[asymmetric_implosion]

That’s an ambitious list.

The switch upgrade is a great idea. I hope these Raytheon switches are the ones developed for the next gen marx generators and LTDs. If they are, the switches are highly reliable and can hold off far more than 45 kV using only dry air. The data from the LTD tests at 200 kV, 1 Hz were very impressive. I know Raytheon recently purchased K-tech. K-tech was involved with the LTD work at Sandia for the 1 MA LTD modules. Moving up to the theoretical limit of the machine would be interesting.

The nitrogen/deuterium mixing should be interesting. N2 is an interesting gas. It will be an interesting study since most people shy away from heavy diatomic gases in a PF. I’m not a fan of oxygen in my PF. My limited experience with nitrogen was mixed. Some days it ran really well and other it was poor. There was no obvious rhyme or reason to the whims of the machine when we never broke vacuum.

I hope the yield number is wrong…66 kJ per shot with a 5 MW plant means 75 Hz operation…so 100 Hz operation after all is said and done. To my knowledge, one one group has operated a PF >100 kA near the 100 Hz level and it required substantial effort and the yield was well below the optimum with soft x-rays. There are substantial problems moving up in repetition rate. I’m fighting a few right now as we move one of our machines from 1 Hz to 10 Hz. It is not as straightforward as one might believe. A combination of heat, residual ionization, anode erosion and chemistry have hampered our efforts to demonstrate something near the 1 Hz neutron yield at 10 Hz. The anode erosion has been a killer. Our little machine at ~60 kA is eroding anode material at ~10 ug/shot. One might laugh at this value but the beam can be highly focused on a small part of the anode. I’ve bored holes over an inch deep into SS304 anodes. According to e-beam scaling predictions (Stygar 1982), a 1 MA PF will be far worse. If someone is going to say use a ceramic, don’t bother. Ceramics are the devil. The anode base must be metal.

The engineering issues have become far worse than the physics problems with encountered. Repetition rate introduces a set of new problems that a few shots a day will not encounter. Just a heads up in case anyone thinks it will get easier once Q>1 is demonstrated.

[asymmetric_implosion]

To give you an idea, PF papers appear at the rate of 100-200 per year since 1990 with some periodic behavior. They tend to spike in years with the Dense Z-pinch Meeting every three years. The percentage of these papers that deal with neutron production and fusion vary but they do make up a reasonable portion of the papers. A colleague is writing a review article on the PF for a journal so he did most of the digging into the history of the PF. I will see if he has a formal list to start with but my guess is he has a set of notes he understands with the key physics papers that make up his review. The paper should appear in a Special Issue of IEEE Trans. Plasma Sci in Dec 2012.

It is sad on some level how little progress has been made since the 1970’s. If you are looking for innovative ideas look to the 1970’s. Some new ideas have come up but most of the key physics and ideas were established long ago. Even concepts of what are referred to as plasmoids on this board are described back in the day. The old rule holds true, science is rediscovered every generation. Think you have an original idea, search papers from 20-30 years ago and you will find someone already did it. PF is old enough this rule holds true in general. Technology advances so more precise measurements can be made but to put things in perspective, LPPX published an image of a plasmoid in the PoP paper, similar images exist from the 1970’s. The resolution isn’t quite as good but 5 ns framing camera images are shown in these papers.

If you want the fast way to get the papers, there are about a dozen groups that publish regularly on the PF worldwide. Most of the university groups post their publications on-line. The NTU/NIE group does this frequently and their archive is pretty up to date with their own papers. This is a resource for expanding the search to the other groups such as the groups in Poland, Chile, US, Pakistan, India, etc. Those publishing in the field are citing the key papers in many of the articles. It doesn’t take long to put together the most important authors. My personal list includes two people that usually do a fine job keeping up with all the published work; Sing Lee and Pavel Kubes. The rest of us in the community cite their work for one reason or another.

