Thursday, December 13, 2007

The Aqueous Homogeneous Reactor

The Aqueous Homogeneous Reactor, almost always referred to in Oak Ridge as the Homogeneous Reactor, was originally designed in Los Alamos during World War II. The advantages of the Homogeneous Reactor were numerous. Reactors were designed by physicists. In a way reactors are the descendants of a physics experiment conducted in Chicago in 1942. That experiment was so successful, that it had a profound influence on the future of reactor design. But chemist had a different take on reactor design. As the History of ORNL states, "It pained chemists to see precisely fabricated solid-fuel elements of heterogeneous reactors eventually dissolved in acids to remove fission products--the "ashes" of a nuclear reaction."

Alviin Weinberg for all his theoretical brilliance was not a physicist.  Indeed his PhD was in biology.  He was also open to what Scientist had to say, as long as they knew what they were talking about. The point the chemists made was simple, as long as you are stuck with the Stagg Field reactor model, you are separating the physics of the reactor from the chemistry of nuclear fuels. The Homogeneous Reactor was a way to get around that. Its design treated the operation of the reactor as a continuous chemical process instead of a physics experiment. The ORNL history points out the advantages of the chemists way of thinking:

"A homogeneous (liquid-fuel) reactor had two major advantages over
heterogeneous (solid-fuel and liquid-coolant) reactors. Its fuel
solution would circulate continuously between the reactor core and
a processing plant that would remove unwanted fissionable products.
Thus, unlike a solid-fuel reactor, a homogeneous reactor would not
have to be taken off-line periodically to discard spent fuel.
Equally important, a homogeneous reactor's fuel and the solution in
which it was dissolved served as the source of power generation.
For this reason, a homogeneous reactor held the promise of
simplifying nuclear reactor designs."

It should be noted that the same advantages also apply to the Molten Salt Reactor. A homogeneous reactor was far simpler than than a light water reactor.  The picture above illustrates the simplicity.  There was no pile.  The core of the reactor is a simple spherical metal pot.    One pipe leads in the other out.   The trick was simple.  Less than 1 pound (454 grams) of U235 was dissolved in a water solution and pumped into the core. The shape of the core brought the fissionable material into criticality, and a chain reaction commenced, as Xenon built up in the reactor the original fluid was pumped out to be replaced by more fuel dissolved in water. As water left the reactor heat could be extracted via a secondary coolant loop, and a turbo generator driven, A further advantage was that fission byproducts like Xenon could be extracted from the reactor fluid.

There was a decided problem, with the homogeneous reactor. It boiled, not with hot water, but with hydrogen and oxygen gases. Radiation from the the chain reaction broke water molecules apart. This lead to the biggest problem with the design, with hydrogen and oxygen in a water filled core, combined with heat and radiation there was bound to be a big problem with metal corrosion, and in fact there was. It might be possible to solve this problem today with heat and radiation resistant carbon composites. 

The biggest problem for the homogeneous reactor was the AEC's light water reactor bandwagon.   The light water reactor was an advanced version of the Stagg Field Pile.  It was something that people in Washington understood.  Forgetting that Wienberg held the patent on the light water reactor, the people in Washington seemed to have thought of Weinberg as a wild eyed dreamer.

Weinberg was by training a biologist.  He saw reactors in terms of evolution.  Weinberg realized what few in his generation did, that there were simpler and potentially far better reactor designs.  Indeed Weinberg understood the problems of reactors more deeply than almost anyone else during his time or since.  Weinberg saw that rather reactors must be treated like processes rather than entities.  That is why he called in chemist like my father.  My father had been an industrial chemist before he came to Oak Ridge.  My father deeply understood chemical processes.   Indeed his own PhD research had been directed to turning chemical analysis into a continuous process.    Had the homogeneous reactor been the only fish Weinberg had to fry, he might have pushed for it further.

Oak Ridge was not the only place interested in the homogeneous reactor. At the same time Oak Ridge was conducting its first homogeneous reactor experiment, Edward Teller had pushed Los Alamos into conducting an experiment with a Molten Salt Reactor as a Thorium Breeder.  By using heavy water, Los Alamos was able to up the Neutron Economy of the reactor to a truly formidable level.   

Except for its corrosion problem, the homogeneous reactor was extremely safe. Even if the core was breached, and hot radioactive water came spilling out, there would not have been a mess. There would have been no molten mass to contend with, no China Syndrome. The radioactive Xenon escape would have been limited because Xenon was being continuously drawn off the reactor fluid as part of the operating process. There would of course have been a mess to clean up, but a mess is a mess, not a disaster. The problem would have been contained in the reactor building.

