A New Tool to Mitigate Global Warming

Industrialized countries use a lot of energy. Even so, today’s increasing use of CO2-producing fossil fuels is now driven mostly by China and the rest of the developing world. We’re living in a one-hump world now — the developing world is catching up, and that means energy use is poised to skyrocket toward European levels. China is already more than halfway there:

Electricity usage graph by ThorCon.

As I sit here in the hot, humid Philippines, the air conditioner beside me — which runs nearly every day of the year and is likely coal-powered — keeps global warming out of the room¹. For now.

Many of my neighbors don’t have electricity, or even a sealed room in which they could put an air conditioner. But we can’t count on poverty to keep carbon emissions down.

Happily, solar and wind power have enjoyed major successes lately. In 2016, solar reached a huge milestone in the UAE when it dropped below the price of coal. Soon, solar power will be the “default choice” for new power plants in all southern climates. Remember, near the equator, energy use peaks in the afternoon when everyone turns on their air conditioners. Of course, this happens to be around the same time that solar panels gather the most energy — supply is usually similar to demand, and everybody’s happy.

But I don’t have to tell you the benefits of renewables — most of you will be aware of that already.

Instead I want to tell you about a “new” category of nuclear reactor called the MSR (Molten Salt Reactor), and how they can help us mitigate global warming faster. Although an MSR experiment (MSRE) was done in the 1960s,³ this type of reactor was all but forgotten until recently, after Kirk Sorensen, former NASA engineer, learned about thorium-based MSR technology and worked hard to raise awareness about it — sending out CDs with technical PDFs to dozens of people, then delivering talks about it at Google and elsewhere. Since 2011, more than half a dozen companies have formed to build them, not counting Chinese efforts.

What comes out of a nuclear “smokestack”? Pure water. Image credit

Think about northern climates, where energy demand is highest on cloudy winter days, just as solar power is at its weakest. Think about areas where the wind is calm or unreliable — calm spells do exist in most places, and could last weeks or be synchronized across large areas. Importing energy from thousands of kilometers south is not a good option — transmission lines are expensive, and people want energy independence. Even in southern climates, people use energy at night.

According to a recent study, powering 80% of the U.S. grid with solar and wind would require 12 hours of expensive energy storage (today’s grid has under a minute of storage²) and a continental-scale grid, all under idealistic assumptions like no transmission losses.

So it would certainly be possible to power most of the grid with wind and solar, but it’s hardly guaranteed to be the cheapest option, and in a world with political resistance to climate change action, the cheapest option tends to win.

Right now the cheapest option is often coal or natural gas. That’s terrible: CO2 can linger in the atmosphere for over a thousand years:

The black line shows the average estimate (from several studies surveyed in Joos et al. 2013) of how slowly CO2 leaves the atmosphere. 59% will go into the ocean and cause ocean acidification; most of the remainder will stay airborne for centuries.

There is a consensus that keeping global warming under the Paris 1.5°C target is highly unlikely. We may even need negative CO2 emissions — actively removing it from the air — to reverse global warming. 80% renewables just isn’t enough. So, shouldn’t we quickly and urgently support every possible energy solution?

I think Molten Salt Reactors (MSRs) are the solution — not as a substitute for solar and wind, but as an important tool to help us reach 100% clean energy faster.

So what are they? MSRs are nuclear reactors powered by liquid fuel salts instead of solid fuel rods. And they could turn out to be a game changer.

Unlike nuclear plants built in the 1960s and 1970s (including Fukushima, Chernobyl and Three Mile Island), “Generation III” plants built today have become very expensive in order to pay for much higher levels of safety, and to pay for decommissioning and waste disposal. They’re so expensive, in fact, that nuclear plants are rarely built anymore — sometimes causing old reactors to be extended beyond their design lifetime. That’s where MSRs come in.

Molten salt in a test tube

In an MSRs in which nuclear fuel is dissolved in hot liquid salt (such as table salt, NaBe, or FLiBe) that hardens when it cools off. No more meltdowns: melted is the new normal.

Why add MSRs to our grid?

MSRs are sometimes called “Generation IV” reactors, but that’s sort of like saying a jet engine is a “Generation IV” propeller. It’s really not — it’s a fundamentally different approach to nuclear energy, with numerous advantages:

1. Very high safety
2. Sustainability
3. Retrofitting fossil fuel plants
4. Nuclear waste burning
5. High efficiency
6. Low cost
7. Load following & energy storage
8. Negative carbon emissions
9. Small fuel, small power plants
10. Process heat
11. Lifesaving Medicine

In detail:

1. Very High Safety

Many people are not aware that traditional reactors, although theoretically more dangerous than MSRs, are probably much safer than you think.

