Key Polywell Questions

Discuss how polywell fusion works; share theoretical questions and answers.

Moderators: tonybarry, MSimon

Post Reply
mattman
Posts: 459
Joined: Tue May 27, 2008 11:14 pm

Key Polywell Questions

Post by mattman »

Pictures are a very easy way to explain things. Recently, I explained the debate over the feasibility of the polywell using a series of pictures. I would be interested in any feedback.


1. Plasma structure, or no structure?

Does the plasma have structure? This includes a “virtual anode” a “fourteen point star” an “edge annealing effect” or any number of other structures. Opponents argue no, proponents argue yes. Any cloud structure changes the radiation rates, ion density, ion focusing, recirculation, ect, ect, ect… This directly affects performance. This is an unsettled argument and is a critical question.

Image


2. Two temperatures or one bell curve?

Do the ions and electrons have different temperatures or not? Two temperatures allow for optimization. The ideal polywell has lots of cold electrons and a few short lived hot ions. This lowers radiation, raise fusion rates, improve structure, improve ion focusing, ect…. The answer is not obvious. Rider argued the plasma is “maxwellian” or “thermalized” and wrote a heat transfer paper to support it. Proponents point to “ambipolar” or “nonthermal” plasmas and papers from magnetic mirrors in support. Where these particles are located matters as well. We need data.

Image



3. Internal magnetic field, or not?

Does the cloud have its own magnetic field opposing the ring field? Opponents argue no, the plasma is “magnetized” and therefore containment is poor. Proponents argue yes. The internal field shrinks the losses at the cusps and improves confinement. This "Whiffle ball" changes the radiation losses, efficiency and power balance. I doubt anyone has a firm answer to this.

Image



4. Ion focus or not?

Can those ions hit at high speed in the center? Mid-90’s work by Wisconsin, Nevins and rider discussed ions cannot be “focused” or reach the “critical density” needed for high fusion rates. Alex Klien left Bussard’s team because of ion injection problems. Not a settled argument.

Image
Last edited by mattman on Tue Sep 17, 2013 5:35 pm, edited 1 time in total.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Re: Key Polywell Questions

Post by D Tibbets »

Point 1:
The structure is obvious in multiple photos. It is not quite that simple though, the structure does not directly imply performance. It is not a question of the geometry baseline, but the change with the claimed Wiffleball effect. The edge annealing effect, ion focusing, and consequences for the ion energy spread (both speed and direction) are experimental questions for which there is little open data. But annealing at least with ininitial radial monoenergetic temperature ion population is simple physics and it would require denial of well accepted physics to argue against it. The real question is how much this effect retards the thermal spread and angular momentum gain of the ions over their lifetimes. Is it modest or major, and how much is needed. As for angular momentum (loss of central focus) Nebel has stated that this is not a show stopper. The presence of a potential well with non square shape and a central anode/ complex shape has been shown in numerous electrostatic machines (Fusors). Obviously a good central focus has numerous advantages, and would allow for the smallest machines, but only if engineering issues are not the size limiter, which may be the case.

Point 2:
There is not temperature separation (I think) if you are looking at the average temperature in the machine.The electron and ion peak temperatures are similar. What is important is the spacial distribution of the ion and electron temperatures. The ion potential well results in maximum ion energy in the central core and minimal (near zero- ties in with annealing) on the edge. The electrons potential well results in maximal energy near the edge and minimal energy near the central core. This has profound consequences for time dependent densities of the two species at various radii and interactions like Bremmsstruhlung radiation. Again, the amount diffusion/ relaxation of this opposing initial distribution into a completely or partial thermal distribution over the lifetime of the respective ions and electrons is the question. Also, how much of this can be tolerated. D-D fusion is much more tolerant of this than P-B11 fusion. Even Rider admitted that D-D fusion breakeven was possible with his completely thermalized assumptions.

Point 3:
I don't think there is any real doubt that moving charged particles will create their own magnetic field and that if the charged particle motions are random (thermalized in both energy and direction) or primarily radial (in and out in equal proportions) that this will exclude the electromagnet B fields is also straight forward. How much this effect will push against the electromagnet B fields is more problematical. ie: how much Wiffleball effect there is., or if Beta can approach 1.

