that offered only a low breeding ratio. I believe that 1.07 was the highest I saw. The ARC-100, based closely on the EBR-II IFR prototype, is a converter, with something like am 0.83 or 0.87 conversion ration. Given the much larger fissionable inventory which fast reactors require, the number of low ratio IFR breeders that can be deployed will be quite limited, and that number will grow relatively slowly.

    High breeding ration IFRs can increase in deployment numbers fairly rapidly. But it will require quite a lot of R&D, and perhaps more than 1 prototype, to bring the high breeding ration IFR on line. I have pointed to the Indian metal fuel fast reactors as an example. The indians already have a commercial breeder program, but expect to take at least 15 years to develop their metal fueled FBR. It seems unlikely that a high breeding ratio IFR can be developed to the point of commercial production within a shorter period of time. Given equivalent R&D resources, it is more probable that the relatively simple DMSR will be developed before the high breeding ratio IFR.

  107. Barry, point me to the report on the successful operation of a high breeding IFR prototype, or at least an advanced design scheme for a high breeding IFR prototype. My understanding is that he only IFR prototype was the EBR-II that was a converter not a breeder. The only prototype IFR designs I have been able to find on the Information Bridge have a maximum breeding ratio of 1.07. My interpretation of this is that the high breeding ratio IFR will need considerable R&D. The results of the EBR-II IFR experiments are documented, and the Indians have read all of the reports. If the Indians need an extensive development program it was because they want to know a lot of things that were not included in IFR R&D reports. Chances are American reactor developers also lack that information.

  108. Charles, you are of course free to make whatever interpretations you will — but it doesn’t make it right, or correctly in context. As such, I respectfully suggest that you have relatively limited knowledge on this matter and are drawing erroneous inferences, whereas the members of the IFRG, with hundreds of cumulative man years worth of experience in operating and researching fast reactors, know what they’re talking about when they speak of high potential BRs and commercial-demo-ready designs. Indeed, the principles of breeding rates in a hard spectrum, and how this is influenced by core/blanket design, smear density, etc. are well established, and as such, were not a major research question that the EBR-II team needed to pursue. That team instead focused on matters like SFR operational/safety performance and metal fuels testing.

  109. Douglas Wise, on 27 September 2010 at 17.00 — Since thermal efficiency isn’t much of an issue, the lower manufacturing and site construction costs+delays for smaller units probably makes the 100 MWe units preferable to larger ones.

    Also siting issues are surely lessened.

  110. Barry, I think that in general, I play a useful role by playing the Devil’s Advocate here. You do not point to the sort of evidence that a rational person would find convincing, in fact you do not point to any evidence at all. You sat “that members of the IFRG, with hundreds of cumulative man years worth of experience in operating and researching fast reactors, know what they’re talking about when they speak of high potential BRs and commercial-demo-ready designs.” But I don’t know what they say about R&D programs. ORNL went through the exercise of documenting in detail the R&D requirements for the creation of a MSBR prototype (ORNL-4812 http://www.energyfromthorium.com/pdf/ORNL-4812.pdf). This document shows that there were no huge technological gaps standing in the way of MSBR prototype development in 1972. it would be rational to look for similar assurance that the high breeding ration IFR could be accessible with a similarly modest effort. Since I lack documented estimates from the IFR community as to the high breeding ration R&D project size, I have to go with the nearest analogy, which is the very large Indian R&D project with similar goals.

    Note that I am not questioning the potential for high breeding ratio. Rather my issue is the size of the research project that would be required to produce a commercial prototype.

    You claim, “the principles of breeding rates in a hard spectrum, and how this is influenced by core/blanket design, smear density, etc. are well established, and as such, were not a major research question that the EBR-II team needed to pursue. That team instead focused on matters like SFR operational/safety performance and metal fuels testing.” Is it your position then that the IFR breeding ratio can be increased from a maximim of 1.07 to 1.65 with no technological issues to be resolved? I would like to have confirmation from a published source if you maintain that.

  111. Charles, I agree that you play a useful devil’s advocate role.

    Is it your position then that the IFR breeding ratio can be increased from a maximim of 1.07 to 1.65 with no technological issues to be resolved? I would like to have confirmation from a published source if you maintain that.

    Yes, it is — I will work through this material on the blog in the coming weeks, with reference to published sources where possible. But most of this is in ANL reports, such as C.E. Till, Y.I. Chang, et al, “Fast Breeder Studies,” ANL-80-40 (1980). This report documented the Argonne contributions to the International Nuclear Fuel Cycle Studies (INFCE).