Statistically, traditional reactors are as safe as hydro and wind power:

(data source)

How can nuclear be so much safer than fossil fuels? Because almost every death linked to radiation is the result of a single disaster in the poorly-designed Chernobyl plant in the USSR, which may ultimately kill over 4,000 people via cancer.⁴ In contrast, coal plants by nature deliver small amounts of poison to everyone nearby — billions of people — contaminating air and water and causing millions of premature deaths (if not more). Coal plants even release 100 times more radiation than nuclear plants. So when fear of radiation shuts down nuclear plants, the result is more deaths and more global warming as oil and coal plants quickly raise output to replace those reactors.

MSRs should be safer still. MSR designers understand that the public has zero tolerance for nuclear disasters, so every aspect of MSR design reflects that. They avoid most of the hazards of traditional reactors, such as potentially explosive high-pressure operation, active safety systems that don’t work without electricity, and water coolant inside the reactor (which, if everything goes wrong at once, could boil away or leak explosive hydrogen gas).

2. Sustainability

Surprisingly, MSRs can be as sustainable as solar and wind.

Traditional reactors mostly burn U-235, which is about as commonplace as gold or platinum. U-235 makes up only 0.7% of natural uranium, and in a traditional reactor, a significant fraction of the U-235 will go unburned. In contrast, there is a kind of MSR called a “breeder” that is powered by thorium and can burn up almost all of it.

Thorium supplies are unlimited. It is so common, and contains so much energy, that, well… imagine an earth-sized planet with a “crust” of 100% crude oil. Our Earth’s thorium contains as much energy as 56 such planets⁵.

One cubic metre of soil or rock contains, on average, 26 grams of thorium containing 2.06 TJ of energy — as much as 56 cubic metres of crude oil. Thorium can be burned in “breeder MSR” or “LFTR”. Image credit

The word “breeder” refers to the fact that thorium cannot burn directly. It’s like wet wood: it can burn, but only if a fire dries it out first.

So thorium is placed inside a reactor, beside the fuel, where intense neutron radiation causes it to transmute into just enough new fuel (U-233) to keep the reactor running indefinitely. The transmutation process is called “breeding”. A breeder must be “jump-started” with a substantial amount of uranium (or transuranics), but runs after that on thorium alone.

There is one major caveat: the earliest MSRs will probably not be thorium breeders; instead they will probably burn “ordinary” fuels like U-235 for the sake of simplicity. This will allow them to be certified by regulators more quickly, so they can start turning on before 2030. Good news, though: First, engineering expertise gained from building early MSRs will make it easier to build breeders later. Second, MSRs use less fuel than traditional reactors.

Note: A common misconception is that thorium reactors are “better”. In fact most benefits of MSRs come from the salt, not the thorium; the biggest benefit of thorium is sustainability. The misconception arises because thorium and MSRs are often presented together as a single technology called LFTR (“lifter”) or MSBR (molten salt breeder reactor). Plus, the “Thorium Energy Alliance” confusingly promotes all MSRs, perhaps because MSRs are a stepping stone on the path to LFTR technology.

3. Retrofitting fossil fuel plants

Wen’an steel plant, Wu’an, China

Businessmen don’t want to destroy their own businesses, so we do not expect that a coal or oil plant would choose to shut down to save the environment. But perhaps we can reach a compromise.

MSRs can use the same general kinds of turbines as fossil fuel plants, so MSRs may someday help shut them down by replacing oil and gas burners with reactors. This avoids throwing away perfectly good turbines, and it saves money on fuel (SMR fuels are cheaper than solid fuel rods, which in turn are much cheaper than fossil fuels).

This is possible because MSR plants, which are more compact than traditional nuclear plants, don’t need skilled operators (no combination of buttons on the control panel can cause a disaster), don’t necessarily need to be near a body of water, can completely isolate all nuclear materials from the turbines, and can run at temperatures hot enough to run standard turbines (though early MSRs cannot provide quite as much heat as state-of-the-art gas plants.)