Tokamaks create a plasma magnetic field which opposes the electromagnets also, but in the Tokamak there is a preferred direction around the torus. So the plasma (the bulk) is not magnetized with movement along electromagnet field lines except in border regions. But the plasma has it's own magnetic field structure due to the preferred movement (thermalized, random due to collisions locally, but the mass of plasma is rotating around the torus in one direction and thus is magnetized- mostly flowing along internal plasma induced B field lines (collisions makes things very messy, but of course collisions are essential for fusion).

The B field geometry at the border is also key, as it is a major problem for macrostability in Tokamaks, while it actually may be a benefit in Polywells.

The Wiffleball is not a cusp plugging process. I see this expressed often and I shudder as it is wrong (based on my understanding).The Wiffleball effect has nothing to do with squeezing the cusps closed. = the central cusp regions. What it does with inflation of the plasma "balloon", is to greatly increase the volume to surface area ratio. This results in the cusp surface area which is constant under given kinetic energy and B field variables, but the surface area away from the cusps increases dramatically in proportion. The balloon still has the same number and sized holes, but the balloon is much larger, the overall surface area and contained volume is increased, thus it takes longer for the balloon to deflate.

The real question, I think, is not the simple physics, but the magnitude and complexities and interactions between density, temperature, B fields, etc that tremendously complicates the issue.

What you might ask happens when Beta=1 is exceeded. The balloon is still inflating, but now the effective cusp size surface area is also increasing , so the total surface area to cusp surface area is noti ncreasing and soon would be decreasing , Think of the opposing Funnel analogy. Also, the plasma may be pushing the electromagnet B field border to the surface of the magnets, which completely demolishes the magnetic confinement.

An important consequence of the Wiffleball is not only the confinement efficiency increase but also the density. The results work out the same. The Wiffleball size is determined by the product of the volume and density in relatin to the partical energy and electromagnet (Magrid) B field strength. This means you can increase the density in a fixed magrid size/ plasma volume without an increase in cusp losses, again because the contained plasma quantity is increased without an equivalent increase in losses- that would have to be replaced with new high energy particles. This means you can generate useful amounts of fusion energy in relative small packages. Even if you could exceed break even in a house sized reactor, it would not be worth much if you could only produce a few Watts of fusion power.

Even a Fusor could theoretically break even if you scaled it up to a size comparable to the Moon. It scales (simple assumptions) as D^4 r^3 / r^2 just like a Polywell. The problem is that the density doesn't change thus the scaling reduces to r^3/ r^2. With the tremendous deficiency that has to be made up , it takes thousands of KM before the scaling catches up. This would make a Tokamak look like a pocket reactor in comparison. The Wiffleball effect claimed for the Polywell allows for the density to also increase within limits (Beta=1) so the D^4 scaling is significant, and the only claimed downside is that the B field loss scaling has to be considered, but this is only B^0.25 so it is easily tolerated. The Wiffleball effect not only improves confinement efficiency relative to machine diameter (surface area- density relationships), thereby reducing the baseline loss numbers by a factor of several hundred to several thousands (depending on whether you are comparing it to a simple biconic cusp machine or a simple Fusor), but also increasing the fusion power production tremendously due to the ~ 1000 fold increase in density claimed. Thus, a useful machine that is profitable at house sizes instead of Moon sizes. With input energies measured in megawatts opposed to Tera watts. Examples : 200 MW out for 10 MW in, instead of 100.0005 TW out for 100.0000 TW in. Or in the case of the Tokamak, perhaps 10 GW out for 2 GW in.

Point 4: Central focus is an attractive feature both for fusion energy production intensity, but also for Bremsstruhlung issues. It effects final machine size and associated loss scaling issues. I think some is essential for profitable P-B11 fusion, though it is again a matter of proportion, not absolutes. As Nebel said, D-D fusion may be profitable even without any effective ion central focus.
In WB6, with the MFP of the average ion KE being multiples of the machine radius, it is easy to argue that thermalization is prevented due to the lifetime of the ions before escape or fusion especially with annealing figured in. In larger machines and especially with increased density the thermalization issues are more challenging. Over the the lifetime of the ions the annealing may not be able to keep up. If the MFP is much smaller than the machine radius, the edge annealing can still occur to a degree but it is trivial if the ions are thermalizing in each pass. A favorable effect of the increased density, though, is that the ion lifetimes would decrease proportionately as they are consumed by fusion, and this would be even more significant if some degree of central focus can be maintained. It is a complex multifunctional problem that is dependent on many assumptions. Only experimental results can test the validity/ accuracy of the assumptions. Note that modeling assumptions/ methods results were different in Dolan's (?) versus Rider's work.

