Fusion power has remained an elusive dream for over half a century, but a practical solution may finally be at hand. We’re joined by Dr. Robert Bussard, former assistant director at the Atomic Energy Commission and the founder of EMC2 Fusion Development Corp, to talk about Polywell — a new type of IEC fusion reactor that he believes will finally deliver cheap, clean fusion energy.
Robert, let’s start out with the current state of fusion research. I think most people are aware that fusion research is a collection of big dollar research projects and government & academia, but I’m not sure how much anybody knows about the progress being made in this field. Can you describe for us the current state of fusion research in the United States and how much money this progress is costing us?
Tim, that’s a very complicated topic because controlled fusion research goes back to 1952. Lyman Spitzer at Princeton university invented a machine he called the Stellarator to try and make controlled fusion happen, and from that point on until 1956 it was a classified program.
It was finally declassified after the Geneva atomic energy conference because the Russians appeared and spoke openly about it. From then on, there has been a continuous government investment in a particular line of research that was adopted by nearly everyone in the Western world.
So far, over $18 billion has been spent in 56 or 58 years and they are no really closer to success than they were at the beginning, except in the sense that they’ve learned more about why things aren’t working as they wish they would.
The problem these fusion programs have is that they’re trying to control & confine fusion reactive ions using magnetic fields, and unfortunately, magnetic fields don’t really confine plasmas. A plasma is a combination of equal numbers of negative electrons and positive ions — it’s the positive ions that make fusion.
A plasma is a neutral thing overall, made up of both positive and negative charges, and magnetic fields will constrain their motion to a predictable level, if you manage to avoid instabilities, but they‘re not able to hold them in place, and that’s a fundamental physics problem that plagues all these Maxwellian local equilibrium plasma machines that everyone is trying to build.
It’s a fundamental physics difficulty that drives the machines they try to build to huge sizes. If you looked at the press releases from the government over the last few decades, you’ll see that these toroidal Tokamak big magnetic donut things are the size of small factories, and cost tens of billions of dollars if you scale them to the size where anybody thinks they might make net power.
They’re simply not economical — they won’t do what the utility companies want, and utilities have been telling them that for 30 years. The government programs go on anyway because it is good & interesting science — but it doesn’t necessarily mean we’re ever going to get to an economical fusion power plant.
I think OPEC and the 1973 oil crisis really highlighted our need for energy independence from foreign oil decades ago. Given that a lot of this research was born from those events, I’m wondering what the timeline is for the big government approach to Tokamak style fusion?
Well, I was an Assistant Director of the Thermonuclear Division of the AEC in 1971 through ‘73 when it was headed by Dr Robert Hirsch. He’s a brilliant man who earlier had worked with Philo T. Farnsworth on the things that we’re now pursuing.
At the time, Hirsch was the head of the thermonuclear fusion program at the AEC when OPEC decided to astound the world by raising oil prices arbitrarily, and suddenly there was an energy problem.
The AEC under Hirsch decided we’d capitalize on that to try and raise enough money to get fusion research really moving in the Atomic Energy Commission, because up until that point it had run it at a relatively small level. It was split amongst five national labs, with nobody really getting anywhere except understanding instabilities and the problems of confining neutral plasmas.
So we went to Congress and created a program that eventually reached something like $800 million a year in 1970-type dollars, because we could say, “look, if you can get fusion to work, you don’t have to keep using oil”.
However, the problem with those types of Maxwellian systems is they all have to use deuterium, the second isotope of hydrogen, and tritium, the third isotope. Tritium is a radioactive material you have to manufacturer by neutron capture in lithium-6, and it’s an enormously complicated process in an engineering sense. However, it’s the easiest probable way to make fusion between ions, and it’s also the only thing that can possibly work with a magnetic Maxwellian equilibrium system.
So we started the program in the early seventies and raised the money through Congress. The three of us who really put that through, Dr Hirsch, myself, and Dr Alvin Trivelpiece, said, “look, let’s get a lot of money here, make the national labs feel happy so they will be able to pursue their own interest in this at levels that they’re happy with and we’ll take 20% off the top to study the things that we know should really be done at our smaller scale.”
The problem was that all three of us left within nine months, and the people who inherited the program thought it was all real and the program should go forward with magnetic tokamaks, and it’s been that way ever since.
Well in 1995 you wrote a letter to most of the physicists and government administrators in the hot fusion field as well as the influential members of Congress on the funding committees and house and Senate saying that you as well as the other two gentlemen had supported Tokamak designs in the 70s for these political reasons you just discussed. I’m wondering what kind of response you got from them.