4. Nuclear waste burning

Traditional light-water reactors (LWRs) use only up to 4% of the energy in their fuel. Through reprocessing, the fuel can be re-used, but even so, much of the fuel is never used and becomes waste that is hazardous for thousands of years (claims about the length of time traditional waste is substantially hazardous range from 3000 to over 10,000 years).

In contrast, MSRs can use up almost 100% of their fuel, and in consequence will produce about 5 times less high-level waste per unit of energy.⁶

So, imagine less of this.

But that’s not all. Some MSRs can destroy transuranics (actinides), so that the final waste stream of an MSR is dangerous for 300 to 500 years (estimates vary), or at least 10 times shorter. Combining these two factors, MSRs can be up to 50 times better.

But here’s the best part: an MSR can be designed to consume nuclear waste instead of producing it. This is possible because transuranics — which is the stuff that is hazardous for thousands of years — is valuable as a nuclear fuel that MSRs can burn.

However, this feature can’t be taken for granted: thermal-spectrum MSRs, such as the ThorCon, won’t be able to destroy much of their transuranics initially. They need to add a recycling step on their spent fuel in which the transuranics are separated and re-used in new fuel. “Waste burner” designs such as Moltex do a better job destroying transuranics the first time through. My impression is that the recycling process is economical, but in order to handle spent fuel safely, companies may need to keep it in storage for a couple of decades while its radioactivity dies down. Arranging temporary storage is not hard, since a few decades worth of spent fuel is much smaller than the reactor that burned it.

“If you take all of that nuclear waste and put it into waste-burning nucelar reactors… you could actually produce enough electric power to power the entire world for 72 years.” — Leslie Dewan, CEO, Transatomic Power

In the long run, the final quantity of high-level waste from MSRs is tiny compared to the reactors themselves.⁷

Similarly, MSRs can be used to destroy nuclear bomb cores.

5. High efficiency

Not only is the fuel used more efficiently, but an MSR also generates electricity more efficiently because it runs at a higher temperature than traditional reactors. Depending on the turbines used, MSR plants will be 40% to 55% efficient at converting heat to electricity, compared to 34% for a typical LWR. This higher efficiency is possible with smaller turbines, since low-temperature steam turbines are very large.

6. Low cost

Low prices are the one thing that can beat global warming faster than pledges and politicians. Proposals to pay taxes for clean energy or give up SUVs face opposition — low prices, not so much.

The only thing holding back MSRs right now is uncertainty. A commercial MSR has never been built before, so there is regulatory uncertainty (necessary regulations aren’t established yet) and no proven financial track record. So, several MSR companies that want to reach the market quickly are using materials and fuels that are already certified and understood by regulators. Even so, one MSR company (ThorCon) estimates that they need to spend over $800 million to complete and test their first working prototype.

Once testing and certification are done though, the first MSR plants will cost less than coal plants because they follow a key design principle: don’t manage hazards — eliminate them.

Old reactors spend a lot of money keeping hazards under control; MSRs just get rid of them. Because of this, they don’t need super-expensive containment buildings or backups of backups of backup safety systems. And their cores are small, often small enough to be built in a factory and shipped by truck, rail or ship, with only a containment building constructed on-site. MSRs need less raw material per watt than coal, wind, or traditional reactors. And once you add assembly-line construction, MSRs might end up as the cheapest energy source in history.

And the fuel? Compared to its energy output, it costs almost nothing.

7. Load following & energy storage

MSRs don’t need backup power plants — fossil fuel burners, pumped hydro storage, big buildings full of batteries, or extra renewable plants located thousands of miles away. None of that. They simply follow the load, producing as much or as little power as needed.

And MSRs can change their power output more quickly than traditional reactors, better matching the demands of the electric grid.

MSRs can complement renewable energy using affordable molten salt energy storage (e.g. Moltex SSR). The idea is to buy fewer reactor modules, and buy tanks of salt instead (this will be clean, non-radioactive salt). When the wind is strong, the nuclear plant generates less electricity and sends excess heat to the salt tanks. When the wind slows down, the plant can use the heat in the salt tanks like an extra reactor, increasing power output far beyond the limits of the reactor(s) alone.

8. Negative carbon emissions

MSRs produce “waste” heat — the heat left over after making electricity — and we can perhaps use it to desalinate seawater. But what else could we do with that heat?

A little-known fact is that halting global warming before we reach 2°C will almost certainly require “negative emissions” technology that removes CO2 from the air. To reach the Paris target of 1.5°C is much harder still — that’s basically IPCC RCP 2.6, which assumes human CO2 emissions will drop below zero by 2100, meaning 100% renewable energy may not be enough.

