New Generation Reactors - Part 1
The shell game continues
I would hope that my previous Substack post [1] offered enough in the way of information and links to other sources to convince you of the need to shut down all the current nuclear reactors. But I did not address the question of future prospects for nuclear power. For some time now, the nuclear and weapons industries have been touting "new", "advanced" reactors that per their marketing statements will eliminate the problems of the past, and even make nuclear power affordable via "small modular reactors" ("SMRs"). It is going to take me three articles to deal with all of that. In this one, I will start addressing the so-called Generation-IV ("G-4") reactors, although most of the SMRs are just scaled-down G-4 designs. In my next post, I will administer the coup de grâce to the G-4s and most of the SMRs, which will leave only reactors based on the Thorium-Uranium233 fuel cycle for a later discussion.
In writing this post, I found it necessary to use a fair amount of technical terminology, terminology that might seem alien to a non-technical audience. I have therefore included a short appendix with background information and terminology to help nontechnical readers better understand the material. If you are not at least conversant with the contents of a university-level introductory physics course for science and engineering majors, I urge you to read the appendix before proceeding.
Most of the information I used from nuclear industry sources came from the Generation IV International Forum, a cooperative involving thirteen nations. I used their technology roadmaps from 2002 [2] as well as their annual progress report for 2021 [3]. (As of September 26 2023 the technology roadmap has not been updated, nor has an annual report for 2022 been issued.)
If the Executive Summary of the Generation IV International Forum (GIF)'s most recent roadmap is read literally, the reason for abandoning all existing reactor designs is simply "to meet future energy demand", and not in any way to correct the faults of Generation I-III reactors. Not even recognizing the self-contradiction, they then went on to set forth four goals for the 4th generation:
sustainability;
proliferation resistance and physical protection
safety & reliability; and
economic competitiveness
These goals, if fulfilled in a meaningful way, would go far toward making nuclear power a viable element in the future energy generation mix. The question before us is whether the six families of designs can achieve dramatic improvement over nuclear power's historic performance in those regards. Note, however, that the list fails to address the problem of what to do with the dangerous byproducts of the fission process.
The six families of designs selected for development by the thirteen national signatories to the GIF include:
Gas-cooled fast reactors (GFRs), are critical, Uranium-Plutonium fueled fast reactors that replace water with Helium as the working fluid. They have the potential for converting some minor Actinide waste to fuel, thereby giving rise to claims of helping to mitigate the nuclear waste problem.
Lead-cooled fast reactors (LFRs) are mostly critical, Uranium-Plutonium fueled fast reactors that use molten lead or lead-bismuth eutectic as the working fluid. These reactors also have the potential for minor Actinide waste to fuel conversion.
Molten salt reactors (MSRs), of two basic types: one with the fuel dissolved in the molten fluoride salt, and the other which uses solid fuel together with molten fluoride salt as the working fluid. Current designs are critical, Uranium fueled fast reactors.
Sodium-cooled fast reactors (SFRs) are critical, Uranium-Plutonium fueled reactors that use molten sodium as the working fluid.
Super-critical water reactors (SCWRs) are the closest thing to existing LWRs of the G-IV designs, and are even touted as being upgrades to current reactors. They are critical, Uranium fueled thermal- or fast neutron reactors, using super-critical water (a state with properties intermediate between liquid and gas, requiring a temperature-pressure regime of 374C, 22.1MPa (705 Fahrenheit, 218 atmospheres)) as the working fluid.
Very high temperature reactors (VHTRs) represent an evolution of gas cooled reactors: critical, Uranium fueled, graphite moderated with very high temperature Helium as the working fluid.
The idea behind the broad notion of "small modular reactors" is to achieve low cost deployment of nuclear power plants with incidental improvements in design to promote safety, sustainability, or some other beneficial improvement. As of July, 2023 there were 15 such systems [4]. (In this I discount the so-called micro reactors with outputs below 50 megawatts.) Of the current systems, 8 are variations on the old theme of critical, U235-fueled reactors and 5 are fast reactors. Thus 8 of the 13 are discountable for reasons explained previously and five, including the best known Terra Power product, are G-4 designs.
It is clearly not a goal of the nuclear industry to eliminate nuclear weapons. From the point of view of doing so, it is obvious that the Generation IV reactors do nothing toward realizing the objective of ending the mining and processing of Uranium. Many of the designs are or can be used to produce Plutonium. Though some designs make diversion of weaponizable material more difficult, the fundamental problem, retaining the ability to refine and enrich Uranium, remains. For that reason, G-4 designs do not fundamentally discourage proliferation.
