The Billion-Degree Energy Solution

Decades of R&D have failed to produce a useful fusion power plant. To succeed, maybe we need to make the problem a whole lot harder.

The world around you is packed with hidden energy. Every thing on top of another thing could fall, roll, or flow downhill, releasing gravitational potential energy. Every element or compound that can burn is suffused with chemical potential energy. Earth’s crust contains vast quantities of unstable elements that emit natural radioactivity, releasing pent-up nuclear energy, and some of those elements break apart so easily that they will initiate a chain reaction if you bring them together. (I’m looking at you, uranium-235.)

And then there’s hydrogen. The primary element in water releases an impressive amount of energy if you can manage to get a naked hydrogen nucleus – better known as a proton – to stick to the nucleus of another hydrogen atom. Proton-proton fusion is the primary energy source in the Sun and, by extension, it is the primary energy source driving all of biology on Earth. Therein lies the enduring appeal of nuclear fusion: If you can get atomic nuclei to stick together in a machine, you can replicate the power of the Sun, and you can do it using an essentially limitless fuel, generating no fumes or carbon emissions.

If you want to make a better life for all of us – a life supported by an abundant, always-on energy source that doesn’t contribute to global warming – then fusion looks like your great enabler.

The bummer about potential energy is that it’s useful only when you can turn it into radiant or kinetic energy, what physicists once called “active energy.” For some types of hidden energy, conversion is not so hard. Stick a wheel in a flowing river and you can drive a generator. Fusion, on the other hand, is a brute. The Sun makes it happen by pressing down on protons with 2 million trillion trillion kilograms of mass. On Earth, the most common approach to fusion is to start with heavier isotopes of hydrogen (deuterium and tritium, which fuse more easily), trap them in an intense magnetic field called a tokamak, heat them to 100 million degrees C or more, and scramble to manage the melee of flying neutrons that emerge.

Achieving controlled fusion is every bit as difficult as it sounds. Scientists and engineers have been working the problem for more than half a century without creating a fusion machine that usefully delivers more power than it consumes. One major reason for these failures is under-investment: Fusion is interesting enough that governments and, more recently, private investors keep sinking money into it. But fusion is complicated enough that all parties hedge their bets, never plunking down the many billions of dollars probably needed up front to develop a viable fusion power plant. Fusion insiders call the strategy we’ve been following “fusion never.”

Much as I’d love to see our global economy running on steady, clean fusion energy in place of coal and oil, I understand the hesitation to invest. If a fusion power plant costs, say, $25 billion to build (like the international prototype ITER reactor), it doesn’t matter if the fuel is free. The up-front capital costs will simply be too high to make fusion power attractive to customers.

Flipping the Fusion Script

Which is why I’ve long been intrigued by a company called TAE Technologies and its determined CEO, Michl Binderbauer. His argument is that the current paradigm for fusion energy is too messy and cumbersome. And his counterintuitive claim is that the only way to make fusion competitive is to abandon the whole current paradigm and switch to an approach that, at first blush, seems to make the job a whole lot harder.

Binderbauer champions an entirely different fusion reaction, one that dispenses with deuterium and tritium. He wants to achieve fusion between nuclei of ordinary hydrogen (protons) and the element boron-11 (B11). Think hitting 100 million degrees C is a challenge? Making p-B11 fusion happen requires temperatures above 1 billion degrees. Tokamaks can’t support such extreme conditions, so most of the existing fusion technology goes out the window too.

To overcome these daunting obstacles, TAE has also been developing an entirely different technology, called a field-reversed configuration or FRC (illustrated below). A tokamak is essentially a magnetic doughnut; a FRC is more like a spinning magnetic tube. TAE has built a large test reactor, called Norman, that is currently testing these spinning plasmas in the company’s southern California research center. The fundamental technology is alive and kicking, though Norman is not nearly powerful enough to hit the target of p-B11 fusion.

One big upside of p-B11 fusion is that it is inherently a lot easier to handle than deuterium-tritium fusion. Remember I mentioned that deuterium and tritium spit out a neutron when they fuse? Fast neutrons are very hard to control. They tend to bombard everything around them, gradually turning the inner workings of fusion reactor brittle and radioactive. (Tritium itself is also radioactive, a further complication.) In contrast, p-B11 reactions spit out protons, which can much more readily be captured and converted into useful energy.

Another big upside of TAE’s unconventional strategy, according to Binderbauer, is that FRC reactors could be mass produced in ways that are essential for reducing costs and deploying fusion energy on a large scale.