As a breeder the Homogeneous reactor held big promise.  Weinberg was able to get the AEC to approve a second Homogeneous Reactor experiment, this one a more complex model with a core and a blanket.  The ORNL history observes: 
"The aim was not only to produce economical electric power but also to irradiate a thorium slurry blanket surrounding the reactor, thereby producing fissionable uranium-233. If this pilot plant proved successful, the Laboratory hoped to accomplish two major goals: to build a full-scale homogeneous reactor as a thorium "breeder" and to supply cheap electric power to the K-25 plant to enrich uranium."

There appears little doubt that the homogeneous reactor represented a major breakthrough. But a breakthrough the politicians and AEC paper shufflers had little appreciation for.  The ORNL history continues:

Initial success stimulated international and private industrial interest in homogeneous reactors, and in 1955 Westinghouse Corporation asked the Laboratory to study the feasibility of building a full-scale homogeneous power breeder. British and Dutch scientists studied similar reactors, and the Los Alamos Scientific Laboratory built a high-temperature homogeneous reactor using uranyl phosphate fluid fuel. If the Laboratory's pilot plant operated successfully, staff at Oak Ridge thought that homogeneous reactors could become the most sought-after prototype in the intense worldwide competition to develop an efficient commercial reactor. Proponents of solid-fuel reactors, the option of choice for many in the AEC, would find themselves in the unenviable position of playing catch-up. But this was not to be.

The ORNL history relates the rest of the story:

After successful operation of the first aqueous homogeneous reactor
in 1954, the Laboratory proceeded with design of a larger
homogeneous reactor on a pilot-plant scale. Whereas the first
reactor had been a one-time experiment to prove yet unproven
theoretical principles, the second reactor, sometimes identified as
the Homogeneous Reactor Test, was designed to operate routinely for
lengthy periods.

The second homogeneous reactor was fueled by a uranyl sulfate
solution containing 10 grams of enriched uranium per kilogram of
heavy water, which circulated through its core at the rate of 400
gallons (1450 liters) per minute. Its fuel loop included the
central core, a pressurizer, separator, steam generator,
circulating pump, and inter-connected piping. Its core vessel was
approximately a meter in diameter and centered inside a 60-inch
(152-centimeter) spherical pressure vessel made of stainless steel.
A reflector blanket of heavy water filled the space between the two

Perhaps the most exotic nuclear reactor ever built, it gave
Laboratory staffers trouble from the start. During its shakedown
run with pressurized water, chloride ions contaminated the
leak-detector lines, forcing replacement of that system and
delaying the power test six months.

In January 1958, the Laboratory brought this reactor to critical
mass and operated it for many hours into February 1958, when it
became apparent that its outside stainless steel tank was corroding
too rapidly. In April the reactor reached its design power of 5 MW.
Then, in September, a hole suddenly formed in the interior zircaloy
tank. Viewing the hole through a jury-rigged periscope and mirrors,
operators determined that it had been melted into the tank--that
is, the uranium had settled out of the fuel solution and lodged on
the tank's side.

By the end of 1958, the AEC considered abandoning the Homogeneous
Reactor Test, and Eugene Wigner came to the Laboratory to inspect
it personally. "The trouble seems to be that the rich phase adsorbs
to the walls and forms a solid layer there," Wigner reported to the
AEC staff, relaying the findings of the Laboratory staff. He
thought altering the flow of fluid through the core would provide
the velocity needed to prevent the uranium from settling on the
tank walls. "It is my opinion that abandoning the program would be
a monumental mistake," he warned, pointing out that the reactor
could convert thorium into uranium-233 to supplement a dwindling
supply of uranium-235.

The AEC allowed the Laboratory to alter the reactor flow and
continue its testing in 1959. These activities were accomplished by
interchanging the inlet and outlet to reverse the fluid flow
through the reactor. Several lengthy test runs followed in 1959,
and the reactor operated continuously for 105 days--at the time, a
record for uninterrupted operation of reactors. The lengthy test
run demonstrated the advantages of a homogeneous system in which
new fuel could be added and fission products removed during reactor

Near the end of the year, a second hole burned in the core tank.
Laboratory staff again patched the hole using some difficult remote
repairs and started another test run. Because of these
difficulties, Pennsylvania Power and Light Company and Westinghouse
Corporation abandoned their proposal to build a homogeneous reactor
as a central power station.

