In my case it is different, because I operate Z-pinches (https://www.researchgate.net/project/PULSOTRON-500k-SE-500KeV-THERMONUCLEAR-FUSION-REACTOR, see Project log) so the reactants are different: deuteron-deuteron or boron-hydrogen. In last reaction 3 alpha particles are generated, two of them are just thrown in the opposite direction from the other, then it would be very useful to know if ejected alphas would carry electromagnetic energy or not because it should be shielded by surrounding plasma. The problem is that an extreme change in new alphas speed will create an EM pulse that affects only to those electrons close to the ions accordingly inverse square distance law. So I think electrons will move in the same direction as the ions, but with different speeds depending on the initial distance, so I think a simulation must be useful to reveals what happens.

Regards -

Hi Javier,

usually the fusion products produced are positively ionised atom cores and they will create an electric field as you pointed out. However, when they reach the chamber wall they should be either implanted into the wall material or reflected back into your plasma. In any case they will take some electrons from the wall material in order to neutralise themselves. Hence, you should expect quite some electric current flowing through your chamber wall.

May be, but the problem begin just between fusion and hitting the reactor wall, because it must be taken in account to extract electricity from it. The question is: what happens with neighbor electrons during ion launch?, perhaps they flow at the same speed than the ion to neutralize the electromagnetic field created by the ion so no electricity could be extracted.

In the A option, ions can be detected using coils and electricity can be extracted, but in the B option it is not possible

Below is my opinion based upon reading of literature, not based upon experimental data.

Looking at only the atomic scale is useful, to a degree, in this situation. The answer lies in the macroscopic scale. The sea of ions and sea of electrons are different seas at these high temperatures (Te and Ti). The electron sea (plasma consisting only of electrons) electrons move independent of the individual ions, typically. And the catch is the word typically. Does your reaction justify treating the fireball plasma as typical? There is no stable plasma, Te and Ti vary throughout the volume of the event.

In this situation, the Debye length is a critical consideration, as is whether the surrounding plasma is electronegative, neutral, or more positive. The overall plasma charge only slightly changes the answer below by increasing or decreasing the 'wake' size of the alpha moving through the electron sea.

Do read Wikipedia about Debye length, as I think that will alleviate your concern. Though, in your device, the Debye length varies throughout the short lived plasma, and the fusion event generates a 'fireball' where the expanding double layer wavefront will have an extreme radial variation in the Debye length.

Assuming a neutral plasma, with the assumption of a large Debye length (much larger than an electronegative plasma), you are correct that a few immediate surrounding electrons will be dragged after the speedy alpha particles, for a split second, thus heating those few electrons. As the alpha particles zoom into the surrounding electron plasma, the Debye length implies the trailing electrons will no longer be pulled along, as the the attractive force between the alpha and electrons is reduced by the surrounding electrons.

Basically, at these temperatures, there is no pairing/coupling of electrons with ions. "Coupling" could imply a partial ionized state of the electron. The electron being bound in orbit around the alpha is an extremely short lived event, due to 'near' collisional interactions with other particles (e and ions), which would fully ionize any partially ionized electron.

Particularly, after an event that results in high speed positive ions.

Any coupling would be very short lived, very. So short, that in any degree of approximation, it can be ignored (unless the plasma is far colder than assumed). The calculation of the Debye length would let you know how short.

The inertia of the electron comes into play here. The time the alpha particle is near the electron needs to be considered. There is a physics equation that can be derived, to tell you exactly how far an electron is coupled with the departing alpha. Solving this equation for the Te and alpha Ti would give the distance and time any trailing electrons are 'coupled.'

The speedy alpha particle moving though the electron plasma would attract electrons, drag them a short distance, creating a wake following the alpha, but no individual electron would ever be 'coupled' with the alpha, due to both particles being "too hot" (moving too fast). The too hot concept is very common, and I have yet to find/read any equations that expressly prove what I have written.

Your concern, of extracting electricity from the directional ion beam created by speedy alpha particles, is greatly reduced by the high plasma temperature (Te of the plasma sea and Ti of just the alphas).

