Transatomic Power – Anatomy of Next
Dr. Leslie Dewan is a co-founder and the CEO of Transatomic Power, a venture capital-funded start-up based on research conducted at MIT. Along with Mark Massie, the other co-founder, Dewan is exploring a design that uses a molten salt fuel that enables materials currently classified as “nuclear waste” to provide the heat source for a steam power plant. As the company’s web site explains, the design requires fuel where fissile isotopes (U-235 or Pu-239) make up 1.8% of the total actinide fraction.
Material meeting that fissile isotope requirement could be provided by lightly enriched uranium or by the reusable material produced after a single pass through a light water reactors and currently stored as “waste” in spent fuel pools and dry storage casks.
Transatomic’s ideas are exciting. Here is a recently posted video designed to generate interest in the concept and perhaps to attract interested and skilled talent to the team.
I’m working on arrangements to interview Transatomic Power leaders for the Atomic Show; you should expect to hear more about this company and its technology in the near future.
Many molten salt experts and nuclear physicists are rather stunned by some of TAPs claims.
TAP is claiming they can breakeven on breeding with 2% fissile concentration, with a reactor that is full of neutron absorbing hydrogen, zirconium, cladding (which they haven’t even specified) and fission products while operating with a big portion of epithermal neutrons.
I’m really excited by the idea of TAP’s reactor design. I’m looking forward to hearing more information about it.
Rod, the primary research was conducted by oRNL, with important contributions by Russian and French researchers. MIT followed the path, and made some R&D advances, but MIT certainly did not originate most key concepts.
I was surprised to hear that TAP’s reactor operated in the thermal spectrum, as I thought most minor actinides had larger cross sections in faster flux regions than they did in the thermal spectrum. Interesting design though.
MRS’s show alot of promise, but I’m not so sure about the salt’s that require highly enriched 7Li. Massive expense and environmentally troubling.
@Charles Barton
You are correct about the locations for the primary initial research.
Rod – Thanks for this post. @Jeff S – it’s well worth going to TransAtomic’s site for their technical white paper (http://transatomicpower.com/white_papers/TAP_White_Paper.pdf). It may give you more information than you want, but I have skimmed it and find it very thorough. You’ll find out a lot about the technology and there’s a lot of engineering detail as well. Their main focus seems to be electricity generation. I hope that they’ll start talking to industrial heat users as well. The reactor outlet temperature is 650 Celsius, which IIUC is a useful temperature for industrial processes. They claim 96% burnup of the fuel, which is a big plus for me.
It is actually a hybrid reactor, with a bunch of fast neutrons that improve the breeding ratio (since this is a U238 reactor) and a bunch of epithermal/thermal neutrons to sustain criticality.
So it is a fast and thermal reactor in one.
Question is does it break even on breeding? to do that would need most of the neutrons in the fast neutron energy region. But that means the thing won’t go critical with 2% fissile. Whereas more in the thermal region the conversion ratio drops off rapidly so you can’t breed anymore.
I wonder if Lockheed’s “skunkworks” Fusion isn’t really seeking a nice fallback as a neutron generator for just such fission breeder designs as this Transatomic power MSR and other similar designs. Low enrichment, smaller fuel loads and less stringent materials requirements with respect to neutron absorption could be a hallmark of many working designs.
I don’t think TAP ever claimed they had a breeder. More like a converter. WAMSR has a conversion ratio of up to 0.97 [page 30], with a burnup of up to 96%. I privately believe engineers are reluctant to design breeders due to the anti-proliferation arguments that sunk the IFR!
WA-MSR has some fast neutrons too, with very low activity in the epithermal region. It is designed to burnup TRU as well as fissiles. [page 13, blue curve]
Even 0.97 is very dubious on these conditions.
We’re talking about a fairly small reactor, perhaps 1200 MWt, so it is a bit less efficient than the big LWR monsters which operate at 0.5 to 0.6 conversion ratio. Then it has large amounts of zirconium and a cladding to be determined, that eat neutrons. Then it has a lot of light hydrogen which eat neutrons. Then it claims to start up on 2% or so enrichment with the fission products in it from high burnup LWR fuel, ie lots of fission products.
