Mark3 operation with ion source

[Andrew Seltzman]

Operation with one of the ion sources in use (air plasma, no feed gas to the injectors).
Grid voltage was in the 10-20kV range, current limited to 5mA.
Ion source in use was on the top left in the photos.
A flat pancake geiger probe registered about 50k+ cpm about 2 feet from the viewport.


Previously, with no ion sources, the fusor would not sustain a plasma below 9mTorr, due to lack of thermionic emission from cooled grid, now it will run stably at 3-4mTorr. The minimum pressure limited by out gassing of the plastic insulator on vacuum hub which probably will be removed or redesigned. A single ion source will completely overwhelm my 40kV 5mA testing power supply if turned up much beyond 1kV on the anode. They can produce significant amounts of ionization, so low pressure stability should not be a problem from now on.

The injectors will run stably in the sub-mTorr range with anode voltages up to 2kV without problems.

I am also having problems with my feed through, it seams to start to spark/conduct at around 20kV, sometimes you can see glow paths(like snakes) coming out of the joint between the 0.75"OD/0.5"ID alumina tube and the 0.5"OD/0.25"ID alumina tube. Not like sudden blue sparks but more like thin snakes of plasma going out from the gap and down the side. The tubes are friction fit, possibly trapped gas escaping and being ionized? Has anyone else had problems with alumina electrical insulators? It had just been pumped down so may be I just need more time conditioning it.


pic1: ion injector running in the 1mTorr range, note the focusing effect
pic2: fusor with ion source running, about 8mTorr, partially impact ionization of neutrals
pic3: ion source running, no grid voltage, 4mTorr
pic4: fusor with ion source running, 4mTorr, most ions from injector alone


Andrew Seltzman
www.rtftechnoloiges.org

[Carl Willis]

Hi Andrew,

Lots of nice eye candy there, evidence of good performance out of the anode layer sources.

I haven't had problems with alumina specifically, but I have found that glass, as an insulation medium in vacuum, tends to be worse than nothing at all. Once a discharge gets started on a glass tube, it leads to local melting and conduction of the glass, which will then exhibit recurring trails or worms of glowing conduction in subsequent operation. So as a first step, have you operated without the alumina at all? It might work better that way. Surfaces, regardless of their composition, are bound to be worse insulators than pure vacuum.

Sorry to hear the plastic was an outgassing problem. It's a daring step outside the normal vacuum-practice repertoire, but if made to work would be a nice dodge around the traditional expensive ceramics. Maybe the inside surface can be treated with Vac-Seal spray resin.

I'm watching your progress with interest and a bit of envy. Hope it sees D2 soon.

-Carl

[Andrew Seltzman]

Carl,

I have seen the save effect with glass too on the Mark1. When I was first using a pyrex tube for insulation, it would be subjected to ion bombardment near the grid and would become conductive, glow, eventually blackening and partially melting. Since switching to alumina ceramics I have not had that problem.

The alumina itself is not being damaged, but it is annoying, to have it spark occasionally. My biggest worry is that although the ceramic is quoted as being non-porous there might be fine holes through which it might spark to the 0.75" quick connect fitting. However this would be a spark going through over 1/4" of ceramic, and although sometimes I can hear the snap of a spark, I have not been able to find any evidence of a spark or arc passing through the ceramic.

What I have seen lately is the plasma snakes coming out of the gaps between the feed through tubes(several concentric tubes of varying diameters). Since the tubes are friction fit, and I have been letting the fusor up to air every time after a run for testing, I believe that It is a trapped gas volume becoming ionized.

Unfortunately in my case due to the cooled grid, the ceramics are required for support and positioning. I am thinking about replacing the 1/2" OD alumina tube with a boron nitride tube ($80 for 6") which is a better electrical insulator and can be machined to provide a pump out channel.

As far as the plastics go, I am suspecting it of out gassing through I don't know yet for sure. The main reason I suspect this is that I had no way of baking out the plastic under vacuum, which is the standard way of conditioning. The plastic also has no shielding from the plasma, though it sees only a very small solid angle.

