Where does the plutonium come from?

new_horizonsLast week I wrote about how the shortage of Pu-238 might impact the exploration of the outer Solar System, but I didn’t much get into where the plutonium comes from. After all, while there are trace amounts of natural plutonium, there certainly isn’t nearly enough to fuel a space probe. So this week it seemed as though it might be worth going over where we get our plutonium, if only to understand why NASA (or DOE) needs tens of millions of dollars to produce it.

On the Periodic Table plutonium is two spots above uranium – uranium has an atomic number of 92 (that is, it has 92 protons) and plutonium is at 94. To make plutonium we somehow have to add two protons to a uranium atom. The way this happens is sort of cool – and there are different routes depending on the plutonium isotope that’s being produced.

To make Pu-239, the nuclide used in nuclear weapons, it’s a fairly simple process. Natural uranium is over 99% U-238, which doesn’t fission all that well. Put the U-238 (which makes up a minimum of 95% of the reactor fuel) into the middle of a reactor, which is seething with neutrons from uranium fission, and it will capture a neutron and turn into U-239. The U-239, in turn, decays by emitting a beta particle to neptunium-239, which gives off another beta particle. Since each beta decay turns a neutron into a proton, these two beta decays suffice to turn a uranium atom into one of plutonium. Thus, a single U-238 atom absorbing a single neutron and being allowed to sit long enough to undergo two beta decays (a few weeks or so) will turn into a single atom of Pu-239. Making heavier plutonium nuclides is just as easy – when Pu-239 captures additional neutrons it turns into Pu-240, Pu-241, Pu-242, and more. Not only is it fairly easy, but it happens all the time in any operating nuclear reactor.

OK – so we can see how simple neutron capture and patience can give us plutonium nuclides heavier than U-238, but this really doesn’t help us to make the Pu-238 needed to power a spacecraft. Making the lighter nuclide is a little more roundabout.

Remember that, through neutron capture, a reactor produces Pu-241. It turns out that Pu-241 also decays by beta emission, creating Am-241 – the stuff that’s used in smoke detectors (among other things). Am-241 is an alpha emitter and it decays to a lighter variety of neptunium (Np-237) which, when subjected to neutron irradiation, captures a neutron to become Np-238. One final transformation – a last beta decay – is the last step to producing Pu-238. This is the reason why Pu-238 is so expensive – making it requires two bouts of irradiation (the first long enough to produce the Pu-241), enough time for all of the radioactive decays to transform plutonium into americium and the americium into neptunium, and several steps of chemical processing to isolate the various elements of interest that are formed.

Although it sounds convoluted (well, I guess it is convoluted), making Pu-238 is fairly straight-forward. The science and engineering are both well-known and well-established, and its production certainly breaks no new scientific or technical ground. But the politics…that’s another matter altogether.

As I mentioned last week, the American Pu-238 production line shut down over two decades ago. Since then we’ve been buying it from the Russians, but they’ve got their own space program and have limited stocks to boot. So this option is not going to work for much longer, regardless of the future of US-Russian international relations.

A recent blog posting by Nuclear Watch suggested that the US might be able to meet its Pu-238 needs by dismantling nuclear weapons and by digging into its inventory of scrap Pu-238 – it notes that the Los Alamos National Laboratory (LANL) documents indicate that over 2000 RTGs’ worth of the nuclide can be recovered from nuclear weapons alone. But I’m not sure if I can accept this assertion, primarily because putting this nuclide into a nuclear weapon makes absolutely no sense. I can’t comment on the “scraps” of Pu-238 that LANL is said to have lying around, and unfortunately Nuclear Watch didn’t provide a link to the LANL documents they cited, making it difficult to check or to comment further. But if there is a Pu-238 stockpile at LANL it would certainly be nice to tap it for space exploration – not to mention the savings in disposal costs.

Yet another way to make Pu-238 is in a liquid fluoride thorium reactor (LFTR) – a reactor that uses naturally occurring thorium (Th-232) to breed U-233, which fissions quite nicely. Additional neutron captures can turn U-233 into Pu-238, which can be chemically separated from the fuel. There’s a lot more to the topic than this, but I covered the topic of thorium reactors fairly thoroughly last year (the first of these posts is at this URL, and there are three others in the same series) and it’s also covered on the Thorium Energy Alliance’s website. There are a lot of nice things about thorium reactors in addition to their being able to produce Pu-238, and it’s a technology that’s been worked out and tested – but the US shows no sign of building any of them anytime soon. India and China might develop extensive thorium reactor systems – but what these nations might do a decade or two in the future won’t do much for NASA in the next few years. The bottom line is that, however promising they might be for future needs, thorium reactors aren’t likely to help us send more spacecraft to the outer Solar System anytime soon.

