Cusp Ion Plug

[BenTC]

Both ions and electrons can be confined by a magnetic field everywhere parallel to the surface in a torus. In a sphere you have to have cusps somewhere. If you try to plug the cusps with electic fields, you have to decide whether to stop the ions or the electrons. You can't stop both.

Thanks Art. That just congealed a few ideas floating around from previous discussions. Let me check...

1. As I now understand it, the magrid magnetic field contains the electrons, except at the cusps, where they jet out, but are recirculated due to the positive magrid. So no need to plug electrons at the cusp.

2. The quasineutrality of the electron jet out the cusps would drag ions with it, except (and I'm speculating here since I haven't seen mentioned) the positive charge on the magrid strips the ions from the electron jet. The neutrality at the cusps would be due to the combination of electron jet and positive charge on the magrid. To maintain neutrality the magrid charge would need to be finely balanaced against cusp jet size, although perhaps the cusp being overall positive rather neutral might be okay.

3. So in the Polywell at the cusp, the ions are being plugged by the positive charge on the magrid.

4. The subsequent (very speculative) image that forms is that all the ions trying to follow the electron jet out the cusp that are stripped gather around the cusp. Ion density grows at the cusp and its actually here that fusion occurs.

5. Actually I guess the ions stripped from the electron jet at the cusp now see the virtual negative electrode again and accelerate back towards the centre.

[chrismb]

I am not sure I really follow the 'logic' of many mechanisms aspoused on this forum. Seems to me that as soon as someone wants ions to go one way they simply have to say "Ah, the electrons go that way so the ions will follow", or vice versa, as if these two components, ions and electrons, behave like cool buddies that like to hang out and so you only have to deal with one or t'other and the rest will follow.

No, they don't. Well, they do hang out together but more like hen-pecked moody husband and nagging wife kinda sticking together and arguing together [neither of whom you can have a rational conversation when they are both riled up], so you have to deal with them as a whole rather than like friends who can be cajoled into going off in the same way. In other words (as the analogies are just getting more and more vauge to deal with concepts that are getting more and more vauge), if electrons and ions are couped up in the same space they ARE A NEUTRAL PLASMA and DC electric fields will be screened by them over a given distance (the Deybe length).

What that means is that this NEUTRAL PLASMA will then flow out of the cusps like toothpaste out of a tube - JUST AS YOU CAN SEE IN THAT BLOOMIN' "TEST PLASMA" PICTURE.

The Debye length is particularly short, especially for the ion densities being idealised for a working Polywell.

Now, if you have some bizarre contructed situation in which the ion 'flow' is in some way co-incident with the electron 'flow' then you can use that effect for useful purposes. For example, to cool down an ion stream (as a moving ensemble, i.e. make them more mono-energetic) then you can introduce a flow of similar velocity electrons and they will exchange their energies. BUT what seems to be beyond any sensible discourse is the ideas here where the electrons are somehow a thermalised mass of cool particles whilst the ions flow with particular directions. There is nothing known in plasma science, to my knowledge, that can make head-nor-tail sense out of this and appears to have been the underlying problem that the examiner had with the Bussard patent, as I read his comments.

[chrismb]

Both ions and electrons can be confined by a magnetic field everywhere parallel to the surface in a torus. In a sphere you have to have cusps somewhere. If you try to plug the cusps with electic fields, you have to decide whether to stop the ions or the electrons. You can't stop both.

Just to cover this, I would be a bit more cautious over my caveats with "If you try to plug the cusps with STATIC electic fields, you have to decide whether to stop the ions or the electrons."

I say that because what the heck those magnetic fields will be doing in those cusps when a plasma tries to exit there is anyone's guess, and maybe some resonance can be tuned to re-stabilise a plasma there, like an anode layer ion source or rotating wall Penning trap with the ion stream pointing back in and with the small electron current loss as you get in such devices being plausibly tolerable.

