EXP: Ion Beam Focusing by a Loop Electrode

[Wilfried Heil]

We did these tests some time ago and although they are not finished yet, I will post a few images, after Jon's recent experiments with solid and perforated spheres.

Our grid was replaced with a single and later a double wire loop electrode of 32 mm diameter. The size was the same as that used previously in the 4-loop grid. This loop formed a single linear ion beam through its center, which appeared very thin and well focused. The operation of the fusor was stable.

The adverse results were a strong localized heating of the shell, where the beam's ends formed two red hot spots, and also of the single loop electrode. The power level had to be reduced to 35 kV at 1 mA (35 W) to prevent overheating of the cathode.

As far as fusion is concerned, the experiment was a qualified failure. Fusion rate was down by a factor of ~10 compared to a 4-loop grid. The reason for this could be seen when looking at the loop assembly from the side: The ion beam formed in the center of the loop, but was then electrostatically deflected towards the stalk, so that the ions would ultimately collide with the stainless steel barrel which holds the loop. This barrel then became white hot after a few minutes run time, while the more distant end of the loop remained comparatively cool.

It was possible to remedy this to some extent by using two parallel loops of wire, set about 3 mm apart. The two loops apparently achieved a better focusing of the ion beam, along a single line. In addition and unexpectedly, they also formed a sheet of plasma extending outwards parallel to the rings.

The red ring around the stalk in the last image is a glowing Pyrex glass tube that was intended for insulation, but becomes semiconductive when hot and then starts arcing.

We may have more results when we find a way to insulate the HV stalk, e.g. with alumina tubing.

Wilfried Heil - Noemi Zudor

[Jon Rosenstiel]

Wilfried, Noemi-

Very interesting, particularly the effect the two loops have on the ion beam. Thanks for posting this.

Jon Rosenstiel

[Richard Hull]

Very nice! This is what we need to see. Different electrodes with fusion results.

Richard Hull

[DaveC]

They look like the simultations I've done, in the past. Nice to see it in real life.

Lensing by very simple divergent field electrodes is quite powerful. It is almost easier to do than to not do.

We have had some initially surprising results with hardware we've built, where unwanted lensing had to be compensated by specific shaping of electrodes.

In general, any port through an electrode will form a lens. Depending on relative potentials between electrodes, the lens is either a converging or a diverging lens

Thus, the inner grid of the fuor creates a multi-faceted lens, of a shape that depends in some detail on the exact grid configuration, but in general a sphere will cause convergence, and the openings at secondary convergence effect.

Thanks for sharing those... very nice photos.

Dave Cooper.

[thorium]

Verry nice pictures!

It looks similar to a penning trap type configuration, less the magnetic field. Would it be possable to use the rings as an electromagnet. Could you please include a photo of the chamber interior with power off and some more data on the runs?

[Steven Sesselmann]

The two ring cathode configuration is not so unlike my S.T.A.R. configuration, although the aperture in my cathode is only 10 mm. It is unfortunate that I can't see the plasma or the star mode inside my cathode, it might have shed some light on what is going on.

I am also not sure if the beams in my reactor are well focused, the focusing in Wilfrieds devise are aided by the spherical geometry of the chamber, my anode is spherical too, but I am not sure how much of the anode the ions can see, as they are separated from the anode by a dielectric.

The advantage of the S.T.A.R. design is not going to be the focus, but rather the ability to ramp up the amps and volts without electron losses.

I am hoping that with enough ion traffic through the cathode, that we can create a traffic jam, sort of like peak hour gridlock

Steven

[Dustinit]

It seems to me that you could make the stalk almost dissapear by putting it inside a tubular resistor. One side connected to shell, one side to HV. Then the voltage gradient on the resistor emulates the electric field and, to charges, simply dissapears.
Now if I only had a ceramic tubular resistor of 100meg.

Dustin.

[MSimon]

Concur,

Most excellent. This is definitely what I would call an advance in the state of the art even if it seems to give smaller fusion output. It is a data point. More data points are always of use!

Excellent work.