There’s one big problem: the technology we need isn’t ready. There are various pilot projects and proposals, but we don’t know which technique will work best, so we should be ready to explore every possibility. If it turns out that the best technique needs a lot of heat, MSRs may be the ideal solution (since wind and solar don’t provide waste heat.)

The Climeworks direct air CO2 capture facility in Switzerland is powered by waste heat.

9. Small fuel, small power plants

MSR fuel is incredibly small and cheap. One kilogram of thorium or U-235 has as much energy as 2.6 million kilograms of coal.⁸ A lifetime supply of thorium literally fits in the palm of your hand. And the waste isn’t that much bigger.

Yes, it’s safe to hold it.⁹

MSR plants can be smaller than traditional nuclear plants, which in turn are smaller than coal plants when you take into account coal tailings ponds. Small plants means lower capital costs and land costs, and compared to wind, money is saved on transmission lines. And as the world population passes 10 billion people by 2100, compactness itself may be valuable on occasion.

10. Industrial heat

Less than half of global carbon emissions come from electricity production. That could change when the world switches to electric cars and electric heating, but there will always be industries, such as steelworks, that need heat instead of electricity. Right now they tend to get their heat from fossil fuels. How will we wean them off?

While traditional reactors run at 300°C, MSRs can provide heat at 600°C, and perhaps hotter in the future. Since electricity generation is rarely more than 50% efficient, getting heat directly from a reactor may end up to be the best and cheapest option.

11. Lifesaving Medicine

Nuclear reactors can produce medically valuable radioactive substances which cannot be produced any other way.

• Bismuth-213, generated by U-233 decay, may be useful for treating dispersed cancers like leukemia (targeted alpha therapy), though my impression is that more research is needed to prove its effectiveness. U-233 is produced by MSR breeders.
• Molybdenum-99 / Technetium-99, a common fission product, is used for medical diagnostics, such as cancer detection.

Nonproliferation

I haven’t listed this as an advantage of MSRs because there still seem to be some proliferation concerns with MSRs. Inspectors will be required to verify that countries comply with the nuclear non-proliferation treaty (NPT) — that they don’t use nuclear power as cover for a weapons program.

Also, good security will be required during fuel transport to prevent fuel from being stolen by terrorists. Nuclear fuel cannot be used to create a nuclear weapon directly; a bomb requires further enrichment, which is usually done with specialized centrifuges. Such equipment seems to be rare, as the NRC notes that the U.S. has only one commercial enrichment facility; nevertheless I don’t have enough information to be confident that terrorists can’t use the technology. Terrorists stealing existing nuclear weapons also seems like a possibility, but they may have to rebuild the bomb starting from the core (the Sum of All Fears scenario) because existing bombs are designed to prevent unauthorized detonation. Spent fuel (also known as high-level nuclear waste or HLW) is less suitable for bombmaking, but I suspect there are still reasons to be concerned about its security.

It’s worth noting that waste-burning MSRs such as the Moltex SSR are good for nonproliferation, because they can destroy nuclear bomb cores and reduce high-level waste.¹⁰

So when can we get them?

It’s often pointed out that nuclear reactors can take over 10 years to build and start, but there are multiple historical examples of reactors being built in less than 5 years. The very first experimental MSR was designed, fabricated and built between 1960 and 1965 with only a few million dollars in funding and without the benefit of today’s engineering tools.

The nuclear experts I’ve seen agree that there are no major engineering barriers to fast construction of simple MSRs (although the MSBR/LFTR faces some significant challenges). The issue, rather, is that there is little political will for anything nuclear. Liberals distrust nuclear power, while conservatives don’t think climate change is an urgent problem. Any delays, therefore, will be caused by political resistance and regulations. Since MSRs are so different from old reactors, they need new regulations to govern them, which in many countries depends on government cooperation. The willingness of governments to establish new regulations depends on public opinion. Thus, those guys who run around telling everyone “nuclear’s never gonna work” are precisely what might stop MSRs from working.

Because of this, the first reactors won’t be built in the United States or the UK; early adopters are likely to be China, Canada, Indonesia, and India. Several companies hope to finish constructing the first MSRs within the next 6 to 12 years.

