Fast Reactors—No Free Lunch

nuke plant cropped

Cooling towers at a thermal reactor plant.

(A slightly technical post; but I find this subject pretty fascinating. This post is more-or-less an extension of my last post, “The Nuke Post” ; I’m going to assume that readers have read that first.)

As I alluded to in my last post and comments, whether breeder reactors are affordable or possible is perhaps a key to the question of the practicality and desirability of nuclear power as we go forward. So, the following is my take on the current state of the field, gleaned from quite a few different sources. Bottom line—1) they do exist and can be built, and 2) we probably can’t afford too many of them.

Breeder reactors are more appropriately called “fast reactors”, so I’m going to use that terminology. And to understand why they’re called fast reactors, we have to step back just a bit and look at “standard” nuclear reactors that are used to generate electricity, the vast majority of which are one form or another of a light-water reactor (LWR). (The “light” refers to regular water, instead of “heavy” water, the latter of which is a different molecular structure that doesn’t naturally occur in large quantities.) These reactors use fuel rods that are enriched to 3%-5% U-235, which are typically interspersed with graphite or carbon control rods, which can be slid in and out of the core to control the speed of the reaction. When they are slid out, the fissile elements (typically U-235) approach critical mass, with neutrons being released that cause, in turn, more neutrons to be released.

Fuel pellet. Reactor fuel is typically enriched to 3%-5% U-235.

Fuel pellet. Reactor fuel is typically enriched to 3%-5% U-235.

Now, in case you’re wondering why the chain-reaction doesn’t accelerate out of control—there are different degrees of critical mass, such as sub-critical, delayed subcritical, and prompt-critical. In most nuclear chain-reactions, some of the neutrons released have a delay that ranges from a few milliseconds to several minutes. In a power reactor this allows control rods time to control the reaction. Fuel in a power generation reactor could never reach a prompt-critical state (necessary for a nuclear explosion), even if the control rods were pulled and left out. The reactor would suffer a melt-down, or even a small explosion, but it wouldn’t be a nuclear explosion, per se. Nuclear weapons typically use plutonium, combined with conventional explosives, to achieve a prompt-critical state.

But back to light-water reactors—when a nuclear chain-reaction begins, the neutrons released have so much energy that they don’t tend to cause fission when they hit other U-235 atoms. They do this much better if they are slowed down, or “moderated”. Thus, the importance of water as a coolant, because water (light or heavy) is an excellent moderator. In LWR’s, water flows through the core, moderating the neutrons, and enabling them to more effectively split other U-235 atoms. A moderated neutron is sometimes called a “thermal” neutron, because its temperature (and therefore energy) has been reduced to that of the coolant. Thus, LWR’s are sometimes called “thermal reactors”.

Idaho National Laboratory's Advanced Test Reactor, a light-water reactor (LWR)

Idaho National Laboratory’s Advanced Test Reactor, a light-water reactor (LWR)

So, that’s that. LWR’s are pretty standard, make quite a bit of electricity around the world, and, unfortunately, leave a lot of high-level nuclear waste behind, as much as 27 tons a year for a 1 GW reactor. Typically about 97% of the original uranium remains in a “spent” fuel rod, along with those fission products I discussed in the last post. In fact, that’s one reason the rod is considered spent—the fission products are all excellent neutron absorbers, and as they build up, the reactor works less and less well, until it has to be shut down and the rods replaced (typically done in thirds, each rod spending three years total in the reactor, but with a third of the rods replaced each year).

This form of operation is considered an “open fuel cycle”; the rods are used once and then disposed of. (In the U.S., no one wants them, so each reactor facility stores them, first in pools of water until their radioactivity subsides some, and then in casks on-site) Now, many consider this to be a total waste, as it is possible for the spent rods to be reprocessed and the fission products removed, and the remainder mixed with new uranium and reused. But, plutonium is typically separated out as part of reprocessing, and proliferation fears caused a ban on reprocessing during the Carter administration in the 1977, and this ban still stands.

So, back to where we began, to fast reactors. They’re called fast reactors because, unlike all of those LWR’s, they don’t moderate the fast neutrons. By letting the neutrons keep their energy, they are much more likely to hit a fertile (but not fissile) U-238 atom, and transform it into fissile P-239. Thus, potentially making more fuel than they use. (Though it’s not some sort of free-lunch—they still need to be fueled with U-238). The cores have to use highly-enriched (17%-26%) uranium, or plutonium, (called “start-up fuel”) and then the core is surrounded with a “blanket” of fertile U-238 where the plutonium is made. That plutonium can’t be used for power until the blanket is reprocessed, though, but I’ll get to this is a bit.

