Focus Fusion On Slashdot

[TallDave]

What reduces problems with the electrons aggravates problem with the ions

Unless the ions are too well confined to be sucked into the cusps.

He mentions that trying to maintain a non-thermal distribution by bleeding off particles will aggravate energy losses that are already too high

Yes, but in WB machines where the electrons go matters.

If I had more spare time I might try to re-calculate Rider's formula there with WB-like assumptions.

Rick has stated they can manipluate the anode height and other variables to achieve optimum conditions. Whether that optimum is good enough for an economic p-B11 reactor is still, I think, an open question.

I am much more hopeful we will get something useful for D-D/D-T. It's quite possible they could be competitive with fission.

Unless it was later than 2005 I'm assuming that WB6 type recirculation was not figured into the equation.

It's been pointed out before that it's hard to reconcile WB-6/7 results with some of the objections made.

WB-8 should tell us a lot.

[Art Carlson]

If I had more spare time I might try to re-calculate Rider's formula there with WB-like assumptions.

What would you like to re-calculate? The bremsstrahlung power? Under the assumption of an EEDF with a hole in the middle to reduce the coupling with the ions? Why don't you just tell us how you would keep that hole from filling in 1000 times faster than the fusion rate?

I am much more hopeful we will get something useful for D-D/D-T. It's quite possible they could be competitive with fission.

Yes, whatever our absolute level of hope, we can all be 5000 times more hopeful that something useful will come out of DT. I have no idea what you mean by "quite possible". Is it also "quite possible" that this whole thing is a wild goose chase?

It's been pointed out before that it's hard to reconcile WB-6/7 results with some of the objections made.

It's been pointed out before that no usable WB-6/7 results have ever been published. What's to reconcile?

[TallDave]

What would you like to re-calculate?

The necessary power to maintain a nonthermal electron distribution given where the electrons are going. (I suspect Rick already has something in this regard, of course.) I think brem will require better simulations that include things like anode height and etc, which is well beyond the scope of what I would ever want to try to tackle on my own (unless I can steal some code...). I haven't really thought about EEDF enough to say anything useful.

I have no idea what you mean by "quite possible".

I only mean D-D/D-T could be competitive; i.e. p-B11 is not necessarily required for the concept to be economically useful.

It's been pointed out before that no usable WB-6/7 results have ever been published. What's to reconcile?

Reasonable inferences can nevertheless be made from what is known. For instance, would you expect neutron counts and good confinement if the cusps were leaking large numbers of ions (as happened when Bussard tried emitter plates at the cusps in his closed-box designs)?

[D Tibbets]

....
Does Rider figure bremsstrahlung on a thermalized electron population, or a dominantly radial electron direction that is cold in the center and hot on the periphery? I believe this is one of the necessary conditions for Polywell to work. If so, Rider's conclusions are presumably correct for the assumptions he made, but does not match the Polywell assumptions.

Rider quickly dismissed using electric fields to reflect the electrons lost through the cusps using much the same arguments as I did: What reduces problems with the electrons aggravates problem with the ions. If recirculation helps anyway, then it helps reduce the cusp losses, but not the bremsstrahlung. The bremsstrahlung power he cites is for Maxwellian electrons, but only after he first shows that the the electron distribution will rapidly thermalize. He mentions that trying to maintain a non-thermal distribution by bleeding off particles will aggravate energy losses that are already too high. I think he has all the bases covered.

Also, a 5 to 1 ratio of P to B11 was quoted. I think (?) I have heard of a 10 to 1 ratio proposed by someone. This would presumably help the bremsstrahlung some but also decrease the fusion rate, so I don't know if there would be any net gain.

The optimum fuel ratio to minimize P_brem/P_fus is n_1:n_2 = Z_1:Z_2.

Is the 'Phi_well = 900 kV' the potential well depth? This seems high if it is.......

Rider goes to a good deal of effort at the beginning to explain why you can't really take advantage of any resonances in the cross section. I'm not sure, but his well depth may have been optimized in light of ion losses through up-scattering.

Does bleeding off high energy electrons waste too much energy? If an escaping electron has an energy of 50.000eV and the magrid is 20,000 volts, then 30,000 eV is lost. But if the escaping electron has an energy of 20,100 eV, then only 100 eV is lost. So, I'm supposing there is some relationship between the rate of electron upscattering to higher energies and the lifefime of the electron. In other words, the upscattered electrons have a progressively shorter amount of time to pick up more energy before they escape (assuming the electrons have ~ the same number of transits befor escaping-eg: 100,000). This might keep the lost energy of most of the escaping (nonrecirculating) electrons low. Or another way to look at it is that the electrons are removed at low cost when those electrons are only a modest distance past the average (mean?) peak of the maxwell distribution, rather than waiting till they reach the much higher energy tail.

