Proposed focus on space propulsion research - why and how?

[ankovacs]

Much of the motivation for polywell fusion research seems to stem from Dr Bussard's emphasis on developing the technology into powerplants first, and into space propulsion later. He specifically referred to 'saving the world', so that polywell fusion reactors would avert an upcoming energy crisis.

I think that this order of emphasis is mistaken. We have already the means for abundant and nearly inexhaustible energy supply - in the form of nuclear fission plants fueled by U238 or Thorium. If an eventual energy crisis erupts, it will be just for the consequence of decision-makers' lack of foresight, not the lack of existing solutions.

Regarding space travel however, existing propulsion systems are woefully inadequate for manned missions in the solar system, and for even satellite missions to other stars. This is where polywell fusion can bring a quantum leap - being the best prospect for surpassing current propulsion limits by several orders of magnitude.

In fact once one thinks beyond traditional idea of the spacecraft engine being inside the rocket's body, and thinks of an externally mounted large yet lightweight design, many of the 'polywell fusion challenges' seem much easier to be worked out. You can read here my draft paper describing how the engine can be constructed from wire segments, with proper voltage levels and currents keeping the wireframe shape and generating electro-magnetic fields:
http://www.broadbit.net/download/polywe ... rch_18.pdf
(first two pages introduce polywell concept, you may want to read from section 2)

If p-B fusion can achieve a good ratio of fusion energy / bremsstrahlung radiation, which is still a big if to be verified, then we can have reason to be optimistic about a practical implementation.
Scaling turns out to require 20m+ polywell grid diameter? No problem, it's still not that much weight with the wireframe design concept. The vacuum chamber is large enough
Voltage levels of 500kV+ are required? No problem, there are no outer walls, so we have no arcing issue. Just need to keep spacecraft far enough from grid.
Is cooling difficult in space? No problem, as we are deflecting momentum of fusion products instead of capturing it. Just need to make sure that the wireframe does not absorb much radiation energy.
Alas, such experiments are hard to do in a ground lab, so much theoretical work is required...

What do you think about retargeting polywell fusion research towards such objective? The article I linked above is just a concept introduction, obviously without scientific details. I am not sure where would be a good place to send it for publication, suggestions are appreciated!

[MSimon]

You will need outer walls. Most of the fuel traversing the system does not react. You will want to recycle it.

[Art Carlson]

There is so much one could say here, beyond the things I've been saying for months already. The most obvious is that your coils won't make a polywell. The degree of "optimism" is also breathtaking. Take the comment "Scaling turns out to require 20m+ polywell grid diameter? No problem, it's still not that much weight with the wireframe design concept." If you want a decent power density, you need a high plasma pressure (and a thousand times higher still if you want to use p-B11). The plasma pressure gets transmitted to the grid through the magnetic fields, which means a flimsy "wireframe" isn't going to hold together.

Run some numbers, and then gracefully withdraw your proposal.

[D Tibbets]

... If you want a decent power density, you need a high plasma pressure (and a thousand times higher still if you want to use p-B11)...

Question- why a thousand fold increase in pressure for P-B11? At appropiate drive voltages (eg. 75,000 KeV 'resonance peak(?)') the graphs I have seen show cross sections close to those of D-D. What am I missing?

Dan Tibbets

[Art Carlson]

... If you want a decent power density, you need a high plasma pressure (and a thousand times higher still if you want to use p-B11)...

Question- why a thousand fold increase in pressure for P-B11? At appropiate drive voltages (eg. 75,000 KeV 'resonance peak(?)') the graphs I have seen show cross sections close to those of D-D. What am I missing?

Dan Tibbets

http://en.wikipedia.org/wiki/Aneutronic ... challenges has some numbers and explanations. The peak cross section for p-B11 is (only) 3 times smaller than the peak for D-T, but the peak for p-B11 occurs at ten times higher temperature. If you are pressure limited, as you often are and as is expected in a polywell in particular, then by dropping the temperature (or mean energy, if you are a Maxwell denier) a factor of ten, you can go to ten times higher density, which ups the reactivity a factor of 100. The D-T reaction also produces twice the energy of p-B11, and the high Z of the boron requires you to keep more useless electrons hot. Put that all together on the back of an envelope and you have the three orders of magnitude.