The proof of the energy-producing is in the fusing, though, and TAE is not there yet. The company is still raising money for its next-generation test reactor, Copernicus, designed to test the science and engineering of a FRC reactor at much higher energies. After Copernicus comes the company’s next-next generation machine, Da Vinci, which is the one that could finally show achieve p-B11 fusion and show the path to a commercial TAE fusion reactor.

That’s a lot of ground yet to cover, a lot of funds yet to raise. Meanwhile other companies are pursuing their own approaches to fusion energy. One of them, the Bill Gates-backed Commonwealth Fusion Systems, boldly announced plans to build a commercial fusion power plant in Virginia “in the early 2030s,” using a variant of the tokamak design. A dozen-ish competitors, operating at a wide variety of scales and seriousness, are making their own attempts. Governments and government consortiums are pursuing fusion, too. No surprise, China is active in this area.

But Binderbauer impresses me, both for his outrageous ambition and his refreshing candor about the steep road ahead. I caught up with him recently to find out where things stand with TAE, p-B11 fusion, and the billion-degree project to unlock the hidden energy of the atomic nucleus.

The following interview has been edited for length and clarity. Even so, it’s a lot longer than my usual column. I hope you agree it’s worth your time.

IN CONVERSATION – Michl Binderbauer, CEO of TAE Technologies

You’ve been pursuing your unusual fusion-energy strategy with TAE since 1998. Where is the company today compared to where you hoped to be by now?

The answer is always, I wish I were further already. Covid changed a lot in terms of efficiency of getting money raised. Luckily, we had enough capital to survive, and we regrouped quickly as soon as California allowed some form of office presence. The silver lining is a lot of learning occurred [during pandemic downtime]. We have substantially increased the performance of Norman with regards to hotter electrons, more stored energy and longer pulses. We ended up pushing the performance on our machine considerably higher, more than double where we were in 2019. We went from 35 million degrees to about 75 million degrees. That’s pretty big.

All this on the back of very sophisticated real-time plasma control capabilities that are courtesy of our work on machine learning and significant advances in power management. Altogether, we can now react to changes in the plasma core on millisecond time scales.

To reach even higher temperatures and densities you need your next reactor, Copernicus?

Yeah. When we built Norman [TAE’s current test reactor, illustrated below], we designed it for 30 million degrees and for density on the order of a factor of two lower of where we are. With clever tweaking, our engineers were able to push the edge of the machine's performance envelope. But if we were to go and push harder with this plasma, we're going to have a blowout!

We've got to build a new machine to get to the point where you can talk about net energy production. P-B11 [proton-boron fusion] is the ultimate goal. The next step on the ladder is to get to the condition of the DT [deuterium-tritium] environment, about a 100 million or 150 million degrees. That’s not far away, but it’s not achievable with the current machine. We need more real estate to pack bigger particle accelerators on. We need to scale the whole machine.

When will that next-generation fusion reactor actually get built?

I’m raising the money. It’s going slower than I wanted. The financial markets at the moment are extremely anemic. What we need is an institutional angel investor — somebody who can write $100 million check and has a time horizon of five, six years at least for return, and has an appetite for more angel-level risk. We need about $250 million for the [Copernicus] machine.

So we are not ready yet, but we're not trading water. The additional time has afforded us the opportunity to retool and test components for Copernicus on Norman, optimizing them before the next machine is built. This has validated our technologies, and led to cost and efficiency improvements. We are finishing all the design work. We have the facilities. It’s going to be in Irvine, about five miles down the hill from where we have the current labs. Now I need to get the capital raised and go into a construction phase.

You also still have to demonstrate the fundamental concept – that proton-boron fusion is technologically feasible. Where are you with that?

We already have a lot of experimental data at very rarefied performance [in the Norman machine], which we can use to calibrate our computational tools. We’re within a factor of two of where we need to be to burn tritium. Then you can extrapolate by a factor of 20 or so and get to boron fusion. Our computational modeling gives us estimates about what we’re going to need to do that. The conviction that we can do proton-boron is very high. I know we can build the technology components we need. We’ve prototyped those. We’ve built particle accelerators at full p-B11 reactor requirements. We see that we can get the quality and performance on the beam we need.

Yet many people remain skeptical about the current commercial fusion efforts, and about TAE’s unconventional approach in particular. How do you respond?

True, nobody has yet operated anything successfully at full sustainment, producing net energy, even at deuterium-tritium conditions. The point people make is, if you haven't done that [achieving deuterium-tritium temperatures and pressures], why are you considering that you could go a factor of 10 or 20 higher? But people used to say that nothing other than a tokamak could get you to the regime that we're operating in today. We’ve totally proven them wrong.

The question I would be perfectly willing to accept is: Can we get there practically? Is the challenge too big? That question stands until we actually do it, but there's no fundamental showstopper.