During the shutdown and repairs, Congress viewed the aqueous
homogeneous reactor troubles unfavorably, and in December 1960, the
AEC directed the Laboratory to end testing and turn its attention
to developing a molten-salt reactor and thorium breeder. The last
aqueous homogeneous reactor test run continued until early 1961.
For months, the reactor operated at full power until a plug
installed earlier to patch one of the uranium holes disintegrated.
Although the homogeneous reactor never found direct commercial
applications, the Laboratory's efforts to test its long-term
usefulness ultimately strengthened its capabilities for maintaining
and repairing highly radioactive systems.

The story then was of a successful experiment that demonstrated teething problems with a new technology. There were metallurgical problems, but the experiment was a success.

Two questions remain. Should the experiment have been continued, and should the Homogeneous Reactor be revived. My answers are yes and yes. The ORNL and Los Alamos experiments showed the Homogeneous reactor to have outstanding promise. The only reactor design that was more promising was the Molten Salt Reactor, but the Molten Salt Reactor did not hold the breeding potential that the Homogeneous Reactor did.

The Liquid Sodium Fast Breeder is still the "official" candidate for breeder of the year. Yet no one has yet mastered sodium technology, and the last Liquid Sodium Breeding experiment, the French Superphoenix, has to be seen as a failure. Alvin Weinberg continued to view the Homogeneous reactor as having a place as a breeder, even after design work on the Molten Salt Reactor had been launched. The Homogeneous Reactor thus is a legitimate candidate for "Breeder of the Year," or the generation or even the century. It can breed thorium with high neutron efficiency. The resulting U233, can fuel Uranium burning Molten Salt Reactors that will in tern breed enough U233 from Thorium to sustain themselves.  The Homogeneous Reactor is safe, far safer than the safest water cooled pile, it is simple, it solves many of the problems that critics of nuclear power raise, it can breed like a rabbit, it require modest amounts of materials, can be operated without refueling shutdowns,  produces hydrogen as a by-product, produces, can produce electricity as efficiently as a light water reactor, and outside the corrosion problem has few vices.  It also would be far cheeper to build.  The Homogeneous Reactor is very flexible in output. It is said that the output can increase from 100 watts to 1,000,000,000 watts with no problems.  For that reason it could be used to balance, and back up intermittent renewable energy technologies.   Given a modest research input, the Homogeneous Reactor could be making a significant contribution to fighting global warming within 20 years.


DV8 2XL said...

Nice article on a much overlooked technology. I came across mention of the Aqueous Homogeneous Reactor by chance and was so impressed that I started and wrote the initial Wikipedia entry on the subject.

While it's true that solid fuel reactors pushed out core programs, my own research into the history shows that the corrosion issues more than anything else was the reason for the halt, in as much as no one at that time could come up with any practical idea how it could be solved, or could say with any real credibility if it could ever be solved. In fact today fifty years latter it's still conjecture.

Not that we shouldn't be looking you understand, but it's still a long shot.

But oh damn it would be great if that horse came in!

Anonymous said...

Charles, could you find out (or ask your dad) which building HRE-1 and HRE-2 were housed in? Were they housed in the same building as the MSRE? (ORNL 3019 I think)

Anonymous said...

A few things:-

(1.) There was a similar long term corrosion problem for stainless steel tanks from the white fuming nitric acid (WFNA) used for certain rocket fuels. They solved the problem by putting in a little hydrofluoric acid or ammonium difluoride which built a protective coating. (This was classified, but it was inadvertently released because they made a patent application - something which also happened with pre-war bomb fuses.) This might do the trick, provided it was compatible with the salts used for fission.

(2.) Recent work has shown that liquid, or better still supercritical, carbon dioxide is a remarkably useful solvent. Not only would it do the dissolving, it would also be a far more cost effective moderator than heavy water (not as efficient, but cheap enough to use in larger quantities). What would happen using this, always supposing that we cound find convenient fissile salts that would dissolve in it? Even though it has covalent bonds rather than being able to separate into ions under radiation and rejoin later like water, we know it doesn't suffer when being used as a coolant. Most probably you get free carbon forming carbon monoxide in short order, carbon monoxide forming directly, and free oxygen mopping up carbon monoxide as soon as it finds it. Would the free oxygen be more or less corrosive than the free hydroxy radicals doing the job in an aqeous reactor? Certainly there wouldn't be any free hydrogen to penetrate and weaken the materials.

(3.) And there is always the possibility of using carbon tetrafluoride (or tetrachloride, with the right isotopes). This would also provide a protective layer on stainless steel, but it might be so hard to find a soluble fission material that you would have to use a slurry, and then abrasion might break the layer down.

Options 2 and 3 might also suffer from having to operate at lower temperatures, if there were problems from higher pressures, but since pressurised water reactors are practical - at any rate with interior pressure vessels, like the CANDU reactor - that doesn't look likely. And even if corrosion problems remained, you could probably use option 2 as a bridging technology long enough to breed a starter stock of fissile material for a molten salt thorium reactor.

