Monday, November 27, 2006

Easy Low Cost No Radiation Fusion

Justin at Classical Values has put up a posts about fusion energy machines way different from the magnetic confinement and heating machines the government is building.

You can read the post here. Eric of Classical values has another post on the subject.

For more details on the physics visit EMC2 Fusion. You can also make a donation there to help the work go forward.

An interesting question is: when was the first steady state (operation times of at least 10s of seconds) electrically operated nuclear fusion machine which produces at least 10s of millions of fusions a second built? The astounding answer? 1959. So far 18 experimenters have produced similar machines including this young experimenter.

The next question is: why have advances been so slow in since then? The answer (and a lot more) is given in this video by Robert Bussard. (note: dial up is going to be incredibally slow as the video is around 1 hour and forty minutes - aproximately 170 mega-bytes) The video tends to the technical and I will have to study it a few times to get all the details. However a fair understanding of high school physics should suffice. Even if you don't understand the physics the general concepts are easy to understand and Dr. Bussard's enthusiasm is infectious.

In any case the idea is to build a fusion device that produces no long lived nuclear radiation and that works with the forces of nature instead of against them. The voltage required to make these devices work is on the order of 10 to 20 thousand volts or less. About the same voltage as you would find in a tube type monitor or TV set. Nothing very exotic. For a full scale power producer it is predicted that you would need about 2 million volts. Well within the range of current technology for small scale devices. Currently the highest voltage used in electrical transmission is 1.15 million volts. Scaling that up to two million volts for production devices should not be too difficult.

Near the end of the lecture (about 1 hour in)Dr.Bussard gets to the heart of the matter by listing the advantages of this type of power plant.

Stop Greenhouse Effect

Eliminate Acid Rain Sources

Decrease Thermal Pollution Sources

Stop Nuclear Waste Production

Destroy Nuclear Waste Inventory

End Water Shortages Forever

Cheap Fuel Free Electric Power

Clean Low Cost System

Fresh Water From The Sea

Practical Space Flight

Global Economic Stability

Cheap, Clean Thermal/Electric Power Readily Available

Fixed Energy Prices Stabilize Economy

Low Value Cane In Third World Countries Becomes High Value Export Product

Third World Nations Can Become Economically Viable

Profitable Industrialization Possible

Destroys World Market For Gasoline

Eliminates Effect Of Oil Cartels

Oil States Suffer Drastic Income Losses
(audience: laughter - ed.)

Desalinization Plants Allow Irrigation Of Arid Lands

Cheap Water Allows Effective Agriculture

Low Cost Power Stabilizes Industrial Nations

Oil Wars Vanish

Mid-East Stabilized by Economics

Third World Becomes Fiscially Responsible
(comment: not likely, more energy does not fix bad government - ed.)

End Use Market Price Ca. $5,000 B In Year 2000 $
(all products the machine can replace - ed.)

Sell/Lease Systems To Supply Energy Plants/Production

Royalty/Lease Fees at 2% of Market Price Equivalent To Ca. 2m/kWhr Surcharge Yields Net Income (Profit) at Ca. $100 B/Year
(which means an estimated electrical cost of 100 mills/kWhr - ed.)

Dr. Bussard says he needs $200 million dollars and five years to build two full scale demo plants. The first year of his five year plan will replicate with improvements his last experiments to get data on the process that can be verified by a review comittee. The First year will cost $2 million dollars.

He says that a computer to do proper simulations on the system would cost $8 million dollars.

Wiki on Dr. Bussard:

In the early 1970s Dr. Bussard became Assistant Director under Director Robert Hirsch at the Controlled Thermonuclear Reaction Division of what was then known as the Atomic Energy Commission. They founded the mainline fusion program for the United States: the Tokamak.
George Miley at the University of Illinois is doing some work in the field. As is Gerald L. Kulcinski at the University of Wisconsin. Here is the U. Wisconsin IEC Fusion page.

A review of the lecture.

Dr. Bussard Talks

An executive summary of Dr. Bussard's Google talk.

