What Bill Gates Has Wrong About “Advanced” Nuclear Reactors

Published Apr 13, 2021

Physicist Dr. Edwin Lyman discusses the safety, security, and environmental impact of proposed “advanced” nuclear reactors.

In this episode

Ed and Colleen discuss

  • Non-light-water reactors and how they are different from conventional nuclear reactors
  • whether they are really "advanced" and whether they offer advantages over conventional reactors
  • the Natrium rector that Bill Gates is funding and what Gates needs to focus on
Timing and cues

Opener (0:00-0:24)
Intro (0:24-2:16)
Interview part 1 (2:16-13:44)
Break (13:44-14:50)
Interview part 2 (14:50-26:54)
Outro (26:54-28:13)

Related content

Editing: Omari Spears
Editing and music: Brian Middleton
Research and writing: Jiayu Liang and Pamela Worth
Executive producer: Rich Hayes
Host: Colleen MacDonald

Full transcript

Colleen: How do we solve climate change? Ask 100 people and you’ll get 100 answers. And if one of the people you ask is Bill Gates, you might hear that we should turn to nuclear energy to help us reach our climate goals. But while generating nuclear power doesn’t create carbon emissions, it does come with a host of other challenges... like affordability, safety, and the unsolved question of how to safely dispose of nuclear waste.

Almost every nuclear reactor operating today is what’s known as a light-water reactor, because they use ordinary water to cool their hot radioactive core. To try and solve the biggest challenges of nuclear energy, the industry is turning away from light water reactors and looking toward new designs that use other materials to cool the core. The industry calls these new designs quote-unquote advanced reactors and claims they will help us build a clean energy future that’s also safe and affordable.

So... are these claims accurate? Today’s guest is Dr. Edwin Lyman, a physicist and director of nuclear power safety at the Union of Concerned Scientists. He just released the report “Advanced isn’t always better,” an independent review of these new designs that cuts through the hype coming from the nuclear industry. Ed wants to make sure we don’t waste money designing and building reactors that aren’t safe and don’t improve on what we already have.

He explains how today’s nuclear reactors work, what’s different about so-called advanced reactors, and whether or not they deliver on the benefits they promise. Ed also tells us about the Natrium reactor being designed by Bill Gates’ company TerraPower... and what he’d say to Bill Gates if they met at a dinner party.

Colleen: Ed, welcome back to the podcast.

Ed: Thank you for having me again.

Colleen: So, we've talked in the past about small modular reactors, the Chernobyl disaster. And today, I want to dig into non-light- water reactors. You just published a technical analysis looking at the safety, security, and environmental impacts of this proposed new suite of nuclear reactors. First off, how are they different from the nuclear reactors that the U.S. currently operates?

Ed: Sure. So, the U.S. currently has 94 nuclear reactors to produce electrical power. And they all use ordinary water as a coolant to remove heat from the hot fuel to convey that heat to a power generation system, which generates steam and produces electricity. These reactors have a main characteristic as they don't use water to cool the fuel, but they use other substances. For instance, you can use a liquid metal like liquid sodium as a coolant, or you can use a gas like helium, or in some cases, the fuel cools itself as a liquid and it cools itself.

Colleen: So, can you run through the new advanced reactors that you evaluated?

Ed: Yes. The first main class of reactors is called fast reactors. And these differ from our existing fleet because they don't have materials that slow down neutrons. So, when a nucleus of uranium fuel is fissioned, it's struck by a neutron and it's split apart, and it releases energy and other neutrons. And those other neutrons will then strike other uranium nuclei, and you have what's called a chain reaction. And that generates a steady level of heat which can then be used to produce electricity. That's how the nuclear reactors work.

But in the water-cooled reactors that we have now, those water molecules actually slow down the neutrons. So, when a neutron is produced by fission, it has a certain energy. But as it collides with water molecules, it slows down. It turns out that makes it...essentially, that makes the fission reactions more efficient. So, you can use less concentrated fuel. So, that's a certain approach to designing a nuclear reactor that we use today.

But in a fast reactor, you don't have a material that slows down those neutrons. So, they remain high-energy, and that has different properties than light-water reactors. In fact, some of the advantages that the developers of these reactors claim, stems from this property of having fast neutrons. And in order to do that, you need a coolant other than water. You need something that won't actually slow down those neutrons. And so, that's why a substance like liquid sodium is used to cool those reactors.

Then there's another category which is called high-temperature gas-cooled reactors. And as the name suggests, they don't use water to cool the fuel, but they use a high-pressure gas, most commonly, helium, which you pump through the reactor vessel. And the gas will then be pumped out after it's heated up, and then used again to boil water and produce electricity.

