The Atomic Show #038 – George Stanford (sodium cooled fast reactors)

George Stanford talks about sodium cooled fast reactors

George Stanford earned his PhD in experimental nuclear physics from Yale University and then spent his professional career doing nuclear reactor safety research at the Argonne National Laboratory. One of his special interest was the sodium cooled fast reactor program. He worked on the Experimental Breeder Reactor II which was the technical prototype for a future reactor proposal known as the Integral Fast Reactor (IFR)

George and I talked about various fuel cycles, their perceived advantages and disadvantages, and some of the politics associated with the IFR. We talked a bit about the passive safety experiments conducted on the EBR II, pyroprocessing of fast reactor fuel, and the uranium use efficiency possible with fast reactor fuel recycling.

We also spoke a bit about the 10 MWe Toshiba 4S sodium cooled reactor, which has been proposed as an alternative electrical power generator for remote areas. The first proposed location for this reactor is Galena Alaska, a small village on the Yukon River that currently depends completely on diesel fuel and kerosene for its energy supplies.

Here are a few related links that might be useful in helping to better understand our discussion.


About Rod Adams

No Responses to “The Atomic Show #038 – George Stanford (sodium cooled fast reactors)”

Read below or add a comment...

  1. Alessio says:

    About recycling actinides in fast reators,personally I don’t think that nuclear waste transmutation in fast reactors is a smart idea.

    Past economic failures of European and American breeders programs suggest us that fast reactor are much more costly and complex to built and operate than light water reactors (and maybe any other kind of thermal one).

    So I continue to believe that recycle actinides in a high burn-up,good neutron economy thermal (not fast) reactor,like pebble beds proposed in South Africa,is a good strategy for a once-through waste transmutation with of course the help of an efficient pyroprocess dry actinides reprocessing technology

  2. Alessio says:

    Sorry,Rod I wrong link.This post was referred to the Atomic Show number 33.I cut and paste there

  3. Those problems were specifically solved by the IFR (eliminating unnecessary systems).

  4. Rod Adams says:


    One of the problems with pointing to previous economic performance of liquid metal reactors is that all of the projects have been done as one of a kind plants without dedicated processes for follow-on and improvement.

    In addition, the projects have been almost entirely dependent on government money with political processes for allocating funds. With that structure, there is plenty of opportunity for people and organizations with reasons to oppose nuclear developments to add to the cost and reduce the performance of the systems.

    I am not capable of determining if liquid metal reactors with recycling could be competitive with other types of energy production systems, but it seems to me that there is a good chance that they could be competitive with at least some of the alternatives if allowed to compete on a level playing field.

    There really are such things as economy of scale, cost reductions due to learning, ways to achieve savings through mass production, and ways to make money by selling accessory products and services – for example, recycling plants can sell used fuel storage and radioisotopes for medical uses or irradiation services.

  5. Alessio says:

    Thanx,Rod for your answer.

    Although it wasn’t my intention to start here the debate,to let Stewart to understand the question I paste a quote of mine from Atomic Show n°33

    “… My own idea is,instead to re-burn them in fast reactors (see integral fast

    reactor program) that are costly to buil and difficult to operate with

    multiple recycles (I suppose due to low burn-ups achievable),to burn Pu and

    minor actinides in very high burn-up (e.g. > 700 MWg/kg HM),good neutron

    economy thermal reactors with only one reprocessing (pyroprocessing) cycle.I

    think that pebble bed reactors developed in South Africa or generally HTGR

    based could have these features

    Finally,I found a lot of Authors are really developing this point of


    thermal actinides transmutation versus fast reactors. (pag.50)

    Clearly,an other approach I thought about is to load pebble bed cores with

    two different kind of Triso,of course in the right proportions:only thorium

    Triso and

    only Pu and Ma Triso in order to “consume” all o nearly all Pu/Ma fuel and

    to reprocess only thorium Triso,if needed,or simply recirculate them in the

    reactor if not needed (low poison or parasitic absorptions?)

    A big question I know,but I’d appreciate any comments,

    suggestions or opinions. Thanks. “

  6. Alessio says:

    Returning to liquid metal fast reactor,I have a question about sodium behaviour.More exactly,what happens in a event of fire of the sodium in the primary circuit and what precautions do they care to prevent it? Thanx for your attention

  7. Nathan Nadir says:


    In the conversation here, there was some commentary on the suitability of U-233 for use in nuclear weapons. I believe that the US has tested such a weapon, but…

    …when thorium is bombarded in a reactor, the fast fraction of the neutrons induces a Th-232(n,2n) reaction which gives rise to Th-231, which decays to Pa-231. Neutron capture in this isotope gives U-232, which has a half-life of just over 70 years. All of the isotopes in the U-232 decay series are relatively short-lived and equilibrium with Tl-208 is rapidly established in the decay series. This isotope is a powerful gamma emmitter and its presence makes manufacture of nuclear weapons using U-233 very problematic.

    Pa-233, which is always present in thorium fueled reactors, and is an intermediate in the formation of U-233, also gives rise to a (n,2n) reaction from the fast fraction.

