Transterrestrial Musings

Comments

Stupid question... there's also a link to Bussard's fusion reactor design. What's the (full) story on that? Is it for real?

Because if it is, y'know, that pretty much changes everything, right there... and that is exactly why I'm pretty skeptical that it works as well as claimed from that last-minute test.

Posted by Big D at June 5, 2007 03:10 PM

there's also a link to Bussard's fusion reactor design. What's the (full) story on that? Is it for real?

He's still stumping for money, right? That should be a clue right there.

P-11B fusion is incredibly hard, and his claimed results (which, as far as I know, haven't been published in a real peer reviewed scientific journal) are still orders of magnitude away from what would be needed in a real reactor.

Posted by Paul Dietz at June 5, 2007 04:59 PM

Paul, I believe that is why he is stumping for money - so that he can build a real reactor. The amount of money he is talking about is orders of magnitude less than what is required to build a Tokamak reactor, which haven't yet produced more energy than they consume.

Posted by Ed Minchau at June 5, 2007 05:51 PM

P-B11 fusion is incredibly hard if you want to do it by thermal methods in a Tokamak type device. Bussard's machine bypasses the thermal fusion problems by using ions of the right energy so as to maximise fusion cross section. Ions are difficult to control with magnetic fields due to their mass. Electrons are much easier so Bussard manipulates electrons with magnetic fields to create electric fields which manipulate the ions. Two neat sidesteppings of the traditional hot fusion problems. The reaction products are charged particles which can generate electricity directly without going via the usual steam generation cycle. A neat third sidestep. Check the Google tech talk by Bussard himself.

As for the Dyson video it seems it may be rather old. It does occur to me that the stratospheric cooling he talks about could probably be measured rather easier than the surface temp or even mid troposphere temp. So does anyone have any *good* historical data on this? Other points - is CO2 well mixed at altitudes of several tens of kilometers where it is very cold? If so why does a gas that is nearly 1.5 times the molecular weight of the major atmospheric constituents and originating at the surface essentially, remain so at great altitudes in the absence of large scale vertical motion, in a field of one g? Are there any direct measurements of this? Is molecular diffusion that effective? Why so little water vapor at high altitudes then?

Posted by Mike Borgelt at June 5, 2007 06:52 PM

"absence of large scale vertical motion"

Hmm .. don't lee waves count? They are known to go to several tens of km high. Or is the well known (and spooky when I first take people into it ... the sudden total absense of vibration and random "waffling" is so unexpected) laminar flow so perfect that there is no mixing?

Isn't the lack of water vapour explained by it freezing, clumping, and falling out?

Come to that ... CO2 freezes at -57C and that's the approximate kind of temperature you get 15 or 20 km up. It's not water though, and water is strange and unusual stuff when it comes to self-attraction and accretion.

Posted by Bruce Hoult at June 5, 2007 10:09 PM

Paul, I believe that is why he is stumping for money - so that he can build a real reactor.

And the fact that he is still stumping for money, even a small amount, shows that the people who have money do not believe him. I mean, if it were really as wonderful as the internet echo chamber would indicate, why haven't they ponied up the money?

P-B11 fusion is incredibly hard if you want to do it by thermal methods in a Tokamak type device.

No, it's incredibly hard, period. Non-thermal methods face ruinous losses as the plasma tries to thermalize. Todd Rider's thesis at MIT basically shot down nonthermal methods for proton-based fuels. He showed the recirculating power fraction to maintain the non-Maxwellian distribution was far too large to be practical.

Posted by Paul Dietz at June 5, 2007 10:09 PM

If so why does a gas that is nearly 1.5 times the molecular weight of the major atmospheric constituents and originating at the surface essentially, remain so at great altitudes in the absence of large scale vertical motion, in a field of one g?

Atmospheric gases even heavier than CO2, such as krypton, remain well-mixed up to the 'homopause' (aka the 'turbopause'), which on Earth is about 80 to 90 km up. The reason is that below this altitude, the mean free path of molecules is short enough that diffusive separation is slower than remixing by turbulence (eddy diffusion).

Come to that ... CO2 freezes at -57C and that's the approximate kind of temperature you get 15 or 20 km up.

It's my understanding that the vapor pressure of solid CO2 is above the partial pressure of CO2 at all altitudes, so any solid CO2 clouds would evaporate. If the atmospheric CO2 concentration were much higher this wouldn't necessarily be the case, which has caused problems with high-CO2 atmospheres as solutions to the 'faint sun paradox' (why the Earth wasn't frozen over in the early solar system, when the Sun was dimmer.)

Posted by Paul Dietz at June 5, 2007 10:17 PM

polar cities in future? see my rant. google term

Posted by danny bee at June 5, 2007 11:15 PM

Bruce,
As you know, most lee waves occur in the troposphere. Which is why Steve Fossett had so much trouble finding lee waves that would take him above 50,000 feet for the record flight which he eventually did. BTW that was some of our instrumentation on that flight and they did show that they got into the stratosphere(recovered the data from our instrumentation after problems with the NASA setup). However I'll stick with "absence of large scale vertical motion" at least on a routine basis. Most times above the tropopause the wind speed drops off fairly quickly and you need some special conditions to get wave high in to the stratosphere.