As mentioned in the last post, these aspects of fast reactors have all sorts of advantages; it makes way more fuel available (all naturally-occurring uranium could be used), and the spent fuel is truly spent, containing only fission products that have dramatically shorter radioactive half-lives. Fast reactors can also be set to burn up transuranics, depleted uranium, or even weapons-grade plutonium.

And, fast reactors are a reality—as many as 20 have been built in various forms around the world, dating back to the 1950’s and 60s. But, there are a number of reasons why we don’t have a whole bunch of fast reactors making electricity and creating more fissile fuel than they consume. First, fast reactors use much smaller cores, and they have to run dramatically hotter than LWR’s. In addition, they can’t be cooled with water, because water is a neutron moderator, and would slow the neutrons down. (Another aside here—everyone says “cooled”, but that’s the same loop that pulls the heat from the core to use it to make electricity, so it’s really part of the power cycle, in all of these reactors). So, most fast reactors use sodium or lead as coolant. Both are transparent to neutrons, and as such don’t moderate them, but lead is highly corrosive at high temperatures, and sodium burns when exposed to air and water, sometimes explosively. (A few other designs exist that are gas-cooled, using helium or nitrogen, but nearly all fast reactors that have been built have used either sodium or lead). So, that’s drawback number one—a very hot core that has to be cooled with difficult materials, some potentially explosive.

Drawback number two—the higher the breeding ratio (or “burn-up”) that a fast reactor can achieve, the less power it tends to make, because the neutrons are off making plutonium instead of making heat. Additionally, scaling up a fast reactor tends to reduce its efficiency, as larger cores tend to trap neutrons and keep them from hitting the blanket. So, most fast reactors have relatively small power outputs, if they are used for power at all, typically 250-300 MW. Compare this to a typical LWR, which might make 1 GW. These two factors together—the difficulty of the sodium cooling, combined with the lower output, are the real reason we don’t see fast reactors. If they cost twice as much, but only make a quarter of the power, and then have more down-time, then the results are capital costs that can be ten times that of a LWR, per unit of power actually produced. Until and unless uranium is very, very expensive, such plants just aren’t likely to work commercially.

Now, drawback number three—fast reactors have to be refueled fairly often. This has made uptime a problem for commercial reactors like France’s two Fenix plants. Fast reactors use up their smaller, richer cores more quickly that LWR’s, whereupon they have to be stopped, cooled, and refueled. In liquid metal reactors, the sodium or lead then solidifies, which makes starting a fast reactor more difficult that starting a LWR. THEN (notice, a whole long string of difficulties here), the fuel in the blanket and the core have to be reprocessed. If this is done in the same facility, then it is referred to as a “closed-fuel cycle” plant, or an “Integral Fast Reactor”, or IFR. Reprocessing is complicated, and can be done in several ways, most involving liquids and organic solvents. Along the way, plutonium is isolated, which is why it is considered by some (but not all) to be a proliferation risk.

Then, in another example of no-such-thing-as-a-free-lunch, fast reactors can be operated on a spectrum from burning up nearly all the fuel in the core, and therefore making it disappear (useful for getting rid of weapons-grade plutonium or high-level waste from LWR’s), to the other end, where they don’t do that but make lots of plutonium in the blanket. At that first end of the spectrum they’re referred to as “burners”, and at the other end as “breeders” (another reason that the term “fast reactor” is more appropriate than “breeder reactor”). Unfortunately, they can’t typically do both at the same time. So if you want all those great burner advantages, you have to forego the creation of excess plutonium.

Now, all of this so far has been about fast reactors that exist. Those other Gen. IV fast reactors that I mentioned in my last post, like the molten salt reactor where the fissile materials flow with the salt, or the Liquid-Flouride Thorium Reactors (LFTR’s), don’t currently exist, and as far as I know haven’t been built, even in experimental form. Some countries are pursuing these designs, but the technical challenges portend to make them even more expensive than ordinary fast reactors, which already tend to be too expensive to be workable.