Let us address the industry's stated goals in turn, starting with sustainability. Sustainability, in the context of nuclear reactors, has an entirely different meaning than in other contexts, referring simply to the ability to extract more energy per unit mass of Uranium than previous designs. Well, not all previous designs, but rather from Generation I-III reactors using U235 as fuel. They do this by modifying the reactor designs to reduce the degree to which fast neutrons are moderated, allowing the U238 present in the fuel to be transmuted to a fissionable form of Plutonium (or other Actinide).
A complete discussion of sustainability must address the topics of resource depletion and waste disposal as well as efficiency of resource use. As of 2009, it was estimated [6] that the world's Uranium deposits would be sufficient for 200 years at the then current rate of usage. More of the material is available, but at greater cost of extraction. So if the use of Uranium were to increase even five-fold, we would be in danger of running out of the economically extractable store before the end of the century. Does that sound sustainable to you? To make matters worse, two of the six types of reactors being developed as G-4 systems use Helium as the working fluid. Helium is used in a large variety of applications, such as high sensitivity medical imaging, and is essential in many scientific and engineering applications. At present, without it being used in nuclear reactors, world reserves are estimated to be sufficient for 300 years. Experience with an VHTR, located at Fort St. Vrain, CO between 1979 and 1989 illustrated that one of the problems with such a reactor is the difficulty of obtaining adequate sealing to prevent loss of the Helium coolant. It seems reasonable to expect that if GFRs and VHTRs are brought on line in significant numbers, the planet's Helium supplies will be rapidly exhausted.
The second element of sustainability has to do with nuclear waste. Nuclear advocates assert that the existence of deadly byproducts of the operation of a reactor is a political problem, but as I argued in the earlier post, it is more than that. The nuclear industry tacitly agrees because they tout reduced waste as a benefit of their fast reactor schemes, and claim that fast reactors are capable of fissioning most of the minor Actinides that are produced as a result of irradiating U238 and its transmuted products in fast reactors. A study by the Union of Concerned Scientists [7] demonstrates that the industry claims of high degrees of minor Actinide utilization as fuel are greatly overstated, that in particular they do not account for the additional raw fuel that must be introduced to transmute fertile minor Actinide isotopes into fissile ones. A report issued by a MIT researcher [8] further demonstrated that the once-through fuel cycle used by current reactors is superior to the various G-4 breed and burn schemes in both economic and anti-proliferation terms.
There is a tendency in the debate over nuclear fuel cycles and nuclear waste to focus entirely on the problem of minor Actinides, that is to say, nuclides heavier than Uranium. But attention needs to also be paid to the lighter fission products. A Handwiki article [9] and the Wikipedia article referenced in the appendix list, inter alia, the long lived fission products for two types of once-through reactors: pure Uranium and 65%U/35%Pu. In the former case, more than 15% of the fission products (by abundance) are radioactive with half lives in excess of 100,000 years; in the latter case the abundance is 20.75%. In a twice-through reactor, the expended fuel is reprocessed to remove fission products, inter alia, then used again. This in effect nearly doubles the aforementioned numbers. The lesson is obvious: The more efficiently the original Uranium is used, the more long-lived radioactive waste is produced.
To make matters worse, the GFRs and VHTRs employ fuel that is encapsulated in multiple layers of chemical- and heat resistant carbon and ceramic cladding, so called "TRISO" fuel. This enables some G-4 design variations based on fuel circulation and use of Pebble-bed cores, leading to possibly increased safety because it greatly increases the temperature required for initiation of a core melt, and increased proliferation resistance because of the great difficulty entailed in removing the Uranium from the pellet. The price for this [10] is a 6.4- to 11.1-fold increase in the volume of nuclear waste.
Safety and Reliability are touted as advantages of G-4 reactors, based on various design features. One nearly common element is the employment of higher gaseous or molten metal coolants, which reduce the likelihood of coolant boiling. But new designs also bring new risks. Staying with the example of coolants, the higher temperatures entail higher rates of corrosion, which engenders requirements for new, corrosion-resistant and high temperature tolerant materials, for which little or no safety and reliability experience exists. Moreover, one of the high temperature coolants - Sodium - is highly reactive and cannot be exposed to either air or water. A leak in a Sodium-cooled reactor presents the possibility for rapidly cascading into a disaster. Speaking of Sodium, its presence in a fast reactor brings with it neutron activation of Na23 to become Na24, a radioactive isotope. So a leak in a Sodium-cooled reactor brings with it the dual dangers of fire and radiation.