The third category is a much different type of reactor than the others. And that's a molten-salt fueled reactor. And the reactors that we use today use a solid material as the fuel, which is something that's desirable. You have a solid material surrounded by a metal cladding. But in a molten-salt fueled reactor, the fuel itself is a liquid. It's a hot molten salt where the uranium and other fuel materials are dissolved in that fuel. Again, that has...according to the proponents of that technology, that gives you certain advantages over solid fueled reactors. But it also has quite a few disadvantages that we can talk about.

Colleen: How long have these types of reactors been studied?

Ed: These reactors, by and large, are very old technology. So, people often call them advanced reactors, but...because that's not really an accurate characterization. We don't really use that in our report. In fact, some of these reactor designs were conceived before the light-water reactor that we use today in the United States. So, they date, you know, from the days of the Manhattan Project in the 1940s. And some of these technologies were actually attempted as early as the 1950s, both in the U.S. and other parts of the world.

Colleen: Ed, I know from reading the report that each of these reactor types has their problems. Let’s narrow down to fast reactors for our audience. What do proponents say about them?

Ed: So again the benefit of a fast reactor is, primarily, the dream for which it was first conceived. Because of the special properties of fast neutrons, the fast reactor, in theory, could operate in a mode where it could generate not only its own fuel, but actually produce a little extra fuel for another reactor. So, this is what's called a fast breeder reactor. And this is one of the reasons why the Manhattan Project scientists who may have had too much time on their hands at one point, dreamed of this reactor.

They thought it would be a way to fuel nuclear power forever without having to use uranium resources, which they thought had to be saved for nuclear weapons. Because it was thought uranium was a very rare commodity at the dawn of the nuclear age and that the U.S. would need all it could get for nuclear weapons. So, if you had a reactor that could actually breed its own fuel, you wouldn't have to compete with that.

And the other potential advantage of a fast reactor is that it has the ability to actually utilize other types of fuels efficiently, in particular, some of the materials in the spent nuclear fuel from current reactors, which has a very long half-life. That means it persists in the environment for tens or hundreds of thousands of years. That kind of material is not utilized very efficiently in current reactors, but a fast reactor could actually convert more of that to energy. So, sometimes, fast reactors are called burner reactors and their proponents tout them as being able to burn spent fuel to take existing spent fuel and essentially destroy it or recycle it. These are the terms that are here.

Colleen: So what are some of the problems with fast reactors?

ED: First of all, it's not...doesn't really live up to its hype in a real system. So, yes, in theory, a fast reactor could operate as a breeder or a burner. But what I discuss in my report is if you look at real systems in a real electricity system and how these reactors actually work, it would take a very, very long time to make a dent, let's say, in the spent nuclear fuel stockpile we have now or it would take a very long time to actually breed fuel in a way that you can say you're being more efficient than the current systems.

So, it's not really realistic to claim these reactors can burn spent fuel or to breed new plutonium in an effective way. So, the benefits aren't that great, but the risks are very significant. Because in either one of those cases, a breeder reactor or a burner reactor will also need to reprocess spent fuel to produce that new fuel for the reactor. And reprocessing is a technology where you take nuclear waste, spent nuclear fuel, and you put it through a chemical process to extract the materials that you want to use in a fuel, primarily, plutonium and separate that from other radioactive waste that you can't recycle in a reactor. And the problem with doing that is that plutonium, in addition to being a potential nuclear fuel, is also a nuclear weapons material. So, when you separate it from spent nuclear fuel, you're essentially making it easier for a [00:08:30] country or a terrorist group that wants to acquire nuclear weapons. It makes it easier to get that material.

And so, therefore, any nuclear power system that uses reprocessing is inherently more dangerous from a nuclear proliferation perspective than the system we have now which is operating light-water reactors without reprocessing the spent fuel. In addition, fast reactors have serious safety issues which the proponents like to gloss over.

For instance, we're all familiar with Chernobyl and we'd discussed that in a previous podcast. One of the main initiators of the Chernobyl disaster was the reactor had a design flaw where under certain circumstances, if the reactor heats up, it actually undergoes a positive feedback. So, the hotter it gets, the more power it produces and you have essentially what's called a massive power excursion that led to explosion of the reactor. That's something you don't want to have in your nuclear reactor. In fact, the light-water reactors in the U.S. today have the opposite behavior. If the reactor heats up, the nuclear reaction tends to shut down.

But fast reactors typically are more like Chernobyl in that if they heat up and that liquid sodium coolant starts to boil, then you actually get more and more fission. So, the power of the reactor can increase by a factor of 100 in a matter of seconds. And so, you have a situation where you can have, again, a core meltdown or even an explosion. So, it's that instability that I worry about with regards to safety.