CANDU has lots less fission products even at end of cycle than starting up a reactor with old LWR spent fuel as-is.
Problem is this. To get a high conversion ratio on U238 cycle, a lot of fast neutrons are needed. With too few fast neutrons you’re into LWR territory of 0.6 perhaps.
But, many fast neutrons means you can’t start up on 2% or so fissile that is available in spent fuel. Fast neutrons don’t do much for maintaining criticality if you only have 2% fissile (ie 96+% U238). That’s where thermal or epithermal neutrons are needed for. You’d need more than LWR enrichment which is unavailable in the spent fuel itself. TAP likely needs extra enrichment fuel for startup to get a reasonable conversion ratio.
There are other issues with TAPs claims. They claim to cut long lived waste production by a factor of over 1000. This needs isobreeding, 0.9 or even 0.97 isn’t enough to make a big difference here. In fact if the burnup is low (high reprocessing rate) you lose a bit of long lived waste transuranics to the waste bucket.
The burnup of 96% is actually not strictly telling the whole story. The burnup in this concept is actually quite low because the fission products have to be processed out to keep the conversion ratio up. That means short burnup, perhaps even below 1%, each pass. Since reprocessing isn’t perfect some long lived fuel material passes into the waste bucket rather than fissioning off eventually in the reactor. So this means it becomes very difficult to get to the long lived waste claims of TAP.
As a result I feel the concept is being oversold and oversimplified. It might be an interesting converter if they can get a cladding to work safely and reliably (the big technical and development issue with this design).
“but I have skimmed it and find it very thorough.”
Mostly, yes but there are some giant holes. Their cladding is “–“. I’ll check with suppliers if they can deliver me some — to see what this magical stuff is.
Seriously, a new reactor with a new chemical and thermal environment that needs moderator cladding with hot hydride inside and molten salt outside in a high thermal AND fast neutron flux… this is a major red flag for me.
The other big hole is the conversion ratio they think they can achieve.
@Cyril R
I understand and concur with your questions with a few exceptions.
1. Fission products. Other than samarium and xenon, what fission products have significant neutron absorption cross-sections? Most fission products are born with too many neutrons and undergo a series of decays that tend to bring their nuclei back into balance for their position near the middle of the periodic table.
2. There is a large spectrum between fast and thermal neutrons (the epithermal region). The behavior of a well distributed moderator like light water reactor is not necessarily similar to one where the moderator is distributed differently.
3. Neutron leakage (which is what I think you are referring to in your statement “a bit less efficient than the big LWR monsters”) is determined by physical parameters like surface to volume ratio, not power level.
On #3:
My understanding is that the goal of TAP is to get a bimodal neutron spectrum. If they can do that, it seems like an extremely useful approach for a converter. Faster neutrons help fission the higher actinides, while the slower neutrons maintain criticality at relatively low enrichment levels.
Number 1, neutron poisons. Neodymium and praseodymium are pretty bad too, and the combined effect of Cs, Mo, noble metals adds up. Most fission products have worse capture than zirconium. TAP has claimed to dissolve spent fuel as is, in fuel salt. That seems a bit oversimplifying (first need to chop off the cladding and hydrofluorinate the UO2 to remove the oxygen otherwise it ruins the fuel salt) but if they do this they get most of the FPs except gaseous.
Number 2, epithermal neutrons. TAP mostly avoids the area between thermal and fast, hydrogen is a quite good moderator but salt isn’t so neutrons that stream through salt stay fastish whereas neutrons that pass through moderator will be well thermalized. Its good that TAP avoids this area mostly because there are some bad capture resonances here.
I’m not sure I agree with the TAP being all that different from some other reactors, in this respect. CANDUs for instance have tightly placed fuel which doesn’t moderate, with D2O in between but not that much and D2O has a low lethargy compared to light hydrogen molecules. So, a CANDU gets some bi-modal spectrum effect with fast fission gains.
Number 3 – yes granted that was not the correct terminology I used. Still TAP has some issues with neutron leakage even if their core is very low power density (which by the way is not attractive for molten salt reactors because the fuel inventory increases with the coolant inventory unlike solid fuelled reactors). The fast portion of the spectrum that they employ is much more likely to leak out. A big moderator reflector would prevent fast leakage but it actually increases thermal neutrons near the edge of the reactor so more fission there which means more neutrons to leak out at the edge… bit counter intuitive that a reflector can increase leakage but this effect is real.
