Magnetically confined plasma fusion was abandoned decades ago. Has their time come?
Magnetic fusion concepts require coils to form the field. One path that is still being proposed is to embed the coils into the fusion-grade plasma because confinement can be nearly steady.
We discuss 3 proposals that hold some promise of being game changers.
- Polywell proposal from EMC2
- CFR proposal from Lockheed-Martin
- Dynomak proposal from the University of Washington’s graduate school in fusion energy research.
These are small companies formed during the 35 years — since American magnetic fusion funding was decreased to a strangulation level.
They are usually near universities, to pursue ideas that are not in the tokamak mainstream of thinking.
Make no mistake, these companies want to be the change agents in the fusion community. They all believe the tokamak is too big, much too expensive and way too brute-force to become economically viable.
Before we proceed with the proposed concepts, review Fig 1, the dictionary of what we are talking about.
EMC2 Jaeyoung Park, CEO
Polywell cusp fusion
EMC2 uses Polywell confinement and images show its WB-8 test device with a coil on every face of a cube, 6 coils in all. A dodecahedron version would have 12 coils and magnetic well with steeper walls.
They propose a 6-coil 3 dimensional cusp design, electron beams are fired into the zero magnetic field central region; the electrons form a deep negative charged region which pulls fuel ions from the edges into the central core.
The highly respected Robert Bussard (1928-2007) formed Energy/Matter Conversion Corporation (EMC2) in 1985 using Defense Department support. Due to this funding source, information is sparse, though they apparently worked on space propulsion methods (fusion rocket engines).
The company lost military support in the early 2000s but regained it later. Due of the nature of its funding, the company website is almost nonexistant, but changes every so often; they prefer to use Facebook. Here is access to a 2014 report.
If all done right, Bussard was certain that a dense high pressure (“high beta”) plasma would form, generating its own magnetic field. With very high field strengths (3 to 5 Tesla) this geometry would act like a Wiffleball and almost completely hold the plasma. Since neutrons from a D-T fuel would endanger the superconducting coils near the plasma, they propose the difficult D-3He fusion. Dr. Park became CEO in 2011.
Charles Chase, team leader,
Compact Fusion Reactor CFR
Lockheed released information from their secret “skunk works” facility of a new design for a very compact fusion reactor.
They opened the facility to the public in October, 2014 and of Aviation Week and Space Technology publicized it loudly. Very little physics information is available on CFR, but pictures showed that it was related to a magnetic cusp configuration. (see, for example, Plasma Physics by Francis Chen)
Fig 5 shows the coil set in the current T4 test stand. Most pictures show 4 coils but this particular one shows that they are of different sizes, which is a significant observation.
Here is the LastTechAge hypothesis: This is a combination of a stable 2 ring quadrupole, with possible magnetic well depth enhancement by the placement of an outer cusp coil.
Fig 6 is from a 2009 publication that describes this geometry, probably for some non-fusion application. We have added a yellow partial ring to the published drawing to emphasize the orientation of the 3 rings in the paper. (a) shows that the two smaller rings are the multipole coils while the large coil on the mid-plane has counter flowing current and is the cusp coil. All three rings are horizontal and the midline is vertical.
Compare this with Fig 7, the image of plasma in the T4 stand. The central axis is horizontal whereas the publication has a vertical axis. The plasma shape on the right (which outlines the field lines) could be either from either side of the smaller quadrupole coil, Fig 6, field drawing (c). This is not a CUSP shape, nor a quadrupole even one operated in a Helmholtz configuration. Notice the plasma wrapping up and around the top of the coil, where the supports are…
So I am betting that the Lockheed CFR is a Quadrupole-Cusp hybrid. This would have confinement potential because of the deep magnetic well in the center (the X point in Fig6).
Lockheed wants to build their CFR with superconducting magnets to replace the T4 coils. The CFR structure would be about 3× larger than the T4 device.
University of Washington
This is a different abandoned magnetic confinement scheme. There are no coils inside the plasma.
Prof. Jarboe leads 3 separate groups. One is a graduate school experimental test of a new way to make a spheromak work better, the others also work toward the development of plasma fusion.
He focuses on the spheromak class of devices, which have strong fusion potential. I classify them as on the “main line” of abandoned paths toward fusion success.
Spheromaks and the related spherical torus (ST) machines were not abandoned because of bad results, but because U.S. support was made a target for political mockery by the ideologues of the 1980s.
That was when fusion had just turned a technology corner and could actually build something that worked. But – fusion leadership were forced to decide on which child to save from sacrifice – mirrors, spheromaks, tokamaks or … (Tell me, Mrs. Jones, do you want to slit the throat of your daughter Sandra, or cut the heart out of your son, Jon? You choose, which would you like?)
Spheromaks have been attractive since the discovery that its plasma containment time rose and fell as the plasma temperature rose and fell. The hotter it got, the better it became.
FIg 9 is from the Italian Frascati fusion laboratory aboratory. These are techniques based on Compact Torus designs. The key to more than a few microseconds of lifetime is a magnetic field that twists about the donut body.
Spheromaks are similar to main-line tokamaks but without massive coil windings that thread from the outside of donut through the central donut hole and back (producing the huge toroidal field stretching the long way about the donut). A tokamak’s coils need shielding from the D-T reaction’s neutron flux. Spheromaks have a self-generated natural twist (“helicity”) to their fields and do not need external help. The device can be much smaller than a tokamak, maybe even not use superconducting coils for any configuration — much cheaper. The downside, perhaps, is that you must stand on the outside and inject helicity into the plasma.
