Taking the WW2 rhetoric seriously
The challenge of climate change is daunting. So daunting in fact that a lot of people are calling for a WW2 style mobilization. While I don’t believe that it will be possible to command equal amounts of the GDP as during the fight against the Nazis, I wonder how our options would change, if we called on the military to help us out.
The world consumes about 580*10^18 J or 580 EJ (exajoule) of primary energy. Primary energy is the name for energy found in nature that has not been subject to human engineering. This measure is primarily concerned with the heat content of raw fuels. Approximately 70% or so of this is lost in the conversion process to electricity or mechanical power1.
That’s the equivalent of 18,400 GW of average power, or 2330W of “primary power” consumption. That figure is eerily close to the energy consumption envisioned in the 2000W society. In this scenario, first world living standards would be achievable with “only” an average of 2000W of “primary power” consumption. Of course, “equity” would require all people to use that much power and allow only a small spread. That scenario is a quite different world from today. The citizens of a lot of countries currently use far less energy than 1000W. But Western Europe stands at 6000W and the US at about 12000W average primary energy consumption.
Without breakthroughs in efficiency, the goal of eradicating poverty will probably require more than 2000W of average power per person. If it’s closer to 3750W, the projected 9 billion inhabitants will require 1,000 EJ or 1 ZJ (zetta = 10^21) of yearly primary energy consumption. If we want to be able to give people access to the life style they seem to consider desirable, we have to become a zettasociety. I don’t know, if they are right in thinking it to be desirable, but for the remainder of this post, I’d like to work with this assumption. And all that energy has to be cheap, reliable and carbon-free. That leaves us with the task of getting 33,750 GW of average primary power.
The zettasociety
Energy is the master resource. When energy gets expensive, everything gets expensive. What we need for the world is a mindbogglingly huge amount of energy.
What an opportunity to improve the world! Today, 80% of the energy comes from fossil fuels. And a lot of that energy comes from underground, hardly visible to us. The majority of the rest comes from nuclear, hydro and biomass.
Given that we have some idea about how to get carbon-free power, the general assumption seems to be that this society will have to be primarily powered by electricity, either directly through the grid, or indirectly through some power-to-gas or power-to-liquid synthesis process for clean fuels. As far as we know, some of that society’s power will also have to go towards powering the sequestration of CO2 from the atmosphere. We are not really sure how to go about that, but we either figure it out or will have to bear additional climate related costs. Biofuels, geothermal and CCS will make up the remainder of total energy demand.
Goldstein-Rose2 has a nice illustration of the energy flows in such a society.
There aren’t that many options to scale the global energy needs. The sheer amount of land and material needed to get even close to the needed/desired amounts of energy with wind and solar alone make me believe that the cost to the ecosystems might be too much. Given that the deserts and the ocean are home to unique ecosystems, tough choices will have to be made. We will have to choose, which ecosystems to destroy and which species to exterminate, if we want to use the solar and wind route at the current state of technology. I do believe that nuclear energy constitutes an ecologically more benign alternative to that.3
The fusion route
In a perfect world, we would have figured out aneutronic fusion by now and could use readily available “vanilla” hydrogen and boron to make electricity with only stray radiation. Companies like Marvel Fusion, LPP Fusion or HB11 would have grown way past a startup stage and would be dominant players in the energy world. At least, we would have fusion plants working on the far less demanding deuterium-tritium fuel. But, alas, we are not there yet.
Even, if a wide range of fusion startups succeeded tomorrow, the 50MWe reversed-field configuration Helion generators, the 200 MWe ZAP Energy Z-pinch modules, 500MWe magnetized-target General Fusion power plants or 1.5GW tokamaks(Commonwealth Fusion Systems, Tokamak Energy) and stellerators would still have to be built and run. Humanity has built millions of ~100kW generators for gasoline conversion using ordinary materials. But building economic power generators is quite a new challenge and we are talking about tens of thousands of reactors. For the very large ones! Hundreds of thousands, even millions, for smaller reactors. For most of these designs, we are taking superconducting magnets, cryogenics and plasma control gear. That’s nothing you would find off the shelf. It’s hard to imagine how this endeavor would become cheap enough fast enough. But that’s what technological breakthroughs do: upend our prior expectations.