I'd still like to see a plot of CO2 vs altitude by direct measurement. Above about 35km this isn't easy or routine. In fact who goes much above 12km nowadays except for the odd U2?

As for p-B11 fusion, Doc Bussard had all the objections in hand in his Google talk and claimed that thermalisation wasn't the problem some make it out to be. I also ask myself why a guy of 79 years with a solid scientific career behind him, much of it in the fusion field, would be making the claims he does unless he thinks he is right.

Posted by Mike Borgelt at June 6, 2007 01:23 AM

I also ask myself why a guy of 79 years with a solid scientific career behind him, much of it in the fusion field, would be making the claims he does unless he thinks he is right.

Oh, I would not be surprised if he thinks he is right. This doesn't mean he is right, particularly when energy investors (and there are lots of 'Clean Tech' venture capitalists in silicon valley now pumping money into energy startups) don't appear to be funding him.

I should note that Bussard has a previous history of optimistic promotion of an alternative fusion concept.

Posted by Paul Dietz at June 6, 2007 05:55 AM

As you know, being the resident science and engineering expert at Classical Values, I have been looking deeply into the Bussard stuff for flaws. I'm an enthusiastic skeptic (i.e. I want to know enough to do the engineering). So far as I have been able to find outside confirmation, what Dr. Bussard says is confirmed. The last bit I did on it:

Since I can't use the proper URL search:

Electron Circulation In A Cubic Polywell "IEC Fusion Technology"

then go to the article:

Polywell - Making The Well

Which is in the May posts.

Pretty much broke the back of my scepticism. If the potential well forms (it does) then all the rest is pretty much self evident. The last time I made a prognostication (about a month ago) I thought that the Bussard reactor had a 1% to 10% chance of working. Now I'd say 10% to 30% (the physics is confirmed the problems are engineering).

Posted by M. Simon at June 6, 2007 09:45 AM

I don't see why that 'broke the back' of your skepticism. IEC schemes have been criticized in general due to thermalization of the ions, not (just) formation of the potential well. Rider (and, independently, Nivens) concluded thermalization would seriously limit performance, even if the well could form.

http://en.wikipedia.org/wiki/Inertial_electrostatic_confinement#Critique

Does Bussard have an article in the peer reviewed literature addressing their critiques?

Posted by Paul Dietz at June 6, 2007 11:19 AM

The physics behind the Polywell is sound. Wether or not it will be able to produce net power remains to be seen, but if the aniosotropic plasma distribution enables the device to minizmize the horrid bremsstrahlung losses inherent to p-B11 fusion, then it should be a matter of proper engineering to get it to net power production.

Posted by Blane Dabney at June 6, 2007 11:22 AM

if the aniosotropic plasma distribution

Rider argues (read his master's thesis) that the ion energy distribution will thermalize 2 to 3 orders of magnitude faster than the ions can fuse, and that losses due to escape of the ions from the high energy tail of the distribution will itself render the concept unworkable for proton-based fuels, even if bremsstrahlung is ignored.

Posted by Paul Dietz at June 6, 2007 11:42 AM

One question in particular has been bugging me about Bussard's proposals.

Early in the talk he describes an optimal configuration for his ion confinement system as being coils arranged in a pattern like a regular solid (forget which one, king of like two tetrahedra mated to each other) with an even number of faces arranged around each vertex.

But he seems to have spent his time building machines arranged around cubes, which have an odd number of faces around each vertex.

Posted by Phil Fraering at June 6, 2007 11:57 AM

Can a physics minor impinge on those who understand better? Bussard's response to Rider's paper is always that the Polywell(IEC in general?) maintain a non-maxwellian distribution. What I want to know is does Rider's paper address this adequately or not? I lack the understanding to look at them both and figure it out myself.
In particular, Rider's paper says:
"It has been suggested that IEC can maintain non-Maxwellian ion distributions...Such a machine would then be highly suitable for use with advanced fuels like... pB11."
So Rider agrees that a non-Maxwellian IEC design would work as Bussard says.
"Velocity distributions in the region of significant density are approximately isotropic."
This is one of Rider's assumptions. As I said, this is over my head. Does this assumption mean that Rider's fusion rates aren't considering the head-on nature of ion velocities in this region?
I am asking because both Bussard and Rider are aware of each other's work. Bussard is emphatic that the Polywell is non-maxwellian. Rider states that because the ions thermalize faster than they fuse all such systems will maxwellianize. Am I just way out of my league or is Bussard suggesting that Rider's fusion rates are wrong(perhaps because of the velocity dist. assumption)?
If I'm just so far over my head it'd take too long to explain just say so as well. I'll take the years needed to understand it myself.