Instead of continuing to modify my last post, I decided to address the fireball double layer separately. The atomic scale of looking at one event can be slightly enlarged to look at a localized collection of fusion/fission events, in the shape of the fireball. In particularly a subtended angle of the fireball, and how that collection of ions develop over time, moving outward, as two separate plasmas, one of electrons and one positive ions, separate, yet effecting each other in very unique ways. So unique, I am not going to detail it, instead refer you to Wikipedia for "double layer."

Your post envisions a trailing electron, while this post on fireball double layer envisions a leading electron layer. A vast difference, mostly one of scale, to me, at least. Details below.

The motion of ions and electrons in this double layer, before the layer contacts the chamber wall, and ions are discharged/neutralized, should be discussed.

The concept of double layer is very difficult to address in your fireball. Most literature deal with double layers as stable plasmoids, where this outer layer is responsible for reflecting incoming magnetic fields and RF waves of higher frequencies. The actual motion of electrons in a stable double layer can be found in the literature, and for your reactor's fireball event, an extrapolation can be done. UCLA has a research department publishing their experiments on fireball plasma physics. Reading some of their articles may shed a little light on this topic.

The fusor electron/ion motion literature can be enlightening for enhanced double layer understanding. But does the fusor motion apply to your fireball? I can not answer with a yes or no. I suspect, your fireball is much more messy, less controlled, than mine.

The basic ion motion in a fireball double layer is below.

The double layer has an electron layer outside of the positive ion layer. Some of these electrons will be attracted to the positive layer, and move at high speed, to the positive ions. And these electrons will go through the positive layer, and reverse direction, due to attractive force of the ions, and go back through the positive layer, and rejoin the outer negative layer. All while both layers are expanding outward, in a fireball.

That your reactor has a fireball, an outward expanding plasma, is a given. That the fireball is not very spherical is a given. Thus, a stable expanding double layer, with a calculable Debye length is not a given. So, the applicability of a fireball's double layer is only a first approximation.

Now, you must decide if the small scale individual alpha/electron is replaced by fireball physics, and do calculations based upon expanding double layer, if the double layer exists for your fireball. I imagine it does, just not as simple as mine. I am considering PIC with varying cell size based upon radial distance from the initial fusion center.

Directional jets, one positive and one negative, is a common form for a fusion reactor to collect plasma energy converting it into electricity. Your post implies a different electricity generation means, one of positive ions impacting the chamber metal walls, creating a discharge current, that can power an external load (rectifier with associated circuitry to convert to AC or DC for grid distribution). Thus, for the expanding fireball double layer, each layer has a different arrival time at the chamber wall. If an outer negative electron layer actually exists, then it would have to be taken into account. But the two layers will have different arrival times.

However, as pointed out, your fireball is not spherical, having nice layer arrival times at the chamber wall. Instead, envision is a sole layer of alpha particles, with their +2 charge, to steal 2 electrons from the metal wall. And perhaps your double layer has a trailing electron layer? The 'wake' mentioned in my previous post. However, I believe this wake only serves to heat the electron plasma sea, and thus the wake is dissipated before it reaches the chamber wall.

So, my two posts have likely greatly changed your perception of your reactor event's by products. I hope it helps.

Well, my idea of the plasma reaching a plate was to measure alpha emissions, not energy generation.

I have measured a similar double layer in some tests, as can be seen in the attached plot. The problem is that the plasma ball generates a strong electromagnetic field that burns plastics as can be seen in the photo, due that events, in the last tests I had several internal blasts in the Pulsotron destroying several parts of it (now it is out of order). I have measured about 13 kiloteslas, which is still low, so I will have a lot of troubles to increase that field. I attached a photo with plastics burning cloud and also the double layer negative-positive measurement.

If the difference between possitive-negative layers is very short a high-speed oscilloscope should be needed (I have two Gigasamples/second), what can be the time difference between both layers?

I will read about UCLA work, do you know somebody there?.

in order to recover a large amount of energy, it should be low frequency (40khz or less) or the opposite: >400MHz. It should be interesting to know what is the approximate frequency and the number of cycles that would reach the electromagnetic energy collector systems

Best regards,

Javier Lopez

Hi Javier,

I have worked with Prof. Stenzel from the UCLA on most of the papers Peter is referring to but I am not sure what fireballs would have to do with your experiments.

cheers

Johannes

I am building two sensors to measure the double layer at 50mm distance, what could be the time difference between arrival of ions-electrons? I need the right oscilloscope, now I have a 2Gsps one

If you have a weakly collisional plasma, you can just use the time of flight between the center of your chamber and the chamber wall in connection with the electron/ion temperature.