No response at all, because the plain fact is what we were saying is the program was completely derailed after we left. Nobody understood that we were doing it to try to raise enough money to scoop some off the top to try new real things, and it became a budget program.
I think professor Larry Lidsky at MIT said it best. He wrote an article in the MIT Technology Review, called The Trouble With Fusion and his article that basically said that when the fusion program became very large in budget, it became a big budget program and ceased to be a fusion research program.
In other words, everybody spent their time worrying about continuing the large budgets year after year, so that they could keep on with their large scale laboratories. That’s a human failing - I mean, people want things to stay the way they are and not be bothered with change.
Well, now that we’ve painted a bleak picture for the conventional approach, I want to talk about your technology. You’ve been working on a form of inertial electrostatic confinement fusion for years now that’s producing big results. Can you start by describing what IEC fusion is?
Yeah, let’s go back. It all begins in 1924 with Irving Langmuir and Katharine Blodgett in the East who wrote several papers in the Physical Review on how to produce negative potential energy wells by having oppositely moving ions and electrons in spherical, cylindrical and slab geometries.
Later, in the 1950s, Philo T. Farnsworth — the inventor of raster-scan television — conceived of a way to use these negative potential wells for controlled fusion. He produced the wells by injecting opposite-sign particles in opposite directions inside of spheres, and then he used a radial electric field to concentrate ions and make fusion in the center of the sphere
Farnsworth was a very ingenious man, and filed very extensive patents in his process in the late 1950’s. Then he proceeded to try to build some of these machines.
The idea is to make a spherical negative electric potential well, and when you drop ions into it, they’ll circulate back and forth like rolling marbles — hitting each other now and then at the center.
As they come to the center, the density increases because of 1/R² convergence, and when they collide with another ion they’ll either scatter out of the center & give their energy back to the well or they’ll create a fusion event.
The fusion products are very energetic, so they leave the system radially and fly out toward the walls. This creates heating on the walls that you can capture with steam pipes and then use to run steam turbines to produce electricity. If the particles are charged, you can also use a grid to capture them for direct-electric conversion.
The point is that an IEC fusion generator is a spherical colliding beam machine, not a Maxwellian mixed plasma machine. The IEC is completely out of equilibrium.
Farnsworth achieved the spherical wells by putting in spherical screen grids — like two sieves back-to-back inside a sphere. He biased the internal screen grids to a high negative potential, which let him accelerate ions from outside through the screens and they would come to a focus in the center.
Farnsworth had a young postgraduate student, Dr. Robert Hirsch, working with him and together they built little machines that gave them record-breaking results, generating 10¹⁰ and 10¹¹ fusions per second on D-T out of these little devices only six to eight inches in diameter. That was the beginning of it.
There was a problem, though. Farnsworth knew — and his patents disclose — that his IEC reactors could never generate net power because the ions had to go back and forth through the sieve-like grids, and every time they went through the grids, they had a chance of hitting the wires of the grids and losing their power.
Remember, the ions had to go back & forth through these grids over a thousand times before they’d create a fusion event in the center of the reactor, and there’s simply no grid that’s transparent enough for that many transits. So ultimately, the machine could never generate net power, despite the fact that it did generate fusion output.
Along about that time, three other people at Los Alamos, Bill Elmore, Jimmy Tuck, and Ken Watson, wrote a paper Elmore-Tuck-Watson, in which they inverted the potential geometry.
Instead of negative voltage on the screens, they put positive voltage on the screens and injected electrons, which would be accelerated inside of the positive screens and make a negative potential well. You could then drop ions inside the screen, and they wouldn’t have to go back & forth through the screens, so the ions wouldn’t hit the screen.
So they solved the problem of ions hitting the screens — but in their design, the electrons would have to make over 100,000 transits before a fusion event occurred, so the electrons ended up hitting the screens instead — which killed that idea.
Ultimately what this means is that no matter how you build these screen-driven systems, you’re likely to produce fusion, but not enough to ever generate net power.
Dr Hirsch was a key developer of this work going back to Farnsworth’s day, and from what I’ve heard he still has one of the machines on his desk at his office in Virginia.
It sounds like Farnsworth’s original idea had a lot of merit, and even produced fusion reactions. The only problem was ions hitting the screens. So what did you do differently to avoid this problem?