    U-233 can always be denatured as well, particularly if one uses “once through” uranium. Under these circumstances one can have a mixture of U-232, U-233, U-234, U235 and U-236 and U-238 all present. This effectively prohibits enrichment to a weapons grade fuel.

    U-236 is of course, a parasitic nucleus, but one can make adjustments for this. The presence of U-236 has the added advantage of assuring that any plutonium that results from the presence of U-238 will be contaminated with U-238. This limits the potential for the diversion of CANDU technology for weapons use. Because CANDU’s are continuously fueled, use for the manufacture of weapons grade plutonium is somewhat more of a risk with them than it is with PWR’s, BWR’s and other thermal reactors.

    U-233 is considered extremely proliferation resistent if used correctly.

    India, which has use thorium reserves, intends to build a thorium based fuel infrastructure. CANDU reactors, with their excellent neutron economy, under these circumstances will function as breeders, albeit with a much longer breeding time than fast reactors.

    I continue to advocate the presence of some fast reactors in the fleet though, mostly as you discussed in the show, to burn-up the even-nuclei. Two nuclei that we can get at in this fashion that are problematic are Pu-242 and Cm-246. (I’d be perfectly happy to let Cm-244 decay to Pu-240.) It is possible too, to achieve criticality, I believe, with Np-237 in a fast situation. The critical mass depending on the geometry is on the order of tons, and this isn’t of practical import, but fast spectra are definitely the way to treat Np-237. Np-237 is in my view, the most difficult nucleus that is likely to arise in the Thorium cycle.

  8. Alessio:

    What would lead you to conclude that an IFR would be inordinately expensive, especially given the elimination of unnecessary systems?

    To address the burnup issue, an IFR has about 20% burnup, and processing does not get more difficult from the first through the fifth cycles. And AFAIK, there is no process for recycling PBMR pebbles.

    Regarding sodium, there are a number of ways to prevent and/or contain fires–lower pressure, helium secondary coolant circuit instead of steam, or simply ensuring that there is enough sodium to cover the core if all the secondary coolant circuit steam were to react with primary coolant circuit sodium. If the core were flooded with water (which is unlikely at best), it would boil off and certainly not cause a power excursion. In short, the result is fire damage and a huge mess, but not much more.

  9. Rod Adams says:


    Thank you for your comment. While I understand some of the reasons why sodium cooled reactors are attractive for certain uses, they have a few characteristics that make them less suitable for the markets that I want to serve.

    These characteristics do not remove them from consideration – Toshiba, for example, believes that their 10 MWe 4S sodium cooled reactors is the right fit for distributed power generation for remote sites like Galena.

    My analysis shows me, however, that the sodium alone – especially for a pool type reactor with natural circulation – adds enough cost to the system to make it significantly (at least 20%) more expensive than the one that I envision. Sodium cooling also adds to operational costs since it must be kept hot enough so that it does not freeze in the piping leading to the heat exchanger.

    The need for a secondary heat exchanger (either for water/steam or for gas) also adds enough to the cost of building and owning the system to cause me some concern. Though heat exchangers are well known pieces of equipment, they are also well known as high cost, fragile devices that need careful attention throughout their lives in order to ensure optimum performance and longevity.

    Finally, if you do plan to use gas as the working fluid with a gas turbine heat engine, the use of sodium provides a more limited turbine inlet temperature than is possible for a TRISO particle based high temperature reactor. In those systems, 900 degrees C is well proven, 950 has a fair amount of supporting data, and 1200-1300 may be within the realm of future possibility. With sodium going through a secondary heat exchanger, it is probably limited to less than 800 C even in the ultimate future refinements.

  10. Alessio says:

    Stewart: thanx for your comment about sodium maintenance

    First,I didn’t say Ifr was an failure,rather than past breeder programs were at least economical unsucceses: I meant in partcular to Japanese Monju and French-Italian-German Superphenix.Moreover,IMHO,Ifr was only a too small scale project to establish “definitive” conclusions on a commercial scale.Of course,research has to go on in this field

    Secondly,you say there is no technology to reprocess TRISO fuels,at least on commercial basis.That’s right,but as I tried to explain above it’s not the point.Primarly we have to burn actinides accumulated in LWR oxide fuels and reprocess them,not TRISO spheres;in the future we could think to Triso actinides transmutation,if clearly economics of TRISO reprocessing are good and an oxide fuel actinides reprocessing technology is available.I think *this* is the most important point.

    Fast reactors should be developed as breeders to improve uranium resources use (even seawater uranium) more than actinides transmuters,IMHO


    If I understood correctly,you say helium could be a better coolant for fast reactors

    I don’t think it’s a good idea,helium can’t perform high power densities (in the order of hundreds of kW per liter,100 times than pebble bed reactors) IIRC needed in any design of fast reactors.Moreover,thermodinamic efficiences aren’t an issue in fast reactors – they are just incredible efficient in natural uranium use.

    Instead,I believe it’s a good choice for low density (kW per liter) thermal nuclear reactors

  11. Rod Adams says:


    Actually, I prefer nitrogen as a coolant for Adams Engines. These are not “fast” reactors, but they will be high conversion ratio reactors operating with fairly low reactor power densities.