[asymmetric_implosion]

Gorgon can model a PF until you reach the implosion. It is, in principle, no better than any of the other large lab codes. Gorgon has been used on Z-pinches for some time now at Imperial College and more recently at Sandia. My group submitted a proposal to DOE to work with the Imperial College guys to model the PF adding onto Gorgon so it can model the pinch using PIC methods, but I don’t think the funding folks are behind it based upon the latest feedback. PF is too “well known” to be worth studying at the 5 MA) is too uncertain. All in all it is the view of DOE reviewers that PF isn’t going to work for the problems of interest for DOE/NNSA. NIF is the only cathedral anyone cares about right now. If your experiments are not supporting NIF and/or not scalable to NIF then please see the door. I have to hand it to Livermore; they marketed NIF beautifully at DOE and around the world. At the HEDLA meeting everyone was buzzing about NIF and what it will do. To be fair, they have done some amazing things with radiative shocks and extremely high pressure compression at low temperature. My primary concern is the cost. NIF cost like $4B while a pulse power machine costs more like $40M.

[asymmetric_implosion]

I’m sitting at the High energy density laboratory astrophysics conference so I found your post funny. No one I know of uses PF devices for so called lab astro problems. The closest thing I know of is using modified Z-pinch to model jets formed by various bodies and colliding them with gas jets or gas blankets. Look up experiments on the MAGPIE generator by Sergey Lebedev and his group if you want to go this route. They started this work a few years back and in my opinion are still the best physicists working the problem.

[asymmetric_implosion]

I will check on a step wedge reference and report back. I am familiar with the step wedge concept from my grad school days as it was included in course work.

Don’t rule out neutrons just yet. I thought the same thing initially and the little buggers ruined a week. If you consider that the neutrons are traveling at ~2E7 m/s and your chamber is something like 10 cm from the source, you have a delay of 6 ns before the neutrons hit the first wall. In your Physics of Plasma paper, you show two x-ray pulses that could be 6ns apart at the peaks. The neutrons produce an ~850 keV photon in some fraction of the reactions with the wall. That photon would give the appearance of a hard x-ray spectrum when it was really nuclear in origin. Another thought is reactions with your copper anode. We have done some calculations that suggest our hardest x-rays might be from both the anode and the chamber. More to do to confirm. A test with hydrogen would lend some insight into the problem. Just double the operating pressure to keep the mass the same as D2.

I was surprised to see the x-ray emission from the anode rim but I’ve recorded shots in most of the common PF gases and the feature is universal. I can’t speak to its absolute strength but some images I’ve collected show the rim is as bright as the base of the anode when looking down/up at the anode. I didn’t use a step wedge to look at the x-ray energy in each region.

[asymmetric_implosion]

Is the tip of the anode blocked as well? There are plenty of images of brems emission from the tip of the anode as well as the base. My working hypothesis is electrons that were once confined during the pinch escape the pinch region when the current falls below the confinement threshold. That explains why they are late in time and why they are hard. Another option is current restriking to the anode after the pinch but I’ve never observed any electrical evidence to support a current restrike near the tip of the anode. An imaging diagnostic would help verify the shield placement. I never trust detectors without images to verify what they are seeing.

I attached an image of an SS304 anode with a hydrogen pinch. The image is integrated over 100 shots at 250 kA. The image is of the >10 keV spectrum. The anode wall is too thick to see the anode base but the tip of the anode is clearly visible. The really striking feature is the lack of a pinch. Too little mass density in the pinch to produce significant brems while Fe does a fine job of converting electrons to x-rays. Put argon in instead of hydrogen and you get a nice visible pinch.

I suggest switching to a gas that does not produce neutrons to verify that your x-rays are not really gamma rays. SS304 vacuum chamber can produce (n,gamma) reactions that lead to 800 keV photons. The reaction relies on fast neutrons so it could explain the time lag in the second pulse. Neutrons are slow compared to photons. Have you considered using a stepwedge spectrometer to measure your photon spectrum? They are cheap and easy to setup. The data analysis requires a little work, but flat response films like GAF film reduce some of the problems. It is perfect for machines that can replace the films between shots. Film to digital reduction does not require developing with GAF as it visibly darkens from white to gray. A modest quality scanner is all that is required to convert from film to digital data.