As far as ions and electrons having equal but opposite responces to conservation attempts in the cusps, I'm guessing that this would only fully apply if the cusp flows are ambipolar, which they are claimed not to be.
Rider's optamistic estimate of a ~ 10 fold shortfall in the energy balance is actually fairly close to what Bussard was saying, up untill the claimed breakthrough with electron recirculation in WB6.

Bremsstrulung radiation could be modeled on the BIG hole in the center. I don't see why this would cost energy. It is dependant on the non thermalized assumption. I capitalized the adjative BIG to emphasize that the core region not only removes some of the area aviable to brem generation in a thermalized model, but also due to convergance and slowing of the electrons in the middle (again assuming some significant degree of non thermalization over the life time of the electrons) which results in a disproportionate concentration of the electrons in the core region. It seems odd that there is (?) a central concentration of both electrons and ions (I assume this correlates to an elliptical potential well -vs- a square well). I'm guessing that it is due to the dynamic time dependant distribution of the electrons, and the convergence of the ions overcomeing the greater time they spend in the perifery.

Brem -vs- fusion rate based on your formula makes sense. But only if the slopes of both curves have the same slope. At 900 KeV the slope of the P-B11 crossection is pretty shallow compared to what it is at say 200 KeV.
Does the brem curve follow the same profile? I'm guessing that this may be one reason why a larger reactor is needed for P-B11 compared to a D-D. The cross section matches or slightly favors P-B11 compared to D-D at higher voltages. Increasing the size presumably could compensate for a lower fusion rate due to higher proton to boron ratios, but a disproportionate decrease in brem radiation.

Why a 900 KeV well to reduce upscattered ion losses? If the potential well equals X eV: then theoretically, any ion upscattered to X+1 eV could potentially escape if it hits a cusp. I assume that the higher electron potential and probably stronge magnets to adequatly containe electrons at that energy would result in a more efficient wiffleball. Did Rider believe in or at least accept the wiffleball argument in his 900KeV calculation. If the wiffleball is real, does it even apply to this required well depth argument?
I'm supposing the ion loss rate would be the same weather you had a 20 KeV potential well or a 900KeV well, everything else being equal. Would this change if it was a thermalized ion population? Are the Maxwell curves different at these different temperatures?

I wonder how many of these questions have been at least tenatively answered by EMC's unpublished data (thermalization times, annealing, cusp energy losses -vs- particle losses, confluence, other transport losses, fusion collision types, etc.)

Dan Tibbets

[Art Carlson]

There are a lot of arguments flying around, so let me repeat exactly where I see the problem. To get fusion, you need hot ions, but the ions will unavoidably transfer power to the electrons through classical collisions, which the electrons will emit as bremsstrahlung. The proposed polywell solution is to manipulate the distribution of electron energies. It does not help to simply remove the fastest electrons because it doesn't really matter whether the electrons lose their energy through bremsstrahlung or "evaporation". What must be done is to reduce the coupling between the ions and the electrons. Since the Coulomb cross section is a sensitive function of relative velocity, the way to do this is to deplete the electrons with the lowest energies, those with speed nearly matching the ion speeds.

Classical collisions among the electrons will tend to fill in this hole in the energy distribution, i.e. will constantly produce new populations of low-energy electrons. To maintain the non-Maxwellian distribution, you need to selectively impart energy to those electrons that have the least energy, either in situ or by extracting them and re-injecting them with a higher energy. Nota bene, we need to "bleed off" low energy electrons, not high energy electrons.

How can this be done? Bussard and Nebel have not ever provided even a vague suggestion of how, and I don't have the foggiest notion either. Rider didn't say this is impossible, he simply calculated that, whatever mechanism you come up with, it will have to be terribly efficient.

[alexjrgreen]

viewtopic.php?p=20032#20032

I also notice that a number of people are trying to make Polywell arguments using classical collision models. The dominant mechanisms for transferring energy between the ions and the electrons are collective mechanisms, not classical binary collisions. Our experience is that you have to do full-up kinetic simulations if you want to understand these mechanisms and their effects. We've been doing that for the past 1.5 years, and we plan to be doing a lot more simulations over the next 2 years.

[Art Carlson]

viewtopic.php?p=20032#20032

I also notice that a number of people are trying to make Polywell arguments using classical collision models. The dominant mechanisms for transferring energy between the ions and the electrons are collective mechanisms, not classical binary collisions. Our experience is that you have to do full-up kinetic simulations if you want to understand these mechanisms and their effects. We've been doing that for the past 1.5 years, and we plan to be doing a lot more simulations over the next 2 years.