Of course none of the fusion concepts has shown even break-even power conversion yet, let alone achieved economic levels of power.4 Given the base rate of failures for startups, it stands to reason that only few of these companies will ever get to the milestone. On the other hand, just one might be enough. For as long as we are focused on getting economic fusion power plants, I think humanity should hedge our bets and look at alternatives.
The fission route
The second form of nuclear power, fission, has different problems. We know it works and have deployed hundreds of commercial units around the world. Cost is the main culprit preventing a wider adoptions. Many proponents claim this is due to regulations demanding safety standards beyond any reasonable level.
Another problem is often only referred to as “fissile startup problem”, but is probably the weakest point of nuclear power becoming the leading global energy source in the really short timeframe thought to be relevant for combating climate change. The reason is rather simple; you need a lot of fissile material before you can start a reactor. But there is only one fissile material to be found in relevant quantities in nature: U235. But in natural uranium, it’s dilute. Only 0.7%. While there are reactors capable of using that as fuel (CANDU), higher levels of fissile are generally preferable, at least from an energy perspective. We know how to enrich natural uranium to desired levels and have built a lot of the necessary equipment. These are a proliferation concern and generally under close scrutiny by international agencies.
Given our current light-water reactors technology, we would burn through all known mineable reserve in a matter of years. Siemers estimates approximately 10.5 Uranium from sea-water might be an alternative, but the price of harvesting it will probably increase the cost of uranium by a factor of 10 to 50. At that level, fuel costs, which until now were only a minor contributor to the price of electricity from nuclear plants, would become a considerable factor.
LeBlanc6 gives an overview of uranium consumption and fuel prices for different technologies and uranium costs.
The global nuclear workhorses are pressurized water reactors. Marvelous facilities. But by design, they are burners and need enrichment facilities to run. Currently, it takes countries quite a long time to build them. And while it’s really impressive how they are made, the world’s heavy forging capabilities might be another bottleneck for attempting to scale the current instantiation of nuclear.
The CANDU-attitude
Pressurized heavy-water reactors are probably a better scalable version of nuclear power. The Canadian CANDU is probably the most famous of such designs. They are featuring hundreds of horizontal pressure tubes with 104mm diameter and 4.2mm thickness 7. Compared to the 300m thick forgings for pressure vessels, far more countries and suppliers could conceivably produce these. There are even ideas for thermal CANDU-style breeders. But of course that would necessitate some amount of U233. India’s IPHWR-700 is probably the most advanced of these reactors and seems to be interested in a fleet deployment of these. These reactors also sidestep another potential bottleneck: enrichment facilities, because they run on natural uranium.
At $1000/kg of natural uranium, a level at which sea-water extraction might be feasible, the raw cost of uranium would be less than $0.02/kWh for these reactors. At $5000 the cost will almost certainly be too high. And that’s assuming that sea-water extraction works at scale. Currently, the feasibility and cost estimates of sea-water uranium are still mostly speculative and might be too much of a bet for our collective future.
If India’s published numbers on the capital cost of the IPHWR-700s are any guide, it might be competitive to deploy these reactors instead of coal power plants in developing countries. Or the price could at least be close enough for developed countries to make up the difference. That could buy the world some time.
Deploying them instead of PWRs might free up enrichment capacities for what I would like to call “the military option”.
The military option
The military of at least 5 “official” nuclear powers (USA, Russia, France, UK, China), 5 NATO members (Germany, Belgium, Italy, Netherlands, Turkey), and 4 “unofficial” nuclear powers (India, Pakistan, North Korea, Israel) are entrusted with handling high-enriched fissile material. Of course mostly in the form of weapons, but there is also navy fuel and some stockpiles of “excess” fissile material, which has not been used for weapons.
The commercial sectors is largely confined to using low-enriched uranium, with some hope of using high-assay low-enriched uranium in the near future. We refrain from using high-enriched uranium for proliferation reasons. And that is probably quite sensible in most situations. But maybe, it’s our only physically feasible way out, that also allows for getting people the huge amounts of energy they demand with a zero-carbon technology.
Let’s take some of the rhetoric about WW2 style mobilization to combat climate change seriously and think about the implications. What, if we allowed ourselves to use materials to combat climate change, that we typically reserve for war? What, if we used the same safety standards that we use during WW2 to solve problems?