[EDIT] Note that I mentioned D^4 scaling. That shoud be B^4 scaling. The density increses as the square of the B field strength, and the fusion rate scales as the square of the density.

Also, the author that challenged Rider's work was Nevins, not Dolan.

Dan Tibbets
Last edited by D Tibbets on Mon Sep 30, 2013 4:59 am, edited 1 time in total.
To error is human... and I'm very human.

mattman
Posts: 459
Joined: Tue May 27, 2008 11:14 pm

Re: Key Polywell Questions

Post by mattman »

Dan,

You added lots of details, fine points and caveats, to basically the same universal questions.

1. If a structure has already been photographed, can this be published?

2. We know that charged motion messes with magnetic fields. Hence, it is likely, that the plasma pinches the cusps. But, we still need to measure this; so we cannot be sure.


I reworded to better describe the shrinking of the cusps. And I forgot to mention the spatial distribution.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Re: Key Polywell Questions

Post by D Tibbets »

The structure has been photographed by EMC2. And on this forum I have photographs depicting some of the structure in a truncated Cube Polywell design, except permanent magnets were used and this resulted in extra cusps going into the surface of the magnets. Still the point, cusps, the corner cusps and to a limited extent the "funny" (region between the corner cusps)cusps are visualized.


The plasma pressure pushes outward everywhere, but in the central cusp the B field lines are essentially parallel to the plasma, so I don't think there would be any pinching. If anything there may be an apparently insignificant widening of the central cusp. With the concvex B field lines present away from the cusps (relative to the center) everywhere else the fields are always pushed outward, thus the inflation of the Wiffleball. This also relates to plasma instabilities (pinches) in Toklamak type B fields. The fields are not always convex towards the center,s o the plasma pressure can become entrapped in small bulges that form conditions where growth of the defect becomes energetically favorable- leading to macro instabilities (a pinch). Even in the Polywell cusps, the B field lines remain convex till the very center of the cusp where the lines approach parallel orientation towards the center. Past this point things reverse relative to the center, but this is past the point of magnetic confinement so is insignificant (unless we start considering fine details about recirculation and electron injection). But considering the density differences inside and outside the radius of the magrids, any Macro instabilities are presumably insignificant compared to what is going on inside. And even here the B fields are convex towards the local plasma.

Dan Tibbets
To error is human... and I'm very human.

Schneibster
Posts: 1805
Joined: Mon Jul 09, 2007 5:21 am
Location: Monterey, CA, USA

Re: Key Polywell Questions

Post by Schneibster »

This style of representation makes sense to me. I have been thinking about polywell in these terms lately and in terms of trying to close the geometry so that energy is not lost from the core to the outside, allowing an ongoing reaction, and I have a Wild Ass Guess to offer:

additional coils oriented at a 45 degree angle to all three convergent sides should be added to each of the vertices of the cube defined by the six polywell coils. These coils would help

1. increase the chaos of the magnetic field at the center of the focus of the rings, and
2. hold the charged interacting particles in, and the charged confining particles providing the field charge out.

I don't insist on this; it's my developed mathematical-to-physical intuition that the magnetic fields curl like that. I might be wrong. But I think I'm not. I think also this is what Bussard saw that made him say Tokamaks won't work. I can see the curls dissipating the energy, no matter how hot you try to make it in the ring. You must have a focus, not a ring. The curl requires it. A ring will always lose energy; a focus will not, there's nowhere for it to bleed to.

And I might be wrong; the additional rings might not be needed, just a better magnetic conductor in the six standard rings. But I think the materials in a focus with the additional eight small rings might not be as critical. It might be more tractable.

Also, the effects of scale make me feel that larger models would not require these corner rings but might not only benefit in efficiency from them, once the complexities of field can be overcome, but be usable in fractal configurations of progressively smaller rings, a lacework of rings of precise size, at first glance random but possessing order dictated by the interfering magnetic fields. I expect the minimum size and which multiples are additively resonant would be determined as a factor of the gyromagnetic radius. Thus there will be a desired temperature which will determine the radius by determining the average velocity. This will vary with absolute size as well, as velocity will determine radius. Tuning this will determine the optimum size. And the optimum size will determine the optimum output. One of course hopes that these optimae don't converge on a G-class sun, but it seems fairly obvious that in fact they do. Clearly we will not be igniting any suns on the surface of the Earth so we must be prepared to tolerate some inefficiency; of course, that too is limited lest it overheat the Earth.