  112. Barry,

    You refer above to “commercial-demo-ready designs”. Other than money, what is needed before such demonstration units can actually be built? Presumably, regulators get involved, but are they likely to slow things down as much as in the case of fully commercial units? Is there, for example, an accelerated path to take for a small one off designs? Alternatively, one might even envisage even longer delays for new technologies.

  113. Luke,

    Thanks for the link. It seems that the manager of the DoE’s Savannah River Site has an excellent plan to allow construction of several different demonstrator reactors. However, it seems that NRC will make things as difficult as possible and none of the power could be released to the grid without involving them. They also seem to be intimating that they won’t be considering licensing any commercial designs except those already on its books for many years. Unless something changes dramatically, it seems that the NRC is committing the citizens of the States to energy/economic poverty.

    Any chance that the UK could steal a march and host a range of demonstrator units? We have potential sites and plenty of fissile material. Do we have enlightened leadership to force through the necessary decisions?

  114. Douglas, in the US it (S-PRISM) would need to be built by government as a demonstration project. This would fast track later NRC approval for commercial rollout. The DOE need to make the decision to do it. Or, internationally, someone (probably Russia), needs to decide to build the first one, and the US needs to agree to a tech sharing arrangements — perhaps with a consortium of interested countries (e.g. Russia, Japan, South Korea, China, India etc.). Also, a facility for converting LWR spent fuel into PRISM feedstock is an essential first step, and should be done in the US. I think the international PRISM and US LWR spent fuel facility are the most likely pathways.

  115. I have a question for George about the metal fuel in the IFR.

    The safety advantage of the metal fuel is that due to properties of thermal expansion, in a loss of coolant situation, the fuel initially heats up and then the reactor goes subcritical. so the metal fuel prevents a criticality accident.

    How does the passive safety of the metal fuel compare to the negative temperature coefficient of a water moderated reactor? The latter also prevents a criticality accident, thus TMI only had a partial meltdown.

    It’s my impression though that the metal fuel negative feedback mechanism is safer than the water moderated one–that in the EBR 2 demos, there was no partial meltdown as in TMI.

    Is this impression correct? and just out of curiosity, is the negative temp coefficent of water moderated reactors technically considered a passive safety feature?

  116. Greg:
    .
    . Let’s not overemphasize the importance of the IFR’s safety advantages. While they are indeed real, the LWRs (and HWRs) already built have proven to be so safe (TMI hurt nobody), that further improvements are really just gilding the lily.
    .
    . In comparing oxide (or carbide) fuel with metallic, the most important consideration is not safety-related at all — it’s the much superior breeding potential of metallic fuel (the neutrons are less moderated).
    .
    . However, here’s a couple of the technical safety considerations.
    – During an unprotected loss-of-flow, the coolant temperature rises rapidly, which introduces negative reactivity feedbacks due to radial expansion of core, control rod drive-line expansion, etc. This brings the power down.
    – The high thermal conductivity of metal means that the fuel has far less stored heat to be dissipated in the event of loss of coolant flow, greatly reducing the temperature swings, making passive convective cooling much more feasible. In the safety demos at EBR-II in 1986, completely passive effects brought the coolant (and core) temperature down to ~200 degrees F below the boiling point of the sodium coolant.
    .
    . You ask,

    How does the passive safety of the metal fuel compare to the negative temperature coefficient of a water moderated reactor?

    What counts is whether the negative thermal-expansion compensates for the positive Doppler effect, and with metal fuel the thermal expansion comfortably overrides the Doppler-induced reactivity increase — not necessarily the case with a water-moderated reactor. (If you don’t happen to know what “Doppler” means here, don’t worry about it. It’s an effect that tends to increase the fission rate as the temperature of the fuel rises.)
    .

    Is the negative temp coefficient of water moderated reactors technically considered a passive safety feature?

    Since this causes something to happen without human or mechanical intervention, it’s a passive feature.

  117. george: thank you very much.

    when I first read about the IFR, the passive safety features of the metal fuel really stood out (they stood out due to my ignorance of the safety features of current reactors).

    so when I read about what seemed (to my untrained eye) a similar feature in water reactors, I was just curious.

    at any rate, the negative feedback feature of the metal fuel does indeed seem an improvement over the water moderated reactor, though as you note, and I agree, we should not make so much of the gen 4 safety features as that might imply that most reactors currently in use are not safe.

    can you explain radial expansion and control rod drive line expansion? are these negative reactivity feedbacks properties of IFRs specifically?