Attached files

[asymmetric_implosion]

E-beam does not need to be born in the plasmoid. Any region of strong electric fields near the pinch region can generate runaway electrons. Most observations suggest that the runaway electrons are generated in low density regions outside what LPP calls the plasmoid region. If this observation remains true, you still have a runaway e-beam problem and the x-rays that go with it. This assumes that a plasmoid can confine all the electrons as the theory under test describes. Again, LPPs recent results suggest that the plasmoid has yet to confine electrons as the x-ray spectrum was very hard; much harder than a 140 keV thermal spectrum suggests. I can’t speak to the theory in detail but the experiment seems to disagree with the theory up to this point.

[asymmetric_implosion]

jamesr wrote: Given that the bulk plasma in the device should not cool between each pulse below ~800C or so, you could just extend the helium gas cooling needed for the anode to the rest of the walls (making them out of something like tungsten), and run it through a brayton cycle turbine.

But as has been said before it would be much better to avoid having to use a thermal cycle and just extract even just 10-15% directly from the xrays to push over the Q=1 and dump the rest as waste heat.

asymmetric_implosion: The 500kV pinch voltage does not relate to the energy of the x-rays. The x-ray energy peak from bremsstrahlung is a function of electron temperature (http://en.wikipedia.org/wiki/File:Bremsstrahlung_power2.svg ). If the electrons are at ~150keV then the bremsstrahlung peak will be around a quarter of this – so around 30-40keV.

Brems in the plasma might agree with a thermal spectrum if the electrons remain confined in the plasmoid but brems from the electron beam will not obey a thermal spectrum. The run away electron beam that impacts the anode is much higher energy. It hast to be to escape the B-field. Literature from as early as 1977 (Krompholz et al Appl Phys 13, 29-35, 1977) has verified this. The mean e-beam energy is near the pinch voltage with electron energies up to 1 MeV in ~100 kA machines. It gets worse as you go up in pinch current. If the electron beam remains the dominant x-ray sources the brems spectrum will be harder than the thermal spectrum prediction requiring a thicker onion. I’ve been fighting this particularly vexing problem of the hard x-ray spectrum for a couple years because it complicates my application for the plasma focus.

If the x-ray emission is dominated by the plasmoid emission, most of the x-rays will be lost in the vacuum spool before the onion if they are separate pieces. Based upon LPP’s recent release of copious x-ray above 100 keV it seems that the beam is still king as 30-40 keV x-ray would have been significantly attenuated by the copper filters. It seems pretty reasonable for a 1 MA pinch that the mean x-ray energy is near 250 keV (typical pinch impedance is 0.25 Ohm with ~ 1MA so 250 kV). To my knowledge LPP doesn’t measure the voltage during the pulse outside vacuum. Pinch voltage is hard to back out without both the current and voltage outside vacuum. Published techniques exist to calculate the pinch voltage if the data is taken. You can use models to estimate the pinch voltage if you wish but I’ve always preferred a maximum data, minimum model approach.

[asymmetric_implosion]

More than double or tripe, probably like 10X, but who cares. The onion is described as a low cost, easy to build solution. If you can convert the photons from fission at 80% to electricity you gain another 3-5% efficiency on a fission plant. That is game changing in the power industry and probably worth more than $100M. That would build a nice PF test facility and demo reactor.