We got rid of the screens! I realized after some time that you couldn’t solve the problem of ions hitting the screens as long as you use screens to provide the electric potential. There’s only one other way you can provide the potential without screens — containing the electrons with a magnetic field.
As it turns out, even though magnetic fields don’t confine neutral plasmas very well, they do confine electrons, because electrons have very little mass. So if you inject electrons into a quasi-spherical magnetic field that goes towards zero at the center and has a big field at the surface, the electrons will come back, be reflected by the field, and continue going back & forth — and they’ll never see a screen.
Now the problem in our design was how to keep electrons from migrating across the fields to hit the magnetic coils that produced the fields and losing power. So you trade the predictable laws of hitting screens, which you can’t solve for a design loss of magnetic fields and coils outside, which you can control by design.
That’s the basis of our patents for Polywell. We use a quasi-spherical magnetic field into which you inject energetic electrons that are trapped inside, go back & forth and create a negative potential well, and you drop the ions in and they never see any screens and circulate until they collide and create fusion events.
So if I’ve got this right, instead of using an electrostatic field to confine the ions, you’re using a magnetic field — and the only way for hydrogen to leave the reaction chamber is to convert into helium. You know, if ions are fed in as fuel & the reactor automatically ejects helium exhaust, this sounds more like a type of fusion engine than a traditional reactor.
Oh, it is. It’s more like a turbojet engine, in the sense that it’s not a device that you load up and ignite. It doesn’t do that. It’s a continuous dynamic through-flow machine where you are injecting electron which makes the well and take the losses. The ions are independent of the electrons. You drop them in, they can circulate and make fusions without the loss.
It’s like a turbojet engine because you inject something in the front end, add fuel in the combustion chamber, then the reaction takes place and the exhaust products go out the back. It’s a continuous through-flow machine, really a type of power amplifier — not an ignition machine.
So the fuel that you’ve chosen for this device is deuterium, right? You’re running a pure D-D fusion reaction in it?
No, no, no. If Polywell works as it’s supposed to and as we seem to have proven it, it will work with any fusionable ions. It will work with deuterium-deuterium (D-D), deuterium-tritium (D-T), d-helium-3 (D-³He), and with hydrogen-boron-11 (pB¹¹)— it just has to be driven at different voltages for these different kinds of fuels.
D-D is the simplest and cheapest because deuterium is available in all the seawater of the world. Believe it or not, every glass of water you drink 1/6000th of every gallon water on Earth is heavy water. You can buy deuterium at any gas welding shop — it’s a perfectly good fuel and generates heavy net fusion power.
However, D-D actually makes a neutron every now & then, so it’s not a non-radioactive fuel. It makes about as many neutrons as you get when you run a pressurized water fission reactor — without any fission products, of course.
In the long run, what you want to do is run them on hydrogen-boron-11 because hydrogen + boron-11 makes three helium atoms and no neutrons. It’s the only aneutronic fusion reaction we know of, and produces three helium atoms — three alpha particles — that escape with high velocity from the center of the machine.
You can put grids outside the machine to electrically bias these alpha particles, slow them down, and use them to make electricity directly without the benefit of turbines. This gives us an interesting long range prospect for clean, aneutronic nuclear power that’s directly converted to electricity at maybe 80% efficiency. That’s the eventual goal.
What kind of power output do you anticipate seeing in the near future from this technology, such as perhaps a successful final prototype?
Well, the next major step in the development program we hope to find support for is about $150 to $200 million, to build a 100 megawatt demonstration plant. That’s not a commercial plant — it’s only a demo plant.
Commercial plants might run at 100 megawatts, but you could scale them up to run as a standard 1,0000 megawatt central station plant. However, smaller sizes seem to be desirable to utilities people because it reduces transmission line losses and makes the grid more distributed.
How does the price tag for this compare watt-for-watt with the big Tokamak fusion projects?
We’ve done studies over the years of what we think our plant costs and what a “big fusion” plant would cost, and we seem to conclude that the actual cost of electric power would drop about a factor of two from the thermal systems that produce steam and run turbines. It might even more if we ran it on pB¹¹ with direct electric conversion.
The problem with pB¹¹ direct conversion is that you have to redesign the plants, because you don’t have any turbines and a lot of the infrastructure that conventional plants have. If you use D-D, can use some of the existing infrastructure because it’s a steam turbine driven system and you’ll have to put the reactors in pits just like for fission. That’s an easy retrofit, though.