[chrismb]

Art,

You're arguing chalk-and-cheese.

*You* might be thinking of a magnetically confined thermalised plasma pressure, and *I* might be thinking of a [questionably] confined thermalised volume, but we are in the minority. The speculation underway here is the magnetic confinement of relatively low temperature electrons that are held centrally and that experience/generate a perfectly balanced electrostatic field that confines free ions, none of which have any rotational momentum and all head towards one single central spot within that electron bunch and pass through it without turbulence or instability.

So there *is* no "ion pressure", it can only be electron pressure, and you only need a tiny number of electrons to confine a vast quantity of ions.

[MSimon]

So there *is* no "ion pressure", it can only be electron pressure, and you only need a tiny number of electrons to confine a vast quantity of ions.

Quasi-neutrality says that the total charge of the ions is about equal to the charge on the electrons.

As to instabilities. What you would like to do is harness them rather than fight them. I have seen at least two independently developed simulations which show that beam formation and oscillation is inherent in the design. If those modes can be enhanced while drawing energy from other modes it would probably be a good thing. I have some ideas.

[chrismb]

Quasi-neutrality says that the total charge of the ions is about equal to the charge on the electrons.

It's probably something I should already well know, but I thought the quasi-neutrality is meant for all electrons (i.e. including as they circulate around the coils) so the number of electrons in the centre will be smaller than the recirculating ions, but enough to form a space-charge region co-centered with the device?

[MSimon]

Quasi-neutrality says that the total charge of the ions is about equal to the charge on the electrons.

It's probably something I should already well know, but I thought the quasi-neutrality is meant for all electrons (i.e. including as they circulate around the coils) so the number of electrons in the centre will be smaller than the recirculating ions, but enough to form a space-charge region co-centered with the device?

The thinking is that there has to be a slight excess of electrons in the reaction space ("inside" the grids) to make the device work. The number commonly given is an excess of 1E-6.

[chrismb]

So, if there's an excess of electrons inside, and there's an excess of electrons outside (recirculating) then won't the whole system, which is therefore net negative all the way from outside the grids to inside the grids, draw in contaminant ions?

[KitemanSA]

So, if there's an excess of electrons inside, and there's an excess of electrons outside (recirculating) then won't the whole system, which is therefore net negative all the way from outside the grids to inside the grids, draw in contaminant ions?

Not if you include the charge on the grid itself. Excess electrons inside, excess electrons outside, large deficit electrons on grid, separate with magnetic fields. Overall charge... neutral?

[kurt9]

From a value added stand point, it is not unreasonable to consider applications such as space propulsion before electricity generation with regards to polywell and other such schemes. Space propulsion has far higher value added service to the marketplace than energy production (which must be a few cents per watt of electricity generated), even though its total market is far smaller. Its like flat panel displays (that can sell for 1-2 thousand dollars per square meter) verses solar panels (which must be 10-20 dollars per square meter to be competitive).

However, there are technical hurtles to be overcome before the polywell or related concepts can be considered for either application.

[chrismb]

Not if you include the charge on the grid itself.

How do you put a charge on the grid if the magrids are floating in free space. To what do you apply the differential potentials?

[93143]

The speculation underway here is the magnetic confinement of relatively low temperature electrons that are held centrally and that experience/generate a perfectly balanced electrostatic field that confines free ions, none of which have any rotational momentum and all head towards one single central spot within that electron bunch and pass through it without turbulence or instability.

The ball of electrons is supposed to be about the same size as the ball of ions. The electrons are high-energy at the edge and mostly radially-directed, except during the brief turnaround at the edge of the magnetic field (which is excluded from a fairly large, nearly spherical volume by plasma currents, giving rise to the wiffleball effect). This electron behaviour is of course idealized; there will be collisions and they will affect the operation of the device.

The electrons slow down as they approach the centre, and speed back up as they recede from it. One of the effects of high-energy injection is to give the electrons enough speed to penetrate the wiffleball and produce a deep radial distribution of negative charge rather than just forming a sheath, which would be (I think) very inefficient at generating a virtual well.