The Bussard Reactor for space propulsion.

A number of links to Dr. Bussard's work. Scroll down.

More good links including links to the Farnsworth patents.

Update: 15 Dec'06 0431z

Mark Duncan in the comments left a link that refers to the Bussard paper given in Valencia, Spain [pdf].

A transcription of the Google presentation [pdf] with illustrations.

Mark has more at Fusion.

Here is a follow up article on the engineering: Reactor Scaling

Hendrik J. Monkhorst did some interesting work on a linear (as opposed to the Bussard spherical design) reactor. Here are a couple of articles one from Science 278 and another one from The University of Florida. Another Monkhorst paper: Science 281. Here is the patent for the Monkhorst/Rostoker design.

Wiki has a nice discussion of the reactions and some techinical details of the various Nuclear Fusion schemes including Dr. Bussard's Boron 11 - Hydrogen reaction.

Update: 11 May 007 0202z

Dr Bussards contract with the Navy has been extended for a year without funding.

Please write your Government and ask them to fund the contract:

House of Representatives
The Senate
The President

and sign this on line petition and send it to your friends to get Dr. Bussard's work funded.

Update: 30 Aug 007 0032z

The US Navy has funded the next phase of Polywell research. This is no reason to let up. The Navy plans a five year program to construct a 100 MW test reactor. With more money they could speed up development. With enough cash a three year time line ought not be difficult. Two years is an outside possibility if we really pour it on.

Update: 20 Sept 007 1012z

If you want to get more into the design details of the Polywell Reactor you might want to try:

IEC Fusion Newsgroup

Details on the design of an open source fusion test reactor.

IEC Fusion Technology blog

Update: 29 Dec 2007 2112z

I should have posted this here months ago. It is a link rich overview of Dr. B's life. He died in early October 2007. The work goes on with Dr. Nebel and Dr. Park of Los Alamos National Laboratories leading the effort:

Dr. Bussard has died.


Here is a report on what is going on at the lab.

Bussard Fusion Update

Update: 19 June 008 0739z

Here are some recent additions you might find useful.

Starting A Fusion Program In Your Home Town

The World's Simplest Fusion Reactor Revisited

Fusion Report 13 June 008

Rick Nebel Updates The Latest News (Dec 2008)


M. Simon said...

Here is the Mark Duncan link:


I have been looking for that paper.


M. Simon said...

Here is the paper:

Bussard on clean fusion

M. Simon said...


Yes the 3 alphas produced are ionizing radiation. However, the shielding necessary is is not very significant. A piece of paper will do the job.

The Bussard design is ingenious in that the alphas are collected by a grid charged to 2 million volts. Thus the reactor looks to the outside world like a 2 million volt battery. Very smart. At 2 million volts a 100 MW plant need only deliver 50 Amps, in the high power vacum tube range - known technology. Since the collector grid is at 2 million volts the alphas are basically at thermal speeds when they are collected. Once collected the alphas are neutralized and are then exhausted to prevent them from filling the reaction chamber.

The cross section problem is solved by the fact that the reaction volume is in the center of the reactor and all the particles involved are focused in that area. The beauty of electrostatic confinement.

Yes, there are engineering problems. However, they are well understood. It will be a challenge. However, a lot less challenging than the design of a nuclear fission (uranium) reactor.

BTW compared to the millions of electron volts the reaction produces, the separation of water into H2 and O2 only requires a few volts. Not very significant in terms of reactor output.

The acceleration is accomplished by the design. No external or auxiliary accelerator is required. Turn on the magnets, raise the voltages, feed in reactants, and energy is produced.

There is some radiation from secondary reactions such as the deuterium - tritium reaction. This could be reduced if deuterium and tritium were extracted from the hydrogen fuel. Not too difficult for the small quantities required. And, the shielding would be much less than the requirements for a uranium fission reactor in any case. A 100 MW (electrical) reactor would require about 1/2 pound of Boron 11 per day and about 1/10th that amount of hydrogen.