Posted by Kevin at June 6, 2007 12:27 PM

Perhaps Bussard is refering to the anisotropy of the ions (when they are away from the center), which Rider doesn't consider the problem. Rider instead focused on the (parallel) energy distribution. That is what he's saying will thermalize much faster than fusion will occur.

About isotropic distribution: this simply means that the ions are approaching the center from all directions. If this isn't initially the case, it will certainly be the case after a few scattering times. And when you have ions intersecting at the center at arbitrary angles (rather than just, say, head on), you can have collisions that change ion energy, leading to rapid spread in the energy distribution.

The fundamental problem is that scattering has a much higher cross section than fusion.

Posted by Paul Dietz at June 6, 2007 12:45 PM

And Bussard argues that Rider has a grossly wrong model of an IEC device, which leads him inexorably to that conclusion. In particular, he argues that Rider has an inverse picture of the plasma density profile. For example, if you go to Rider's critique of IEC systems in general, (page 13), right off the bat his first assumption is at complete odds with Bussard's description of the polywell. There's no "broad, flat central region" in an IEC. In fact, Bussard's descriptions have density thickest and Maxwellian at the edges (low energy), with decreasing density in the middle until you get pretty close to the focus at the exact center.

Second, Rider's last assumption is that quasi-neutrality exists in this "broad, flat central region". I don't think even Bussard would argue that Maxwellian wouldn't take place if it was a neutral plasma. This is something Rider just assumed about IECs, not proved. In effect, it looks like Rider assumed initial conditions that guarantee Maxwellianization.

Keep in mind that every one of Rider's points against, including Bremstahhlung and "high energy ion escape" is predicated on the theory that thermalization will occur.

So although Bussard doesn't have any problem with Rider's thesis, he does argue with its applicability to an IEC polywell. He hasn't presented any peer-reviewed papers to back that up (that I know of), and has apparently taken the attitude that just getting good data from WB-7 will take care of that objection (I think that's a mistake because any number of mundane things could go wrong between here and there, science notwithstanding). So right now it's just Rider's assumptions vs. Bussard's contention that those are bad assumptions.

But just by reading Rider's critique, it's clear he made some assumptions that don't apply to the theory of how a polywell works.

Posted by tom at June 6, 2007 12:53 PM

There's no "broad, flat central region" in an IEC.

Rider points out that density actually doesn't matter to his argument, to first order (there's a coulomb logarithm in there, but that's a weak effect). As central density increases, fusion and scattering rates increase similarly. The density affects the absolute rate of the two effects, but the ratio -- which is what damns IEC -- is unchanged.

Second, Rider's last assumption is that quasi-neutrality exists in this "broad, flat central region"

What keeps electrons out of that region? With all that positive charge they should be strongly attracted there. But even if this assumption is wrong, it doesn't change the ion-ion scattering much (may even make it worse, I think). So this fails to save the concept.

Keep in mind that every one of Rider's points against, including Bremstahhlung and "high energy ion escape" is predicated on the theory that thermalization will occur.

No, Rider demonstrates that thermalization of the ion energy will occur much faster than fusion. It's a conclusion, not an assumption.

Posted by Paul Dietz at June 6, 2007 01:02 PM

I just realized that I didn't mean two tetrahedra mashed together, but an eight-sided figure whose name I can't remember.

Posted by Phil Fraering at June 6, 2007 02:23 PM

Regular octahedron?

Posted by Paul Dietz at June 6, 2007 02:23 PM

Probably so.

I've been reading more about this; apparently these designs are meant to be "truncated cubes" of some degree but there are only magnets for the faces of the cube and not the triangular faces that are where the vertices of a cube would normally be.

Posted by Phil Fraering at June 6, 2007 02:46 PM

First off, thanks for clarifying that for me. Makes Rider's position make sense even to a mostly lay person like myself. I have seen a pretty good explanation though for why energy loss from scatter may not apply. I found this explanation on the nasaspaceflight forums.
"Ions, on the other hand, can start at zero velocity at the edge of the well and reach fusion speeds by the middle. So whether they are freshly injected or on their 20000th transit, their liklihood of fusion is the same."

As I understand this, the ions that have lost energy from scattering against each other are still in the center waiting to collide/fuse with incoming full energy ions. Perhaps I'm again just in over my head, but does Rider's paper address this? If incoming ions are traveling fast enough to fuse with static/low energy ions are his assumptions violated?

Posted by Kevin at June 6, 2007 02:51 PM

The problem isn't ions that have lost energy. The problem is ions that have gained energy. This 'high energy tail' will escape from the potential well and constantly drain energy out of the system, energy that has to be replaced. Rider computes how much energy this is, and concludes it's far too large for p-11B fusion to be practical.