Have you done any spectral analysis for alpha/electron recombination? I am looking for high sampling rates over 100,000 per second, for my fusion chamber.

The time difference depends on the initial velocities of both electrons and ions (as Johannes mentioned the Te and Ti), if you know it. Otherwise, you can measure it with finding the V=r/t, where t is the delta time from ignition to peak of the electron or ion measurements, and r=50 mm. At 50 mm, it will not be much time difference, but you should be able to measure it with the 2Gsps oscilloscope (I would not risk the oscilloscope).

Do you know what voltage you must protect the oscilloscope from? Do you have a high voltage probe? A red probe, about 1 foot long, or 2, and 1-2 inches in diameter, to handle the stray surface voltage creep, and an internal resistor of 10 or 100 meg, I would think. And then hope for the best. Given the damage rate already in evidence, ruining an oscilloscope ...

I would not be risking an oscilloscope for this measurement. Low cost RPi ADC comes to mind, where losing $50 of an embedded computer is not an issue.

I refer you to the below cluster explosion experimental techniques for other methods.

Cluster explosions caused by femtosecond short burst laser beam articles give a methodology for measuring both the faster electrons and the slow ion wave fronts arriving at the sensor. While I enjoyed reading their papers, and plan on using similar, it will be much later this year, or next before we need that type of data. They give all the math they used as well.

The UCLA plasma departments number over 20, two in non physics, engineering of diagnostic probes, and Johannes collaboration there, are quite interesting reading. I found the fireball work the most interesting, some real basic plasma science, highly overlooked for decades, is getting done at that lab.

While the experimental setup for the UCLA fireballs is not really an "explosive expanding fireball", it might still give insight into diagnostic techniques, was my thought in posting.

Plasma abrasive is a real severe issue in explosions. Do look up what the explosive developers use in their testing. I understand metal probe tips abrade within seconds to mere nubs, and become useless after 1-3 trials, and must be replaced.

What type of plastic is in use? I can not imagine anything would not be damaged over time. PEEK is first choice, and it has 2-3 better plastics now, but no plastic handles over 500 degrees. It all softens at prolonged high temperatures, and deforms shortly thereafter. I am most interested in better, so if you know of something, please post.

I do know if I was doing explosive plasma experiments, then sapphire protective sight glass, to sapphire optical fiber, with ball lens, and high speed digital cameras (10 million frames a second?) might be the order of the day to capture visual and spectrum of the initial ignition loci. The reason for the sight glass is to prevent damage to the fiber end. A cheap replacement cost might be normal glass, as given the explosion degree, the sapphire sight glass will get damaged. That is if you think there is anything of visual (IR to UV) interest at the ignition point... if you know where that is, and it is a constant location.

Well, I made an initial test that could be used to measure the time of flight of ions and electrons. I used a high voltage probe so the oscilloscope survived (this time), but the scale was wrong because the pulse voltage was extremely high, but the copper sheet did not increase the temperature, so the energy goes electromagnetic. The low voltage power supply blown the fuse and I had to stop test...again. -It is very usual that Pulsotron-3 blown not robust parts-

Have you tried to put a big resistor into your measurement circuit? You can also use the current pulse for TOF measurement...

I made the test at end with the right channel voltage but had a lot of troubles in the journey (2 other fuses blown, a dc/dc converter is dead and all the arc covered by expensive accurate pt100 blown under fire), but the test is done. It appears a great negative pulse then a positive one at about 10khz. It can be exactly known the current through voltage loaded at the capacitor. The output voltage is fixed but not the energy so I suppose that using a larger capacitor more energy could be obtained.