The next question might be if a dual spherical grid would improve the focus i.e something like the MIT guys did without the dropping resistors between inner and outer grids. Of course fabrication is going to be an interesting question to say the least.

Here is a link to the MIT construction details:

http://ssl.mit.edu/publications/theses/ ... chCarl.pdf

As my plasma physics PhD friend says: all the calculations in the world are of little help with plasmas because they can never be properly simulated. The best you can hope for is something suggestive from the simulations. Once something works equations will be developed. Proved against experiment. In the mean time the way forward is cut and try.

[MSimon]

A carbon coating might do the trick.

Some research into how carbon composition resistors are made might be in order.

[Steven Sesselmann]

Dustin,

I am not sure if I quite understand what the benefit of a tubular resistor is, you may need to elaborate.

In the mean time, I think I have found a solution to my focus problem, which will also solve the problem with Paschen breakdown.

By lining the tubes inside with a conductor, leaving only a short 10 mm break between the lining and the cathode, the mean free path of the electrons will be too short to cause Paschen breakdown, but still more than long enough to accelerate the ions to 100 KV.

I predict (fingers crossed) that I will be able to inject a high current ion beam into the cathode, drawing virtually no current. Need a week or two to pull the reactor apart and make the modifications.

Steven

[Dustinit]

Page 104 and 105 of this doc show ion paths due to the stalk field effect better than I ever could.
http://ssl.mit.edu/publications/theses/ ... Thomas.pdf
If you imagine a number of concentric circles inside a fusor (sperical) from the inside edge to outside if the inner shell they would represent a topographical map of the ideal electric field gradient. The stalk or HV feed distorts this field. It only requires a small amount of field distortion on each pass of a recirculating ion to chaotically upset a stable recirculation and cause grid or stalk collision. In the concentric circles you have drawn in you mind for a smooth gradient they will be equidistant . The stalk crosses all of the circles and is effectively equipotential along its length creating a topographic ashtray for a cuban cigar. Putting the stalk inside a resistive casing with the applied voltage across it gives a smooth voltage gradient on its length aproximating the electric field gradient.

This sketch may help visualise.
Dustin.

[waltsphotos]

Excellent experiment that I would call a linear fusor, I think this easily suggests a linear design of a fusor is possible based on the "Electrode port lensing effect" Dave mentioned. (consider a linear fusor as a singlur cross section of a spherical fusor, so that the center Electrode is two negitive loops with a positive area/loop to either side of the center Electrode loops). This experiment also shows that a linear fusor would likely not produce much fusion (More experiments are in order with an acctual linear fusor design, something I might attempt)

Considering the Electrode Lensing Effect, Dave mentioned, and results from this experiment, Consideration of a mesh Grid is in order. A Heavy mesh grid would increase the Ports for ion beams and thereby increase the number of converging Ions, and possibly increase the fusion. A heavy mesh grid might also have lower Voltage requirements. Downside, increased grid losses, and grid heating/melting.

in any case more experimentation with this Electrode Lensing Effect is in order

[Chris Bradley]

Is there any good way to deter the electrons from forming those hot spots (I strongly presume it is ejected electrons)?

If there were some ceramic material installed at those points then I would guess that it would become charged up and repel the electron beam to some extent, but then would it also attract the positive ions out of the beam? Or might if form a screened double sheath around the charged ceramic.

[DaveC]

Regarding Chris' question about electron emission "hot spots".

The inner grid, usually operats BELOW field emission levels.

Both Thermionic emission and Field Emission depend very strongly on the surface work function. This is the energy barrier the emitted electrons must overcome before they are free to leave.

All the very high melt temp metals have work functions in 4's to mid 5's ev. This keeps field emission at very low levels, until surface fields are above 100 kV per mm. With 25 kV operating potential and 0.25 mm (0.010") radius grid wire, one is approaching the low edge of this level of surface field.

The same high work functions, mean the gird surface temperature must quite high before significant current densities are achieved.

There is also a third and quite likely larger component of electron emission, which comes from secondary emission... when ions strike the grid wires. Of course this also causes the grid heating. So cooling the grids removes both this and the thermionic emission.

Typically,the fusor grid operates in a regime where the electron emission is a mix of these three sources.