Rider extended this work in his Ph.D. thesis, where he demolished a much larger class of non-equilibrium fusion schemes. Maybe this is why his MIT advisor Lidsky (who had published a well-known critique of conventional DT fusion in 1981) went away from fusion to work on fission in his final years.

Posted by Paul Dietz at June 6, 2007 02:59 PM

All predicated on the existence of a "neutral plasma", which supposedly doesn't truly exist in the center of a polywell.

(population of x fast electrons + x ions existing in y polywell volume) =/= (neutral plasma ball size of y volume) This is akin to assuming that a soap bubble of 2cm radius weighing .005 gram consists of a homogenous distribution of .005 gram worth of soap, water & air in a 2cm radius sphere. Not so!

Rider assumes it, doesn't prove it. Bussard disagrees with the premise.

Other way around, that's the center of the potential well. So electrons are plentiful in that region. It's thermal ions that are not, and more to the point, there can't be a thermal distribution of ions radially any more than there can be a "thermal" distribution of ions in a linear accelerator when the voltage is on.

Keep in mind that every one of Rider's points against, including Bremstahhlung and "high energy ion escape" is predicated on the theory that thermalization will occur.

No, Rider demonstrates that thermalization of the ion energy will occur much faster than fusion. It's a conclusion, not an assumption.

It's a conclusion based on his (probably correct) math resulting from his assumed model. His model was an assumption about how an IEC machine would work. If the model is wrong, as Bussard claims, and Rider doesn't present evidence about why he picked that model--maybe it's self-evident to plasma physicists, one of which I am obviously not--there's no way to know if the conclusions apply or not.

Posted by tom at June 6, 2007 03:00 PM

It seems there is some serious money going in to p-B11 fusion:
http://www.classicalvalues.com/archives/2007/05/latest_fusion_n.html

Different configuration from the Bussard machine.

Posted by Mike Borgelt at June 6, 2007 03:56 PM

All predicated on the existence of a "neutral plasma", which supposedly doesn't truly exist in the center of a polywell.

Um, no, completely wrong. The thermalization of the ions via ion-ion scattering does not depend on the existence of a neutral plasma. It merely depends on the interactions of the ions, which have to occur if you are trying to get them to fuse at all.

Posted by Paul Dietz at June 6, 2007 05:14 PM

Different configuration from the Bussard machine.

A configuration from that group (I assume it's the one they're talking about) was also savaged in the literature. Nevins and Art Carlson (well known on usenet) pointed out similar problems would occur with its nonthermal ion distribution.

http://www.sciencemag.org/cgi/content/abstract/281/5375/307a
http://64.233.167.104/search?q=cache:Ys3rqGMd698J:sci-phys-plasma.caeds.eng.uml.edu/1998/07-98-18.htm+carlson+boron+plasma+fusion&hl=en&ct=clnk&cd=6&gl=us

Posted by Paul Dietz at June 6, 2007 05:24 PM

"The thermalization of the ions via ion-ion scattering does not depend on the existence of a neutral plasma. It merely depends on the interactions of the ions, which have to occur if you are trying to get them to fuse at all."

If it truly was that simple I don't think Dr Bussard and others would persevere because this is a simple fundamental physical process. I cannot believe they have simply overlooked this.

I am minded that all sorts of things have been predicted to be impossible by the best people working in the fields at the time(heavier than air manned flight, gas turbine propulsion for aircraft for two) which a short while later turned out to be possible. Nothing wrong with the proofs, just the assumptions that went in to them.

Posted by Mike Borgelt at June 6, 2007 07:40 PM

All predicated on the existence of a "neutral plasma", which supposedly doesn't truly exist in the center of a polywell.

I don't think you're getting the point here. I'm not arguing that Rider calculated something wrong here--I don't have the time or math skills to check his work, and I'm not accusing him of shoddy work. But Rider necessarily made certain assumptions in his modeling. For example, one assumption is that particle velocity distribution will be isotropic. In fact, all his base equations in Ch 3 (where he "disproves" IEC), are predicated on this assumption.

Is this a good assumption in the strong spherical electric field of an IEC? I don't have the knowledge to say. But it wasn't addressed in the paper. It's not true, for example in a linear particle accelerator. Like I said, maybe it's obvious to physicists why that's trivial.

Regardless, any paper like this is based on assumptions, and those assumptions are certainly worthy of questioning.

Posted by tom at June 6, 2007 07:44 PM

Has Rider updated his pre-polywell paper? I think to a degree Rider's 1995 paper is not applicable/germain to the polywell concept. I have been reading up on the issue since I found the google tech talk, I have not found anything on the net Rider has written recently, say... since the google tech talk, that applies.

I think its fair to say Riders paper was quite good, it properly described why previous IEC efforts had failed. No?

Posted by Roger Fox at June 6, 2007 10:54 PM

I think Paul is saying that since the ions come from all _directions_ (isotropic), they can exchange energy in a way that adds extra energy to certain ions and that then makes a "long tail" of high energy ions and they are the problem, they cause energy loss, and Rider calculated this loss is bigger than the energy production from fusions.