Look at the movie two or three times to get all the details. It is really a brilliant design.

Also have a look at the papers referenced in the post. They flesh out some of the details that are unclear in the movie.

Anonymous said...

I am repeating a question posed to M. Simon on another blog:

If the technology is sound, why wouldn't venture capitalists and/or charitable foundations such as Google's not funding this? The amount needed is a drop in the bucket compared to the $ chasing other types of technologies.

I can point to many sources of capital that would be interested, starting with the Google Foundation, Kleiner Perkins, and Sequoia Partners.

Have these scientists approached any of these folks? If not, why not?

--Lumpy Gravy

Anonymous said...

Lumpy Gravy,

If the technology is sound, why wouldn't venture capitalists and/or charitable foundations such as Google's not funding this?

The simple answer is that venture capitalists seem have a policy that nothing shall be financed the first time. If the concept has been demonstrated and working facilities exist they will gladly jump on board. But they absolutely do not want to be the first one into the boat.

Wramblin' Wreck

Roga said...

Some answers:

"The issue really is going to be charge transfer. These 3 helium nuclei are going to want 2 electrons aweful fast. You will probably need a to boil off some electrons around some collector plates (like old vacuum tube controls).

Yes, electron boiling off. As the poster after you said, you're firing the nuclei into a matched-bias grid, which means that they'll be almost perfectly "cooled" relative to the electrons in the grid when they hit, so they will readily accept those electrons. Which is how the current is generated: you basically create an electron "vacuum" as the nuclei pull electrons out of the biased grids, which is then replaced by the circuit's ground.

"The other issue will be the center of momentum collision cross sections. You are going to need the protons going fast to get similar momentum. This means some sort of accelerator system. You will need to pump energy into the water to get the H2 to separate from the O. Then you need to ionize the H2, so you have 2p+. Then you need to accelerate and focus these so that the cross-section for fusion-fission is high enough.

The acceleration is done by the electron cloud in the center of the magnets. Basically, you can confine lots of electrons in a magnetic mirror much easier than protons because the restoring force is equal to the charge. Since electrons weigh 1/1000th what protons weight, for any given velocity they are deflected 1000 times more easily by a magnetic field.

The protons "see" the electrons as a negative charge and RUSH towards it. Now the best part - since you can very finely control the voltage on the electrons (this is the same technology that makes your TV work), you can make sure that, when those protons hit the middle, they're all going almost exactly the same speed.

"All of that takes energy budget. You have to pour energy in. Ok, then you get the fusion reaction. How are you extracting the energy out? Most of this energy is going to be released as kinetic energy, and possibly some gammas, and recombination xrays. You are going to need to absorb the EM stuff, and re-emit it as sound/heat in the material. The alphas will carry away lots of KE, but harvesting this may be hard. You might need to work on this in a water bath to get conduction.

Nope, EM radiation is only thrown off here as side reactions. The vast majority is locked up in the alpha particles. These are harvested by the voltage-matched grids. To understand how this works, imagine that the alphas are balls being thrown up against the side of a building. Say you build a balcony on the building at the very crest of the balls' flight paths, so it will come to rest exactly on that ledge, and all the kinetic energy in it will be stored until you need to use it. That's the way the biased collection grids work. Since we know the range of kinetic energies for the alpha particles (i.e., how hard the balls are "thrown"), we can figure out the voltage that will just exactly catch them (the height of the balcony), without losing barely any energy.

And just like the ball on the balcony, we've stored the kinetic energy in another form (electrical energy). Well, actually, the last part isn't quite accurate. What actually happens at the grids is that the protons rip electrons away from the metal in the grids. The only way to replace these lost electrons is to suck them from another part of the metal, so a current forms. And since the grids are at a huge positive bias, what the protons are essentially doing is forcing electrons to climb a huge positive "mountain" to replace their lost cousins.

This works any time you hit metal with protons. The think is, in order for the circuit to do work, it has to be at a positive bias. And in order to hit something that is positively biased with a proton, you have to shoot it really hard. The harder you shoot it, the more bias you can put on the grid, and the more work each electron can do.

