Vacuum Basics

How to get to vacuum, what the classes are, and what is needed for what job.

Vacuum Basics

Postby Doug Coulter » Thu Jul 15, 2010 6:41 pm

So you want to get to a vacuum. eh? Well, it's not super easy or super hard, depending on what you need.
Let's begin by defining classes of vacuum quality:

Coarse -- better than a vacuum cleaner, or most hold down fixtures, but nothing special, maybe a few millibars. One atmosphere is a bar, or 760 torr, so millibars are a little smaller than torr, usually interchangeable till you get to the tuning phase of a process. I think in bars because all my metric gear does.

Medium -- well below one millibar but not really high vacuum.

High -- equal or below e-6 millibar

Ultra high -- well below e-9 millibar.

As that last one is pretty tough, it's a good thing that you rarely need it unless you are studying surface contaminations of less than a mono-layer of atoms and such things. It's still orders of magnitude higher "pressure" than even earth orbit, however, just to give a sense of scale. Man has never approached the level of about one atom per cubic meter seen in intergalactic space (where there aren't gas clouds that is -- we can just about get to their pressure at all).
Cyclotrons run at about e-6 mbar, because they need some gas to ionize, and it's a tradeoff with beam losses to collisions, while big guys like CERN tend to be closer to e-11 mbar -- and pay real good money and use lots of the finest gear to get there and stay there.

As a point of reference, single and dual stage oil sealed mechanical pumps do the first two fine. Anything more -- you need another high vacuum pump in the chain of some sort. You will not get to e-6 mbar with any mechanical pump, or even reliably to e-4 mbar no matter the nameplate ratings. Truly. A really good recently rebuilt two stage pump will let you make neon signs that will then be filled to maybe 10 millibars or so...and not much else in physics. Or freeze dry food if you don't mind wrecking your pump by putting a lot of water vapor through it.

Note, anyone truly serious about all this should obtain the free Kurt J Lesker catalog and read every word on every page of tech notes at least twice. Sign up for it -- you'll wait some, but you'll want to get some things from them anyway (gaskets at least), and they have a premier explanation of all this, I'm just paraphrasing from that here, more or less. They want you to be successful when you buy their products, to cut support costs, and they have the straight dope there -- almost the best there is, and certainly for the price :) Another good source for the straight skinny is John Strong's book, "Procedures in Experimental Physics". What was "rocket science" or "brain surgery" when that was written is now mainstream, and he does a good job describing things in much more detail than we'll reach here. For a lot of things, the older books are better for that reason, and because what they cover was new, they're actually trying to communicate it to you, rather than impress you with bafflegab and assuming you know what I'm sure most of the writers of the newer material don't know themselves, never having had to actually make one themselves.

All real vacuum systems are operating at some sort of equilibrium between the effective pumping rate, and the rate gas is let in. And even a totally leak free system has what we call outgassing, where gas ab or adsorbed on surfaces comes off to interfere with things at some rate, so whatever "base" or best pressure you can get to is a balance between what a pump can take out, and what's being let into the system, or just coming off the walls of it.

Vacuum people define a quantity called "mean free path" which is very useful as a guide. It's the mean distance any one gas atom or molecule would travel without hitting another one in your vacuum. This is usually a more important measure than pressure is. Assumptions are that N2 is 3.1e-8 cm diameter. You can kind of think of this as the pressure vs system size where things are transitioning from a fluid that uniformly presses on tank walls, to a bunch of little balls moving mostly ballistically. One upshot is that it's very easy to pump to the point of mean free path being commensurate with the sizes involved, but it gets harder fast when the atoms are in ballistic mode, mostly clinging to the tank walls. The old joke is water will sit on a tank wall, fly off, go across the tank and interfere with your process along the way, stick to the other wall, and stay there from attoseconds to weeks before repeating. Once you get to long mean free paths, it has to randomly find its way to the pump to be removed.

Here's a table of mean free paths in N2, as calculated by John Strong. This came out on the large size, but that matches how important this is,
so I didn't edit it. This ties in with the way Paschen's law works, so is important in plasma work too.

MFP.gif
From John Strong
MFP.gif (9.53 KiB) Viewed 2614 times


The equation is: mean free path (cm) = 1/sqrt(2*pi*n*d^2) Where n is how many per CC and d is the diameter of what is there.
So for smaller molecules, the mean free path is larger, and the opposite.


Here's some of Lesker's words on mean free path.

The outgassing concept is important to understand as it affects the resulting purity of whatever it turns out you want to have in there -- it's always getting a little polluted due to outgassing of what was in there last (like you opened the tank and now have damp shop air clinging to everything inside). Gas from surfaces, especially water, is the enemy of purity and a pretty evil one. An atom or molecule can get kicked off one tank wall, fly across the tank, interfering with your process on the way by, then stick to the opposite wall for a time from pico seconds to weeks, only to repeat that when it feels like it. Once vacuum starts to get to long mean free paths, there is no mechanism "pushing" atoms into the pump at all, they have to find their way there at random, or as the Kurt J. Lesker words on this say -- vacuum doesn't "suck". Pressure pushes, but only while mean free path is much shorter than the tank dimensions -- with few or no inter-atom collisions, there's no push, just pure or nearly pure ballistic trajectories.