    Based on my computations, however, the overall energy system power density is competitive with light water reactors that have steam plant secondary systems.

    When it comes down to it, for economical operation, that is a more important consideration than the power density inside the reactor itself.

    Thermodynamic efficiency is not the big reason that I like higher turbine inlet temperatures – it is the power output that gets me excited. If the same machine can be operated at a significantly higher turbine inlet temperatures – say an increase from 900 C to 1000 C, the power output of that machine can increase. That means that the same machine, with just a little tweaking or analysis, can produce more sales, leading to significantly improve profits over the course of a year.


  12. Alessio says:

    Of course,regarding actinides transmutation in high burn-ups (700/800 MWg per kg of HM,I suppose easily achievable in the future for ad hoc applications) fertile free (non uranium) fuels pebble bed/HTGR reactors,it is not worthless to note that we don’t need a multi recycling strategy – only one reprocessing and burn cycle is needed for a complete (or nearly complete) consumption of LWR actinides,without the trouble of higher production of actinides as happens with MOX in LWR reactors

  13. Alessio says:

    A last question about breeding vs burn-up in thermal/fast reactors

    Extract from wikipedia:

    ” All commercial Light Water Reactors breed fuel, they just have breeding ratios that are very low compared to machines traditionally considered “breeders.” In recent years, the commercial power industry has been emphasizing high-burnup fuels, which are typically enriched to higher percentages of U235 than standard reactor fuels so that they last longer in the reactor core. As burnup increases, a higher percentage of the total power produced in a reactor is due to the fuel bred inside the reactor.

    At a burnup of 30,000 Gigawatt days/ton heavy metal, about thirty percent of the total energy released comes from bred plutonium. At 40,000 Gigawatt days/ton heavy metal, that percentage increases to about forty percent. This corresponds to a breeding ratio for these reactors of about 0.4 to 0.5. Namely, about half of the fissile fuel in these reactors is bred there. [5]

    This is of interest largely due to the fact that next-generation reactors such as the European Pressurized Reactor and AP-1000 are designed to achieve very high burnup.[6] This directly translates to higher breeding ratios. Current commercial power reactors have achieved breeding ratios of roughly 0.55, and next-generation designs like the AP-1000 and EPR should have breeding ratios of 0.7 to 0.8, meaning that they produce 70 to 80 percent as much fuel as they consume…”

    If I understood correctly,we can achieve with LEU in LWR higher conversion factors with higher burn-ups,although perhaps not breeding

    Now,I know it is possible to achieve breeding in thermal spectrum with thorium as fertile element and with uranium 238 only in a fast spectrum;my question is: is it possible to achieve breeding or at least high conversion factors (near breeding) with enriched uranium in high burn-up thermal reactors (no thorium blanket)?

    This is an important feature,given the fact that not only new generation of LWR like AP-1000 or EPR use more enriched uranium (so higher burn-ups),but for example PBMR achieves burn-ups in the order of 100 GWd per HM tonne.In that case,which could be a typical conversion factor?

    Clearly,I supposed that in conversion factor definition we include plutonium fissile isotopes generated and fissionated in situ,not only that present in the fuel at discharge

  14. Rod Adams says:


    You might enjoy reading an article titled Light Water Breeder Reactor published in October 1995.

  15. Alessio says:

    Thanx,Rod. Yes,I just knew about Shippingport plant

    But,if I understood correctly,Shippingport used uranium-thorium cycle.My question was more about breeding using only low or high enriched uranium,even in very special and exotic configurations (never developed neither in a small scale,AFIK).Definitely,it could be interesting to know what are the limits in conversion ratios for an only uranium (LEU or HEU) reactor system

    The only fact I found about,correct me if I wrong,is conversion ratio should be lower than

    (eta -1) value = about 1.1 for uranium,assuming no neutron losses from the reactor.

    Currently,typical conversion ratios values are in the order of 0.6/0.8 in LWR and HWR

  16. Rod Adams says:


    The conversion ratios will vary during the life of a core, and they will be dependent on a number of factors including parasitic neutron absorptions in both fuel and not fuel material, neutron leakage, fissile material depletion, core temperature, etc.

    There are people who spend their lives playing with models and working out details of various cycles.

    For my money, however, a more interesting way to spend one’s life is to work on refinements to new generations of operating reactors where the ideas can be implemented and tested on machines that are actually doing people some good. It is kind of the difference between designing microchips in an academic or laboratory environment and working for Intel or AMD.

  17. Jim Baerg says:

    Hi. I recently listened to this podcast, long after it became available. One question I’ve had about liquid metal cooled reactors is why so much attention to sodium?

    I understand that lead or a lead-bismuth mix could also be used & I wonder why the lack of fire hazard with those materials didn’t make everybody drop sodium from consideration?

  18. Harold Shattuck says:

    Suggestion: Encase final ash after burning all the actinides in glass and drop the cylinders down very deep abandon oil and gas wells and seal with cement. They wouldn’t surface for probably millions of years, if ever. With a danger life of less than 500 years this approach would be super fail safe.