"Basically I am trying to say even if it could work, there are quite a few technological hurdles to overcome.

True, and I'm not convinced that all the physics hurdles are worked out either - particularly, I suspect there will be some fundamental roadblocks with current arcs and steady state ash removal. And some of the engineering challenges are doozies too - supplying the superconducting current and its coolant without interupting any current arcs, getting rid of all the voltage sinks in such a high-bias machine, and most of all feeding the beast with an even enough stream of gas to keep it going, but not to choke the reaction with too many ions.

M. Simon said...


Most excellent!

Anonymous said...

These comments have been invaluable to me as is this whole site. I thank you for your comment.

Anonymous said...

I wonder if the fusion simulation would distribute well enough for a Folding@Home kinda project...instead of an $8 million computer, a bunch of volunteers donating spare cycles... I'd certainly sign up!

M. Simon said...

Anon asks:

I wonder if the fusion simulation would distribute well enough for a Folding@Home kinda project...instead of an $8 million computer, a bunch of volunteers donating spare cycles... I'd certainly sign up!

It should distribute well enough.


1. compute a small volume for the various functions.
2. when all the volumes are done, distribute the results, update the volumes.
3. rinse repeat.

In principle the code should be able to be run distributed. As to if it is written that way or optimized? Unknown. I'm speaking specifically of Dr. Bussard's code.

As to actually doing it. We are not that organized.

However, if you are up to the task feel free to volunteer:

IEC Fusion Newsgroup
IEC Fusion Technology blog

Anonymous said...

the BOINC project ( provides free software for volunteer computing, so if the simulation software exists, it shouldn't be too hard letting the world contribute...

Anonymous said...

I wonder if there are any trademark registration issues involved with the advent of clean nuclear fusion?

brian wang said...

In regards to the $8 million computer for simulation.

there are several new computers this year and early in 2008.

Clearspeed has 12 teraflops for $840,000 and 1 teraflop for $70,000

AMD has 500 gigaflops for $1900 and both AMD and Nvidia should have dual precision teraflop GPGPU in early 2008 for $1500-2500.

M. Simon said...



AST said...

I thought this was going to be a paean to solar power. I suspect that we'd have an easier time launching a series of solar collectors into orbit around the sun in orbits variously angled to the Ecliptic with the ability to store energy and send it to earth in the form of microwaves.

Containing plasma with magnetic fields reminds me of the time my son stuff our cat into the sleeve of his sweatshirt so that he could comb tangles out of her fur. I suspect that the one real way to contain such a reaction is through gravity, but maybe in the next 14 billion years, nature will provide us with a better one than the sun.

M. Simon said...

The Bussard reactor does not confine the plasma with magnetic fields. The magnetic fields serve to contain electrons and shield the grid.

Confinement in Inertial Electrostatic Confinement devices is done by electrostatic fields which in many ways are similar to gravity (the force is perpendicular to the field) but millions of times stronger.

You really can't take a snapshot view of the technology and understand what is happening. It is simple but deep.

Anonymous said...

'Containing plasma with magnetic fields reminds me of the time my son stuff our cat into the sleeve of his sweatshirt so that he could comb tangles out of her fur. I suspect that the one real way to contain such a reaction is through gravity, but maybe in the next 14 billion years, nature will provide us with a better one than the sun.'

Tokomak design contains plasma using magnetic fields.

Rhapsodyinglue said...

"Third World Becomes Fiscially Responsible (comment: not likely, more energy does not fix bad government - ed.)"

Though it may sound overly optimistic, this fits with a widely held view in social and political sciences that is often referred to as the curse of natural resources. The idea being that countries that suddenly find themselves sitting on top of huge resource wealth without first having had the luxury of developing rule of law and stable societies often end up with despot governments and/or civil wars.