But there are yet two other photo-electron sources: Xray and UV bombardment. As the operating voltages rise, the Xray stimulated emission will increase, but currents from neither of these will approach the first three sources at most common fusor operating potentials.

So the grid emits electrons by at least 5 different means.

The "hot spots" on the grid surfaces, are almost certainly areas of lower work function and/or, place where ion or photo bombarment are concentrated. Oxide layers are good candidates here. Oxides can also become stimulated emitters and may glow where the stimulus is heaviest.

A fusor that has been plasma cleaned with an inert gas, for sufficiently long, should be more free of these surface conditions.

Dave Cooper

[Chris Bradley]

I was also thinking of the ion-neutral interactions in the beam itself and the electrons that get knocked off those. These would run straight along the beam direction, as it is the shortest direction (strongest field line) in the e-field.

[DaveC]

The energy and momentum relations must be satisfied for any collision. The probability of an absolutely dead-on (zero angle) collision is very low. Therefore a radial velocity component will exist and both particles will exit the beam...scatter...

Lots of options... and lots of possible results... very likely that we see all of them.

Dave Cooper

[Wilfried Heil]

What I can offer here is an overview of the fusor as a whole. The loop grids have been replaced with the normal grid again and the viewport is too sputtered to take an image of the interior with no discharge.

The chamber is spherical with an inner diameter of 84 mm and the single wire loop has a diameter of 32 mm. The operating conditions were the same as with the 4 loop grid, around 15 mTorr depending on voltage.

What the experiment shows would not be seen in a simulation, unless it were specifically designed for this, namely the needle sharp focusing of the ion beams.

This is not a result of the spherical fusor. The beams will form in the same way between the flat flanges over the ports as well. The grid was turned so that the beams ímpact on the spherical portion because we did not want the plasma torch to enter the leak valve or UHV valve. As it appears, the beams form only in those regions of the chamber where the Paschen product p*d is highest. This automatically focuses the ions into a tight beam, where they can reciprocate and create more ions along their path.

>It looks similar to a penning trap type configuration, less the magnetic field.
Exactly. It confines ions, but not electrons.

>I was also thinking of the ion-neutral interactions in the beam itself and the electrons that get knocked off those. These would run straight along the beam direction, as it is the shortest direction (strongest field line) in the e-field.
That seems to be correct. Where the tips of the beams touch the chamber, they heat the wall (16mm thick!) to a red hot glow within an area of at most 5 mm diameter. These beams that hit the wall are electrons and probably charge exchange neutrals. They can be moved around easily by 20-30 degrees with a NIB magnet from outside the chamber.

The beams do not move as a whole, but are drawn into a short line by the magnet. This indicates that the electrons from the beam have a range of different energies, depending on where they were created in the gas through ionisation by the recirculating ions.

Electrons from the grid would all have the same energy and no specific preference in direction.

[Chris Bradley]

This would be true of the nucleus, but not an electron freed from its Coulombic grip. An electron would experience an acceleration of 6.15e+16m/s.s in a field of 350kV (I'm treating the fusor as a radius of 100mm).

So that would mean it can accelerate from stand still at the centre and reach the outer in 1.8ns (using d=1/2(at.t)).

Now if we look at a deuteron with a lab energy of 35kV, it has a lab velocity of 1.8e6m/s. Let's say that the target electron gets scattered completely sideways (that is, orthogonally to the radial beam - I'll use 'radial' as a radial from the centre of the fusor) and picks up this full projectile velocity.

But it can only get 1mm sideways in 1.8ns at 1.8e6m/s.

So electrons can be scattered completely sideways and without any initial radial velocity, but they are accelerated radially (that is, away from the fusor centre) so quickly that they would not leave the confines of a beam with a diameter bigger than 2mm even if they had the full [circumferential/sideways] velocity of the fastest moving particles in there.

Or to say otherwise, an electron formed in the beam will take on such an extreme parabolic motion from an initial zero radial velocity ina field of 350kV/m that you'd not be able to visually tell it had been moving at right angles to that beam when it started.

best regards,

Chris MB.