Jamesr: thanks for the ref, NIST XCOM is my standard data base for these calculations. For a standard SS304 vacuum wall, you lose sensitivity below 40 keV. For a pinch, mean x-ray energy is near the pinch voltage. For a 2 MA PF, the maximum pinch voltage is around 500 kV so you can expect many x-rays around 500 keV. If you do a better job of pinching as expected from LPP you could get up to 1 MV as the mean energy. Go up in current and the pinch voltage increases along with the mean energy of the x-rays. Consider the anode converts most of the runaway electrons to x-rays in existing PF devices. If the anode is Be, your electron energy is likely lost to heat. If the plasmoid is the dominant converter it will be B doing most of the converting to x-rays. There is a reason bremms x-rays sources use high Z converters. One might argue that less x-rays is a good thing. According to Zapkitty these x-rays are supposed to push you over the top to Q>1. It seems pretty convoluted to me.

As I already stated above, fission produced 700 keV photons nominally. It seems that the two system are in the same ball park. The x-ray spectrum is a bremmsstrahlung spectrum and it favors lower energy but it is not stretch that LPP will be producing x-ray that are already produced on operating systems that would make excellent test beds for the “onion”.

[asymmetric_implosion]

Photoelectric effect is used all over the place to detect radiation but never in a power conversion configuration. It seems like a daunting problem considering the plasma interacting with the first wall of the onion leading to eddy currents and other electrical noise that will screw up the collection of photo electrons. The onion seems easy enough to test now. In fact, it seems that it should be tested before the PF. If LPP shows a viable means to convert x-rays/gamma rays (more layers in the onion is all you need unless you are worried about Compton scatter), it could be implemented in nuclear fission plants. That would provide a revenue stream in licensing or sales that could support the PF development while improving efficiency of the electrical production in the near term. It seems the whole idea was approached backwards. Use money from the easy thing to support the hard thing, if the onion is as simple as producing photo electrons and gathering them.

[asymmetric_implosion]

I hope the photovoltaics work as described. Is there test data to support the 80% efficiency or are these theoretical calculations? If they do work why aren’t they deployed on every nuclear power plant? Tons of gamma rays are produced and could be converted to useful electrons.

I hadn’t considered using the heat for buildings and such. The loss seems significant but probably better than a turbine. The Russians have some data on this since they used fission plants to heat towns.

I guess the big question is will a power conversion cycle be needed at all?

[asymmetric_implosion]

I find it surprising that neutrons are produced in lightning mainly because of the fuel. I hope someone can nail down the reaction. I would be interesting. To my knowledge no one has observed neutrons from nitrogen pinches or argon pinches. It might be interesting to test air in a pinch and see if neutrons are produced. The real problem seems to be dose rate and shot rate for a plasma focus.

[asymmetric_implosion]

Brian H wrote: a_s;
“probably thermal cycle (same old same old).” Not a watt, unless you count such possible applications as heating nearby buildings with warm air, etc. The ancillary equipment to extract heat and convert it to electricity is not a) present, b) economic, or c) usable with the low-grade (low-temp) output of the FoFu.

As for the photovoltaic “inefficiency”, examine the “onion” design. It has perhaps thousands of layers/stages to extract a very high percentage of the X-ray energy.

And as with Z-pinches, sacrificing wires or metal cylinders or pellets etc. with every shot is, IMO, a complexity bridge way too far for continuous output. Murphy’s power grows exponentially with number of components.

Low temp? I’ve seen numbers of 1000C floating around this site. That is a pretty good temperature for a thermal cycle with modest efficiency (~40%). I’ve looked at the onion. It will absorb x-rays but how much will turn into useful electrons? That is where the efficiency typically comes in. You can absorb 100% of the sunlight but only about 10% is converted to useful electrons. X-rays may convert more efficiently as they are well above the band gap but you are still limited on efficiency. I hope it is greater than 40%. I know heat will be generated by a FoFu power plant so the question is why not use it? Turbines on the 5 MW scale are small and efficient. A helium turbine in particular is very compact and efficient. It adds more electrons to the grid. Some cooling is going to be required of FoFu-1 so why not put that heat to good use rather than dump it into the air needlessly.