I will note that BWRs have similar breeding ratio as PWRs even though BWRs have half the power density. To really get a big gain in neutron loss reduction would need a very low power density which means an enormous reactor.
Cyril, would you say that an IFR-type concept (with pyroprocessing employed also in order to fabricate IFR fuel from LWR spent fuel) is more promising than the TAP concept for LWR spent fuel consumption? If so, why?
Off-topic: what breeding-ratio do you expect is achievable in the current Russian BN reactors, specifically the big ones they are planning to build? AFAIK they use oxide fuel, which I understand is not very good for achieving high breeding ratios.
Related: what breeding-ratio could have been achieved in the French Superphoenix reactor if it had gone on to be operational with the purpose of maximum breeding? Similar to the Russian BN?
And what about the Monju reactor in Japan? Similar?
For whatever reason, this has turned into a masterpiece of PR. Maybe the money they can raise will mean they can turn this from hype to reality. I feel a little bad for people like Kirk Sorenson, and David LeBlanc, that seem to actually know a lot more, but don’t get anything like this attention. Apparently coming out of MIT counts for a heck of a lot.
Thank you for the link. It is an interesting read.
@SteveK9
It certainly doesn’t hurt. One of the great advantages of a place like MIT is that there is a strong alumni network with many successful grads that have a predisposed interest in participating in or funding a venture by other grads. (I’m fortunately enough to be a graduate of a school with at least some of that capability.)
It also does not hurt that the current Secretary of Energy is from MIT. I’ll refrain from mentioning another fairly obvious reason that TAP gets more attention than David LeBlanc.
@Cyril R – thanks for all your comments on AI and especially on this thread. You help me to discover just how naive I am about real nuclear engineering. But then, revealing my ignorance should help me get educated….
I’d like to see a lot of reactor designs actually getting built, tested, and marketed. The energy and fuels markets are very large, and there’s room for a lot of different designs. Evolution and adaptation work by successive cycles of build, test, and select.
I get what you’re saying, Rod, and it has nothing to do with LeBlanc being Canadian.
Rod & Cyril R.,
Concerning question #3 above, on fission products and neutron capture, don’t forget about the build-up of gadolinium-157 in used fuel >~10 year old vintage. I know this to be a major limiting factor with the DUPIC AIROX process which basically simply vents off the xenon gas and re-clads the slightly used LWR fuel. DUPIC (utilizing LWR fuel <10yrs old) combined with once through thorium is capable of ~20GWD/ton of re-conditioned LWR fuel compared to the 7GWD/ton for natural uranium in conventional CANDU fuel; plus the bred U-233 stockpile fuels the next generation of LiFTRs.
That’s where I see a huge advantage of SMRs: being smaller / cheaper / faster to build, the cycle time from design idea -> build / testing -> deployment –> results –> design update can be considerably less than what would be possible with large-scale reactor design/build projects. Under the SMR paradigm, many units of many different designs deployed around the world coupled with much shorter innovation-cycle times will result in much faster evolution of nuclear technology. [Of course, the above statements presuppose a friendly public-policy environment.]
Evolution by artificial (market) selection is how the free market works it magic in delivering innovative technological solutions that meet customer needs. We are going to need A LOT of nuclear innovation if we are going to decarbonize the global economy while simultaneously lifting billions of people out of energy poverty over the coming decades!
What is it that produces the Gd-157?
Recently I have put some of my thoughts on TransAtomics’ TAP concept (formerly WAMSR), as compared to other recent upstarts, in writing for my own reference, which I would like to share here.
While also proposing liquid fuel, TransAtomics’ TAP concept replaces graphite moderator with zirconium hydride, also a solid (ZrH1.6).
Since hydrogen is by far most effective at slowing neutrons, this allows a very compact high-power density reactor (cited as 86MWth/m3, in contrast to only 4MWth/m3 for MSRE).
Indeed, TAP’s 1,200MWth (520MWe) places it well beyond the 300MW limit of SMRs.
And while the concept succeeds at combining compact size with a large source of power – not unexpected for reactors with circulating rather than stagnant fuel – it comes up short in other respects.