The U233-Th cycle, which for non-obvious reasons has attracted a lot of attention and enthusiasm on online platforms, runs on high enriched uranium (HEU). U-Pu breeders will have to handle RGPu, which has above 60% fissile content and is suspected to be at least in principle being usable for a “fizzle weapon”. Using fuel at higher levels of enrichment makes the creation of breeding reactors far easier, well, at least feasible.
A breeding reactor is capable of converting more fertile material (Th232 or U238) into fissile material than it is using during operation. And there is likely enough fertile material in the earth’s crust to last us until the sun goes red giant.
Breeding reactors have the nice property (purely for the question of energy) of exponentially growing the amount of fissile material. The rate of growth depends on the design, the operation mode and the reprocessing. The term system doubling times describes how long it takes a reactor to produce enough fissile material to start a second unit. The theoretical system doubling time is proportional to the initial amount of fissile material and inversely proportional to the breeding ratio, power level and the percentage of initial fuel that gets used up. The real system doubling time is also dependent on how accessible the fuel is and how it can be reprocessed. This is why a metal-cooled fast reactor has system doubling times in the range of 30-40 years, despite the high breeding ratios they are achieving.
Molten Salt Reactors are a specific type of breeding capable reactors. The more reprocessing you do, the higher the breeding ratio. Because of the chemistry of the salts in the fuel and blanket salt, getting the bred uranium out of this salt can be done via a cheap fluorination process. By design, reprocessing can be done continuously while the reactor is online. They can be run in either the thermal or the fast spectrum.
In general, the reprocessing requirements for thermal reactors should be higher. This is because a lot of the fission products tend to capture neutrons in the thermal spectrum. Given that there is only a marginal surplus of neutrons in the Th-U cycle in the thermal spectrum fission products have to be removed continuously, fast and thoroughly. There are too few neutrons in the U-Pu cycle in the thermal spectrum to achieve breeding.
These thermal molten fluoride salt plants could be rolled out quickly. LeBlanc6 sketched out a plan for getting to huge amounts of deployed units.
Step 1, get enough HALEU to start the reactor. This will likely involve less than 2000 kg U235 or about 10 tonnes of 20% enriched fuel.
Step 2, run the reactor surrounded by a ThF4 containing blanket salt. Don’t transfer any of the produced U233 to the fuel salt. Fuel processing of the fuel salt to remove fission products beyond gasses and noble metals is optional.
Step 3, as the 235U and some produced Pu fissions off and/or fission products build up, add HALEU to the fuel salt to maintain criticality. While this is happening, approximately 300–400 kg of U233 will be generated in the blanket (which can be left in the salt or fluorinated out and stored temporarily).
Step 4, after about a year of operation, shut down and remove the remaining HALEU from the fuel salt by simply fluorination. The reactor can be restarted with the U233 from the blanket in a clean carrier salt and continue running indefinitely on the Th-U233 cycle. The moderator would have to be changed on a regular basis.
For a thermal 1GWe 2-salt reactor, about 400l of fuel salt and 2860l of blanket salt have to be reprocessed per day, for the 1-salt 1 GWe reactor, 26350l have to be reprocessed. Some have argued that the reprocessing requirements are too high to be feasible5 for these thermal reactors. The disposition of the graphite poses additional concerns.
The alternative is the usage of fast reactors, where fission products tend to capture less neutrons and additionally more neutrons are available per fission reaction. Siemer5 gives an overview of the breeding ratio as a function of how much reprocessing is done.
The downside of fast reactors is of course that more fissile material is needed for startup. About 5.5t of fissile instead of 400kg for the fast fluoride salt reactors.
Additionally, the structural materials are subject to higher radiation damages. To get into a breeding mode as fast as possible, these reactors would have to be started with about 80% enriched U235. This puts significant demands on our uranium enrichment capacity.
This capacity is measured in kgSWU. If we accept a typical “tail” enrichment of 0.2%, 1kg of 20% enriched uranium will need 42 kgSWU, 1kg of 80% enriched uranium will need 190 kgSWU. Currently, there are 67 000 000 kgSWU/y of global capacity, which equals about 1600t of 20% enriched HALEU or 350t of HEU. One large PWR needs about 130 000 kgSWU/y. If we are really serious about solving climate change, doubling the global enrichment capacity seems like at least a feasible proposition. Assuming this, we would be able to start about 100 fast molten salt reactors per year or 320 of LeBlanc’s thermal reactors.