Many would argue that countries in the Middle East and Africa need to be weaned off easy petro dollars just as much as we need to wean ourselves off their fuel. If the rulers of these countries didn't have huge influxes of money from outside they would logically have to figure out how to enable stable internal economies through educating and empowering their people to create wealth rather than simply digging it from the ground.

Stable economies are ones that depend upon the productivity and contributions of a large educated middle class. Petro dollars from overseas create governments that aren't accountable to their own people.

M. Simon said...


Electricity has a somewhat different effect than oil.

1. Labor saving machines empower women.

2. Night lighting improves safety.

3. Computers bring education.

4. Industry is empowered.

And yes. Bad government is the prime cause. When the US goes marching in to spring an unannounced election on a country it would help if the Americans could improve electrical supplies quickly. The current 3 year lead time for a plant is too long.

Neil said...


I have been struggling, as a non-physicist, to get a complete picture of the Polywell design. Your descriptions and discussions have helped a lot. To help myself further, I have put together a summary that tries to fill in some holes in my understanding.

I offer it to you for your consideration and, if you're in a generous mood, comments.


1. The Polywell reactor is contained in a large vacuum filled chamber. I'll call this the outer chamber.

2. The walls of the chamber are conductive.

3. A roughly polyhedronic (e.g. cubic) space at the center of the chamber is delineated by flat electromagnetic coils, each forming a face of the space. I'll call the space within the magnets the reaction space. It's on the order of a couple of meters across.

4. Think of a doughnut for each magnet's shape. The magnet conductor loops form the body of the doughnut.

5. The magnets are somewhat squished to conform with the shape of the reaction space. For a cube, they are squished into a squarish shape.

6. The magnets come close to, but don't touch each other.

7. The magnet conductors are superconducting. Naturally they are very well insulated thermally.

8. The conductors carry extremely high DC currents, forming very powerful magnetic fields. It follows that the structure of the magnet assemblies must be very strong.

9. The same magnetic pole of each magnet is pointed to the center of the reaction space.

10. The magnets are equal strength, so the fields at the center of the reaction space cancel out.

11. Magnetic fluxes flow in closed loops, passing through the holes of the doughnut toward the center, bend back, and pass between the magnetics.

12. Where the fluxes pass between the magnets they squeeze each other, forming "cusps".

13. Enclosing the magnet assembly is a metal box or grid. I'll call it the cage.

QUESTION. Is it a solid box? A grid?

14. The cage is at an extremely high positive potential, a couple million volts. Let's call it (P).

QUESTION. Is it about 2 MV?

15. The magnets are each encased in a conductive housing.

16. The magnetic housings are at a high positive voltage, tens of thousands of volts.

17. The housings conform with the shape of the magnetic fluxes.

18. The outer chamber wall is at ground potential.

19. The cage is far enough from the magnets that its potential doesn't significantly affect the potentials of the magnet assemblies. It is also far enough from the chamber walls that there is no chance of arcing.


1. There is a powerful electron gun (one per magnet) that injects electrons toward the hole of each magnet.

QUESTION. Where is the electron gun? Just outside the magnet?

2. As the electrons leave the guns they are attracted to the magnets due to the high positive potential of the magnet housings.

3. The magnetic fluxes divert the electrons through the magnet holes toward the center of the reaction space.

4. The electrons tend to follow the fluxes, curving back through the cusps and out of the reaction space, where they loop back through the magnets again.

5. Because of the constriction of fluxes in the cusp, electrons have a relatively hard time getting out of the reaction space. Because of this, and because the net magnetic field is zero at the center, electrons eventually accumulate at the center of the reaction space in a swirling cloud.

QUESTION. Do the electrons just bounce around in the center or do they follow a (significant) geometric path?

6. As electrons accumulate at the center, they tend to distort the magnetic fields there in a way that constricts the cusps further. To the electrons the boundaries of the reaction space look rather like the inside of a wiffleball. As they fly around inside it they usually bounce off the inside of the wiffleball, but sometimes they pop through a hole (cusp). The wiffleball-like boundary is called a magrid.