First off, TAP relies on pure Li7-fluoride carrier salt, like most other MSRs, albeit without any BeF2.
As such it faces the same “unobtainium” procurement and corrosion issues, and assumes the standard Hastelloy structural material for both reactor vessel and moderator tubes.
TAP also loses half of its delayed neutrons in the HX circuit, like other MSRs with circulating fuel, which impacts controllability and safety to some extent.
Moreover, by eliminating beryllium from the carrier salt, it also loses delayed neutrons from the Be9(g,n2a) photo-nuclear reaction – the equivalent of which in Candu reactors, D(g,np), is valued for the large effective delayed neutron lifetime, much longer than delayed neutrons from fission products.
In terms of tritium production and containment, TAP will also need multiple serial HX loops, like ThorCon’s concept (perhaps only three instead of four if the “solar salt” is avoided).
The corrosion issue could get much worse for TAP than other MSRs in case a leak develops in one or more moderator tubes during operation, because reaction between the fluoride fuel salts and ZrH would produce hydrofluoric acid (HF) – far worse than any fluoride salt.
In some ways, TransAtomics’ TAP concept is the reverse of Moltex’ SMSR: instead of having fuel salt in closed tubes, they put the ZrH moderator in tubes around which the fuel salt circulates.
But the issue is similar in that heat transfer out of closed tubes is very limited, hence requiring a great number of thin tubes, to avoid excessive ZrH temperatures.
Thousands of tubes in fact, in both TAP and SMSR.
If nothing else, one must consider the impact on maintaining all those tubes leak-free over many years of operation, as well as providing the means for detecting leaks and managing retrieval/replacement of failed tubes.
The moderator in graphite-moderated power reactors reaches high temperatures even in designs where the nuclear fuel is cooled in a separate circuit (vis. RBMK, max. graphite temp. 750ºC; max. water coolant temp. 350ºC).
Temperatures as high as the graphite would make a great medium for heating steam for a Rankine cycle, comparable to high-efficiency fossil-fired plants
Unfortunately the graphite can’t be pumped around a heat exchange (HX) loop.
In terms of core lattice design, this inhibits optimisation of the reactor lattice, where fatter ZrH tubes would have been preferable, were it not for the inefficient heat transfer and limited temperature tolerance of ZrH.
Secondly, the consequences are much worse for TAP than for SMSR, since a large amount of Hastelloy in a thermal spectrum (TAP) is a far worse drain on neutrons than in a fast spectrum (SMSR).
Thirdly, as already noted earlier, zirconium itself is a fairly strong neutron absorber – with the large bulk of ZrH moderator in the TAP reactor adding significantly to the Hastelloy problem (Hydrogen is also a strong neutron absorber, but Zr91 is about 3.6 times more so, and Zr92 only slightly less; Some of the chromium, nickel and molybdenum isotopes in Hastelloy have absorption x-sections about 10 times larger than Zr91).
Nevertheless, TransAtomic claim that their analysis indicates that TAP could operate with fuel enriched only to 1.8% U235 (i.e. 2.5 times NU) and a fissile conversion ratio (CR) not far below unity (i.e. close to an “iso-breeder”).
That is a remarkable result, if true.
There is a lot of discussion about the choice of zirconium hydride as the moderator. This is a major element of this design. But the other major change is the use of a LiF-uranium-fluoride salt. This allows them to load a lot more uranium in their core than is conventional in MSRs. I’ve not seen a specific salt composition in their white paper, but it will be a lot more than one would normally think about. Loading a lot of uranium into a core allows lower enrichment–this is the principal behind the DOE’s reduced enrichment research reactor program.
I don’t know about Kirk, he’s been fairly quiet, but David LeBlanc seems to be doing well enough, with its Terrestrial Energy outfit facing much more reasonable challenges due to CNSC more reasonable regulatory framework and not trying to be the perfect actinide burner reactor, focus on making much more efficient usage of new LEU fuel first, then after they have an operational reactor in mass scale production they could calculate a profile akin to liquid MOX fuel for their IMSR. Dr LeBlanc does state their design can keep the transuranics in until fully fissioned, so at least they should be able to take transuranics rich fuel for startup.