So, two last things. First, while I don’t currently view nuclear power as a wise path forward, for all the reasons I keep mentioning, I do think that there is a place for fast reactors to be used as burners. The U.K. is currently considering building GE’s new S-PRISM designs to help it deal with its 120 tons of near-weapons-grade plutonium at Sellafield, and the U.S. is prepared to let GE build one as a demonstration plant at the Department of Energy’s Savannah River site . These plants will produce some power, though at high cost; they are really better viewed as a way for countries to pay a price to burn up high-level waste of various kinds. Russia is taking a similar path, having recently announced that it will convert its BN-600 to a burner configuration to help burn up Russia’s 40 tons of plutonium left over from the Cold War.

And, lastly—these developments in the nuclear industry are occurring against the backdrop of steadily dropping prices for solar and other renewables. Nuclear prices have tended to go up and up, while the price of solar has tended to go down and down. If these trends continue, nuclear power is likely to remain technologically fascinating and yet not common. And that’s a result I can live with.

Links– Great technical info on fast reactors— “Fast Neutron Reactors” on the World Nuclear Association website.

Image credit: drakodav / 123RF Stock Photo
Image credit: Wikimedia Commons
Image credit: U.S. Nuclear Regulatory Commission

6 thoughts on “Fast Reactors—No Free Lunch

  1. CodySan

    Loads of good info here, Mr. B. Thanks for assembling it all. Perhaps you’ve mentioned this in some of your previous posts–I haven’t yet read all of them–but it seems to always come back to too many people using too much energy. There is no technology that can indefinitely sustain our current population (much less a greater future population) at the current level of energy use. How, then, we can reduce the population and/or energy use seem to be the key things, and this boils down to changing behaviors on a worldwide scale–not an easy task. Although making the issues clear–as you’re doing here–is an important step in the right direction.

  2. Taborri Post author

    With regard to energy, this is where efficiency comes in; efficiency is where the “free lunches” are. I do think we can sustain our current population, but as you say, we are going to have to change behavior on a worldwide scale, and it won’t be easy. A combination of renewable generation, more efficiency, conservation/less waste, and attitude changes with regard to the role of consumption will all be required. Along with circular systems (recycling, soil nutrients, etc.) and a transition, somehow, to some sort of a steady-state economy, along with more wealth equality worldwide. Plenty for everyone to work on, but we have to do it, I just don’t see any other alternative. Eventually maybe we humans can decouple our economy from its effect on the planet, but it isn’t going to happen anytime soon, if ever, so until we get it fixed we need to be extremely cognizant of the fact that we live on a finite planet.

  3. Damien RS

    Interesting! I hadn’t seen the “don’t scale well” claim before, nor the “burner or breeder, not both at once” point.
    Quibble: I think with fast neutrons it’s not that they don’t breed or cause fission when they hit, but that they don’t hit much. Quantum applies: high energy means small de Broglie wavelength, size, cross section, etc., while a slow thermal neutron is much fuzzier (large wavelength) and more likely to interact with a nucleus. It’s like trying to hit a baseball with a baseball, but if you throw your baseball fast enough it’s the size of a pea instead.

    Quibble 2: Traditionally fast reactors and breeder reactors have overlapped a lot, for the reasons you say. But the thorium fuel cycle apparently supports thermal breeding.

    Is “ten times more expensive” a real number or made up? I know breeders are more expensive for the power they make, but I don’t know if a 300 MW breeder would be 2x the cost of 1 GW LWR, vs. breeder power in general being 2x LWR power. (And that’s with something like mass production of or experience with LWRs.) Small reactors should be cheaper in lots of ways than bigger ones, just from needing less materials (including the almighty containment dome) offsetting some of the internal complexity.

    There’s also even more experimental designs, subcritical or accelerator driven ones, where a spallation beam or a forced fusion reaction provides neutrons causing fissionables to fission. Likely cost I have no idea.

    I’ve seen it said that one reason thorium wasn’t pursued was precisely because it’s not good at producing known bomb-quality material, when doing so was a goal of nuclear bomb states. And non-bomb states were largely getting technology from the bomb ones or in cooperation with them; AIUI reactors are mostly US/France style LWRs, Soviet style graphite-moderated reactors (cheap to build, cheap moderator, cheap to fuel, may catch on fire) or Canadian CANDU heavy water (cheap fuel, more expensive reactors [pure heavy water is expensive, and its a worse moderator so you need bigger reactors], not what we needed given uranium and LWR prices.)

    So it might be that thorium isn’t so much unworkable as unworked due to desire for proliferation.