This leads to a nice rule of thumb -- make it short, fat, and shiny. Just as you would if you wanted to collect light coming from a lot of directions, but didn't have a mirror light won't stick to. A rough surface has many times the nominally measured surface area to trap gas -- that's the shiny part. Short and fat should be obvious - if your pumping lines are long and thin, the gas can just bounce from wall to wall in there, some coming back to the chamber, some getting into the pump and gone, for quite a long time. And that time can be long indeed from the standpoint of we mere mortals who just want to get on with an experiment -- days and longer.

One upshot of all this is that you will need a pumping system that will produce far lower base pressure than you want to actually have. The ratings for those are at zero flow -- something you will likely never live long enough to see in your system, as it will just about never fully outgas, it's a half life sort of thing.

Also, the purity of what you want in there will be a function of how fast you let it in, vs the outgassing rate -- so if you want even 90% purity, your inlet and pumping rates have to be 10 times the outgassing rate at the pressure you are running at. This can be somewhat daunting, and things that may eventually reach low pressure and are cheap (think the modified refrigeration compressor trick, or other things like that) just don't cut it in real life. Sorry to give you the bad news, but maybe it saves you some time.

For example, my fusor runs at around 1 e-2 mbar, with a base pressure of 3 e-9 millibar, so I can have pretty pure conditions in there. But that takes a 520 liter per second turbo drag pump and large backing pump on a completely leak free system to get to that base pressure. Nearly everyone with any success at fusor pressure levels has at least a good mechanical two stage forepump, and has a diffusion pump for the high vacuum stage to get to decent purity. This has been well established by pro cheapskates and scroungers and if there was an easier way, they'd all be doing it -- but they aren't. Because nothing less gets the job done. As luck would have it, diffusion pumps are getting cheap as the rest of the world goes to the turbos, so they can be scrounged. Most of them, with the right oil, will get to e-6 or lower without even a cold trap, with the expensive oil, some push e-9, so that can be a solution on the fairly cheap. But diffusion pumps won't work unless their output line is kept to the pressures you will need a 2 stage mechanical oil pump to get to. So you need two pumps at a minimum.

A turbo-drag pump will run fine with a lot more pressure on it's output line than any diffusion pump will, and this allows for the forepump to be a lot cheaper, and oil free itself, so when these start hitting the surplus market, that's going to be a boon for us all. Right now, most of the turbos out there lack that last drag stage and need really good forepumps -- which is why they are on the market cheap, and also the older types tend to need an expensive external controller and cable to run - so the word is "avoid these on ebay" unless you know more than I do about how to kludge them into operation. And I design multiphase variable frequency and voltage and current drives in my sleep. Or write them into a few lines of microprocessor code. My time is still worth more than the money I'd supposedly save doing that sort of thing.

Now to move on (all the above will get covered in much more detail later on) to gaskets and sealing. There are basically two systems in wide use out there, O rings (QF, ISO) and metal seals (KF). The O-ring systems are nice for being reusable -- you can make and break the seals many times before needing to buy a new one, but they limit just about any system, even with "perfect" fast pumping, to the e-8 millibar range. Viton rings are usually used, as they stop air cold, but let some water right through the rubber. Buna stops water, but not air so well, and for the life of me I don't know why no one does a composite. Also the O rings have a definate temperature limit they can handle, and after that they degrade and give off gas. The advantage of the copper gaskets is that they leak just about zero...but you can only use them once if you follow the directions and tighten the flanges fully. I don't do that, and with a bit of touchy-feely can get about 3-4 uses out of them before they have to be replaced.

So many practical vacuum systems use a composite plan -- Viton seals for doors and access, and copper for the rest. Mine all do it that way, it's a decent tradoeoff. Unless you always need ultra high vacuum and have unlimited money, you're going to wind up there too - this advice is not lightly given.

Now, about that outgassing issue. As a particularly hip Pfeiffer rep once told me -- anything that puts energy into the tank helps. Heat, UV photons, ions, electrons etc all bash gas atoms off the tank walls and help them find the pump. But there are limits. Unlike electron vacuum tubes, we can't bake our systems to red heat for this, and then add a getter (we can add the getter if we want to pay for it each time, or add a third, ion pump). Viton will only take 100-125c or so without degrading, and an all copper sealed SS system might handle 300 or even 400 C if the copper is silver plated, but that's only about half the red heat they use for sealed-off devices. So it takes longer for us than for them, and our gaskets set a limit to what we can do there. A glow discharge in the same gas you'll want in there later anyway seems about the fastest way to flush all the contaminants out, with internal quartz lamps about second there -- and when you turn those on the tank pressure can double in a few tenths of a second, indicating how well they work. Doing that for say, an hour, gets things pretty hot, and pretty clean.

//////////*************************//////////

I have noticed at least two distinct approaches to this building of vacuum systems. The very young and the very old are in a hurry, and tend to find ways to just buy things new or refurbished -- the young because waiting a year's worth of scrounging seems like forever, and indeed it is as a fraction of their adult life so far, and the very old because, frankly, there's just not that much time left. In middle age, people tend to be more patient and scrounge for years to put one together -- and often tend to forget the price of all the gasoline they spent going to hamfests and so on where they didn't get anything useful, or worse, bought stuff they never did use, but had to store anyway, when they boast of how much they "saved" doing it that way. A good idea is to figure out where you are coming from around this concept, and plan accordingly. Some people love to scrounge, and that's a life some find well worth it (it can be a lot of fun after all). Some just want to get on with doing real experiments after they get this part of things handled. Decide which you are, and you'll be lots happier in the long run.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Return to Vacuum Technique

Who is online

Users browsing this forum: No registered users and 3 guests

cron