So here’s where we stand. The US stopped producing the Pu-238 needed to run our deep-space probes and we’ve pretty much used up our stocks of the material. In the intervening years we’ve been buying Russian Pu-238, but that won’t be available for much longer, leaving us high and dry. There may be scraps of the material – possibly even stockpiles – at various DOE facilities, but dismantling nuclear weapons is probably not going to do the job. Over the long run thorium-cycle reactors might be a great way to make it, but these reactors aren’t operating anywhere in the world today and there are no American plans to build any of them anytime soon. That would seem to leave us with only three options – re-start our Pu-238 production line, find another way to make (or obtain) the material, or confine ourselves to the inner Solar System. As I mentioned last week, I sincerely hope we don’t go the last route. So let’s see what we can come up with – and let’s hope we don’t leave the solution (and decisions) too long.

Tags: , , ,

15 Responses to “Where does the plutonium come from?”

  1. G.R.L. Cowan September 30, 2013 at 11:14 PM #

    “Am-241 is an alpha emitter and it decays to a lighter variety of neptunium (Np-237)” –

    That is true, but the slowness of this decay, and the requirement for three neutron captures — by 238-U, by 239-Pu, and by 240-Pu — makes it an impractical means of producing 237-Np.

    Much more neptunium-237 is produced by *two* neutron captures, first by 235-U and then by 236-U.

    • Dr. Y October 1, 2013 at 11:47 AM #

      Good point – thanks!

  2. Mark October 1, 2013 at 6:55 AM #

    I started Googling and I found this:


    This is a DoE FAQ from 2005. Page 3 includes a table of Pu-238 stocks as of that year. A footnote mentions that apparently some nuclear weapons do include Pu-238 RTG’s as power sources. It lists 28 kg at LANL, 11 kg at Idaho National Laboratory, and “less than 20 kg” at PANTEX, and says of that 25 kg is needed for “national security use” but the rest would presumably be available to NASA. This is from 2005 so the Pu-238 from these sources may already be included in NASA’s planning projections.

    This GAO report:


    From the 80s mentions that only one type of nuclear weapon in the active arsenal includes an RTG. Annoyingly, it doesn’t say which one. This 2003 report:


    Also mentions “hundreds” of RTGs stored in a vault at PANTEX. Given that we don’t know the size of these RTGs, that “hundreds” may actually mean a relatively small amount of Pu-238.

    So, there was definitely a lot of Pu-238 available at various DoE complexes from dismantled weapons, but they don’t say which weapon contained it, so we may have already dismantled all of them and shipped the material to NASA.

    • Dr. Y October 1, 2013 at 11:46 AM #

      Thanks for the citations and links – you were able to find more than I could. And after reading this I have to say that I stand corrected. In all honesty, my imagination was a little too limited in that I was trying to think of how Pu-238 could be used as a part of the fission process and I never stopped to consider its use as an RTG. I appreciate your providing these links – thanks!

      Given that, I agree that it would be nice for the government to scavange what Pu-238 they can to help power our space probes, while we work to get the production line up and running again.

  3. Mark October 3, 2013 at 6:59 AM #

    I’ve been doing some more Googling. I’m not sure about this, but it looks to me like all of the scrap Pu-238 has actually been used already except for possibly the PANTEX material. According to the first link I posted, all but the PANTEX stuff and ~6 kg of the rest of the material available in 2005 had already been allocated to various missions. I found a reference in another doc (whose link I’ve misplaced) that seemed to be saying the PANTEX material is too dilute for use in new RTGs, and would require blending with ultra-high-purity Pu-238 made using the Americium process to be useable. Nobody seems to be talking about actually producing any Pu-238 using that method, so the PANTEX material may not actually be suitable for NASA use. Mr. Coghlan has access, I would assume, to documents I have not yet found, so perhaps he would care to elaborate?

  4. A. DeVolpi, PhD October 7, 2013 at 12:29 PM #

    Regarding your review of Pu-238 production and applications, the “liquid fluoride” reactor pathway can be expressed more favorably in terms of technical feasibility, product yield, and public benefit. In a briefing in September at FAS headquarters, I pointed out that a small molten-salt reactor, fueled with uranium (or thorium) could efficiently produce not only Pu-238, but simultaneously other radioisotopes in short supply, such as medical Mo-99, as well as tritium and He-3— the latter needed in homeland-security applications.

    A fluid-fueled molten-salt reactor in the range of 10 to 100 MW(th), sited on a government reservation, could more than satisfy national and international needs for these rare and valuable radioisotopes. It could generate them by means that are at least two orders of magnitude more efficient and less expensive than alternatives. And this can be done while reducing the risk of nuclear proliferation.