The foregoing was intended to illustrate the fact that new designs bring new risks, and that the increase in safety in one respect can bring about an increase in danger in another [11]. One is entitled to ask whether the nuclear industry is simply shuffling risks, as in a shell game.
In its 2014 Technology Roadmap, the G-4 Forum described a three-pronged approach to ensuring safety in its new designs. The use of Probabilistic Risk Assessment (PRA), simulation, and integration of lessons learned from the Fukushima Daiichi accident all factor prominently.
It is good that attention is being paid to the Fukushima accident, but we are left wondering why the Roadmaps do not discuss the safety and reliability records of earlier systems employing similar designs. There are examples of earlier Sodium- and Lead-cooled reactors as well as a VHTR that should be studied to inform the design process. Taking Sodium-cooled reactors as an example, the US has had some ominous experiences, first with the US Navy's second nuclear powered submarine, the USS Seawolf, SSN-575, and also with two Sodium-cooled land-based reactors, in 1959 [12] and 1966 [13]. In the case of the Seawolf, the problem was with the maintainability of the reactor, which was so bad that the reactor was removed from the submarine after only two years in service. The 1959 and 1966 events were core meltdowns.
As to PRA, we have already alluded to its limitations, and in fact, the 2003 G-4 Technology Roadmap acknowledges them, at least with regard to passive systems [14]. Then there is climate change. Large reactors are preferably located along rivers or adjacent to large bodies of water to provide a cheap source of secondary cooling. To guard against loss of auxiliary power due to flooding, weather statistics have been employed to guard against, say, a once in 100-years flood. Similarly, to guard against a loss of secondary cooling capability, the same sort of statistics have been employed to guard against, say, a once in 100-years drought. The problem is that the process is nonstationary: What was a once in a hundred years event a generation ago may now be nearly the norm. While it is possible to make some crude predictions concerning the current climate, it is unreasonable to expect predictions about the future to be accurate, because the future depends on what mitigating measures are adopted, to what degree, and when. If the design is carried out so as to be responsive to the entire range of worst-case conditions plus margins for added safety, the cost becomes prohibitive.
Simulation can be a very useful tool for exploring the behavior of a system under a range of conditions, but the accuracy of the simulation depends critically on the accuracy with which the parameters are known, and in the case of Monte Carlo simulations also the degree to which the various probabilistic elements are representative. Thus simulations become tools that evolve in accuracy only as more becomes known about the systems they model, and become fully useful only once the system modeled has accumulated operational experience. To use an example, a flight simulator is useful in training pilots due to its having incorporated all the measured responses of the actual aircraft, thus enabling it to respond to pilot responses to various simulated emergency situations.
The logical way to overcome many of the limitations of PRA and simulation as safety tools is to accumulate experience using a prototype system. This, in conjunction with exhaustive testing of subsystem prototypes would provide much of the data required to accurately identify at least the most likely failure mechanisms. But alas, as will be discussed further next time, the G-4 schedule does not allow for this.
Appendix on Basic Concepts and Terminology
A "fast"reactor uses the high energy (fast moving) neutrons from fission to produce additional fuel, by means of a process called "neutron activation". For example, when bombarded by fast neutrons, U238 atoms are transformed with high probability into U239 atoms, which can decay to form Pu239 via Np239, or by absorption of another fast neutron become U240, which decays with high probability to form Pu240 via Np240. Pu240 in turn absorbs a fast neutron to form Pu241. Both Pu239 and Pu241 as well as U235 are said to be "fissile", which means that they undergo fission with high probability when bombarded with slow, aka "thermal" neutrons. U238 is an example of a "fertile" nuclide, as it possesses simultaneously a low probability of undergoing fission via a slow neutron interaction and a high probability of being transformed into a fissile material via a fast neutron interaction. In the nuclear community, the phrase "cross section" is used in analogy to probability here. I use the latter term to convey the notion that we are discussing phenomena which are inherently probabilistic. Fast neutrons can be converted into slow ones by interacting with atoms which are neither fissile nor fertile. When a fast neutron collides with a, say, carbon atom, the atom speeds up a little and the neutron slows down a lot. Graphite, a form of carbon, is used in reactors to slow the fast neutrons down to promote fission. The process is called "moderation" and the material (here, graphite) is called a "moderator". Deuterium, a nuclide of Hydrogen, is also a good moderator and is used in place of graphite in so-called "heavy water" reactors (as distinguished from "light water" reactors which use H2O vice D2O as the working fluid).