Z-pinches: I’m not saying that MAGLIF will ever be a viable power plant. I don’t think NIF is viable either. I’m simply saying that NIF and MAGLIF will reach breakeven. It will be a demonstration and nothing more. FoFu will demonstrate Q>1 well before a power plant is built if economically feasible. Using best guesses for numbers, NIF is producing something like 10 kJ of fusion energy per shot (DD) using 4 MJ of laser energy (I believe the bank is 400 MJ with 1% conversion to photons). MAGLIF will use 20 MJ of pulse power to get something like 10 kJ of fusion energy in the near term. FoFu uses more like 56 kJ to get ~1J of DD. So the math suggests that NIF has a physics Q of 0.0025, MAGLIF will couple no more than 33% of the stored energy to the load so MAGLIF is 0.015 and FoFu using the same 33% coupling to the pinch as MAGLIF has a Q of 5.4E-5. If all things were equal MAGLIF would get there first but NIF has more funding and they are running “bad loads” right now to test the models. When they get serious later this year I expect they will get to Q=0.5 pretty quickly (yes, physics Q). Q=0.5 puts NIF on par with JET which has demonstrated a Q=0.7 some time ago. MAGLIF will suffer from something like 10 shots per year so unless the models are spot on, it will take some time. FoFu, the smallest project with the least funding, will come in third on this list. One can argue with the exact numbers but they should be order of magnitude correct. Resources are likely to shape the results and NIF has the most resources by far. The first gate will be the physics Q>1 followed by the engineering Q>1. FoFu has the most straightforward road to get from physics Q>1 to engineering Q>1. NIF and MAGLIF have some serious hurdles after they show physics Q>1. The big question is whether a plasma focus (a two stage Z-pinch) can be made to show Q>1. I look forward to the answer over the next year or so.

[asymmetric_implosion]

James: FF will only burn a tiny amount of fuel per shot. My guess is it will be on the same order as an optimized NIF. NIF may not be perfect and it will not be a viable fusion reactor but physics breakeven is likely. Engineering gain is another story…

Brian H: MAGLIF does not require a complex B-field. The magnetic field is the same as FoFu field (solenoid-like field). The difference is the magnitude. FF uses very weak fields while MAGLIF uses an ~1 T field. Both systems use flux compression on some level. The key difference in the SNL experiment is the B-field due to flux compression is strong enough to limit electron thermal conductivity to allow the plasma to stay hot. Flux compression is a proven method to increase B-field and gain some of the benefits described by Lerner’s theory without the formation of a plasmoid. The conversion of MAGLIF to MHD conversion is do-able. Of course people will use the neutrons to extract heat and that is less efficient theoretically than MHD conversion but MHD conversion has not been implemented on the large scale (to my knowledge). MHD pumps move liquid metals but that is a far cry from what is being proposed on FoFu-1. A simple rogowski coil may be enough to extract the beam energy which is efficient but the rest relies on photovoltaic-type technology (low efficiency) and probably thermal cycle (same old same old). I don’t think anyone can claim to know the exact ratios of these three outputs until it is tested.

Time will tell who gets there first but if I’m betting NIF is first. FoFu may beat Sandia but I could see SNL getting their next for one reason: resources. MAGLIF could produce an intense neutron source that NNSA (folks funding NIF as well) would like to have.

[asymmetric_implosion]

MAGLIF is far from a NIF boondoggle. They did the testing up to this point on a dime (in big science terms). Z is already paid for so the proof of concept experiment is all that needs to be completed. If they have the funding, ~$5M, they could probably get it working in a couple years. The real problem is shot time on Z. You need to get in line a year or more in advance and the Z operations folks don’t like neutrons. I know they have done some tests with the metal shell with hydrogen. The magnetic field coils are about ready. If they are in the queue, they could have data by end of the year showing a Q of 0.5-0.8.

Bad news though, NIF will hit breakeven first. It is dubbed too big to fail by NNSA and others in DOE.

[Author]

Posts