I agree that the energy transfer rate from ions to electrons could be much higher than the classical lower limit. Likewise, the rate at which the electrons thermalize could be much higher than the classical lower limit

[alexjrgreen]

viewtopic.php?p=20032#20032

I also notice that a number of people are trying to make Polywell arguments using classical collision models. The dominant mechanisms for transferring energy between the ions and the electrons are collective mechanisms, not classical binary collisions. Our experience is that you have to do full-up kinetic simulations if you want to understand these mechanisms and their effects. We've been doing that for the past 1.5 years, and we plan to be doing a lot more simulations over the next 2 years.

I agree that the energy transfer rate from ions to electrons could be much higher than the classical lower limit. Likewise, the rate at which the electrons thermalize could be much higher than the classical lower limit

Or much lower....

[Art Carlson]

viewtopic.php?p=20032#20032

I agree that the energy transfer rate from ions to electrons could be much higher than the classical lower limit. Likewise, the rate at which the electrons thermalize could be much higher than the classical lower limit

Or much lower....

Wrong. There is no way to make less water flow out of a bucket by drilling another hole.

[alexjrgreen]

I agree that the energy transfer rate from ions to electrons could be much higher than the classical lower limit. Likewise, the rate at which the electrons thermalize could be much higher than the classical lower limit

Or much lower....

Wrong. There is no way to make less water flow out of a bucket by drilling another hole.

If the classical collision argument doesn't fly, what model are you now using?

[Art Carlson]

The dominant mechanisms for transferring energy between the ions and the electrons are collective mechanisms, not classical binary collisions.

If the classical collision argument doesn't fly, what model are you now using?

In a plasma you have charged particles whizzing past each other. There's no way to change that. And when charged particles pass each other, they exchange energy and momentum. That's as basic as physics gets. There is no way to avoid energy exchange through Coulomb collisions.

Sometimes, however, you can neglect that transfer because some other mechanism of transfer is so much larger. A tokamak with classical confinement would be pretty neat. Unfortunately non-classical energy losses are so much larger that you don't need to spend much time talking about classical losses.

That's what Rick is saying. Transfer through collective mechanisms are much larger than transfer through binary collisions. They dominate. I don't know how to calculate how bad it really is, but I know it is at least as bad as the classical rate, and for p-B11 that is plenty bad enough.

[alexjrgreen]

The dominant mechanisms for transferring energy between the ions and the electrons are collective mechanisms, not classical binary collisions.

If the classical collision argument doesn't fly, what model are you now using?

In a plasma you have charged particles whizzing past each other. There's no way to change that. And when charged particles pass each other, they exchange energy and momentum. That's as basic as physics gets. There is no way to avoid energy exchange through Coulomb collisions.

Sometimes, however, you can neglect that transfer because some other mechanism of transfer is so much larger. A tokamak with classical confinement would be pretty neat. Unfortunately non-classical energy losses are so much larger that you don't need to spend much time talking about classical losses.

That's what Rick is saying. Transfer through collective mechanisms are much larger than transfer through binary collisions. They dominate. I don't know how to calculate how bad it really is, but I know it is at least as bad as the classical rate, and for p-B11 that is plenty bad enough.

If that was true, Rick wouldn't have got past the peer review. One of these assumptions has to be false.

[Art Carlson]

If that was true, Rick wouldn't have got past the peer review. One of these assumptions has to be false.

1. viewtopic.php?p=20032#20032 was not peer reviewed.
2. As far as I know, he has never made a similar statement in a peer-reviewed publication.
3. Peer review is not perfect.

[alexjrgreen]

If that was true, Rick wouldn't have got past the peer review. One of these assumptions has to be false.

1. viewtopic.php?p=20032#20032 was not peer reviewed.
2. As far as I know, he has never made a similar statement in a peer-reviewed publication.
3. Peer review is not perfect.

http://cosmiclog.msnbc.msn.com/archive/ ... 18741.aspx

First of all, our work has been peer reviewed. An independent panel of experts has looked at these results. I don’t believe that there was anyone on the panel who has less than 40 years experience working with magnetic confinement. It included senior professors and people who have managed the fusion program. We asked them for their honest opinions and that’s exactly what we got. We are proceeding with our program in line with their recommendations.
Secondly, the talk-polywell blog has a large variety of people who post there. There are Phd plasma physicists as well people from the general public. I think that’s a good thing. Science needs to be accessible to people.

The peer review panel were as well informed, as experienced, and as skeptical as you are...

[MSimon]

Art you are leaving out that while the particles are whizzing passed each other their "point" of maximum interaction is at "1/2 r" where the velocities are equal and opposite. Every where else (if it is a beam machine) the electrons are fast while the ions are slow and vice versa.

In addition you have the "dwell time" at turn around where the slow particles can thermalize at the slow speeds.

At least that is what I have gotten from reading what is out there.

Will it work? TBD.

It all hinges on what the relaxation time to Maxwellian is. And a couple of other things.