Unfortunately, about 30% of the arc was burned but I repaired it but I had again internal explosions in the Pulsotron-3 and I had to stop (switching it off while it makes shorcircuits and blasts was too much for me)

Here is the plot (the blue line)

I would like to make the test using a larger capacitor but also without capacitor as suggested Johannes, perhaps tomorrow, but I have to repair the Pulsotron-3 first

Also the burned arc photo during its autopsy

Congrats on getting the e- and positive ion pulses. You certainly got a TOF for both peaks. These are separate peaks, which leads me to ask, did you see the trailing edge of the electron peak be very sharp, and 1/3rd of the way back up, the positive ion curve began, below zero. That is, the electron wave built up on the leading edge, and after the peak had a very sharp drop off. This means the electron expanding wavefront had an initial build up, and then few electrons were striking your measuring device. That is my interpretation. Another possibility is the oscilloscope probe was not a match for the oscilloscope input, a possible impedance or other mismatch, resulting in the curve shape?

The trailing edge of the positive peak is also sharper than the leading edge, I think.

Do integrate the area under both peaks. For total charge comparison. With the negative peak trailing edge being sharp, this integration will be difficult, to separate the negative peak from the positive peak. With the two peaks so close together, I assume the chamber is small, for the velocity of the ions, so the peaks merged together. Electrons being so much lighter, I was expecting the electron peak to be well separated from the positive peak.

The negative peak magnitude is about 25% larger than the positive peak. Seems normal, in that the electrons bunching together, having a smaller wavefront thickness, than the positive ions, would cause this. The thickness of both peaks is meaningful how? Just the curious again. I expected the peaks to be of different thicknesses, as electrons are easier to accelerate, they would leave in a bunch, much at the same velocity, compared to an assortment of different positive periodic elements in the positive peak, which would broaden this peak. Do you know the each element in the positive wave? Seems their AMU are very close to each other. Perhaps it is just one element? The flight path is likely too short for any separation to occur. Or the means of acceleration was not highly dependent on their AMU, likely due to ionic charge level.

Curious about the aftermath oscillations, after the positive max peak. Reflections perhaps from the far chamber wall? Explaining those will be a challenge? The negative signal, instead of zero, is also of interest? Charged the chamber wall with electrons?

How did you ground the oscilloscope? The chamber? For high voltage measurements, grounding the oscilloscope properly to avoid damage to it, like blown fuses, can be a challenge. Making grounding changes can make the damage worse, so do ask some experts. The oscilloscope manufacturer would be the first one I would ask. And ask the maker of the probe, if you are using it within reasonable purposes, or should you try another model. Just re-iterating Joannes point.

Last point, thanks for the picture of the burned arc. Curious it burned in the middle. That must tell you something of why, and thus how to avoid it. I would think. Asking an expert in such would be enlightening. I just know the burning I would expect to be at one end or the other, where the current exited (e.g. electrons entered). Being in the middle ... perhaps there is some other piece of equipment near that location? That is my first thought. A resonance buildup is my second. A winding error is my third. I am not expert in this area, so these are just my guesses. Getting a hot spot in the middle is what drives my guesses. It could be something else? Does the picture also show burning at the left end?

Well, it is clear that can be troubling due to coupling with trigger return of the scope. I was going to improve the test to use separate oscilloscopes for each capacitor and better ground decoupling (to improve the ground star distribution to avoid magnetic coupling), but I had a high inrush current in the capacitors and blew up wires and discharging resistor. Still, I do not understand how the 47kohm 5W resistor could be blown. Also, in the last test the main capacitor bank coupled through the plasma and received also powerful ray when wires were blown (it is similar to an opening switch discharge) and I had to dismount to test the capacitor bank. I checked that 70% of the capacitors survived.

It is very clear that I have to buy a lot more powerful capacitors.

I do not mind to blow one or two capacitor banks if being successful, but I am losing data, I should need some help making more sensors and diagnostics

I can make the integral to check the areas of the positive and negative phases again the DC voltage at the recovery capacitor at the end of the test.

About the burning of the center of the arc is due after the confinement the plasma ball explodes radially hitting directly the center of the arc. I have to protect it more.

Also, I should do the test under vacuum, to check the electromagnetic expansion without ionizing air but I do not expect many differences in the compression because I made tests under 60 gigapascals pressure with similar results but it would change the electric recovery. I should prefer high pressure because it helps to cool down the system.

I have bought a lot of very powerful and high current diodes that could help to improve the tests.