Siemer proposes a mode of operation, where LiF-ThF is discarded during the reprocessing phase along with the fission products. After a few hundred years (yes, taking a very, very long term view here), this discarded waste will be an excellent reservoir of enriched Li-7 and Th. The waste, which amounts to about 4m³ per GWye could be stored in WIPP, given that it would be considered “military” waste or in HOLTEC storage pads.
This scheme requires huge amounts of Li-7, if it is supposed to be deployed at the relevant scale. But given that fusion power plants would need considerable amounts of enriched Li-6, it looks like we are in need of lithium enrichment capabilities anyway. Until deuterium-tritium fusion power plants become available, the “tail” Li-6 could be rented out to manufactures of batteries for mobile applications to decrease the weight of their batteries a little bit. When the cars using the batteries get recycled, the Li-6 could be reclaimed.
The doubling times of MSFRs seem to be on the order of 40 years.8 That is too long to take advantage of the exponential growth in the timeframe relevant to climate change.
An interesting alternative might be the molten chloride salt reactors (MCFR). Some recent work9 has shown that a U-Th MCFR could achieve doubling times of 7 years, if 200l of fuel salt per day were reprocessed. This proposal is a little bit odd, because it uses an “unusual salt”, with the far less common fuel cycle in an unconventional spectrum, requiring some “in between” amount of reprocessing. Such breeding MCFR would rely on Cl-37, for which there is currently little to no enrichment capacity. Again, this scheme would create huge amounts of HEU, something which we were not willing to use for civilian applications so far and have only entrusted to military personnel.
If we assume that it is feasible…
to convince governments to use their military to manage energy related HEU and become dominant actors on the global energy market
to realize reactors with such doubling times.
to build Li-7 and/or Cl-37 enrichment facilities in unprecedented volumes
to double the global uranium enrichment capacity and funnel all of into into starting such reactors with HEU
to mine approximately 3 times as much uranium as we do today
to mine thorium
to build new reactor capacity as fast as we get the fissile material
it would take the world 26 years to have enough fissile material to run in excess of 13,500 Th-U MCFRs. Given an efficiency of ~40%, we are taking about the very same 33750 GW of thermal, “primary power”, which are required by the wish of “first world living standards” for 9 billion people. If more power were needed by then, we could continue this scheme. After 30 years, we would be at 18,000 Th-U MCFRs, after 40 years we are at 30,000 reactors. Once the world’s energy needs are satisfied, it should be possible to transition from these breeder reactors to civilian reactors, for example to Elysium’s MCSFRs. These would be started by blending down the HEU and transition these reactors to a U-Pu breed&burn operation mode, where only RGPu has to be handled. Civilian engineers are generally entrusted to do that. The vast amounts of depleted uranium that got created as tail of the enrichment process could power the world for quite some time. If we allowed for the use of U233 in the CANDU-style PHWRs in the civilian sector, these could become sustainer reactors, but their relatively low operating temperatures would probably make this economically so unattractive that these reactors would be replaced with different designs at the end of their life.
Siemer summarized the advantages of fast molten salt reactors as follows:
its compact size and simplicity relative to alternatives. There is no solid fuels and/or moderators. This means that it should be relatively cheap to both build and operate. There is no need for fuel assembly fabrication, handling and less concerns about durability, shuffling, transport.
Its fuel cycle is genuinely sustainable – no fuel shortages “forever”.
management of radioactive waste should be relatively simple/cheap.
it has high operating temperature (~700°C), a high heat capacity working fluid and therefore higher electricity generation efficiencies and more direct process heat applications.
it has a nonreactive, high-boiling temperature working fluid, which reduces both the probability and consequences of accidents (spills, etc.).
after the startup and transition period, there is no need for uranium enrichment facilities and the handling of HEU.
fueling them would generate far less environmental impact (e.g., mine tailings, etc.) than of any nonbreeder reactor concepts
Additionally, molten salt reactors have the ability to buffer thermal energy in insulated tanks what makes them capable of cycling daily. In this way, it might be possible to satisfy peak demand, which is probably 1.8-2.1 times as high as average power, without requiring more fissile material. Multiple nuclear powers could work on this in parallel. Russia, China, the US and India are obvious contenders, but this technology could also be really interesting for France and the UK. While France will have to overcome stark resistance from its eastern neighbor to which it is bound by EU treaties, the UK could actually be in a good spot, at least geographically for delivering a lot of energy to Europe in that way.