Anonymous said...

After doing some searching, it appears that there is a salt available to try option 1 above: uranyl fluoride. However, searches showed that people are currently using supercritical carbon dioxide to dissolve uranium in the form of organic complexes. That suggests that simpler salts wouldn't dissolve (or they would have been tried), but those complexes wouldn't hold up in a reactor environment. So, either you could get salts to dissolve if you spiked the supercritical carbon dioxide with water (maybe even light water would work, if you didn't have to use too much), or you would have to use it for a slurry. But if you are going for a slurry anyway, why not just put a lot of powdered carbon in it, e.g. sugar charcoal, to get a fluidised bed and not have to bother with the high pressure environment? So at the moment we have three things that look potentially realistic (subject to further testing):-

(1.) An aqueous homogeneous reactor with heavy water and uranyl fluoride, spiked with heavy hydrofluoric acid to protect the pressurised stainless steel reactor vessel.

(2.) A fluidised bed reactor, with a powdered sugar charcoal moderator and something that wouldn't be broken down as fissile material and fluidising gas.

(3.) A variant CANDU reactor with a supercritical carbon dioxide moderator (probably spiked with carbon monoxide to head off corrosion by free oxygen).

Anonymous said...

As a Nuclear Engineering Student at UMass Lowell with aspirations of starting a company in the nuclear industry, this fascinates me.

Charles Barton said...

Shawn my father worked on this reactor. I can still recall reading a story about it in the Oak Ridge Newspaper.

Anonymous said...

I may have found another contender for a pseudo-aqueous homogeneous reactor: supercritical sulphur hexafluoride with uranium hexafluoride gas in it, maybe not even enriched, with the fluorine content providing most of the moderation (only, would it be necessary to use certain isotopes of sulphur? I'm guessing not, as this wasn't mentioned as an issue in the uranyl sulphate fuelled experiments). On the face of it, supercritical carbon dioxide with uranium hexafluoride gas would have worked, as the latter does not attack the former, but once things got ionised insoluble compounds might build up and settle out. There's no risk of that with the twin hexafluoride approach, particularly if it is spiked with a little surplus fluorine, as the original compounds would tend to reform. You would only need to clear out the fission products, which is manageable as the fluid moves through the reactor vessel to the reprocessing unit(s) before much can build up. That does leave the question of whether the reactor vessel could take it, but I suspect that the protective fluoride layer idea would still work. It definitely looks worth further investigation - again, particularly if it worked as a breeder using a thorium blanket.

Charles Barton said...

Mr. Lawrence You are definately getting over into the things done better with MSRs territory. The aqueous homogeneous reactors used heavy water as their carrier, and heavy water is far superior as a moderator to fluoride. I think my father who did research on the chemistry of both the aqueous homogeneous reactor and the MSR would shake his head at the idea. There are numerous advantages to going with the MSR rather than the aqueous homogeneous reactor.

Anonymous said...

Charles Barton, we are talking two different things. Of course heavy water is a superior moderator - when you've got it. Likewise a Molten Salt Reactor is clearly superior to an Aqueous Homogeneous Reactor - once you can operate on that scale, have solved the chemical engineering issues, and (for breeder reactors) have built up a starter stock of fissile material.

However, I was thinking in terms of a bridging technology to get started, both to help with the R & D and to get early energy production and/or fissile material breeding. Even though these moderators and this approach would be less efficient in engineering terms considered in isolation, they would probably be more cost effective starting from scratch.

It's a bit like the old joke about a man trying to buy a shirt who tells the shopkeeper that shirts are half the price down the road.

"Then why don't you buy them there?"

"Because they haven't got any in stock."

"Well, when I haven't got any in stock, my shirts cost half too."

For what it's worth, I think at this stage the supercritical carbon dioxide variant of the CANDU reactor is marginally the best way to start, because it doesn't quite have to start from scratch (there used to be quite a few problems, but the Canadians already solved those); all up there are fewer outstanding set up problems. However, if the corrosion problems can be solved and suitable moderator/fuel combinations can be found, a variant Aqueous Homogeneous Reactor becomes the lead contender - for this role. But yes, once all those other issues have been dealt with and you're in full operations mode, then yes, your suggestions are the way to go.

Anonymous said...

My apologies if this obsessiveness is getting boring, but...