I have no experience whatsoever in the field and haven't read Rider's papers.

Posted by mz at June 7, 2007 05:14 AM

But Rider necessarily made certain assumptions

He made a number of assumptions. And he reached a number of conclusions. Not all the assumptions are necessary to reach every one of the conclusions. The quasi-neutrality assumption, for example, is not necessary to reach the main result (although I think you can argue that quasi-neutrality has to hold or else the space charge limit on the density rules out power anywhere close to what Bussard is claiming he will achieve.)

For example, one assumption is that particle velocity distribution will be isotropic.

As it will necessarily be in the polywell, or indeed any IEC fusion scheme, for reasons I explained. If the ions are not approaching the central point from all directions, they soon will be, since scattering (which happens much faster than fusion) will cause them to be traveling away from the central point in all directions after only a short time.

So, this is not an assumption, but a necessary consequence of how the underlying physics works.

Posted by Paul Dietz at June 7, 2007 06:40 AM

I think Paul is saying that since the ions come from all _directions_ (isotropic), they can exchange energy in a way that adds extra energy to certain ions

Another point I didn't make, but Rider did: if the ions are gaining energy by falling into the potential well from its 'edge', then the energy gained will be proportional to the ion charge. The boron ions will gain five times the energy of the protons. When they scatter at the center, the protons will gain energy and the boron nuclei will lose energy, on average.

Posted by Paul Dietz at June 7, 2007 07:33 AM

I recently got to talk to Bussard on a radio show where he was interviewed. What he really needs at this point is to get a WB-7 machine built to prove the concept. That's probably a few million. He and Ligon have said there is interest but it's not clear how much.

And the fact that he is still stumping for money

Well, to put this in perspective, a similar concept from a Paul Allen-funded startup just got $40M in VC funding.

Here's their patent.

Bussard hasn't been looking long; the results are only about a year old. Also, Bussard apparently wants to do this as an open-source project that will benefit everyone, rather than "making one guy really rich" as he puts it. That probably doesn't bring investors running.

Re Rider, the Polywell was an answer to the IEC questions he raised.

Tom Ligon, who worked with Bussard, has been very open and available to answer questions. Here's a 60+ page thread at Nasa Space Flight in which he addresses a lot of the technical questions.

P-11B fusion is still speculative. Barely worth talking about till we get a working D-D net power machine.

Bottom line, the Polywell concept seems to have a fair chance of being do-able. It's not a done deal, but it's not pie in the sky either. And if it works, it will probably change the world significantly.

Posted by TallDave at June 7, 2007 08:56 AM

Oops, here's that NASA forum thread link.

Posted by TallDave at June 7, 2007 08:57 AM

Does Bussard have an article in the peer reviewed literature addressing their critiques?

I asked Bussard this question too, as it would be a good way to address this question. Apparently this effort was begun earlier this year, but is on hold in preference to trying to get a WB-7 built, which is probably better proof.

Posted by TallDave at June 7, 2007 09:00 AM

Here's Ligon's exegesis on Rider's paper (:

If the ions would live sufficiently long and make enough passes without fusing, they would no doubt thermalize, or at least lose energy. This will also happen if the density is allowed to get too high. The key is achieving a density profile where high-energy collisions occur in the center at fusion energies, very low energy (and thermalized) collisions occur in the high-density ion turn-around zone near the magrid, and very few other collisions occur. A narrow range of optimal ion lifetime versus fusion density exists for any given size of machine and set of operating parameters (too high, thermalize, too low, too few fusions). Larger sizes are required for that optimal density to produce net power.

The thermalized turn-around zone is a key to killing two of Rider's objections. Rider believes A) the plasma will maxwellianize, and B) because it is maxwellian, some of the ions will upscatter in energy by collisions and so be able to exit the potential well complety. The outer collision zone does have maxwellian properties, but at very low energy levels, so it essentially removes the scatter in the energy of the fuel ions on every pass, essentially resetting the population to all near zero kinetic energy. Thus, thermalization keeps the machine from thermalizing.

Rider thinks the machine will have excessive bremsstrahlung losses. Bussard and Krall counter that the central region Rider thinks will cause bremsstrahlung due to mutual repulsion is actually a convergence zone for both electrons and ions. The electrons are making a virtual cathode that attracts ions, the ions make a virtual anode that attracts electrons, and the whole zone is never all that far from a neutral plasma. Control of virtual anode height controls the bremsstrahlung problem.

All of this says you need very good control of the ion population. WB6 and that puff gas system did not offer such control ... I suspect the system put in roughly the amount of gas needed, initially, but the amount continued to rise. Fusion occurred in the short period during which the right density profile existed. A successful machine will need to hold that condition.

It may be that it will occasionally be necessary to stop the reaction and clear the machine of junk gas, which may include fusion products in a power reactor.