Colleen: So, Ed, are these fast reactors, are they just theories or have any of them been built or any pieces of them built to do actual testing?

Ed: Yes. Fast reactors, there's been a fascination with them in the U.S. and other countries. So, actually, the first fast reactor was built back in 1951 here in the United States. And there have been a number of demonstration and test fast reactors in the U.S., in the Soviet Union and now Russia, in the United Kingdom, in France, in Japan, in India, where most of them are located. And the actual record with these reactors has been extremely mixed.

One problem with using liquid sodium as, you know, any mischievous high school student has ever tried this can tell you is that it doesn't...that sodium metal doesn't play very well with water. In fact, if it comes in contact with water or air, it can actually catch fire. So, you're using a potentially flammable coolant in your nuclear reactor. And so, the developers have to put in all sorts of extra safety systems to make sure they can detect if there's a sodium leak that could actually lead to a fire. In fact, this is really one of the biggest issues that affect the reliability of fast reactors around the world.

But the reactors, by and large, had not been demonstrated on the scale and using the same fuel and safety systems that are being proposed for the generation of fast reactors that are being talked about to be built today.

Colleen: So, I'm assuming, with the climate crisis upon us, could the potential risks justify the enormous public and private investments needed to get them up and running?

Ed: Well, that's one of the fundamental questions here is obviously, we're facing this potentially devastating climate crisis and we need to evaluate every possible tool that could help us to deeply decarbonize the energy sector as rapidly as possible. So, nuclear, overall, nuclear power is a potential option for doing that. But that doesn't mean nuclear power should just be given the benefit of the doubt and any cockamamie idea that someone comes up with is something that the federal government investors should throw billions of dollars at. Because it will take many billions of dollars and probably a couple of decades at least before any new reactor design could be actually commercialized and have a hope of being safe and reliable.

So, one of the reasons why I pursue this report is to examine some of the claims that are being made about these reactors. Because if there's no real benefit to pursuing a radically different type of reactor design than the one we have now, then...if there's no real benefit, then those investments wouldn't be justified. So, I think it's very important for the public to date to make sure that they know what the facts are and where claims are being made, what the actual subtlety is or caveats or fundamental truth about those technologies is being discussed. And so, people are not misled into supporting very speculative reactor designs that are essentially high-risk, low-benefit technology.

Colleen: Well, you know, Ed, you're making me think of Bill Gates. He's touting these technologies as a promising solution for meeting our climate targets. And, you know, he's a successful guy. He has a lot of power to persuade. I mean, if I set up a dinner party for the two of you, what would you want to talk to him about?

Ed: Yeah. So, I wanna tell Bill Gates that he really needs to go beyond what his advisors may be telling him. Sure, he's obviously a very successful person and not an idiot. But when he talks about nuclear power, he really demonstrates that there's some gaps in his understanding. I get the feeling that he has not done an independent review for himself of the projects that he's funding. In particular, through a company called TerraPower which Gates founded and which he's the chief investor in.

This company is developing a sodium-cooled fast reactor called the Natrium. And this is moving forward because the Department of Energy selected the Natrium design as one of two that will be part of what's called the Advanced Reactor Demonstration Program, which was created by congress in 2019 as a public/private partnership to build two advanced demonstration reactors by 2027. That's the very aggressive goal. And the Natrium reactor won one of those awards. So, they are pursuing a demonstration reactor.

But, again, the sodium-cooled fast reactor, has a lot of liabilities and not a whole lot of benefit. In fact, the Natrium itself, because of the fuel it's going to use, has even less benefit than a fast reactor would, in theory. Because, again, one of the real...only advantages to taking on the additional risks associated with a fast reactor is you could get this benefit of breeding plutonium, expanding or producing your own fuel, and reducing the need for actual mined uranium.

But in the case of the Natrium, it turns out it would probably take two to three times more uranium to generate a kilowatt/hour power from that reactor than from current light-water reactors. So, it would actually be less uranium efficient than current reactors. So, why would you invest billions of dollars in developing a technology when it doesn't even meet that basic test of having the benefit that you thought it would have?

Colleen: So, one of the goals of your report is to provide the technical analysis so that policy decisions that are being made are well-informed. What are some of the recommendations that you make in the report?