If you get your funding and build Copernicus, could you turn it on and achieve deuterium-tritium fusion right away, as a demonstration of your FRC technology?

Initially we’re going to run on hydrogen [not tritium], so Copernicus won't do any ignition. The reason is cost. If you want to add [radioactive] tritium, you have handling issues, maintenance issues, you need shielding, there’s regulatory overhead. That machine would go up to $1 billion plus immediately. We know enough about the nuclear physics that we can run the machine on a deuterium and hydrogen mix and verify the plasma performance.

I'm totally certain that we know how to do this. We've got the basic recipe, we have a design for the machine, and we're confident that it's going to achieve its mile marker. Where I have to be a little careful is, there is very likely going to be tweaking to do. Some of the hardware we are building extrapolated from today’s machine may require a rebuild. People make mistakes too in the design, or fabricators make errors. I'm concerned that some stupid mistake will set us back by six months. That is probably the highest risk on Copernicus.

The other big question is not just can you do fusion, but can you make it affordable? How do you propose making it a cheap, useful part of the future electric grid?

We started from that question, because this isn't science for science’s sake. This is for an applied end product. I think that's where a lot of people in the field didn't bother to look. I mean, look at ITER [the $25 billion test reactor under construction in southern France]. That thing is not going to lead to a practical machine. You will need massive economies of scale and improvements in fabrication techniques and probably tweaking on the raw material to bring prices substantially down. The thing has to be small enough or you can’t have mass production. This has to be a fundamental constraint that you have to solve for in your design. We ended up with hydrogen-boron fusion for that very reason.

Most fusion research focuses on using a tokamak – a magnetic doughnut – to contain a deuterium-tritium plasma, and then managing the neutron radiation. It’s an established technology. What makes you think your alternative, FRC approach could be cheaper?

If you have a lot of shielding and decay of critical reactor infrastructure [due to neutrons from deuterium-tritium fusion], that's going to hugely impact pricing performance. The other thing is the magnetic field that's used in a tokamak is very inefficient. We are about 10 times as efficient, and fusion output scales to the square of efficiency. So if you have the same magnetic field, we're going to have a hundred times more fusion output. This is just basic physics.

The US government has a long history of funding fusion research but also setting regulations that pose challenges for commercializing fusion energy. With the current green-energy backlash, are things getting better or worse for fusion?

There is bipartisan support for fusion in both Congress and the White House. The regulations governing fusion safety are a great example. Two years ago, the Nuclear Regulatory Commission decided to regulate fusion separately from fission [as in nuclear power plants] and treat commercial fusion similarly to how it treats particle accelerators. It’s important because these rules aren’t one size fits all. If you can build a device like TAE’s that generates less radioactivity, then the regulations scale to match that. Last year Congress wrote that into law with the Fusion Energy Act. It provides regulatory certainty for the fusion industry and for TAE specifically, as our aneutronic [ie, without neutrons] fusion is actively referenced in the text.

We’re already seeing support for fusion from the new Administration. The new Energy Secretary recently released orders with his initial direction to the department. The top innovation priority he called out as a major focus was fusion, followed by high-performance computing and AI.

What about support from the commercial side? Do the growing energy demands of AI and the tech industry's support for nuclear power offer an opportunity for fusion energy as well?

Absolutely. Major cloud providers and AI developers will be addressing an urgent need for abundant clean power and will be trying to bring as much new capacity to the grid as possible to achieve their goals for scaling data centers and supporting even larger AI models. As an outsider looking in on this, I feel like it almost won't matter whose AI tech is the best, but who has the electrons to run computations. And there's all these other uses of power that are coming up, including EVs plus all the electric transportation.

Given the aneutronic and physically compact nature of TAE’s approach to fusion, we also benefit from being able to site an energy generator very near to population centers or other industrial installations.

Reading new stories about fusion energy always feels like whiplash: starry-eyed dreams about the potential of fusion, contrasted with cynical dismissal that it could ever work. How do we break through that divide and turn fusion energy into something real?

From its inception, TAE has been thinking about commercially competitive, grid-scale fusion. As we are marching closer to the final steps in our plan, we have more data, more operational mastery, and now machine learning-based control systems to drive the transition from an experimental system that demonstrates net-energy conditions [Copernicus] to a first system that actually makes net electrons [the planned Da Vinci machine].

P-B11 makes this transition much more realistic, because we don’t need to develop new materials or shielding technology that can limit the damage from neutrons. We are also actively developing strategic partnerships with companies and entities who can help deploy fusion power plants and eventually scale a commercial roll-out. We believe it’s critical to develop these relationships now, as it will take time to develop mutual trust and the technical understanding necessary to be successful in delivering commercial fusion.