I know these plastics are used in UHV applications. On MST here at UW-Madison, these types of plastic insulators are frequently used with out problem and we idle at 1E-7Torr. Although there are 3 turbos and 2 cryos pumping on it.

What I'd really prefer is a 1.25" long nipple and a 1/4" thick plastic break to reduce surface area. There are also better inexpensive plastics like PEEK, PPS, Ultem, or torlon, but they are more expensive. Acetal co polymer is the cheapest vacuum rated plastic you can get and it probably should have been baked out.

Andrew Seltzman
www.rtftechnologies.org

[John Futter]

Andrew

Oh for a RGA on the beast.

I think that what you are seeing is water vapour coming out of your alumina tubes. It is a beast to get rid of. Back filling with nitrogen or even dried air when coming back to atmospheric will reduce the water molecules particular attraction for alumina.

As for the plastic, we use PETP (cheap stuff easy to machine and it will seal against conflat flanges to make vacuum tight insulation breaks) to make the ion source to beam line insulator on our implanters. These are operated at 30kV plus with no problems, with the beamline 4" ID stainless tube around 6 feet long, all its needs is a single 360l/sec unit to get the beam line to mid ten to the minus sevens while operating, the pressure rise due to the gas load in the ion source (ultimate pressure without the ion source running around 2 by ten to the minus eight millibar). Ion source exit aperture one eight on an inch in diameter and penning ion sources operating around 1-4 milliamp 2kV.

I'll post a closup pic when I get time.

[Steven Sesselmann]

Andrew,

Nice work as always...

I guess you have already checked for this, but any dirt or carbon based impurities in your chamber quickly forms a conductive layer on the ceramic. A light sandblast with aluminium oxide powder will remove it.

I recently bought this ebay item for a similar purpose, Item number: 190303904124

Keep up the good work..

Steven

[John Futter]

Andrew pics as promised showing conflat plastic insulators around target chamber, +40kv ion source plastic insulator, and target chamber target bias insulator to -40kV

Edit
info on PETP

· High mechanical strength, stiffness & hardness
· Very good creep resistance
· Excellent wear resistance (comparable or even better than Nylon grades)
· Very good dimensional stability (better than polyacetal)
· Excellent stain resistance
· Better resistance to acids than nylon and polyacetal.
· Good electrical insulating properties
· Physiologically inert (suitable for food contact)
· Good resistance to high energy radiation (Gamma and X-rays)

[Richard Hester]

Glass is pretty much verboten for insulation in any environment subject to heating or ionic bombardment due to alkalai/alkaline earth constituents (sodium, potassium, calcium, barium, etc). Both pyrex and normal soda glass have these constituents in spades, so you get spectacular conductivity when they are heated up, especially when high voltage is involved. With some coaxing, you can actually plate sodium out of the heated envelope of an ordinary light bulb. This experiment is described in Stong's book on procedure in experimental physics. For insulation purposes, fused quartz and aluminum oxide work best. as far as ordinarily available materials are concerned.

[Andrew Seltzman]

The general class of PET materials are polyester variants, however there are probably differences between the types. What is the last P in PETP and how is it different from standard PET?

After talking to others about the use of acetal, I have come to the conclusion that the Acetal copolymer I am using is probably not out gassing. There is just a leak somewhere else that I haven't fount.

Andrew Seltzman
www.rtftechnologies.org

[John Futter]

Andrew
more info here http://www.engineeringplastics.co.nz/pr ... talyte.htm
and attached is a PDF

[Andrew Seltzman]

If the vacuum cube is removed from the core and pumped down with a turbo it will only reach about 1mTorr. If it is then sealed off the pressure will steadily rise at about 1mTorr every 6 seconds. It will climb past 1Torr if left for a sufficient amount of time, the rate seams the same at 1Torr and 1mTorr. The cube has not been baked out. It will pump down to 1mTorr much quicker the second time, but the pressure will still climb after it is valved off.

Am I dealing with a leak or out gassing of water?