[Art Carlson]

Usually I have to repeat myself a lot here, but I'm not sure I ever said this before:

I have been arguing that the cusps have to be quasi-neutral, so that they will contain an equal number of ions and electrons. If there is a potential difference between the cusp throat and the wall, the electrons will be reflected/contained, and the ions will be lost. One might ask, then, How can the plasma on the one hand be quasi-neutral, and the ions and electrons on the other hand still go their separate ways? I believe the separation occurs in the Debye sheath at the wall. There is a beam of neutral plasma extending almost to the wall. Only in the last lambda_D (0.1 mm at these conditions?) is the bulk of the voltage difference dropped, reflecting the electrons and slamming the ions into the wall with a vengeance.

[chrismb]

I hope I am not out of place to suggest that it sounds like you are getting close to re-inventing well-established plasma-wall-interaction physics. (I've heard MHD physcisits doning this occasionally, whilst industrial-low-temp-plasma type physicists just look on!!)

(For the electron temperatures and densities presumed for Polywell, I rather think the Debye length will be of the order of nm, but I can't be bothered to run that calculation.)

Initially, as a plasma contacts a solid, electrons are dominantly lost until the plasma becomes positive. In the case of a Polywell plasma where it may already be being held at magrid potential, that step will, simply, probably be side-lined as a 'step'.

After that, it is much as you say; electrons are mostly reflected back into the quasineutral plasma at the sheath edge (excpeting for a few highly energetic ones). A sheath exists between the plasma and the wall and ions enter this region at the Bohm velocity (having been accelerated through a 'pre-sheath' quasi-neutral region). Ions go through pretty much the whole pd in that thin sheath region and are lost/neutralised.

The question, then, is; what is the maximum e-field gradient in the sheath that can tolerate that rate of neutrals being formed by the ions hitting the wall (some will embed, some will sputter the wall and contaminate the plasma) and stil maintain a non-discharge sheath structure? As far as I can see, it can't and it will simply be a gas discharge under most conditions (exactly as appears to be depicted in "that photograph")

I would expect any industrial plasma person to take one look at this setup and simply say "er... why not float the wall?!" because no net current is drawn, the ions impact the wall with lower energy (less sputtering) and lower flux, and the total pd across the sheath is a function of the electron temperature which, supposedly ( ) is low.

The plasma between the magrid and the presheath should, I think, be quite neutral (in the normal parlance). The presheath is certainly quasineutral and begins to accelerate the ions out to the wall.

If you try to create magnetic structures to slow this ion-sheath transit flux down, you can succeed [see all the prior attempts at plugging mirror machines] but only that you will slow ion flux down and not eliminate it. Electron loss is not a concern. In fact, I may be giving succor to a generally poor setup: The electrons will be cooled by this wall interaction and will reduce overall electron temperatures in the plasma in the rest of that cusp region. Forget about around-and-about electron recirculation, it is a fantasy. But electron population and low temp can be maintained by managing the wall interaction.

[Art Carlson]

@chrismb: It's a pleasure to meet someone here who understands something about sheaths. My area of expertise is plasma-wall interaction generally and Langmuir probes specifically. It's just that I seem to spend so much of my time here explaining things like quasi-neutrality over and over again, that I never get to interesting questions like how the potential structure around the cusps looks (inside the plasma ball, too, not just out to the wall).

[KitemanSA]

I keep hearing about neutral plasmas, in the reactor and flowing out the cusps. But Dr. B. and others have stated the plasma is only QUASI-neutral. They have contended that there is sufficient excess negative charge in the reactor to keep the ions in. They have also stated that the polywell, as corrected by WB6 works. The data presented, limited though it be, supports their contention, they say. Dr. N. has stated the the Polywell is NOT ambi-polar, and he stated it at a time when Art was trying to demonstrate that neutral plasma had to flow out the cusps. This suggests to me that he was trying to correct Art, and that neutral plasma does NOT flow out the cusps.

Until you folk can generate data to negate their contention, I figure you are just flapping your lips.

Have fun at it.

[chrismb]

It might well be the case that what they observed was quasi-neutrality at various points. My question; for how long did they observe it?