I tested at least that the energy recovery is correct by using scope probes in differential mode and star ground lines configuration. Still, I do not know how the 47kohm resistor was burned, but now we use a lot more powerful ones, cooled and using large cross-section conductors

I recall when I went up the learning curve for High Voltage Capacitors, Pulse Power Capacitors, and Tank Capacitors (same thing, just a different name). Their construction is vastly different from typical/standard caps. I do not know if you have gone up this learning, so I mention more below.

Pulse Power Capacitors

Pulse Power Capacitors are rated differently from regular capacitors. Start with the "High Voltage capacitor" spec sheet parameters. There are 7 spec parameters for HV Caps as compared to typical caps of just 3. It's not a difficult learning curve, but you should be aware of the extreme difference in construction method, internal geometry, materials, etc. About 3 times more complex than typical capacitors. Understanding the reasons for capacitor failure will lead you to better capacitors. The 7 parameters deal with this failure aspect, though it's not readily apparent as how, until you find a design requirement paper that explains each parameter and why it was added to the tank capacitor spec sheet.

Caps can tests fine with a capacitor tester, but the cap can still be damaged, number of burn through spots is high, so the next test the cap fails. Why? Testers do not use high voltage or high current, which are the two main causes of failures. So, testers can give false positives. Again, understanding how the foil, poly sheets, oil, leads each have different or combined failure modes will be key to capacitor type selection.

Building your own is likely cheapest. Wind them consistently tight is key. Training takes about 10 times, before the burn through rate drops. Buy capacitor metal foil, rounded on the edges and double smooth (shiny) on both sides, for very flat, consistent thickness to prevent hot spots that lead to burnout spots. Double poly sheets is better, I feel, or triple, depending out the required capacity rating. The consistent thickness of the special cap poly sheets is critical, to avoid burn out spots. Certainly must be oil filled, to dramatically reduce cap failure rate. The oil aids in equal spacing between the foil sheets, the poly sheets, so fewer hot spots develop, that change to burn through spots and then failure. A few cap manufacturers can make these, most can not. I would expect 1 in 10 you contact can make them.

Anything off the 'standard' shelf will always have a high failure rate, except pulse power, tank, capacitors will last a long time, and survive dozens, if not hundreds, or thousands of events. You want tens of thousands events ... eventually for energy production. If not hundreds of thousands, where best would be millions, if you can achieve that MTBF. For energy production, you must.

Use caps rated at least twice as high as the maximum voltage you are expecting. Even higher is better. The web now has several design specs for pulse power caps, hard to find these pages, but well worth it.

High Current Issue with Resistor and Diodes

There are many types of resistors, and my opinion would be based upon the type you are using. And how, where it was blown. I suspect the issue is with the "interface" surface between the metal, near zero resistance, conductor interface surface to the resistive media surface, where abnormalities in the manufacturing process might produce defects that do not withstand a steep inrush current. That is, the end of the resistor is blown, not the middle. Second suspicion is the middle of the conductor is blown due to high heat from the inrush current, or more likely from the current over time building up heat, where the short duration event means near the end of the event. A larger cross section, and stronger outer body with superior heat conductance might prevent it. 5 watt is not much rating for what you are doing. I would go with 50 watt or much greater.

What is the purpose of the resistor? Limit current? Set the output voltage? Both? Zener diodes to direct over voltage current to ground might help? Get the new, heavy duty type with equal volume P and N material?

Sensors and Diagnostic Tips to avoid Data Loss:

Change from Oscilloscopes to Arduino ADC

The Arduinos brand is likely too slow, at 8 or 16 megahertz. Look alike brands go up to 70, or even 170 megahertz, and likely more today. An Arduino has five ADC GPIO pins for sampling 5 different circuits (voltage or current). For faster sampling just use one ADC pin per Arduino. Protect the Arduino from over voltage and current with zener diodes (5.1 volts) and even optoisolators. Optoisolators are cheap to replace, just use a socket. Have the Arduino send their data to a Raspberry Pi, which then sends it to a Laptop, for storage and real time display, perhaps using NI's LabView software. Post event analysis with LabView timing diagrams will be easier.

If Arduinos are not fast enough, then get dedicated ADC circuit boards, though they are more expensive.

Using Arduinos as a front end to a laptop display oscilloscope software is far cheaper than buying 20 data storage oscilloscopes.

There are thousands of web pages that document the above, with an abundance of available open source software, free to download.