Ed: One of the primary recommendations is that the Advanced Reactor Demonstration Program I just mentioned with the goal of building two commercial reactors that would be connected to the grid and generate power by 2027, that that program actually be suspended. Because I don't think that the safety data is available yet to support that kind of deployment. It's possible these, one or both of these reactors, will be built at ordinary utility sites. Utility is going to expect that the reactor will operate at full capacity without significant reliability problems. So, in order to have a reactor like that, essentially a commercial reactor that's ready to generate commercial power, you need to have a basis for licensing that reactor in that mode, and also enough technical information to be able to understand the problems with operating it and to operate reliably. And I do not believe that the existing record is here to support the safety analysis.

So, the Nuclear Regulatory Commission, which is the regulator of nuclear power plants in the United States will have to license those reactors. And they're going to be faced with the question of whether they should license a commercial-sized reactor based on the spotty safety database that exists. In the past, the NRC has said, "Well, if you want to build a new reactor type that hasn't operated commercially, in order for us to license it, you're going to need to probably build a prototype, and to run that either a smaller scale or full scale." But you run in a mode that's not for generating power, but for doing safety testing, for qualifying fuel which is always a very important safety measure to make sure that the nuclear fuel is safe under the conditions that it will be used, and other critical testing to support licensing that commercial reactor.

And so, I feel like they've skipped this development step, and the NRC doesn't have the information they’ll need to really make a safety finding to license these demonstration reactors. So, I've argued that the program should be slowed down, that the NRC should have the opportunity to consider what additional data it'll need to license those large reactors which would most likely involve building a prototype where, because you're going to be using it for safety testing to address uncertainties in the safety of the design, you're going to want to have additional features that may not be in the commercial reactor.

For instance, many of these reactor designs don't include a conventional containment. The nuclear reactors we have operating today in the U.S. have reinforced concrete containment structures, which are designed to prevent leakage of radiation in the event of an accident, and even in the event of an explosion like what we saw at Fukushima Daiichi in 2011, that they would provide some protection against that. But a number of these reactors won't have any real containment at all because the developers argue that the reactors are so safe they won't need one. But that is a statement that really has to be verified. And so, you'd want to use a prototype to test certain scenarios. But because you don't know how they're going to play out, you'd want that prototype to have a containment even if the commercial design doesn't have one. So, the prototypes could look a lot different than the designs that are being pursued under this program. And I've argued to slow it down.

Another conclusion is that I don't think there is due diligence when the Department of Energy has awarded these demonstration project awards. I'm not confident that the process is really examining whether these reactors have all the benefits that they claim they do. And I think there has to be a better vetting of reactor designs before billions of dollars of public money is committed to these reactors because we don't want to subsidize the development of unsafe reactors.

Colleen: If nuclear power needs to be part of the climate solution, why not continue to use what we have? I understand the reactors that we have are aging out. But why not either shore those up or use the same design that we currently have where we wouldn't have to go through the lengthy and costly development phase?

Ed: Yes, that is the baseline we have is the operating light-water reactor fleet as well as what are called the evolutionary design changes that is building off that experience and trying to do better, but having the same fundamental design using water as a coolant. And without having to take a position on what the role of nuclear power should be in deep decarbonization, you can ask the question, "Are these advanced reactors better?" Or if you're going to invest tens of billions of dollars in new nuclear reactor designs, would it make more sense to focus on the existing technologies on how to improve them with respect to safety and cost? So, that's really the baseline. But the operating reactor fleet has its problems. And, you know, we've seen what happened in Fukushima Daiichi at a light-water reactor. They clearly are susceptible to core melt accidents, especially in the event of a severe earthquake or flood. And so, they have been somehow discredited in the eye of the public.

And also, they're costly, the current fleet. In some parts of the United States, nuclear power is no longer economical because not only natural gas is cheaper, but also wind and solar is cheaper than nuclear power under certain conditions. And so, a number of operating plants are not economic anymore. And as far as new plants go, those have turned out to be extremely cost-prohibitive. And so, the only two nuclear reactors under construction in the United States right now in the State of Georgia are running at twice the original estimated cost, up to, I think, now $28 billion for that project. And they're taking twice as long, at least, as originally planned.

So, the light-water reactor has lost some credibility. And I feel like that's what's driving the messaging from the nuclear industry today is they feel like they have to show the public they have something different and they can do better with something different. But the problem is that this... we're not talking about messaging here. We're trying to win public hearts and minds it's not what needs to be done. It needs to be done as to address these fundamental safety and cost issues of nuclear power. And just doing something different for the sake of the fact that it's different is not necessarily the best approach.

Colleen: Well, nuclear power is certainly a subject that is fraught. And I appreciate your expertise and the clarity that you bring to the issue. Ed, thanks for joining me on the podcast.

Ed: Thank you. It's been a pleasure.

Return to top


Related resources