I repeat this is all intuition from my attempts to visualize the magnetic field. But I also repeat this visualization is what Bussard did that was interesting. I think these kinds of visualizations divide the insightful from the pedestrian.
We need a directorate of science, and we need it to be voted on only by scientists. You don't get to vote on reality. Get over it. Elected officials that deny the findings of the Science Directorate are subject to immediate impeachment for incompetence.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Re: Key Polywell Questions

Post by D Tibbets »

Adding additional magnets complicates things, and I'm not certain what effects they may have. In essence the Polywell is a simple design and it's elegance is partially due to this. Others have added external magnets to make the cusps into twisting convoluted shapes, but no analysis of the effects are available, except what is mentioned in the patent application where some design failures are pointed out.

That WB5, which presumably had better electron confinement, was a dismal failure pointed out that not only was electron confinement critical but so was maintenance of a potential well without local peripheral electron collections in the cusps that would attract ions near their outer edge of their normal travel. There is a compromise between the efficiency of the electron magnetic confinement and the ion electrostatic confinement. The Polywell is advertized as the best compromise that almost works. Still, with the advantages of the Wiffleball/Beta=1 operational conditions, it is not good enough. The recirculation gains introduced with WB6 seems to have been the final eureka moment. This ~ 10 X improvement in the electron input costs was critical. Note that also, the significance of ExB drift of the electrons was also recognized as a significant issue, thus the separation of the magnets, which hurts the cusp magnetic confinement, but again, this seems to be the best compromise.

Why Bussard considered Tokamaks to be failures was based on ion magnetic confinement, not electrons.

ExB drif or diffusion is fundamental problem for magnetic confinement. In Tokamaks and most other efforts, the plasma is neutral. This means there is no electrostatic confinement of any species. Both electrons and ions undergo ExB drift across a magnetic field. But the gyroradius of ions are ~ 60-120 times that of the electrons , so each collision of the ions results in gyroradii jumps ~60-120 (or 180 for tritium) times greater. The collision frequency is the same so the ions will transport across the fields that much faster. To limit this you must decrease the density (bad because it also decreases the fusion squared) in order to slow the collisions that leads to ExB drift. Alternatively you can increase the B field strength to reduce the gyro radii jumping/ collision length, and/ or increase the distance that the ion needs to traverse to reach the magnet surfaces. Both solutions drives the machine to greater sizes. As I said, the electrons do this but with ~ 1/60th the inertia/ momentum, the gryo radii are that much smaller, so each collision results in smaller ExB jumps. With the same density, temperature and B field strength, it takes the electron about 60 times as much time to traverse the field. Containment is 60 times or more greater. This is why Bussard said magnetic fields were fine for containing electrons.

With a neutral plasma with replacement of lost plasma also neutral, the ion loss will result in left behind electrons, this builds up a space charge which quickly leads to a balancing of charge, ie electron losses will match the ExB limited ion losses. I think that momentum plays a role in this balancing. The ions tug along an electron close by ~ 60 times better than the reverse. This is mentioned in why the Polywell potential assumes the shape it does. If the situation is changed though, the results are much different. If excess electrons are injected, a negative space charge contains the ions so that ion magnetic containment becomes almost insignificant. Ions are not exposed to ExB drift issues. The electrons are, but so long as they are replaced rapidly enough the ion confining potential well is maintained. In fact under
Polywell density, temperature, and B field conditions, the electron losses through ExB drift are trivial compared to the electron cusp losses. In the patent application ExB energy losses (electron losses) were only ~ 1 to 10% of the cusp losses. I don't know if this ratio applied before or after recirculation was factored in.

Thus cusp electron losses dominates everything (at least with D-D fusion). In a Tokamak, if the macro instabilities are ignored, there is little if any cusp losses and ExB losses (ion ExB losses)dominates the loss picture.

Granted that the Polywell trades cusp losses for ExB losses and does so with confinement times perhaps 10,000-100,000 times worse (for the electrons), but it does so at densities of ~ 1000 times greater (due to the Wiffleball density enhancing effects). And because fusion scales as the density squared) the fusion rate per unit of volume is ~ 1,000,000 times greater, and the final triple product result favors the Polywell (provided it works as advertized). Add to this the central focus effects, if present; and the larger percentage of plasma participating in the fusion process (thermalization issues and a higher average temperature) results in a speculative tremendous gain for the Polywell relative to Tokamaks.