7. Also, as they accumulate, the electrons form one side of a powerful electrostatic field located between the center of the reaction space and the very positive magnet casings. The cloud serves as a cathode. The casings serve as anodes. The field
between cathode and anode is called the potential well.

8. A few electrons do leak through the cusps. That is in fact a good thing because it reduces the tendency of the electrons to turn into heat.

QUESTION. Why don't electrons that escape the cusp just fly to the cage, where they are absorbed? Is it too far, even relative to the much less positive magnet casing? Do the fluxes hold on to the electrons? Or do they in fact complete the circuit for power plant output?

QUESTION. I assume that the cage is too far from the magnets to affect the potential well significantly, even at its extremely high potential. True?


1. Boron and hydrogen (B, H) atoms are released into the reaction space (just inside the magwell boundary).

QUESTION. Are they injected with some kind of a gun, or are they just spritzed?

QUESTION. Where are they injected - between adjacent magnets? At one place or several?

2. As they enter the reaction space, the B & H are immediately ionized by the magnetic fluxes.

4. The B & H ions are accelerated by the well potentials toward the center of the reaction space. They reach a very high speed (in fact so high that their mass and momentum are increased relativistically).

5. The B & H usually pass through the center (one end of the well) and head toward the opposite side of the magrid.

6. They are slowed and stopped by the positive potential of the other side of the well. They then fall back to the center.

7. The ions fly back down the well thousands of time, until by chance a hydrogen ion (proton) hits a Boron ion squarely.

QUESTION: what happens when two B ions or two H ions collide? Is either case as common as a B/H collision?

8. The two particles fuse. Ultimately forming three He ions.

9. The He ions have a couple MeV's of energy. Its energy is a little less than the cage potential, so call it (P-).

QUESTION. Is it about 2MeV?

10. The He ions have so much energy they fly right out of the magrid.

QUESTION. What keeps the He ions from hitting the magnets?

QUESTION. The electrons, and B and H ions inside the magrid don't interact with the He ions significantly?

QUESTION. There isn't so much friction at the center that it eventually clogs up with deionized B & H and impurities bounced off various structures by the He ions?

11. As the He ions fly toward the cage, they lose almost all their energy to the cage fighting the cage's extremely high positive potential (P). The cage absorbs this energy. In fact it's this energy that allows the cage to maintain its very high
positive potential when there's a load between it and ground.

12. Since the cage potential (sustained by the energy absorbed from the He ion) is at (P) and the He ion is at (P-), it is still attracted to the cage. The He still has a little (i.e. thermal scale) energy as it hits the cage.

13. The He ions absorb electrons from the cage.

QUESTION. How does this work? Do the ions just grab electrons from the surface of the cage and drift away (since they are now uncharged particles)?

14. Because the He ions absorb electrons, there is electron flow. Note that for a 2 MV DC circuit, the current for a 100 MW powerplant would be (just) 50 A.

QUESTION. Does it therefore follow that enough B/H must be fused to carry 50 A of current? Is that synonymous with the power output rating for the plant?

QUESTION. Does (P) represent a point of equilibrium for the fusion energies, the geometry of the magrid, and the load? If so, does that mean there are only a few possible plant sizes?

Is the equilibrium point naturally stable? How narrow is the range of stable operation for a given size?

QUESTION. Is it going to be tricky lighting these plants up and turning them off for maintenance?

15. The He is evacuated from the chamber. Somehow. Whiskbroom?

M. Simon said...


You have things correct. As to the questions. Way more than I can comfortably handle in one sitting. In addition there are still a lot of unknowns - places where no answers are currently possible.

May I suggest visiting:

IEC Fusion Technology blog

There are links on the sidebar to active discussion groups Working Groups. Join and ask your questions there (just a few per post).

Also be sure to read:

Don't be a luser. Read this first.

To get a feel for proper etiquette.

nukemhill said...

Thanks for the comment in my blog, M. Simon. And thanks for the link to this article. I've added it to my list of links. Lots of info to absorb!