Ion Gun test rig

[Starfire]

Drool!

[Doug Coulter]

John,
those are real impressive results indeed. I've been playing around with all the same stuff, including Einzel lenses, and my main problem with them was beam blowups at high space charges, even with electrons moving fast, and therefore kind of spread out in space for a given current. The one I made can have all 3 electrodes at different voltages if I want, and I was able to get focal length to vary over a wide range in the normal Einzel hookup-- I did this pre-acceleration to get that happening, as after I put some serious kv on the electrons, the lens was kind of a null effect device at reasonable lens voltages.

But it worked fine before one final stage of real high voltage acceleration, kept the beam in the tube till it got there which was my goal then. Once it got really tuned, I could even let in some gas to see the beam without messing it up totally, which was really cool looking (since I have a tank of neon, I used that, of course). The fun thing I did with that by accident was to reduce (?) some of the glass target I used to see the beam spot -- pyrex glows fine when you hit it hard. Looked fine in the tank, but when I took it out there was this big white spot where I'd affected the glass, some kind of powder which was deliquescent and after wiping off the result after a couple of days (it was just a wet place by then), no visible damage at all. Very weird, but fun. I don't think I implanted any neon in it, but who knows? And it wouldn't explain the powder. Something funny happened for sure.

I do sometimes wonder why any design data for magnetic lenses is so scarce out there, I've been looking and looking and not finding very much that's specific enough to be useful. Seems like the microscope guys like their trade secrets, or I'm looking in all the wrong places. There's one good overview in the Scientific American amateur science projects disk, that's about it, and that's not really enough, but it is helpful for just getting a feel of what's needed at what energy. All the e microscope papers I've found kinda leave out some of those details...either it's so easy it's not interesting, or something they want not to reveal it seems.

I've not tried this type source myself yet, AndrewS described it quite well,and IIRC he provided (or heck. maybe I just found, I just don't recall) some interesting links where you can affect the beam pattern out of the raw source by moving one of the pole pieces back and forth a little, from converging to diverging, maybe he'll pop in on this and let us know that link again -- it was to some small company that makes them, in an animation they had on their site, it was pretty cool. It might help you get into your lens aperture better.

Be sure and let us know results! I think this kind of thing is too often not paid enough attention to here.
And ion sources I've invented have been my nemesis so far, they work great but don't last long -- everytime I fix one failure mode, another appears. I did a neat one I still like that used a tiny capillary tube and a very short gap to a hole in a charged plate, that allows very low tank pressures -- we ionize before the gas has time to expand to the much lower tank pressure, and it makes cold ions due to the post expansion, with shock diamonds in the beam -- looks real cool, but has a number of failure modes, mainly due to metal loss off the electrodes -- even if you're careful to keep it from getting on the insulators, it makes the gap too big at some point. So my current design on this one only lasts some tens of hours at full power.

For completely other reasons I've been wanting very low pressure ions too, and very few neutrals, so I am awaiting any results with high interest! Of course, I'd love to be able to ion implant some things myself, I already do sputtering and evap here for fun and games. I want to find a source that doesn't need another pump to keep my main tank pressure low as I want, and I want that very low.

So far, my best results have been with another type of source, one that uses an oven magnetron I finally managed to get running continuous instead of pulsed at about 5% or less rated power, which is plenty, and it gave me about a 10x factor lower pressure than I could get with any DC/magnetiic field, and I tried some real large fields of both where other problems appear. I've since added ECR magnets to that, and plan to get a good measurement soon, but upcoming HEAS means I've stopped work for the moment and am getting ready for that. And I'm on to the next rev in design anyway, now I want to do this right in the tank, instead of in a quartz tube in the microwave cavity, and I think I can if I can seal the tuning rod well enough.

Note, if you're getting your window hot you'll have trouble at some point. I do two-three things, depending. One is just put a cheap piece of thick pyrex inside the tank to shield the window, It's cheap and can be dunked into acid to eat off anything you may deposit on it, not something I'd do with my fancy hinged window/door. It you ruin it, not a big deal. I've also used even cheaper mica I get from McMaster-Carr, as "india glass" but the last couple batches haven't been real clear -- they are also where I get the thick cheap round pyrex glass. You can also of course put some metal screen there with a charge such that you just repel off any particles, but that may mess up your experiments by making a field in the tank. But all these things are far better than a shattered main view-port and an air inrush accident.