    My technical papers at the Winter ANS meeting, the Thorium Energy Alliance — and subsequent presentations at Argonne National Laboratory and the National Nuclear Security Administration — provide ample technical description of a fluid-fueled MSR dedicated to production of rare and valuable radioisotopes while reducing potential nuclear proliferation risk. And, did I mention that it would make money?

  5. Jay Coghlan October 7, 2013 at 7:30 PM #

    You write “But I’m not sure if I can accept this assertion (that Pu-238 can be recovered from dismantled nuclear weapons], primarily because putting this nuclide into a nuclear weapon makes absolutely no sense.” The plutonium isotope of choice for nuclear weapons is of course Pu-239. But I’m talking about recovering the Pu-238 from the radioisotope thermoelectric generators (RTGs) in nuclear weapons.

    The Department of Energy had this to say in its 2005 Draft EIS for the Proposed Consolidation of Nuclear Operations Related to Production of Radioisotope Power Systems (a final EIS was never completed):

    “Another source of available plutonium-238 is milliwatt radioisotope thermoelectric generator (RTG) heat sources removed from nuclear weapons as part of the ongoing weapons dismantlement program. A milliwatt generator is a very small RPS designed to produce a fraction of a watt of electricity, and it has been incorporated in nuclear weapons design since the 1960s. As the weapons are dismantled, a total of about 3,200 heat sources are projected to become available between Fiscal Years 2009 and 2022. These heat sources are located at Pantex and LANL…” Page 2-5, http://energy.gov/sites/prod/files/nepapub/nepa_documents/RedDont/EIS-0373DEIS-2005.pdf

    And please note that there should be even more dismantlements under New START and perhaps future arms control agreements or treaties.

    The same draft EIS projected <20 kilograms of Pu-238 remaining in 2010 (page S-6), overwhelmingly supplied by RTGs from planned dismantlements at Pantex. But unfortunately the Pu-238 scrap recovery line at the Los Alamos National Laboratory (LANL) needed to harvest those RTGs has apparently never been put into operation, despite a start up originally scheduled for 2005. In fact, LANL claimed in a 2008 site-wide environmental impact statement that it was capable of recycling/recovering up to 18 kilograms of Pu-238 per year, far more than needed to take care of the nation’s needs. Page 3-58, http://energy.gov/sites/prod/files/EIS-0380-FEIS-01-2008.pdf

    But there is one needed intermediate step. As Mark noted in his comment the Pu-238 recovered from RTGs is not of sufficient purity for direct use in new RTGs. However, LANL also claims it can “[p]rocess, evaluate and test up to 55 pounds (25 kilograms) of plutonium-238 per year in production of materials and parts to support space and terrestrial uses.” Ibid. So why can’t LANL also be processing the Pu-238 that it should be recovering to the level of needed purity?

    Therefore I conclude that rather than talking about new production in reactors we really ought to be pressuring LANL to recover and recycle Pu-238 for the nation’s needs, at a tremendous savings to the taxpayer. But then again, as a scathing Albuquerque Journal editorial put it just yesterday, “Bureaucratic ineptitude [is] entrenched at LANL.” So, in my humble opinion, those concerned about the future of U.S. space exploration (as I am) should be telling Congress and the Administration to make Los Alamos Lab get its act together, and for the U.S. to accelerate nuclear weapons dismantlements.

    Jay Coghlan
    Nuclear Watch New Mexico

    • Mark October 8, 2013 at 8:04 AM #

      First, thank you for your reply, Mr. Coghlan. However, your first link appears to be broken.

      Regarding purifying the Pu-238, that depends on what this material is contaminated with. If it’s contaminated with other plutonium isotopes, purifying it is impractical; the only realistic option would be upblending with high-purity Pu-238. I would not expect them to be even discussing the Am-241 production route if upblending were not absolutely necessary – and I note that your second document repeatedly refers to the need to “recover, recycle, and blend” the plutonium. I have to run to teach right now, but later I will see if I can dig up the link I misplaced about that.

  6. Jay Coghlan October 8, 2013 at 7:48 PM #

    I apologize for the broken link. A hyphen was missing. Please try

    Jay Coghlan

  7. Mark October 8, 2013 at 8:28 PM #

    Thanks! So, reading this document, I’m afraid I was right. The Pu-238 in the weapon RTG’s is contaminated with other plutonium isotopes, to a degree that makes it unusable without upblending. See page 443 in the pdf. For those who don’t want to go digging, the plutonium in the RTG’s was originally 80% Pu-238. The weapon RTG’s are old enough that a lot of the Pu-238 has decayed, but the other plutonium isotopes have much longer half-lives, so the proportion of the plutonium that is Pu-238 has dropped to about 60%. This is not useable without upblending with high-purity Pu-238. So this can supplement reactor production but cannot replace it.

