It’s little wonder that nuclear fusion has long been one of humanity’s great hopes to address the world’s energy problems. Two isotopes of hydrogen fuse together and release millions of times more energy than an equivalent amount of fossil fuel. Indeed, just fifteen kilograms of fusion fuel provides about as much energy as 100,000 metric tons of coal. On top of that, it’s carbon-free, and, unlike fission, its reaction products do not take the form of long-lived radioactive waste that must be stored. Apart from a spray of neutrons, which do not go beyond the reactor’s containment structure, the only direct byproduct is helium, which is inert. All things considered, fusion power would be world-changing.
Of course, that’s just in theory. Making it actually work has proved horribly difficult. The challenge is to create and sustain a tightly compressed, multimillion-degree ball of hydrogen plasma that wants to expand. It wants to explode. It wants to radiate and cool, shutting itself down. But if we could contain it—keep it hot, keep it dense—natural fusion would proceed continuously, and we could use the heat it produces so prolifically to generate electricity, just as we use coal-fire heat now.
My career at Los Alamos has allowed me to contribute to a wide range of nuclear science and technology. I’ve worked on fundamental physics, specialized materials, nuclear weapons, and various external signatures for monitoring activity inside a nuclear reactor. But perhaps my favorite line of research has been my work on something called inertial-confinement fusion, or ICF.
There are two primary ways to try to contain a multimillion-degree plasma (that is, a gas so hot that electrons detach from their atoms). One way is with a powerful magnetic field. The other way, used in ICF, is a spherical implosion driven by lasers to compress the plasma, at least briefly. Unfortunately, in decades of experiments, the power required to maintain the magnetic field or drive the implosion has exceeded the power produced by the resulting fusion. Scientists have long been trying to achieve something we call “ignition”: getting more power out than what we put in.
In the United States, there is a major ICF research facility that was built expressly for this purpose, the National Ignition Facility (NIF) at Lawrence Livermore National Laboratory in California. The scientists there work on ignition in earnest, and some of us at Los Alamos collaborate or help with certain aspects of the problem, such as redesigning the fuel capsules or developing specialized instrumentation to see what’s happening—generally, what’s going wrong—inside the imploding core. (We also use NIF experiments to inform us about fusion processes that impact the design and performance of nuclear weapons.)
NIF works like this: 192 extraordinarily powerful ultraviolet lasers converge from all directions into a container that acts like an oven, cooking a fuel capsule only a few millimeters in diameter. The cooking causes the outermost part of the fuel capsule to blast away, producing a powerful recoil implosion that compresses and heats the fuel enough to initiate fusion reactions. The whole thing is quite exquisite, really.
And yet the manner in which it typically fails is utterly ordinary. The compression just isn’t symmetrical enough. It’s like squeezing a peach with your hand: you can never do it perfectly, and bits of fruit come squirting out the gaps between your fingers. In other words, the basic challenge of fusion, containment, is the problem. And that’s just for the plasma; it is also a struggle to contain the radiation. The heat from fusion generates a kind of x-ray glow, and the x-rays do what x-rays do: go right through things and fly away, taking lots of valuable energy with them—energy we ICF scientists would rather see remain in the core to keep the fusion reactions going.
Just fifteen kilograms of fusion fuel provides about as much energy as 100,000 metric tons of coal.
Don’t get me wrong: ICF works, in that fusion happens, but full ignition remains challenging. We have been edging closer with a series of clever experiments, but things have not always gone as planned, even from the start. The imperfect-compression problem is thornier than we expected. And although Livermore recently achieved an impressive marginal ignition yield, it’s unclear whether or not they can get the rest of the way, much less exceed the ignition threshold by enough to convert an encouraging result in an ICF experiment into a practical power source.
Several of my colleagues have long believed that ignition would never happen without at least containing the x-rays. That much, fortunately, can be done: The fuel capsule can be redesigned to include a shell made from a heavy metal, much like wearing a lead apron at the dentist. I find this heavy-metal shielding to be an ingenious idea; however, it is not sufficient to reach ignition (not yet anyway), and it introduces two additional problems. First, in the heat of the implosion, the heavy-metal layer will not remain solid; its atoms may start to mix into the fuel, thereby effectively diluting the fuel’s ability to support fusion. (This is an important consideration, and I will return to it.) Second, most of the standard diagnostics used to measure what’s happening inside the core—the diagnostics that tell us how unevenly the core is compressing—do so by observing x-rays, which would now be blocked.