    I (just) found a 2009 INL article by R Wigeland that discuss it but have not read it yet.

  118. Greg:
    Reactor dynamics is not my specialty, but here’s how I see it. When the whole core expands, everything gets farther apart, which means that “neutron leakage” increases — i.e., more neutrons can escape from the core without causing fissions — decreasing the reactivity.
    That would apply to any reactor. The IFR’s advantage lies in the fact that the (positive) Doppler effect is considerably smaller than the (negative) expansion effect..
    The control-rod assemblies in an IFR (such as GE’s PRISM) would be anchored at the top of the reactor vessel, quite far above the core, while the core itself rests on the bottom of the vessel. When the coolant temperature increases, the heated drive trains expand more than the less-heated vessel does, resulting in a net insertion of neutron absorber into the core.
    Perhaps the Wigeland paper you refer to is the one here:
    http://snipurl.com/1dqafs
    which has a considerably more detailed discussion.

  119. Geroge S Stanford,

    Thank you for your clear statement:

    Let’s not overemphasize the importance of the IFR’s safety advantages. While they are indeed real, the LWRs (and HWRs) already built have proven to be so safe (TMI hurt nobody), that further improvements are really just gilding the lily.

    That seems reasonable and realistic to me.

    Taksing thsi into account, I believe the reasons for developing Gen IV are:

    1. To reduce the cost of electricity (this is by far the most important reason from my perspective, and should be the number one stated objective)

    2. to improve the rate at which nuclear power can be implemented all over the world, so we can avoid building new, and replace existing, fossil fuel plants

    3. to improve reliability of supply over the longer term (a long term, not a short term, requirement).

  120. More on the ‘balloon storage’.

    Some across in another science forum had this reply. This is all a bit confusing for me, because being non-technical I hate it when ‘experts’ I respect disagree with each other. I don’t have the time to dig deep enough into the physics to decide one way or another. All I can do is play ‘Devil’s Advocate’ and ultimately admit I don’t know if reasonable enough objections are mounted either way.

    ***
    “When I read the title, I immediately gave a bit of thought to the thermodynamic issues but I don’t quite agree with those objections. If the deep water is no warmer than the environment at the surface, good design could make it efficient. Further, if the temperature is significantly lower down below, the difference could even be exploited to draw a bit of thermal energy from the environment, while you’re at it. It all depends on how feasible a good design would be. ”

    Then this…

    “according to this (i can’t vouch for its accuracy though, it just seemed reasonable to me), the temperature at depth is between 3 and 6 degrees, and on the surface between 4 and 28 degrees. so if done on the equator, there could be some potential to that idea.

    http://residualanalysis.blogspot.com/2010/02/temperature-of-ocean-water-at-given.html

    personally i’d be more worried about friction and viscosity losses over such a large length of piped air, i think.”

  121. EN, without seeing the context of the commenter you quote, I can only guess, but it looks like they are just thinking in terms of the temperatures of the two different heat sinks (surface and deep water) and ignoring the large pressure difference.

    I put some numbers on this in this comment. Here’s another way to think about it.

    Imagine pushing a large beachball under water. It takes a lot of effort, working against the bouyancy force. You do more work as you push it deeper. Part of the work winds up as thermal energy in the air in the ball. Quite a lot of it in fact. The work you do to push it down to 750 m would heat the air to ~730 C, if the ball were insulated and the heat couldn’t escape. But its not insulated, and the ball doesn’t get hot, because the heat just gets transferred to the ocean, and lost.

    Think about this at the scale of a grid connected storage system. Unless the whole submerged system is insulated, you’re throwing away a lot of the energy you used to push the air underwater. This energy is “recovered” by using gas to reheat the air.

    The energy associated with water temperature differences of 10 or 20 degrees which your correspondent is thinking of are irrelevant in the context of energy losses associated with many hundreds of degrees. Without calculating it out, I would also expect it would dwarf the pumping losses.

    I hope that helps conceptualize whats going on here.

  122. Pingback: Into Eternity: Fact or Fiction? « Environmentalists for Fission Energy

  123. Pingback: Nuclear power at Ed's Home

  124. Pingback: Eine internationale Weltraumbehörde | Dr. Herzschmerz sucht seine Göttin

Leave a Reply (Markdown is enabled)

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s