You can put a D-D power system down next to an existing steam plant, cut into the steam lines, and shut off the either the oil boilers or the fission reactors to run it on fusion. That immediately saves you money, because the fuel is a very important part of the cost of electricity.
This current project has been underway since 1986, right? How much progress were you able to make in that period of time?
We had 20 years of small scale research in our laboratories with the $21 million of government money has been invested in it, and over 200,000 man hours by technical staff working on this project.
During that time, we’ve tested 15 different prototypes of these machines, and we‘ve managed to define, understand and solve all 19 of the critical physics issues that we found — and we solved them one at a time.
It’s been a very long & tedious process because our funding was always very small. The Navy, which supported us, couldn’t ever put the right amount of money into it. If they had, it would have raised eyebrows on Capitol Hill and generated complaints from the DOE about the allocation.
In any case, they said “we can only fund you at a small level — can you do anything with that?” Well, we did. It took us 20 years, but we did finally succeed in solving the last basic physics problem during our final test. All the physics is done, we were ready to do engineering development.
The Navy discontinued funding for your project, and I’m wondering what their reasoning was after 21 years of support?
The reason our funding died was because of the Iraq war. The war budgets have been just consuming everything in sight in Washington — and when it came time for the annual budget process, the total Navy R&D budget was cut by 26%.
So the Navy didn’t just cut us — there were cuts all across the board. One of the things that was cut out was a thing called the Navy energy program, and we were a victim of that.
We managed to find some friends in the Office of Naval Research who kept us alive for about nine months, but that was it, and at end there was no more money.
After they pulled funding, I understand that you kept running tests up until the very moment the power was shut off and that you had some promising results that emerged from later analysis on those final tasks.
Yeah, it wasn’t the power of being shut off. We were looking at our budgets and we had to pay leases on our lab space and we had to commit to yearly leases and we couldn’t do that with the budget monies that were left.
We had a plan to close down by the 1st of November, and we’d just started closing the lab and getting rid of the equipment when we ran some tests on a big machine called WB5, which illuminated us on some things we should’ve seen for 10 years and didn’t.
We realized we had missed a critical point in the old problem of electron losses and quickly designed & built a final machine called WB6 our solution to the problem.
By then we were approaching 1 November 1st, and we still hadn’t tested WB6 in heavy fusion conditions. But I told the team, “we have to do this, we have to finish this”, so we kept right on working past the time to shut the lab down.
On November 9th & 10th we ran the machine for four times, and finally produced fusion from D-D at a rate 100,000 or more times higher than had ever been done by Hirsch and Farnsworth at the same voltages. We realized that we’d solved the electron loss problem. Finally, at last it was solved.
If you had the solution, what made you stop when you did? What ultimately led you to shut down the lab?
On November 11th, we tried another test run, but the hasty construction caused a short in the magnet coils and the thing arced and blew — it didn’t blow up, but arced & melted down. We didn’t have time or money to rebuild it, so on Monday the 14th we started shutting the lab down.
In six weeks we took the whole lab to zero and it all disappeared. We didn’t even know our final results for a month because we didn’t have time to reduce the data until December.
When we finally reduced the data, we looked at it and said, “Oh My Lord, look what we’ve done! It’s actually worked….the last piece is there & the puzzle is solved.” That’s it. Quite ironic.
About Our Guest
Dr. Robert W. Bussard was an pioneering American physicist specializing in nuclear fusion energy research. He was the recipient of the Schreiber-Spence Achievement Award for STAIF-2004, a fellow of the International Academy of Astronautics, and held a Ph.D. from Princeton University.
In the early 1970s Bussard served as Assistant Director under Director Robert Hirsch in the Controlled Thermonuclear Reaction Division of the Atomic Energy Commission. They founded the mainline fusion program for the United States: the Tokamak.
In addition to his contributions to the US fusion energy program, Bussard was also a pioneer in the field of aerospace nuclear propulsion in the NEPA program, and later influenced the Project Rover & NERVA nuclear-thermal rocket programs.
Bussard is also widely known for conceiving of a novel space drive for interstellar travel, named the Bussard Ramjet in his honor. It has been popularized in science fiction by authors such as Poul Anderson, Larry Niven, Vernor Vinge, Jerry Pournelle, as well as being mentioned by Carl Sagan.
Dr. Robert W. Bussard passed on October 6, 2007 at age 79. His legacy continues on at the EMC2 Fusion Development Corporation, online at: http://www.emc2fusion.org/