In fission reactors, it's all about the neutrons. Neutrons, along with protons, are the stuff that makes up the nuclei of atoms. Lighter elements tend to have fewer neutrons than protons in their nuclei, or equal numbers; for example Helium4, Carbon12, Nitrogen14 and Oxygen16 - all very stable - have equal numbers of protons and neutrons. This is true up to about the mass of Iron; Fe56 has 26 protons and 30 neutrons. As atomic weight increases, neutrons increasingly outnumber protons in the nucleus. U235, for instance, has 92 protons and 143 neutrons. Thus, when a U235 nucleus fissions, a large number of free neutrons are produced. That's what makes a chain reaction possible: One slow neutron causes a single fissile atom to fission, releasing heavy fission products and a number of neutrons, more than one of which cause fissile atoms to fission, a process that builds rapidly and exponentially unless the neutron flux is carefully managed.
As alluded to above, an atom of one element can be transformed into an atom of another via absorption of a neutron. This is true of many elements, not just the really heavy ones. For example, Na23, the sodium in ordinary table salt, is relatively easily converted to Na24 via neutron activation. And Na24 is radioactive. Remember that for later. Also, unlike the proton, which has a proven half life in excess of the age of the universe (if indeed it is not stable), an isolated neutron will decay into a proton and an electron (also an electron antineutrino, but that's irrelevant to the discussion); its half life at rest is 10.25 minutes. Thus if a neutron happens to be thermalized (slowed enough to be in equilibrium with the surrounding matter), there is a 50% probability of its decaying within 10.25 minutes. When that happens, the resulting proton (hydrogen ion) will acquire an electron and become a (neutral) hydrogen atom. If the neutron is inside something, say a steel coolant tube wall when it decays, the resulting H atom will be trapped within the structure and will influence its chemical or mechanical properties. Over time, as hydrogen accumulates, this can become problematical.
The figure above, drawn from Handpedia [15] and Wikpipedia [16] respectively depict a part of the transmutation process in a fast reactor and the statistical distribution of nuclide fission products for four different fuel mixes. The transmutation chain is partial in the following respects: It does not depict the transmutation to higher Curium isotopes; to Berkelium or Fermium beyond Curium, nor does it show the alpha decays of the Uranium and Plutonium isotopes, which give rise to numerous, generally non-fertile and often radioactive daughter species such as Radon. Both references list fission products including chemically toxic ones such as Arsenic and also the various radioactive species. Depending on the initial fuel mix, up to nearly 22% of the fission products can be radioactive with half lives in excess of 100,000 years.
Notes
[1] https://stephenschiff.substack.com/p/starving-the-nuclear-weapons-beast
[3] https://www.gen-4.org/gif/jcms/c_177525/gif-2021-annual-report
[6] https://www.scientificamerican.com/article/how-long-will-global-uranium-deposits-last/
[7] https://www.ucsusa.org/sites/default/files/2021-05/ucs-rpt-AR-3.21-web_Mayrev.pdf
[8] https://dspace.mit.edu/bitstream/handle/1721.1/17027/54495851-MIT.pdf?sequence=2
[9] https://handwiki.org/wiki/Physics:Radioactive_waste
[10] https://art.inl.gov/NRC%20Training%202019/04_TRISO_Fuel.pdf
[11] The G-4 2014 Technology Roadmap repeatedly cites use of higher temperatures and non- aqueous coolants as features introduced in response to lessons learned from the Fukushima Daiichi accident.
[12] https://blog.ucsusa.org/dlochbaum/nuclear-plant-accidents-sodium-reactor-experiment/
[13] https://blog.ucsusa.org/dlochbaum/nuclear-plant-accidents-fermi-unit-1/
[14] G-4 2002 Technology Roadmap, page 69
[15] https://handwiki.org/wiki/Engineering:Breeder%20reactor
[16] https://en.wikipedia.org/wiki/Fission_products_(by_element)
Reposted with format changes only 3 Feb 2024