    OTOH, it took all of 18 years to go from “nuclear fission exists” in 1938 to a British commercial power plant in 1956, which is pretty sobering compared to everything else. (Especially fusion, the basic physics of which were worked out *before* 1938…)


    As for needing nuclear or not, that might depend a lot on country and geography. I could see Mexico easily being a renewable-only state. But Sweden’s already on half-hydro, half-nuclear for electricity, and then there’s all the heating and transportation they need, and they don’t have Norway’s coast for wind. And your cloudy day pause post really handwaved away the storage problem, when it may be as problematic as, well, getting affordable breeder reactors.

    I don’t see why individuals need an opinion on “do we need nuclear?” other than keeping a mind open to “maybe!” and making sure risks/costs are accounted for without being unscientifically alarmed or unmindful of the costs of alternatives. For most of us, for perhaps anyone at the moment, the problem is too big and uncertain to have an answer.

    1. Taborri Post author

      Good points throughout. A few clarifications/responses– first, these posts are my attempt as a non-nuclear-physicist to understand the field, and they reflect my understanding after reading a chunk of material on the subject. But, there could well be something that I don’t understand correctly, or that needs clarification. So, if anyone out there is (or knows) someone with genuine expertise in the field, they could probably provide some additional insight.
      The “don’t scale well” and the “ten times more expensive” (per unit of power) are both things I’ve read, and both points seem to match what we see in the world. Though, like you say, it’s hard to make fast reactors cheaply when they aren’t “tried and true” like LWR’s. As to the actual cost of power from something like the BN-600, I don’t know (nor do I know if the price of power in Russia is a good indicator of the cost of making it).
      Now, the “burner or breeder but not good at both” I did NOT see directly written, but seems to match everything I’ve read. (Though, that being said, just because something is written somewhere doesn’t make it true, another reason I’d like to get input from someone who knows these specifics. ) On the whole, I think I have the basics right, and I think that the cost/complexity of these plants combined with their effectiveness at making power is the key reason we don’t see more of them.
      I have also read that thorium supports thermal, but I’m not sure how that would work since the thorium isn’t fissile. (India is working on some sort of three-stage system with different reactors at each stage, and thorium used in one of them (the final?)).
      Grid-storage– I think we’d all be surprised at how much demand could be shifted to daylight hours if we all put our minds to it. Though, even in my “Cloudy Day Pause” scenario you end up with lots of storage, either in EVs or in thermal mass in buildings, it’s just not “grid-scale” storage, per se. This field is definitely one where the answer will likely lie with a large number of smaller solutions.
      And lastly, though I’m a fan of the market, I do think that we need to have an opinion on nuclear power. (though that opinion may be that we just don’t know yet). If we don’t have an opinion, then the market will decide, and since many intangibles can’t be priced into the system, it might be better if society helped direct. For instance, whether or not it’s wise to have extremely complex systems as we go forward into an uncertain future (systems that won’t work well in hard times) isn’t something that the market will take into account.
      Thanks for your comments, though; much to think about. -tb

      1. Damien RS

        Allegedly Th-232 captures thermal neutrons better than U-238 does, also U-233 generates more neutrons than other fissiles, making thermal breeding viable. You need outside neutrons sources to get started, whether that’s a supply of U-235 or Pu-239 fissiles, or more exotic sources as I mentioned. Afterwards I think you could self-sustain with the bred U-233. The U-233 is contaminated with high-gamma U-232 making bomb-making hard. It obviously takes a lot more neutrons to get up to Pu-239 or other transuranics, so less of that sort of waste, and the Pu made is likely Pu-238, useful for space missions.

        OTOH it takes 27 days to go from Th-232 to U-233, 10x longer than Pu-239 breeding, and the intermediates are neutron absorbers, so you need more regular processing and storage (thus ideas of molten salt reactors, so you can cycle the liquid without having to pull and crack fuel rods), and the cycle does generate Pa-231 as 33,000 year actinide waste. I couldn’t find anything on burning up that like heavier actinides.

        I assume you’ve read Murphy’s storage posts?

  4. Taborri Post author

    More good info; I can see why thorium is/will be difficult.
    Storage–I have read Tom’s posts. His posts deal with storage for the electrical grid/system as we know it, which will be very hard. Thus, the need to spot-price electricity and let the market shift demand to when it’s cheapest. This will require a true paradigm shift in terms of how people think about grid power. Fortunately, it’s already underway and can be achieved gradually over time.

Comments are closed.