  8. Jay Coghlan October 9, 2013 at 2:54 PM #

    There is no question that the Pu-238 from old RTGs would have to be upblended for use in new RTGs. I do question that production of new Pu-238 in reactors is needed to achieve that.

    To repeat from my previous comment:
    “As Mark noted in his comment the Pu-238 recovered from RTGs is not of sufficient purity for direct use in new RTGs. However, LANL also claims it can “[p]rocess, evaluate and test up to 55 pounds (25 kilograms) of plutonium-238 per year in production of materials and parts to support space and terrestrial uses.” http://energy.gov/sites/prod/files/nepapub/nepa_documents/RedDont/EIS-0373-DEIS-2005.pdf page 3-58. So why can’t LANL also be processing the Pu-238 that it should be recovering to the level of needed purity?”

    Mark says “If it’s contaminated with other plutonium isotopes, purifying it is impractical; the only realistic option would be upblending with high-purity Pu-238.”

    The Los Alamos Lab has aqueous processing capability for plutonium. Leaving aside issues of institutional competency, what are the technical reasons why LANL cannot sufficiently purify recovered Pu-238 to the needed 80% level? Would not reactor-produced Pu-238 also have to undergo a roughly similar processing step?

    To be clear, I am no plutonium technician, but it seems logical that it would save taxpayers money to recover and recycle Pu-238 instead of new production, and perhaps lower transuranic waste disposal costs as well. Further, the supply of potentially recoverable Pu-238 should increase with accelerated dismantlements of nuclear weapons.

    • Mark October 9, 2013 at 4:50 PM #

      The problem is that the Pu-238 is contaminated with other plutonium isotopes. You can’t separate two isotopes using chemical methods because they are chemically identical. You have to exploit non-chemical properties – separate them using centrifuges or gaseous diffusion. It’s the same method they use to make enriched uranium for reactor fuel.

      This is much, much harder and more expensive, and would require new, highly specialized facilities. We can’t just feed this stuff into a centrifuge cascade – I’m no expert, but if I recall correctly from discussions of modifying centrifuges to handle uranium recovered from reactor fuel rods, significant modifications are needed to cope with different isotope mixes, and I suspect that the chemical differences would also be extremely problematic (plutonium does not play well with other elements). And we won’t be able to extract all of the Pu-238 – I’m not sure what the efficiency would be off the top of my head, but centrifuges typically turn a lump of average purity material into a small lump of high-purity material and a large lump of low-purity material.

      So I’m about 90% sure they would need to build a specialized centrifuge cascade to handle it, which would be very expensive and have no other purpose besides purifying this material. Just making more Pu-238 in a reactor would be cheaper, especially since we would have to do that eventually anyway – there doesn’t seem to be much Pu-238 available in these old weapon RTG’s.

      • Mark October 9, 2013 at 4:52 PM #

        Sorry, I meant to add: the reason this isn’t necessary for Pu-238 that comes out of a reactor is that the material the reactor spits out is contaminated with lots of nasty stuff, but most of it is other elements. We can remove the other elements using (comparatively) cheap chemical processes, leaving behind the Pu-238.

  9. Gord November 17, 2013 at 10:30 PM #

    I gather NASA funded a study, about using a TRIGA (or similar), done by Center for Space Nuclear Research. http://www.nasa.gov/pdf/636900main_Howe_Presentation.pdf Date in the file is April 2012. I don’t see anything about “cooling” the neutron spectrum around this feeder loop. That would probably help with the resonance production. I had been thinking along similar lines for making U-233 from Th-232. Not at a TRIGA however.

  10. i'm not here January 27, 2014 at 7:52 PM #

    Mark gets an ‘A’ on this subject on the following…

    Mark October 9, 2013 at 4:50 PM #

    “The problem is that the Pu-238 is contaminated with other plutonium isotopes…”

    Thus the need for breeder reactors where short burn up times (3-4 months) of a higher percentage and purer Uranium produces the necessary purer PU that can be successfully exacted for bombs and heat generators.

    PU developing in a typical nuke power station is highly contaminated after a 3 year burn life which makes extraction costly and creates much nuclear waste in need of long term storage.

    Sodium reactors are just another way to produce ‘new’ PU, the same as the passe breeder reactors. Still requires a reprocessing plant alongside either type of reactors, for extractions and the resulting proverbial waste products. Works great if there are no accidents, you have a safe and secure facility and taxpayers absorb all costs and risk into the future.

    I don’t see the beneficial trade off for space exploration unless the elites are searching for another planet to escape to.

Leave a Reply