Johannes Gruenwald - I mentioned fireballs as they have double layers. Measuring the TOF of the double layer of an expanding fireball was my point. I recall the UCLA research may be of static fireballs created via a spherical screen ... my recollection is hazy as I did the reading of UCLA fireball research 6-8 years ago.

Also, other published research on fireballs, like cluster explosions, particularly senior thesis and postdoc papers, explain in exact details the diagnostic equipment, methods, and much more, including TOF measurement issues and how to overcome them.

The TOF is not the only issue, but also the impedance. I measured different impedance of the negative and positive layer. It is an expected behaviour as long as negative ions has different mass and speed than positive ones

About TOF, there is an FFT using phase instead of frequency: https://www-leland.stanford.edu/group/Zarelab/publinks/686.pdf

I look forward to your next report. Thanks for the link to the HT-TOF MS paper.

Well, in this plot (in pulsotron 500k in project log), it was obtained following plot:

https://www.researchgate.net/profile/Javier_Lopez27/project/PULSOTRON-500k-SE-500KeV-THERMONUCLEAR-FUSION-REACTOR/attachment/5ebc509eead4db000155408a/AS:890821104644096@1589399710202/image/S3033.png, it can be seen that the magnetic sensor resonates at the LC corresponding to energy recovery coils instead to energy discharged of the capacitor bank, si it is clear that main capacitor bank and secondary ones are coupled trough generated plasma when it expands after the compression.

I am sorry due I am slow, but I have to study dozens of megabytes of data after every test. Fortunately, I only make 3 tests every week

Below is my initial feedback, and later, when I have reviewed the graph, I might have more.

You do have your work cut out for reviewing the data. Very necessary work. Explaining the stages and variations within each stage, very hard to do.

Plasma can act as a battery, putting current back into surrounding electrodes, sensors and coils, that created the plasma. Quite an issue, as the voltage and current are quite high, and can burn out external electrical components. Yes, diodes, big ones, would be desired to block the reverse current, or zener diodes to dump the over voltage current to ground (but you want to collect most of that energy...). Along with noise capacitors for the grounding the high frequencies.

Plasma can also act as a capacitor, charging or discharging, not with the expected exponential curve. Or inductor. And a resistor in a resistor network.

Plasma can do all of these four functions, at the same time, or varying during the discharge. Not necessarily in the same circuit path, but there is more than one circuit path the plasma participates in. Each path must be mathematically modeled separately, yet simultaneously. Think a large matrix problem statement. Given your plasma states, deriving the equations that go into this matrix, and solving... very hard problem. Good luck.

In the simulations diodes not works well. I am making several tests, but I should make some high power plasma impedance measurement during the discharge. There are not papers about that

I were reviewing the data obtained last week to compare with this week and I discovered a very interesting issue:

In one of the tests the magnetic field reached 62 kiloteslas, but after that, the magnetic field remained at 25 kilotesla level during 5 microseconds, that is a lot of time. It is like the plasma created its own magnetic field. I observed a similar phenomenon during Pulsotron-1A tests a lot of years ago, but not at so much high level

62T is high, and without an easy discharge path, the stored magnetic field around the device (and inside) will take time to decay. It is similar to a car's ignition coil, only that coil does have a high discharge rate. Once an external magnetic field is created, by any method, like the fusion 'explosion' being asymmetrical, the plasma will oppose that magnetic field, by creating eddy currents on it's outer layer, which in turn creates a magnetized plasma, with it's own B field. This B field will vary over time, opposing the external field, which both decay in sync. Given you are capturing ion motion in the external coils, to create electricity, these coils with a current in them, will create a magnetic field... The LPP fusion project has does much computer simulation to match their experimental data and apparently have solved the math model of what drives the fusion event and ion jets that explode outward, that their device converts to electricity. The LPP papers may interest you, particular how they solved issues, likely the same issues or similar that you face.

I am surprised due in the first two tests made to try ignition, it was received more magnetic field than without fuel. We did not expect that behavior, but we will need a more powerful oscilloscope (and new sensors and new capacitors and engineers)

What are the LPP papers?, I read one where explained it was produced 1e10 neutrons

Tomorrow I have a lot of possibilities to blow several pyrex glass coils

https://lppfusion.com/ also called Focus Fusion