Dan Tibbets
To error is human... and I'm very human.

Schneibster
Posts: 1805
Joined: Mon Jul 09, 2007 5:21 am
Location: Monterey, CA, USA

Re: Key Polywell Questions

Post by Schneibster »

Thanks, Dan. I'll think about that a while.

I find math is less useful when dealing with magnetics. Visualization is everything but it's not "like" anything except perhaps the way a swimmer feels the water. I'm looking for usable intuition; you can't design if you can't see it in your mind. I have designed a lot of switchers, and it's as important to understand how the fields affect the input and output coils, and link them, as how the semiconductors switch. The semiconductors make it function; but the coils and caps do the work. The semiconductors just have to do their work efficiently enough that they don't burn.
We need a directorate of science, and we need it to be voted on only by scientists. You don't get to vote on reality. Get over it. Elected officials that deny the findings of the Science Directorate are subject to immediate impeachment for incompetence.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Re: Key Polywell Questions

Post by D Tibbets »

Schneibster, I notice in a previous post that you mentioned maintaining fusion by better retaining the energy in the machine. While interpretation is uncertain, it seems that you are thinking in terms of ignition. This is a goal of some fusion concepts but defiantly not the Polywell. The Polywell is not dependent on ignition which depends on the temperature being maintained by the heating from fusion products. The opposite is true. The fusion products quickly escape- whether neutrons or high energy ions like protons, tritium, Helium 3 or 4, without significantly heating the plasma. This is a result of the cusp architecture , the charged fusion particles escape before they can thermalize with the rest of the plasma. This is due to the proportionately much greater MFP of these high energy particles relative to the fuel plasma. There is some debate about how much thermalizing of the fuel plasma occurs within the average lifetime of the particles, but with KE's of perhaps 10-100 times greater, the comparable MFP is hundreds to thousands of times greater. The fusion products simply leave the system before they can transfer much energy to other particles. And with proper design, they leave primarily through the cusps- not hitting the magnets. Nebel pointed this out. This not only implies that ignition is not possible, or even desirable in this system; but also, that these fusion products lends themselves to direct conversion schemes outside of the reaction space. This is not the case for ignition machines like Tokamaks, the intent and consequences are far different.

The Polywell has been described as a simple power amplifier. Input energy is applied (mostly in the form of high energy electrons) and fusion power comes out. If the input power can be controlled well enough, then you are in business. Electrostatic acceleration of electrons and secondarily ions is a simple process, much different than the oven like heating and subsequent self heating of ignition machines. The relative long term confinement times in Tokamaks has advantages, but also disadvantages.

Consider some numbers. In a Polywell a fusion ion might have a confinement time of perhaps 1,000 to 10,000 passes (based on the Wiffleball trapping factor- the electrostatic potential well is too feeble to stop the fusion ions). The fuel ions may complete upwards of a million passes due to the potential well electrostatic confinement (provided they do not fuse first). The fusion ions are traveling perhaps 10 times as fast as the fuel ions, so confinement time (number of passes / passes per second) results in confinement times perhaps 10,000 times less than the fuel ions. Combine this with the increased MFP which is perhaps ~ 10,000 times greater results in ~1/ 100,000,000th the number of possible energy exchanging collisions relative to a Tokamak at the same density. Even the claimed 1000 fold increased density claimed for the Polywell results in only about 1 percent of the heating from the fusion products relative to Tokamaks. It is insignificant and of course hopeless if ignition was required for the Polywell. Fortunately, it is not needed nor desired.

Dan Tibbets
To error is human... and I'm very human.

Schneibster
Posts: 1805
Joined: Mon Jul 09, 2007 5:21 am
Location: Monterey, CA, USA

Re: Key Polywell Questions

Post by Schneibster »

D Tibbets wrote:Schneibster, I notice in a previous post that you mentioned ...

>polywell explanation with details<

Dan Tibbets
Thanks Dan! Great stuff. I'll read this over and try to absorb it and modify my expectations accordingly. You're just the kind of person I was looking for. I'll have questions, I'm sure. Image
We need a directorate of science, and we need it to be voted on only by scientists. You don't get to vote on reality. Get over it. Elected officials that deny the findings of the Science Directorate are subject to immediate impeachment for incompetence.

Post Reply