Rider's final objection was dead-on right ... for HEPS, PXL-1, and WB-5. Cusp losses of electrons in a box machine are too high. They are irrelevant in a magrid machine. Electrons lost out the cusps come right back in.

Posted by TallDave at June 7, 2007 09:43 AM

TallDave: looking through that 61 page thread (admittedly skimming parts due to length) I saw no substantial rebuttal of Rider's key technical points. Perhaps you could point to the page and message you consider to have addressed the criticisms?

IMO, Rider's key point (about thermalization of the ion energy distribution at the interaction point) is devastating to the concept, and the responses have been muddled at best.

Posted by Paul Dietz at June 7, 2007 09:48 AM

The outer collision zone does have maxwellian properties, but at very low energy levels, so it essentially removes the scatter in the energy of the fuel ions on every pass,e ssentially resetting the population to all near zero kinetic energy. Thus, thermalization keeps the machine from thermalizing.

This makes no sense to me. Thermalization is a process increasing the entropy of the ions. If what is happening in the outside resets the ion distribution, it isn't thermalization, but rather some refrigeration process that actively removes entropy. Where is this entropy going?

Posted by at June 7, 2007 10:37 AM

Sorry, that previous comment was mine.

Posted by Paul Dietz at June 7, 2007 10:40 AM

"P-11B fusion is still speculative. Barely worth talking about till we get a working D-D net power machine."

Disagree. The main point here is that the reactor in a D-D fusor will fall apart in a pretty short time, from neutron bombardment, and will be radioactive when it's decommissioned (sp?), and a hell of a lot of the energy is wasted because fast neutrons can't be effectively tapped for energy.

P-11B, however, creates a large proportion of its energy in fast alpha particles.

Posted by Fletcher Christian at June 7, 2007 04:32 PM

All right, everyone.

Rand Simberg's comment sections are exclusively for logical fallacies. Y'all need to take this factually based discussion elsewhere.

Like, say, physorg.com or some such place.

Posted by MG at June 7, 2007 06:47 PM

Heh. Just you wait and see. In 6 years the super weak sunspot cycle 25 will begin to produce 1960's type temperatures AND we'll be thisclose to practical nuclear fusion. The Reavers on the Left will commit mass suicide. Bwahhaaahaa!

Posted by Mike M. at June 8, 2007 04:05 AM

No Mike, it will be Global Cooling again, because we didn't act in time against Global Warming.

Posted by Mac at June 8, 2007 09:26 AM

Shamelessly copying this response from Tom Ligon on the NasaSpaceFlight.com Forums. This is the clearest explanation I've seen for why Rider's fusion/scatter rates are inaccurate and/or don't apply to Bussard's Polywell configuration.

"The ion velocities in the central region are very high. They don't have much time in which collisions can take place, compared to the time spent outside that region. Yes, no-fusion collisions are more likely, but not enough will occur to cause much thermalization on any single pass. Ions entering this zone all have about the same energy ... the non-fusion collisions are primarily elastic, and so the amount of momentum transfer in any given collision is not great.

The ions reach the outer zone near the magrid where KE is low. They're slow, and back up. Here they do thermalize, all to low KE. This tends to equalize their KE. The spend far more time here than in the central convergence zone. Thus, on every pass, the machine will tend to "anneal out" the kinetic energy distribution acquired by core collisions.

Any ion which does happen to upscatter sufficiently on a single pass that it escapes the potential well will, in fact, be lost. This can be minimized by making a deeper potential well (running at a higher voltage) and generating the ions at some depth into the well. You can stop the milk from sloshing out of the bowl by using a deeper bowl."

I don't understand the physics enough to say if this model circumvents Rider's assumptions or not, but it sounds like others here do. To me it seems like Tom is explaining why Rider's High Energy Tail calculations are inaccurate? Anybody else able to explain this for a laymen like myself?

Posted by Kevin K at June 8, 2007 09:55 AM

Actually, Rider explicitly addressed the Polywell and the possibility of moving the ions to a region of higher potential energy, in his Ph.D. thesis (page 90, I think). He concluded this cannot do what Ligon claims it does, for basically the reason I explained: the irreversible process of spreading of the ion energy distribution continues to occur; moving the energy from kinetic to potential and back again doesn't change this.

I suspect Ligon's error is assuming all the ions go out to the same distance. But this only happens if they have the same energy. Ions of substantially different energy will reach the low KE states at substantially different radii, where they won't be interacting.

Posted by Paul Dietz at June 8, 2007 10:02 AM

Posted by Paul Dietz at June 8, 2007 10:51 AM

The main point here is that the reactor in a D-D fusor will fall apart in a pretty short time, from neutron bombardment,

Sure, tokamaks have similar problems. Still, a D-D net power machine is almost certainly going to be built before a P-11B, just because the energies for the latter are so much higher.

the irreversible process of spreading of the ion energy distribution continues to occur; moving the energy from kinetic to potential and back again doesn't change this.