This brings me to a crazy idea I had about ten years ago that actually paid off.
My colleagues were developing an imaging system to see neutrons, rather than x-rays, coming out of the core, including neutrons that scatter off of ions of fusion fuel before escaping. I wondered: What if we looked for the products of those neutron-scattering reactions? Could we somehow observe those? And if we could, what might they tell us about the conditions inside, such as the effect of heavy metals mixing into the fuel?
While everyone else was naturally concerned with what was happening inside the core, we focused on what was happening just outside.
The scattering of neutrons in the capsule captured my interest. Each time a fusion-borne, high-energy neutron from the core fusion region careens into a fuel ion outside the core, there’s a chance that the newly energized ion will subsequently induce another fusion reaction out of equilibrium, by which we mean a very energetic fusion reaction. I started to focus on what we call “reaction-in-flight” (RIF) neutrons: those that emerge from that second, higher-energy fusion reaction.
Only about one in ten thousand neutrons produces a second reaction. Some of my colleagues thought it was a waste of time; we would never see RIF neutrons. My brain knew they were probably right. But my gut must have thought otherwise, because I just wasn’t ready to give up on it.
RIFs at NIF
Maybe the reason I couldn’t just let the RIF neutrons go is the fact that they are sort of fascinating in their own right. The two isotopes of hydrogen used in fusion, deuterium and tritium, collide and reconfigure in a way that spits out a neutron. A deuterium nucleus (D) is made from a proton with one neutron, and tritium (T) is a proton with two neutrons, so the D-T collision involves a total of five particles. For reasons specific to nuclear physics, D-T collisions result in a helium nucleus (made from four of them: two protons and two neutrons) plus a stray neutron. That’s the emitted neutron my colleagues and I wanted to observe from high-energy fusion reactions. We succeeded. It was an incredibly challenging measurement but one that turned out to be incredibly useful.
In a normal D-T fusion reaction, the energy of the outgoing particles comes from the fusion energy itself. The helium nucleus comes away with about 3.5 megaelectronvolts (MeV), enough to help heat up the plasma (this heating is the useful energy produced by fusion), and the neutron has 14 MeV. That’s a lot. So, I was interested in what would happen if the 14-MeV neutron hit something else. If it hit another D, for example, probably somewhere outside the main fusion hot spot, and that D then careened into a T, there would be another fusion, and this time, the energy of the neutron that emerged would depend not only on the fusion reaction itself but also on the speeds and trajectories of the collision that produced it. If you do the math, it turns out that the second neutron—the RIF neutron—can have up to 30 MeV! This particle would be both incredibly energetic and, potentially, incredibly useful, so while everyone at NIF was naturally concerned with what was happening inside the core, our team was concerned with what was happening just outside in the non-burning plasma.
So, in an ocean of 14-MeV neutrons, I wanted to see if we could detect a comparatively tiny number of much higher-energy RIF neutrons. Following a brilliant idea from my Los Alamos colleague Bob Rundberg, now retired, we rigged up a series of metal foils to capture RIF neutrons. Two foils in particular do the heavy lifting—a thulium foil, which collects neutrons in a particular way at a threshold energy of 15 MeV or greater, and a bismuth foil, which does something similar above 22 MeV. In both foils, capturing a neutron means changing an atom of the metal from a stable isotope to a highly radioactive isotope with a known half-life. When the original isotope bismuth-209, for example, captures a neutron, it can spit out four more neutrons to become bismuth-206, thereby going from a stable nucleus to one with a six-day half-life—a distinction that’s hard to miss. Still, we needed a very specialized detector and a sophisticated analysis technique to find our RIF-derived radioactivity signal amidst the overwhelming background created by the majority of neutrons that emerged at “only” 14 MeV. But it worked. We were able to quantify the particular slice of radioactivity we were looking for and thereby determine how many RIF neutrons hit the foils.
Now we can use RIF neutrons—the same neutrons most experts thought it would be folly to even try to detect—to make important measurements.
