Diffusion Pump inconsistencies

As far as I know, backing a diffusion pump with a rotary backing pump is absolutely standard. That how I pumped down my vacuum line when I was doing my Ph.D. in physical chemistry back in the 1960's - just like everybody else who was working with vacuum lines - and that's what everybody did where I was doing my post-doc a few years later.

When I was working on electron microscopes at Cambridge Instruments in the 1980's, their standard set-up relied on a turbomolecular pump backed by the same sort of rotary vacuum pump. A turbomolecular pump cost a bit more than a diffusion pump but needs less cooling.

Where the electron microscopes were coupled up to big specimen chambers, the turbomolecular pumps were bit too slow, and we used cryopumps instead - still back by rotary pumps (albeit tolerably big ones).

I've heard it claimed that you can back a diffusion pump with a water pump, which only gets you down to about 15 torr, so I'd say that you should go ahead and buy your diffusion pump.

-- Bill Sloman, Nijmegen

If you want a clean vacuum, DO NOT USE an oil diffusion pump. The whole chamber fills up with pump oil. Look into sorption pumps, ion pumps, getter pumps, cryopumps; and, well, if your vane pump is also known as a turbo pump, you've got that covered. :-)

Good Luck! Rich

Thanks! I can beat 15 torr, so no problemo! I'll buy it this week, but first Rich Grise said it fills chamber up with oil. I'll wait to see what he says. ;)

Anon

of

no

rr?

Ohhh, I wish my budget was for a turbomolecular! I don't like ion / gutters because they're one timers, need to replace the material, right? This is for a garage project, so would ion pump be afordable to buy the materail to absorbe the gases every run. Maybe I should learn more about ion pumps.

As far as oil in the camber, I know what you mean, but oil's a lot better than oxygen as far as deposition. I think silicon molecules love to stick to oxygen molecules more than oil. So long as the pressure gets to at least 1e-4 torr.

I'm not going to use a liquid nitrogen cryotrap to prevent backstreaming because it's a garage project, but I read that polyphenyl ethers make a cryotrap unnecessary. I thought that maybe placing a lengthy metal tubing coil cooled a bit would trap enough backstreaming oil. You know, similar to what hotrodders use to cool the gas prior to carburator. In hotrodding, the gas flows through the metal coil, and the coil is inside a canister with ice. My only concers is how much metal surface is allowed in a vacuum because atoms are continually escaping the surface in a vacuum. You know, if there's too much surface then the vacuum pumps can't keep up with it.

Thanks, Anon

Diff pumps will start to pump in the 1 torr range, enough to get themselves down under 1e-3 torr where they are happy. Just make sure you don't have any leaks that will keep it in the transition region any longer than necessary since this is when you get the majority of backstreaming. Use pentaphenyl ether (Santovac 5) oil, it is very good about not creeping along surfaces so the backstreaming is usually limited to line of sight to the pump inlet. Hang the pump from an elbow off the side of your chamber and that does just about as good as a refrigerated baffle, or fabricate a metal disc to sit over the pump inlet spaced an inch or so above the bottom of the chamber so only the bottom of the disc has line of sight into the pump inlet. You also need a trap (zeolite, copper wool, something) between the rough pump and the diff pump to stop oil from the rough pump from migrating into the diff pump where they will promptly pollute the chamber.

Try to get a copy of O'Hanlon's book A User's Guide to Vacuum Technology, it is very readable and will fill in lots of gaps in how to set up and operate a high vacuum system.

----- Regards, Carl Ijames

Ohhh, I wish my budget was for a turbomolecular! I don't like ion / gutters because they're one timers, need to replace the material, right? This is for a garage project, so would ion pump be afordable to buy the materail to absorbe the gases every run. Maybe I should learn more about ion pumps.

As far as oil in the camber, I know what you mean, but oil's a lot better than oxygen as far as deposition. I think silicon molecules love to stick to oxygen molecules more than oil. So long as the pressure gets to at least 1e-4 torr.

I'm not going to use a liquid nitrogen cryotrap to prevent backstreaming because it's a garage project, but I read that polyphenyl ethers make a cryotrap unnecessary. I thought that maybe placing a lengthy metal tubing coil cooled a bit would trap enough backstreaming oil. You know, similar to what hotrodders use to cool the gas prior to carburator. In hotrodding, the gas flows through the metal coil, and the coil is inside a canister with ice. My only concers is how much metal surface is allowed in a vacuum because atoms are continually escaping the surface in a vacuum. You know, if there's too much surface then the vacuum pumps can't keep up with it.

Thanks, Anon

Our pumps go down to 2e-7 torr, one of ours will hit -8 now and then but I suspect an inconsistency with the ion tube.

we operate our 2 pole turbo's up to 1000 hz. :)

We always put a liquid nitrogen cooled trap between the diffusion pump and the rest of the line; I had two, more to protect the oil in the backing pump from the bromine in the vacuum line than to protect the vacuum line from the oil in the diffusion pump.

Electron microscopes mostly make do without this - in any event the specimen chamber of an electron microscope contains mechanical moving parts which have to be lubricated so there's always oil vapour around.

Liquid nitrogen costs about the same as milk per unit volume and it's rather easier to get it delivered to your house or garage; you need a big vacuum flask that will take about a week's supply but they aren't all that expensive - a quick Google suggests something around $500.

-- Bill Sloman, Nijmegen

Lots of good suggestions above. Someone said you can back a Diff pump with 15 Torr. This might OK but backing and starting the diff pump are two different things. Your rotary vane pump that gets down to 0.1 Torr will be fine. (You need some valves so that once the rotary pump has got the system down to 0.1 Torr it can be switched over and used as the backing pump on the diff pump.) Other wise you need two pumps. With only one mechanical pump you can get the system to work but you have to do some careful valving.

You might look into the text "Building Scientific Apparatus" by Moore Davis and Coplan. I would definitely use a nitrogen trap with the system. You might also look into cryo pumps. These can't pump things like neon, hydrogen and helium, but if you can flush your system with nitrogen then they are very clean pumps.

Good luck, George Herold

AFAIK, the getter pumps and ion pumps got purged pretty well when they baked out the bell jar.

Unfortunately, I don't remember what vacuum pressure they were going for, but with the cryopump it was good enough for molecular beam epitaxy.

Well, actually, I wasn't talking "cryotrap" - I'm talkng about a pair of cannisters about 6" diameter and about 12-15" long, full of Zeolite. They'd hang dewars around them, fill the dewars with LN2, and when the LN2 stopped boiling, they'd open the valve. You could hear the air being sucked out ot the chamber, and the LN2 would start boiling again. When it got down to some level, they'd close the valve and start using the other pumps. When they'd let the chamber come up to air (to change experiments or hatever), all the air would come out of the sorption pumps, and they'd bake them out.

It seemed to work; the company was so rich that they built a full-on gym and a couple raquetball courts. Every three months they had a Profit Sharing party, and they'd bring a keg of beer into the lunchroom. ;-)

Unfortunately, my wonderful boss quit one day presumably for more money, and his replacement was such a moron, I arranged to get myself fired (so I could collect unemployment), then found an even better job (well, not as exciting, but an infinitely better boss.) ;-)

Cheers! Rich