I don't know, it seems reasonable that the spread won't keep spreading out if they continually undergo large shifts in energy as they circulate up and down. At the top, their energy necessarily has to be low, and so that "lowness" gets smeared across all of them.

Ions of substantially different energy will reach the low KE states at substantially different radii, where they won't be interacting.

If you check out the recirculation models, you'll see the ions are supposed to bunch up at the recirculation points. This also might be ameliorated by the fact the energy differences would cancel out based on their position vis-a-vis the bottom of the well. But it probably does need to more accurately modelled.

Posted by TallDave at June 8, 2007 02:43 PM

Well, entropy for a particular portion of a system doesn't have to increase; after all, you can clean up your room, which seems to violate thermodynamics only until you consider all the hydrocarbons you had to disorder to achieve the ordering of your room.

In this case, the entropy is getting sucked up by the magnetic field (which itself of course requires a greater increase in entropy in the form of whatever process produces the electricity that powers it) that re-orders the ions.

Posted by TallDave at June 8, 2007 02:51 PM

In this case, the entropy is getting sucked up by the magnetic field

No, that can't work. The magnetic field doesn't have the internal degrees of freedom to carry the entropy. Anyway, carrying away the entropy means that something else has to be getting hotter (since at nonzero temperature entropy is accompanied by nonzero energy). Where's the heat going? The magnetic field doesn't carry heat.

the magnetic field (which itself of course requires a greater increase in entropy in the form of whatever process produces the electricity that powers it)

In principle the magnetic field needs no power at all, once it has been created. So fundamentally this can't be a consideration.

Posted by Paul Dietz at June 9, 2007 06:36 AM

If you check out the recirculation models, you'll see the ions are supposed to bunch up at the recirculation points.

And if the ions have different energy, the recirculation points will be at different radii.

Put another way: if the kinetic energy of two ions differs by delta-E, their kinetic energies will continue to differ by delta-E at any point in the system they can both reach. There will be no point where they both have kinetic energy much less than delta-E.

I'll add another objection: if ion-ion interaction at large radii is occuring at a significant rate and transfering significant energy between ions, then it is also transfering significant momentum between ions. Adding momentum to an ion at large radius will tend to add angular momentum, unlike scattering at the center of the well (since the angular momentum goes as the cross product of the momentum and radius vectors). So this will tend to diffuse the ions into trajectories that miss the potential minimum at the center.

Posted by Paul Dietz at June 9, 2007 06:48 AM

In principle the magnetic field needs no power at all, once it has been created

Hmmm? It's exerting force on the ions and electrons. No free lunch.

That orders them, and entropy requires that power come from something being disordered somewhere else.

And if the ions have different energy, the recirculation points will be at different radii.

It won't matter if they keep hitting each other. Based on the modelling, those radii should intersect, so they end up nearly monoenergetic.

So this will tend to diffuse the ions into trajectories that miss the potential minimum at the center.

The virtual cathode should be strong enough to pull them into an area where energies are high enough to fuse.

Posted by TallDave at June 9, 2007 09:58 PM

Hmmm? It's exerting force on the ions and electrons. No free lunch.

Oh good grief. A magnetic field does no work on charged particles!

That orders them, and entropy requires that power come from something being disordered somewhere else.

Obviously not. A magnet can be designed that uses no power in steady state. You've heard of permanent magnets and superconductive magnets, yes?

Anyway, Rider addresses the issue of entropy being carried off by EM fields in his Ph.D. thesis. He shows mathematically that they can't change the entropy of the particles (see page 123), at least in the non-relativistic limit.

The virtual cathode should be strong enough to pull them into an area where energies are high enough to fuse.

Conservation of angular momentum keeps them from getting too close to the center, if they have nonzero angular momentum. The high density region at the center requires the particles be on trajectories with near zero angular momentum.

Posted by Paul Dietz at June 12, 2007 01:34 PM

Oh good grief. A magnetic field does no work on charged particles!

Yes, but creating the field does require work. No free lunch.

Obviously not. A magnet can be designed that uses no power in steady state.

Sure. Presumably you are not claiming permanent magnets violate entropy?

Again, not if the well is deep enough.

Posted by TallDave at June 12, 2007 09:49 PM

Here's a video that may or may not elucidate.

http://www.youtube.com/watch?v=ao0Erhsnor4

Posted by TallDave at June 12, 2007 10:23 PM

Here too is the back and forth on another IEF fusion effort by Tri Alpha, which cites Bussard's work. Very mathy, with a somewhat related discssion of the thermal equilibrium problem.

http://www.sciencemag.org/cgi/content/full/281/5375/307a

Posted by TallDave at June 12, 2007 10:45 PM

I think it's fairly obvious a magnet can reduce entropy without doing work (other than the work to create the magnet). Consider a magnet sitting in a pile of iron shavings. The shavings are ordered along the field lines. You come along and push slightly on some of the shavings, disordering them slightly (obviously you are doing the work, not the magnet). When you stop pushing on them, the magnet moves them back, restoring their order (less frictive losses etc).