As my colleagues and I would soon learn, we could infer a lot about the conditions in the plasma just outside the burning core by comparing the number of neutrons captured by each foil. Because of the foils’ different energy thresholds at 15 and 22 MeV, the relative numbers captured give us a good indication of the energy distribution of the emerging RIF neutrons. From past experience at NIF and elsewhere, we knew what the distribution ought to be, so it would be possible to compare the measurements with theoretical calculations. We dutifully worked out the math and found that the number of RIF neutrons depends sensitively on what happens just before the fusion reactions that produce them.
In particular, after a D (or T) is hit by a 14-MeV neutron, but before it collides with a T (or D), it will be slowed down somewhat as it passes through the surrounding medium. The degree to which the D is slowed is referred to as the stopping power of the medium. We found that we could directly connect our RIF neutron measurements to the stopping power and the stopping power to the key conditions, such as density, inside the exploding fuel capsule during the fusion burn—exactly what I had originally hoped to do.
Now, remember I told you that adding a heavy-metal layer around the fuel capsule to prevent radiative energy losses runs the risk of some of that metal mixing into the fuel and diluting it? Well, that kind of mixing has a very strong effect on stopping power: the greater the amount of mixing, the greater the stopping power. Therefore, the RIF measurement system is extremely sensitive to mixing. Going forward, that means if any NIF fusion experiment underperforms expectations, we can use RIF neutrons—the same neutrons most experts thought it would be folly to even try to measure—to make a reliable estimate of how much of that underperformance we can blame on mixing, even when x-ray data are unavailable. And obviously, understanding the reason for a failure is essential when trying to remedy it.
I realize I’ve painted a rather rosy picture here: a measurement challenge overcome against the odds. And that’s true, but of course, it’s never so simple in practice. For a while, we were getting stopping-power results that didn’t make a lot of sense, and it took some time before I figured out what was going on and how to deal with it.
By way of explanation, allow me to take you back to high-school chemistry. Perhaps you remember a lesson in which you were supposed to put an atom’s electrons into their proper orbitals: two in the first energy level, eight in the second, and so on according to a specific progression. (This orbital-packing business is the reason for the shape of the periodic table: two atoms on the top row, hydrogen and helium, then eight on the second row, etc.) The reason behind all this is a quantum-physics effect which prohibits any two electrons from sharing the same quantum state. In an atom, that means the electrons must progressively occupy different orbitals; in a plasma undergoing fusion, where the electrons are not attached to atoms, it means a force arises to resist too many electrons being crammed into too small a space (where they would be in danger of sharing the same quantum state). This force is called “degeneracy pressure,” and it arises from the successively higher energies—and therefore stronger push, or pressure—that each electron must take on to avoid sharing the same state as some other electron nearby.
We physicists are well trained to handle everyday situations when degeneracy pressure is too small to matter (like for the pressure of a gas inside a balloon or inside an engine), and we also learn to handle situations when the degeneracy pressure is much larger than anything else, such as in a white dwarf star, which our sun will eventually become. But as luck would have it, NIF fusion experiments tread on the borderline, where electron degeneracy pressure doesn’t overwhelm ordinary gas pressure, but it isn’t so small you can simply ignore it either.
My brain knew it was a long shot. But my gut must have thought otherwise, because I just wasn’t ready to give up on it.
I won’t lie: fixing the math to account for partial degeneracy pressure was not easy. But once completed, the results made perfect sense. Changes in stopping power could now be convincingly associated with the transition region between gas pressure and degeneracy pressure. Moreover, we got a sort of bonus. Not only do RIF neutrons provide an indicator of conditions such as density and mixing, but they also tell us how much degeneracy pressure is at work—an important property of NIF implosions that can’t be directly determined by any other means. This knowledge should certainly prove useful in ongoing fusion research, and the only reason we can quantify it is because of a successful effort to measure something we all thought (myself included, to some degree) was probably impossible.
This is really my point: You never know what unexpected benefits (or challenges) will appear around the next bend if you don’t get close enough to look. And sometimes, especially when you’re feeling stuck on some complicated problem, the best thing to do is just pursue some aspect of the problem that catches your attention. You can move a short distance in some direction and you find your landscape changes subtly; the problem looks a little different from where you stand now. Luckily for the RIF project, this change in perspective was enough for us to see our way around the obstacles. Sometimes, whatever your crazy idea is, you just have to listen to your gut and give it a try.