The ions are introduced inside the grid at the edge of the wiffleball. They are cold there. They then notice the slight electron excess in the volume below them and fall into the well. Result? Fast, nearly monoenergetic radially-directed ions near the centre of the wiffleball, slow thermal ones at the edge. The ions will start to thermalize due to collisions in the high-density core, but Bussard predicted that collisions in the cold, high-residence-time-per-pass edge region would retard global thermalization sufficiently to allow a power reactor to run in a partially-relaxed steady state. Any angular momentum picked up by the ions should be largely cancelled by this annealing effect, since most of the high-energy collisions will take place in the convergent region at the centre and thus the angles involved will mostly be small.

Maxwellianization in the intermediate acceleration region, with issues such as two-stream instability, shouldn't be a big issue, as the circulation time for the ions is smaller than the thermalization time at the energies encountered over most of the circuit, and the two-stream region has lower density than the core and lower density and residence time than the edge.

This setup results in a situation where the ions are slow where the electrons are fast, and vice versa. The energy transfer rate is slow enough, due to the mass differential, that ion-ion and electron-electron collisions will dominate thermalization.

The ion focus at the core causes a secondary positive well to form at the bottom of the negative one. This means that the actual bottom of the well is a hollow sphere around the core, and thus that electrons will be moving somewhat faster in the core than at the radius of this sphere. Controlling the height of this "virtual anode" is one of the critical issues for making p-11B work.

It may be possible to substantially increase the performance of this system by utilizing a technique called POPS (Periodically Oscillating Plasma Sphere) or something like it. This takes advantage of natural beam bunching effects to amplify the plasma focus by driving a compression/expansion oscillation with RF forcing (warning: I haven't spent nearly as much time on trying to understand POPS as I have on Polywell, so this description may be slightly inaccurate).

Alignment of this geometry is not appallingly difficult. Spherical electrostatic geometries are fairly stable; Gauss' Law makes up for a lot.

That's the general idea. And yes, at this point it's largely speculation and repeating of Bussard's talking points. I don't know where you got your version...

How do you put a charge on the grid if the magrids are floating in free space. To what do you apply the differential potentials?

They're not floating in free space. How would you build something like that? They're connected to the support equipment with insulated standoffs. You use those standoffs to electrically connect the grid and the electron guns to opposite ends of a high-voltage power supply. Problem solved.

If you've got trap grids and such, it's obviously more complicated than that, but the basic principle is the same.

[D Tibbets]

... If you want a decent power density, you need a high plasma pressure (and a thousand times higher still if you want to use p-B11)...

Question- why a thousand fold increase in pressure for P-B11? At appropiate drive voltages (eg. 75,000 KeV 'resonance peak(?)') the graphs I have seen show cross sections close to those of D-D. What am I missing?

Dan Tibbets

http://en.wikipedia.org/wiki/Aneutronic ... challenges has some numbers and explanations. The peak cross section for p-B11 is (only) 3 times smaller than the peak for D-T, but the peak for p-B11 occurs at ten times higher temperature. If you are pressure limited, as you often are and as is expected in a polywell in particular, then by dropping the temperature (or mean energy, if you are a Maxwell denier) a factor of ten, you can go to ten times higher density, which ups the reactivity a factor of 100. The D-T reaction also produces twice the energy of p-B11, and the high Z of the boron requires you to keep more useless electrons hot. Put that all together on the back of an envelope and you have the three orders of magnitude.

Ah, I see. You are using D-T fusion as your baseline rather than D-D. But, D-T fusion is at least several hundred times more efficient than D-D fusion at comparable voltages that a Fusor and presumably a Polywell can achieve ( a thousand fold greater at energies a Tokamac can achieve), so you need (or I need) to make an adjustment to talk about realative comparisons with these different baselines.

Or the comparison can be fliped. An example comparision could be: 'Net positive P-B11 fusion in a Polywell is impossible at the drive energies, pressure, and sizes obtainable; D-D net positive fusion is iffy at best; D-T net positive fusion is a piece of cake in comparison.'


Dan Tibbets