The formation of a persistent sheath in equilibrium is not instantaneous (as I alluded to) and, with a finger casually cocked in the air, on a timescale similar to the pulses as I understand them to have occurred.

Could Polywell generate quasi-neutral flows and, even, fusion for a ms with a MJ pulse going through it? Sure, why not, no reason to doubt that. Is that the same thing as saying it worked as planned?

[D Tibbets]

It might well be the case that what they observed was quasi-neutrality at various points. My question; for how long did they observe it?

The formation of a persistent sheath in equilibrium is not instantaneous (as I alluded to) and, with a finger casually cocked in the air, on a timescale similar to the pulses as I understand them to have occurred.

Could Polywell generate quasi-neutral flows and, even, fusion for a ms with a MJ pulse going through it? Sure, why not, no reason to doubt that. Is that the same thing as saying it worked as planned?

I guess that depends on your attitude towards the time frame that seperates pulsed or steady state conditions. The WB6, and WB7(?) results are based on transient conditions when beta passed through 1. I understand this time frame was ~0.25 milliseconds in WB6. I don't know if WB7 was able to extend this time. According to Bussard this was long enough to approximate steady state conditions, as this was ~ 1000 times as long as the relavent physical processes. A 0.25 ms time frame would represent ~ 5 thousand electron passes, and ~ 500 (?) ion passes in WB6. You might dissagree that this is long enough, but I can say with confidence that I have no idea. I'm accepting Bussards contention. Certainly many processes require only a few (>5) passes/cycles to approach steady state conditions. I suppose the speed of sound in the plasma may effect the formation a debye sheeths to. Since the plasma is quite hot, I assume the speed is quite fast.
If you are challenging predicted thermalization times, I suspect 0.25 ms is not long enough to aquire diffinative data. Such challenges should be met with longer run times - fraction of a second to several seconds (?) that seem to be one of the targets of the new Navy contracts.

Again, I'll admit that debye sheeths confuse me Taken in isolation it is easy to follow the arguments, but when the debye sheeths are concidered as a string of adjacent regions, each effecting the makeup of the next region the interplay becomes murky. I assume that the statement that ions will reverse electrons in the cusps, while the ions continue on is based on the much larger inertia of the ions within these isolated domains. But, does this take into account the net negative charge towards the center of the machine which is trying to impead the ions (whether taken as individuals or as clusters within sequential debye lengths) while pushing the cusp electrons outward? This would reverse once the magrids were reached and the positive potential on the grid is exposed to the charged particals. And, all of this assumes that the ions and electrons are entering this region in equal numbers, with a neutral plasma being left behind. Both situations are in question.

I'm not sure of this argument, but if there is significant inertia exchange between ions and electrons within debye sheeth lengths of each other in those local conditions outside the magrid, wouldn't escaping electrons be slowed more and tend to be more suseptable to recirculation, while the ions would not pick up as much energy from the positive magrid? If this effect is real, is it significant?

And, finally, assumeing the ions do not reach the cusp nearly as often as the electrons, due to the electrostatic containment. The electron numbers might be thousands of times higher than the ions, so inertia exchange would be different. It would seem that this degree of excess electrons would pull out more ions from below the cusps, but I'll invoke Gauss's law as a counter (due to the multiple cusps surrounding the WB acting as a hollow sphere- perhaps a leaky sphere with large areas between the cusps, but still with a significant effect). Debye arguments may be a counter- counter argument. I don't know. Also, concider that 9/10 electrons will be recirculating in the cusp, so only 10% of the net electron flows would have net outward tugging on the ions when thy passed close enough to an ion that the local effect was dominate (Debye length?). Does debye sheeth conciderations allow for ths apparenty hugh realative excess of electrons to exist without arching? Am I completely confused ?

Dan Tibbets

[D Tibbets]

Could Polywell generate quasi-neutral flows and, even, fusion for a ms with a MJ pulse going through it? Sure, why not, no reason to doubt that. Is that the same thing as saying it worked as planned?