Helicity injection – that is what Prof. Jarboe’s Dynamak (his HIT program) is designed to do.
Fig 10 is the injector part of his Dynomak. (The rest is a big empty tank!) His students inject into the chamber a toroidal plasma with currents that spiral about its donut-like body.
This is a government funded research activity. So… why is the Dynomak in our Paths Not Taken posts? There are several reasons – because it is a diversion from the tokamak line of study and because the DOE is loudly preparing to cut more facilities.
- The U.S. is focused on tokamak research, for no reason other than it was the most advanced method in the late 1980s. This is where our ever-diminishing support goes.
The U.S. appears about to chop all the unnecessary parts of fusion research from its care. Meaning, Obama’s administration is focusing on completing the task started by merry men under Reagan and Bush-I’s authority, continued by Clinton’s minions, and ignored by Bush-II (after all, there are no possible war profits, so why worry).
So, though it is a crucial technology for our future well being, Jarboe thinks that his research program is close the the “So long, see ya” moment. This is why he hit the news media this year. And he does need support, so do many other university programs. So do the NSTX-U effort at PPPL and the DIII-D tokamak studies at General Atomics. (It really looks too late for MIT’s Alcator-Cmod)
Put it together
The downside for both the EMC2’s Polywell and Lockheed’s CFR designs is the embedded coil structure. Cusps and multipoles have been tested for over 50 years, as suggested in Fig 1. Cusps do not contain plasmas well, they provide magnetic flow paths to the outside; and both types would be embedded in the fusion medium. Internal ring geometry was abandoned after it proved out as a viable containment geometry, because
(A) Coils must survive plasma at at least 150 M degrees and very high pressure (high beta). This is more than 100 times the temperature of the sun. Material in the fusion plasma flow path would not survive.
(B) The easy fuel is to D-T. This is what is proposed for a fusion/fission hybrid reactor to cook fission waste into usable fuel for a second tier of reactors. Fission fuel is self poisoned pretty early but can be rejuvenated by the heavy neutron flux from D-T interactions. Survival of unshielded superconductors in such a neutron bath would be problematical. Even D-3He generates a neutron flux. Can embedded superconductors work in that environment? (The idea needs experimental testing, at least.)
Did EMC2 and Lockheed invent a SciFi Impervium shield that will hold against a welding torch that is 100 to 500 times hotter than the sun? How about a neutron repulsor? Or maybe they invented a PR method to attract new funding sources? These are sharp physicists; maybe I am missing something? Maybe. But even if so, their 5 year to reactor pronouncements generate echos of that “On line by ’79” thing at old KMS Fusion.
Charles J. Armentrout, Ann Arbor
2014 Dec 29
Listed under Technology… a post in the Fusion-Other thread
Have a comment? Click on the title of this post, go to bottom, let us know.
Related posts: Click the INDEX button under the Banner picture
“Did EMC2 and Lockheed invent a SciFi Impervium shield that will hold against a welding torch that is 100 to 500 times hotter than the sun? ”
I see no mention in your post that the polywell structure presumes a very high DC potential on the coils and supporting structure in the region of high electron density. The flow path of the electrons tends to avoid the coils and supports, and are concentrated strongly in the central region of the apparatus. The ions are attracted to those electrons.
Yes, it does. And your are right, my sentence about the deep electron well also talked about the high voltages on the rings — somehow it got edited away. Thanks for pointing this out.
This situation is a very similar to an argument used by the the MFTF-B design team. The huge potential wells designed for the end cells would hold the plasma in the center cell, no matter where where it “wanted” to go. If you treat a fusion grade plasma as a collection of single particles, this looks good. MFTF-B was turned off before it was turned on (if you read my report on it) so we will never know. And tokamaks were mirror devices with their ends wrapped about to meet, so no losses at all were expected.
Cusp fields have high magnetic field paths leading from the null field point right out of the machine. Mirror trapped ions will pitch angle scatter into the loss cone, no matter what. These paths are between all the coils, so the coil fields will not slow them down. Collision scattering is much higher frequency event than collision fusion, so guess which must win? Those that do not scatter out, will loop back around the cusp coils and hit the supports.
High beta plasmas are a different kind of beast, not a collection of single charged particles. External fields are shielded out and this strange fluid goes where it pushes toward. I worked on an “octupole” a machine with very stable confinement, but the supports could not be electrostatically shielded from the plasma flow. And our plasma was pretty cold. I think that when the polywell is pushed to the fusion temperatures with plasma pressures > 10 atmospheres of more, we will see reality strike.
That said, the proof will be in the final results on all the devices discussed. EMC2 has good PR and they attract wonderful support. So I think they will have the chance to show me wrong. The final “good device” will just work. It will function independent of all the analysis to the contrary, and function well, not particularly because initial naive pictures said it a walk in the park. The race is definitely on, may the best concept win.
“Collision scattering is much higher frequency event than collision fusion, so guess which must win? Those that do not scatter out, will loop back around the cusp coils and hit the supports.”
The whole point of the high DC potential is they do not tend hit the supports or coils, and especially not the coils. The estimated number of passes an electron makes before striking a surface is over 100,000 passes. Obviously I have no inside information, or I’d be under an NDA.