[John Futter]

Doug
Thanks for the comments.
I'm painfully aware of most things you mention, and because of that we will not be operating the ion guns above 2mA on target until we get a mesh screen to take most of the charge. The sreen will be fine SS mesh which may glow hot to yellow hot later on.
Andrews info on the anode layer source gave the links to the Russian web site that showed the focusing effect of moving the outer pole piece.
Early days yet
1 we have to shift the einzel lens - Simion CAD run required to save time
2. Mesh screen on the target window to keep window temp below 100 C
3. Install pneumatic gate valve above turbo connected to alarm out on high vac gauge
4. redo anode layer source with screw thread adjustment of outer cathode for tuning beam
5. improve gas feed in ion source so that at higher gas flows it doesn't setup an arc in the gas feed tube (quartz)

edit
Over the weekend I has a few ideas to make the assembly of the anode layer source alot easier
1 I couldnt source a NB magnet with a hole in it and drilling a 4.1mm hole through it without breaking it took ages so the magnet will in future rest in recesses top and bottom to locate it.
2. outer cathode to be on a screw thread to optimise
3 gas feed to be metal tube into bottom of ion source not into middle of gas space
6. model extraction electrode (Simion) for annular ion exit to improve extraction from 25% to at least 90% of active ions. as well as finding the right values of voltage / current in the ion source for various gas loads.

Up until now most of our ion sources are cold cathode penning type. These where introduced to us by our principal scientist Andreas Markwitz who did his Phd under Bauman in germany who invented / optimised this sort of ion source. We get around a hundred hours of operation before having to service these (new front and back cathodes) and typically these run at around 2mA ion+ current output around 85 to 95% of the arc current in the source. The higher the current in the ion source the sooner you have to service it, integrated ion current is directly proportional to cathode wear.

Again The extraction electrode is the secret to getting only excited ions to move to target the neutrals do not see it and the hole in the centre of the extraction electrode is small (3-4mm) allowing differential pumping to the rest of the system.

System was sitting at low ten to minus seven millibar today ready for the Argon test but I/ we had other things to do maybe tomorrow-- will try to take better pics of ion beam trajectory.

Edit 2
Bear in mind the first ion source tested on this rig was a negative gas ion source and it is now being modified and will go on the system between iterations of the ion layer source. It involves changing out all the Hv psus for negative versions

[Doug Coulter]

On the magnets with holes, try any of amazingmagnets.com, magnetsrc.com or even scitoys.com.
I've gotten the ones in the pics below from amazing, and had fine dealings with all three of the above.

Here's the source I'm currently testing. It's a CW microwave driven, ECR source in a cavity, which has a quartz tube going through the cavity which has gas inlet at one end via capillary tube (2" by .005" hole), and is open to the tank on the other end. This lights up down to about 1e-5 mbar and below, at very low magnetron input/output power. Best I've done so far, and it all runs cool and looks like being real reliable.

However, on the first test we found that the ECR field we needed also bashed all the nice ions into the quartz tube wall before they reached the tank -- no good. So we added a little piece of Cu tubing inside the quartz right at the exit point and put about -3kv on it (from the maggie power supply via ballast, a few mA were drawn at about 3kv on this), grounded the capillary tube, and bingo -- ions coming into the tank very nicely, and making a big difference in fusor operation over a foot away from this coupling. I doubt however, that my "e lens suction design" which was a guess, is optimal at all, and we'll try a couple more things before rev two, which will have the cavity in the tank, skip the quartz tube and maybe not have this problem anyway.

In that case, we'll just make the cavity bottom out of screen wire, so the ions can get out over a wide angle anyway, and may not need the electrostatic "suction" element. I have to add though, that this added a good factor of being able to run lower pressures yet with some ripple on the magnetron output which would let it "go lights off" at low pressures easier. Much better with a little bit of keep alive DC in there too.

Getting an oven maggie to go CW and not fry kind of took some convincing, but we have that going on now -- and it only takes around 20ma HV current for this to run well, and with somewhat reduced filament voltage. It doesn't even oscillate well unless the cavity is tuned right, but when it's right, it's right and everything stays cool for long operations, and it stays right from time to time. We have noted that some maggie tubes mode hop a lot, and we had a selection to pick from here to get one that works really good despite some slight tuning errors. We get them from dumpstered maggie ovens, more or less free.

I used a hall effect magnetometer to get the field right on that magnet yoke -- you can slide one of the pole pieces and magnets back and forth until you get the numbers you want -- when you do, stick a shim in there and it's done for good. The magnets make about a 2.5 order of magnitude difference in the lower pressure limit, and we don't use differential pumping here yet (trying to avoid the need, as that would be past what most of this group will accept, even me, and I have the extra pumps).

At any rate, in a "lousy geometry" experiment with this about 18" away from a fusor grid and not pointed quite at it, we got real good results right off, and are now able to run the fusor at about half the pressure we could before -- I have high hopes that with good geometry (this setup could almost not be worse than it is for what we're doing with fusors), we'll be able to keep a fusor lit off at the lower limit the ion source will run at which is where we are heading here as part of our overall research.

As I said, this runs real cool and there's no evidence of any sputtering onto insulators after hours of continuous running, so it may be of interest to you. Looks like good high ion currents are possible, but I've not done quantitative measurements yet. It does work better than other sources we've tried, subjectively, but we've not tried the type you're working on yet either.

We will, of course. Real soon, as your results are in the same ballpark, and with less gear to clutter up the setup required.