Footnotes

[1] (Trigger warning — anecdotal information) In daytime, the average difference between my home office and the outdoors is about 4°C. Global warming has made the Philippines at least 1°C warmer on average, air conditioning makes up well over half of our electric bill, and we sometimes continue air conditioning at night, so global warming causes about 25% of our electricity consumption.

[2] I couldn’t track down the amount of total deployed storage, but 0.336 GWh of new storage was deployed in 2016 (source). Since the U.S. uses 528 GW of electricity, that’s about 2 seconds of storage. Since energy storage is a new phenomenon associated with renewables, it is unlikely that total storage exceeds twenty seconds.

[3] Kirk Sorensen gave a talk discussing the history of 1960s nuclear/MSR development.

[4] Some of the 4,000 deaths have not happened yet, although if the LNT hypothesis is true then 5,000+ additional cancer cases may occur across Europe for a total of 9,000+. These additional deaths are impossible to confirm, since the potential increase in cancer rates is far too small to observe. It’s interesting to compare this with the dam collapses starting at Banquiao, which killed 171,000 people and displaced 11 million — but did not cause intense worldwide fear of hydro-electric dams.

[5] Wikipedia cites three estimates of the abundance of thorium and the average of the three is 9.2ppm by mass, which is 25 grams per cubic metre weighing 2.7 tons. Based on the energy density of thorium, the 25 grams contain 1.97 TJ of energy, which is roughly the same as 53 cubic metres of crude oil. This is only the abundance in Earth’s crust; I gather that thorium has a much lower concentration in the upper mantle (see Workman & Hart (2004): Major and Trace Element Composition of the Depleted MORB Mantle, Table 2).

[6] This is based on a presentation by David LeBlanc of Terrestrial Energy, who says MSR burners use 1/4 to 1/6 as much uranium as LWRs. Other sources imply MSRs use uranium relatively more efficiently than this, but I do not understand how the efficiency of LWRs is calculated and I would appreciate it if a nuclear expert could comment on this issue.

[7] For example, ThorCon cans (which are a tiny fraction of the size of a power plant) have a volume of 490 m³, and they produce less than 0.5 m³ of “ash” per year. Since this ash will be virtually harmless after 500 years, it will never exceed 250 m³. The ash requires extra volume for shielding, but since the reactor itself also requires shielding, it’s fair to say that the HLW can never be larger than the reactor itself.

However there are a couple of caveats. First, it should be noted that the fuel’s intense radiation induces radioactivity in the reactor module, which becomes “intermediate-level waste” (ILW). Since the quantity of high-level waste (HLW) is so tiny, there is likely to be more ILW than HLW. There is also “low-level waste” (LLW) — other objects that picked up radioactivity from spending time in a radioactive environment, including hospitals where radiation is used for diagnostics and treatments. LLW is really not very hazardous, but it’s kept away from humans out of an abundance of caution.

Another caveat is that MSRs, especially non-breeder designs, won’t completely burn up all their fuel, because the reactor needs a certain amount of fuel inside to keep working. As I mentioned, leftover fuel can be recycled and placed in a new reactor.

[8] This can be derived from MJ/kg energy densities on Wikipedia, assuming 100% burnup of both fuels and not counting the weight of oxygen used in coal combustion. Note that “uranium” technically contains about as much usable energy as U-235 but the energy is much harder to extract, requiring a breeder reactor. I believe a thorium breeder will prove easier to build, but until that happens thorium can only provide a minor boost, e.g. in the ThorCon, thorium will be 80% of the fuel mix but provide 25% of the power. Natural uranium is 0.72% U-235 so without a breeder, its usable energy is only 19,000 times that of coal.

[9] The most important thing to know about radioactivity is that it is inversely proportional to half life. Thorium is radioactive, but because its half-life is 14 billion years, its radioactivity is very small despite having a nasty decay chain.

[10] By the way, plutonium in normal spent fuel is not generally considered a proliferation risk, because it contains enough Pu-240 to reduce the explosive power of a nuclear weapon by 10 to 100 times (if I understand this document correctly). If we don’t build new reactors to burn our previous waste, 65% of this Pu-240 will have decayed away after 10,000 years. Some would argue this creates a far-future proliferation risk, but I personally think that if human civilization lasts 10,000 years, it must have already figured out how to avoid rash decisions like building nuclear weapons or world-ending superviruses.

I’d like to thank George Moore of MIIS for speaking with me about proliferation issues, and to Ian Scott of Moltex Energy and the folks at ThorCon for answering some questions.