Given that these reactors would be handled by the military, the presence of soldiers could also help to reduce the O part of O&M, which seems to be a huge fraction of operational costs of commercial nuclear power plants.
Cost of operators isn't as large a portion of total costs as many assume
Current staffing at US nuclear facilities includes very large, heavily armed security forces.
Existing facilities weren't designed for efficient protection
O from O&M
P. 6https://t.co/UvX8AGaHDH pic.twitter.com/jjrIGWxeJ7— Rod Adams (@Atomicrod) January 23, 2022
Summary
So what is the plan outlined in this article to get to a zero-carbon world?
Build out new PHWRs to keep coal power plants from being built
Use the freed up enrichment capacity and build more like our lives depend on it.
Use that enrichment capacity to start fast molten salt reactors. Ideally with short doubling times. These are built and operated by the nation’s respective military engineering units.
Leverage the high temperature to produce hydrogen for fertilizers, shipping, heavy-duty trucks and air planes. This could be stored as ammonia.
Colocate energy intensive manufacturing facilities (e.g. water desalination, aluminum, steel, etc.) with these clusters to provide millions of high-quality jobs.
Build methane synthesis plants to feed into the already existing huge infrastructure for it. Methane power plants all over the world could use it directly, gas-to-liquid processes could replace oil for transport and heating applications.
This might create a market for captured carbon and make the usage of CCS equipment, like the Allam cycle, profitable.
Once there is enough capacity for the world’s needs, transition to “civilian reactors”, which don’t use HEU.
All of this would of course be easier, if a global, uniform price on carbon emissions existed that didn’t disproportionately affect developing countries and the poor everywhere. Considering the magnitude of the challenge and the political and technical issues facing this idea here, getting a global, binding agreement on carbon prices looks like an easy task.
I am well aware of the fact that this is “the bad version” of what might become a good plan in the future. It considers the aspects of scaling, prosperity for another 8 billion people, fuel, reliability, dispatchability and hard to decarbonize sectors. That’s more than most plans focused on degrowth and intermittent energy sources can claim. It has also the advantage of using technology that exists or seems feasible. The downside of course it that it would hand over huge responsibility and economic power to the military and we would create unprecedented amounts of high-enriched fissile material, which does pose a threat. Just like during WW2. It is somewhat mitigated by the fact that military personnel is handling this material, but you cannot discount this risk entirely. You cannot take the effect of this plan on the questions of the likelihood of nuclear war and terrorism lightly.
I sincerely hope that humanity can come up with a better plan satisfying all these constraints or that a breakthrough in fusion voids the necessity of worrying about energy. Our posterity’s lives might very well depend on that.
There is some debate about how to add the contribution of renewable energies to the measure of “primary energy”. ↩︎
Goldstein-Rose, Solomon. The 100% Solution. Melville House. Kindle-Version. ↩︎
If humanity ever gets around to colonize the solar system, power satellites or lunar solar power might also be viable, benign alternatives. But that doesn’t seem likely in the relevant timeframe. ↩︎
The National Ignition Facility has come really close, though. ↩︎
SIEMER, Darryl D. Why the molten salt fast reactor (MSFR) is the “best” Gen IV reactor. Energy Science & Engineering, 2015, 3. Jg., Nr. 2, S. 83-97. Link ↩︎ ↩︎ ↩︎
LEBLANC, David. Molten salt reactors: A new beginning for an old idea. Nuclear Engineering and design, 2010, 240. Jg., Nr. 6, S. 1644-1656. Link ↩︎ ↩︎
FIORINA, Carlo, et al. The MSFR as a flexible CR reactor: the viewpoint of safety. American Nuclear Society, 555 North Kensington Avenue, La Grange Park, IL 60526 (United States), 2013. ↩︎
LIAOYUAN, He, et al. Th-U Breeding Performances in an Optimized Molten Chloride Salt Fast Reactor. Nuclear Science and Engineering, 2021, 195. Jg., Nr. 2, S. 185-202. Link ↩︎