I have reason to doubt it. Based on multiple aspects of my understanding of the system. The ions in the cusp regions are at the top of their potential wells, or even above it (upscattered ions) so thay are moving slowly. Also. the dominate motion of these ions would be nearly parellel with each other. Very highly upscattered ions may be bouncing off of field lines close to the center of the cusp and hit each other head on, but I would expect this to be extreamly rare. If significantly thermalized, a very fast upscattered ion my catch up and hit a slower ion, but again I suspect this would be vary rare, based partially on the assumption that the repeated upscattering collisions nessisary to do this are impeaded by the annealing process.If there are fast neutrals flying around, fusion could occur, but neutrals of nessesity have to be rare beasts (arcing issues). The ions that have picked up some energy outside the magrid from the positive.potential (I'm confused about how much energy the ions will pick up in this manner because the magrid positive potential is only ~20% greater than the negative potential well. I suppose debye lengths enter into it again ). These ions may hit the vacuum vessel wall and fuse with embedded target ions. Even if the ion population within the WB do not have any confluence, the increased density and speeds inside will favor fusion by factors of millions, billions, trillions...(?).
[EDIT] I should add that at the end of the test the breakdown of confinement may produce chaotic conditions, but I think the time dependant neutron production in WB6 correlated to the time when beta was passing through 1.

Dan Tibbets

[alexjrgreen]

There is a beam of neutral plasma extending almost to the wall.

You might want to take the Faraday cage between the device and the wall into account.

[D Tibbets]

There is a beam of neutral plasma extending almost to the wall.

You might want to take the Faraday cage between the device and the wall into account.

WB6 had a Faraday cage seperate from the vacuum vessel wall, which may have been used primarily to protect sensers placed in the vacuum chamber. If the WB7 setup has a seperate faraday cage, it is not visible in the picture. The grounded vacuum vessel wall should serve as a faraday cage.

Dan Tibbets

[Art Carlson]

I keep hearing about neutral plasmas, in the reactor and flowing out the cusps. But Dr. B. and others have stated the plasma is only QUASI-neutral. They have contended that there is sufficient excess negative charge in the reactor to keep the ions in.

I may occasionally use the term neutral, but quasi-neutral is always what I mean. The idea is that you might have all kinds of potentials, which, of course, are always associated with an imbalance of charge density. But any time you ask how big the ion density is, it's OK to use the same number as you do for the electron density. There is no contradiction between Bussard/Nebel and me on this point.

They have also stated that the polywell, as corrected by WB6 works. The data presented, limited though it be, supports their contention, they say.

In my book, data is at least one number. I have never seen a number published for the loss rate of ions or energy for any polywell machine under any conditions.

Dr. N. has stated the the Polywell is NOT ambi-polar, and he stated it at a time when Art was trying to demonstrate that neutral plasma had to flow out the cusps. This suggests to me that he was trying to correct Art, and that neutral plasma does NOT flow out the cusps.

If a polywell is driven in a way that is not ambi-polar, then the losses will not be ambi-polar. I am sorry if I have not always made this clear.

I must admit I don't always find it clear how Bussard/Nebel thought/think the polywell should be run. It seems instead of perfect electron confinement they require (or at least think they require) a loss rate on the order of the electron thermalization time. They can presumably set the electron loss rate by adjusting the magrid potential. I'm not sure how the thermalization time compares to the cusp loss rate (without recycling = electrostatic confinement). If it's faster, then they have an (even bigger) problem.

At this point I was going to start talking about ion confinement, but that might be moot. The first question, which I'm not sure has been answered yet, is how energy efficient the external recycling of the electrons has to be in order to make a reactor, assuming you cannot live with thermalized electrons.

[alexjrgreen]

In my book, data is at least one number.

Surely two? One would be a datum

[alexjrgreen]

It seems instead of perfect electron confinement they require (or at least think they require) a loss rate on the order of the electron thermalization time.

Does this help?

Collisional Equilibration