Andrew Seltzman
www.rtftechnolgies.org

[John Futter]

Andrew

It doesn't sound like a leak due to the faster pumping on subsequent pumpdown.

Dare I ask how clean it is??

If left the cube will rise to around 100 torr (vapor pressure of water) and stop.

baking in your domestic oven with the cube lightly wrapped in clean Al foil to keep grease from the oven off it at 160 degrees C for a couple of hours then reassemble while hot.

The bake out will remove most of the water vapour and some of the hydrogen dissolved into the Stainless.

For UHV apps at work we use scotchbrite then acetone then ethanol and into a massive ultrasonic bath with decon90 for an hour then rinsed with de-ionised water then into a hot air oven at 160 for two hours and assemble hot.
PS no scotchbrite anywhere near knife edges of CF's. These are cleaned with a special polish that doesn't leave outgassing debris.

[Andrew Seltzman]

The cube was wiped down with isopropyl alcohol and assembled with viton o-rings. The oven I have access to is gas powered and kind of dirty, so I would like to avoid it. I am going to buy some bakeout heating cord and wrap the cube in it to heat it up during pump down.

What is the maximum temperature a viton o-ring can handle under vacuum?

Andrew Seltzman
www.rtftechnologies.org

[Dan Tibbets]

Speaking from telescope mirror cleaning recipies, I hope you are using 90% or higher isopropel alcohol. 70% isopropel alcohol (rubbing alcohol) often has other things in it- oils, fragrences, preservatives,etc. I'm not sure how pure even 90% is (90% alcohol and 10% deionized water?). Is 99% isopropel alcohol aviable? Acetone is probably more pure (99.5%?)


Dan Tibbets

[Andrew Seltzman]

I am using the 70% rubbing alcohol, but it is the pure type, just alcohol and water. It works well and is easy to come by. Unfortunately it does have the 30% water.

I'll try something more pure as well though I tend to avoid acetone since it dissolves some plastics.

Andrew Seltzman

[DaveC]

Andrew -

Without knowing the constituents of the residual pressure, my guess is the same as John's - probably mostly water. So... heat here, and continued pumping are your friends. You probably need a full day under vacuum... to get reasonably dry.

Viton O rings will outgas (a lot for a while) for several hours. One typically has to beware of using too much solvent when cleaning them the first time. Typically a wipe with a lint free wiper (Kimwipes or similar) using rubber gloves on your hands so that finger oils are dissolved by the solvent and transferred to the surfaces you're cleaning..
Bakeout temps are good up to about 150 C..... But is you have windowed ports, I would opt for a bit lower temps just for caution about thermal stresses. Pyrex and fused silica are pretty robust, though... so this just my own cautiousness showing.

Reagent grade Isopropyl, Methanol and Acetone are best for cleaning. Each has its advantages. The "anhydrous-ness" of them is their big advantage over the domestic 70% isopropyl.. which can never dry a surface.

Thanks for the pix of your latest work.

Dave Cooper

[Andrew Seltzman]

Ok, I cleaned out the cube and ports with 99% isopropanol and used a bakeout tape to heat it up under vacuum to about 170F. While at that temperature, even with a turbo pumping on it, it would not get below 2mTorr. After it was let to cool back to room temperature it pumped down to about 3E-4torr but after being valved off it increaced to about 95mTorr in about 6 min and 150mTorr in 12min.

The only non-metal component is the viton o-rings and they are not greased. Could the viton be out gassing this much? They are the standard Mcmaster viton o-rings. Would baking out the o-rings under vacuum help? I have heard of this being done.

However the pressure rise seams to be completely liner between 1mTorr and 1000mTorr with a rate of 1mTorr/6sec so I'm probably dealing with a leak.

Andrew Seltzman
www.rtftechnologies.org

[Andrew Seltzman]

It was indeed a leak. I left the o-rings on a paper towel prior to assembly, and it left an almost invisible fiber across one of o-rings which caused the leak. It was measured to be 2E-5 torr-L/sec with a leak checker.


Andrew Seltzman
www.rtftechnologies.org