Posted by TallDave at June 13, 2007 06:34 AM

Yes, but creating the field does require work. No free lunch.

You are being repeatedly obtuse, and are completely mixed up on this issue.

Creating a magnetic field requires work. It does NOT require the creation of entropy. The energy stored in a magnet's magnetic field is, in principle, entirely recoverable. Charging a ideal magnet is a reversible process, like compressing an ideal spring.

Now, the power plant that created the electricity that charged the magnet may well have created entropy. But this has nothing to do with the entropy that is created in the plasma, nor can it be used to excuse that entropy creation. In thermodynamics, one is allowed to draw the 'box' defining the system however one likes. In this case, we can draw it around the fusion device, but not around the powerplant. The electricity flowing into the box to charge the coils has no entropy. When I asked you 'where does the entropy go', I wanted to know what is flowing back out of the box to carry that entropy.

We can see you are completely mixed up in another way. The entropy that was created when producing the electrical power that charged the magnet is a fixed quantity of entropy; once the magnet is charged (assuming it's an ideal electromagnet with no resistance) no additional entropy is produced. But the Polywell claims to be a steady state device, and the scattering collisions will create an arbitrary amount of entropy if you run it long enough. Clearly the latter can exceed the former -- just run the device long enough.

Sure. Presumably you are not claiming permanent magnets violate entropy?

Since I have no idea what 'violate entropy' even means, I cannot answer this question.

> Conservation of angular momentumkeeps them from getting too close to the center,

Again, not if the well is deep enough.

If the particle has nonzero angular momentum, then the lateral part of its speed at distance r from the center will be proportional to 1/r. So, simple conservation of energy tells us you're wrong, since the particle's speed (and energy) must diverge to infinity as it approaches the center of the well.

What this means is there's a fixed lower limit on how close it can come to the center, and you can solve for that distance knowing the shape of the well, the particles energy, and its angular momentum. For a fixed depth of the well, the more angular momentum, the farther away from the center the close approach point must be.

Posted by Paul Dietz at June 15, 2007 12:31 PM

For a fixed depth of the well, the more angular momentum, the farther away from the center the close approach point must be.

Tall Dave's point was that for any given angular momentum a sufficiently large potential well can have the required velocity and density around the center for the particle to potentially fuse. The ratio of well depth to angular momentum does not make fusion impossible. It's just a parameter that tells the size of well that is needed versus the average angular momentum that can be built up before fusion.

I'd have to brush up on magnetism to comment on the whole entropy back and forth. I would point out though that permanent magnet field lines drive the electrons right through the magnets themselves. Bussard has pictures with the appropriate scorch marks on them from this(I can't find them at the moment but can dig it up if anyone wants). That was also the 'big' moment Bussard mentions in the google talk were they, as a fusion team, realized what would be obvious to someone working with magnetism. They changed the magnet configuration so the field lines didn't cross anything to cause losses and got the results that have them so excited.

Posted by Kevin Klassen at June 18, 2007 07:22 AM

Paul,

Sorry if you think I'm being obtuse.

Creating a magnetic field requires work. It does NOT require the creation of entropy.

I don't think anyone claimed otherwise. It does, however, create potentials that can act to buffer entropy, as in the example I gave.

Since I have no idea what 'violate entropy' even means, I cannot answer this question.

Sorry, I thought it was fairly obvious in context, but I'm happy to explain. It means that presumably you would not argue permanent magnets reduce total entropy, in violation of the commonly accepted rules of how entropy works.

When I asked you 'where does the entropy go', I wanted to know what is flowing back out of the box to carry that entropy.

And my answer was: it doesn't have to. Like the filings around the magnet, the potential energy in the magnetic well can act to restore them to prior levels of entropy.

But I asked Tom Ligon this same question, and he had a slightly different take: entropy is irrelevant because 1) it's not a hard and fast rule that entropy in a system must increase 2) the tendency toward minimum energy is more important anyway.

Tom also says Rider picked the conditions to fit his conclusions, which others have noted as well.

For a fixed depth of the well, the more angular momentum, the farther away from the center the close approach point must be.

Bussard actually addresses this in the orginal speech: the well has very steep sides and a large bottom. Controlling the virtual anode height (keeping it below the level that it blows out the well) controls whether you have enough collisions for fusion, according to Bussard. As Kevin notes, it's just a parameter.

Posted by TallDave at June 22, 2007 12:53 PM

BTW, it is assumed waste heat would be produced and the system would need to be cooled, so maybe that's a better answer as to where the entropy goes than my initial take (there's some discussion of how this will be done here, where several engineers are working on designing a Polywell device that anyone could build, with some help from Tom Ligon).

Posted by TallDave at June 22, 2007 01:01 PM

Post a comment