I was a little worried that the sulphur in my earlier suggestion might absorb too many neutrons unless the uranium were enriched, so I tried to check it out. There wasn't enough on the internet to settle the matter one way or the other. A hint in other work where people were considering phosphate fuels suggested possibilities based on phosphorus. From this I got "These high-energy neutrons can be subjected to 4 basic actions:... 2. Elastic Collision. The neutron collides with another atom and leaves with most of its energy intact. Sodium, Aluminium, Silicon, Phosphorus, etc.", so it qualifies! Also, phosphorus pentafluoride is a gas with suitable physical properties, i.e. it is in the most oxidised state and has decent critical pressure and temperature (unlike silicon tetrafluoride, which that list might also suggest), so I can adapt my earlier suggestion to read phosphorus pentafluoride for sulphur hexafluoride throughout - again, always supposing that lacing the fluid with fluorine would provide the vessel etc. with a protective coating against corrosion.

DV8 2XL said...

Rehabilitation of the aqueous homogeneous reactor is a favorite pastime of nuclear engineering undergrads. Many have attempted to solve the problems inherent to this design.

They have all found that there is no combination of materials, pressure, and chemistry at any given power density that will yield an economically viable power reactor.

Some work is being done in Russia on this type of system for isotope production, but to the best of my knowledge, they have not yet come up with anything substative to date.

As Charles said up thread, one is still better off looking to the MSR system for power generation from a liquid phase nuclear reactor.

Anonymous said...

"They have all found that there is no combination of materials, pressure, and chemistry at any given power density that will yield an economically viable power reactor" [emphasis added].

See above - economically viable isn't the object of the exercise for me, a bridging technology is. For instance, if it took ten years to get a properly understood molten salt thorium breeder reactor built, it could well be worth while running some research style simple reactors both to build up the knowledge base needed and to build up a starter stock of fuel.

If I thought these speculations could lead to a serious industry base in themselves, I would have serious reservations about discussing them in public, for reasons of proliferation if nothing else.

DV8 2XL said...

Look P.M. I don't want to start a fight so I'll say my piece and you can take it or leave it.

AHRs are a dead end - beyond satisfying a certain purely scientific curiosity, there is not likely to be anything of real value to the MSR effort that can be learned from them. As I mentioned the Kurchatov Institute in Russia, has been pissing around with the ARGUS reactor - an AHR since '81 without coming up with anything new, because there isn't anything new to come up with.

Anything one could accomplish with this design can be done better with some other type of reactor. AHRs are just not worth the trouble. You are wasting your time.

Anonymous said...

There seem to be some cross purposes here, so I had better spell out where I am coming from.

My only reason for bothering to post is the intellectual fun of it. Nevertheless, I do see some practical value in this approach for those who don't have the resources to do it right - hence, bridging technology.

I 100% agree that "[a]nything one could accomplish with this design can be done better with some other type of reactor" - if that option is available. In fact that is why I suggested that it would be better to use a CANDU reactor if possible, with the variant of a supercritical carbon dioxide moderator for those who couldn't get the heavy water but who could get the expertise for a CANDU reactor (I have grounds for believing that that would work as a moderator, albeit less efficiently and at less efficient operating temperatures).

But suppose they can't even get that, and can only get limited access to fuel? To take a concrete instance, if Iran is serious about working towards a power delivery system, this sort of thing could help with that - the "working towards" part, not the actual achievement of a final operational system. (Obviously not for making weapons, since the realistic scale is too small - I don't think it presents a proliferation problem.) That is, for a country or group in that position, unable to tap into what it takes to do it right, it could be worth it as a stepping stone - even though the final stages would take further resources. The alternative is to wait even to start until everything else comes right, and then face the delays that the bridging technology could have headed off. Mostly what they would get is expertise, with (as I noted) the chance of building up a starter stock of fuel for later breeder reactors. Yes, those are better obtained in other ways - if you can do that. Not if you can't.

It's about doing the best you can with what's available. The fact that there are better ways is irrelevant if you can't access them. You should see some of Atkinson's ingenious ways of making an internal combustion engine without using Otto's patented - and "right" - way! This sort of thing is sometimes called "misapplied ingenuity" - "a triumph of misapplied ingenuity" when it actually does what it was supposed to.


Blog Archive

Some neat videos

Nuclear Advocacy Webring
Ring Owner: Nuclear is Our Future Site: Nuclear is Our Future
Free Site Ring from Bravenet Free Site Ring from Bravenet Free Site Ring from Bravenet Free Site Ring from Bravenet Free Site Ring from Bravenet
Get Your Free Web Ring
Dr. Joe Bonometti speaking on thorium/LFTR technology at Georgia Tech David LeBlanc on LFTR/MSR technology Robert Hargraves on AIM High