My university has a sputtering machine which is this crazy expensive piece of equipment that has to have a really strong vacuum pump and wacky copper seals and if it loses power for even a minute it has to spend 16 hours pumping it’s vacuum back down.
I know people talk about how a perfect vacuum is like near impossible, but why? We can pressurize things really easily, like air soft co2 canisters or compressed air, which is way above 1 atmosphere in pressure, so why is going below 1 atmosphere so hard? I feel dumb asking this as a senior mechanical engineering student but like I have no clue lol.
In: 3143
Imagine a warehouse, filled to the brim with glitter. A lot of the glitter is pretty easy to remove, but there will always be some glitter particles left somewhere.
A perfect vacuum is the same, it is incredibly hard to remove all the molecules in the air, to do a perfect vacuum
Below one atmosphere is also pretty easy. We can get ‘close’ to vacuum, no trouble.
But imagine trying to pressurize a CO2 canister, not to *roughly* two atmospheres, but *exactly* two atmospheres. Accurate down to ten molecules in the whole space. Suddenly you have to keep very close track of every possible place even single molecules can come from.
Things like microscopic cracks in your components, sticking to the surface of your components, or even downright being absorbed. Even metals will sublimate a very tiny little bit, and some even do it so much that they’ll easily ruin your vacuum.
If anything, this precision gets *easier* under high vacuum. You know that every molecule in the tank is one molecule too many, as opposed to trying to somehow measure 2 atm accurate down to the parts per quintillion.
One technology you might be using to get to low pressure is a cryo pump. Basically you have a very cold surface and any air molecule that happens to bump into it “freezes” onto the surface. Now imagine what happens when you cut the power: the plate warms up and the air molecules are released back into your vacuum chamber. You turn the power on and the plate gets cold again, but now you need to wait for all those air molecules to randomly bump into the plate which takes time. And the fewer particles you can afford to have left the longer you need to wait.
There are many reasons. You have problems with off gassing for example, at low pressures things do not tend to stay solid and will start evaporate, including a lot of rubbers and plastic. But the main issue with very high vacuums is that the air stops behaving like a gas and more like individual particles. You can imagine pulling back a piston in a vacuum chamber, the place previously occupied by the piston will be free of any air molecules even though there is a few random stray molecules in the rest of the chamber. You have to wait for these molecules to bounce around and enter the new volume. Depending on the size of the opening and the pressure inside the chamber this may take minutes for the pressure to equalize. There is just nothing to push the air into the new void.
Your sputtering machine needs so long to pump down again because it’s likely using a cryo pump, which needs a helium compressor to be constantly pumping. On extended power loss, you need to regen the cryo pump, which takes hours. It’s likely not the actual pumping down that’s taking that long, it’s prepping the cryo pump.
One of the issues is when you get into high vacuums there there aren’t a lot of molecules left, and your pump can’t remove them from the center of the container.
When you remove them on one side of the container by pump or whatever you’re waiting for entropy to push molecules from the rest of the container towards the place you can pump them away. In normal atmospheres this isn’t an issue, but getting close to ‘perfect vacuum’ this takes time. And in that time new molecules can leak/offgas/whatever other ways back into your container.
(note: “not a lot” is relative here. The space between galaxies are pretty good vacuums but still have about 1 trillion molecules per m3)
Hydrogen – just based on the physics of turbo pumps and diffusion pumps alone, hydrogen is near impossible to actually remove from space entirely. One of the ways UHV chambers which analyze hydrogen are made is by using either titanium coated insides or making the whole vessel out of titanium which picks up hydrogen quite easily so you can get down to super low levels of hydrogen
Vacuum is hard, and there will never be a perfect vacuum in a 3d space.
Let’s start:
Have a chamber, with vacuum pumps.
To begin, roughing pumps can remove a lot of the air, but not completely, because at low vac, air is a fluid, and flows.
After Roughing pumps have done their job, you switch in (generic) Roots blowers, to try to feed tge roughing pumps. That might, with perfect roots blowers, get you down to 10 to the -3 Torr.
For big chambers, now you need a vacuum booster, which is basically a big cone, surrounded with cooling pipes, to coerce random air molecules to get a bit heavier, acd be able to let the Roots Blowers suck out.
Now we are at the level of literal molecules, not attached to one another, so pumping can’t happen.
At this point, you open up HUGE ports, into your Vacuum vessel, and spray through almost a Christmas tree, extremely low vapor pressure oil, that captures floating molecules of any gas, concentrates them at the bottom of the diffusion pump, and uses roots blowers/roughing pumps to get those few molecules out.
But not every air molecule, or even singular element, will drift into the diffusion pump in a reasonable time. Now we are at 10 to -6/7 torr.
There are still individual atoms/molecules floating randomly around. They don’t flow with any other, and we can’t cool them to make them sink. So there is not a perfect vacuum.
Even if I could create a perfect (and I mean PERFECT) Vacuum, quantum particles would randomly appear/disappear.
I used to be a vacuum engineer, I spent like 8 years mostly designing vacuum chambers lol. One issue is the pumping, yes. You are trying to move air molecules from a space where there are basically none to a space that’s at 1 ATM. There’s really no way to do that absolutely perfectly, there will always be some atoms left behind. Imagine you have a steel container half filled with motor oil (I’m using oil because it won’t just evaporate over time lol). You want to get 100% of the oil on one side of the container and absolutely none on the other side. How do you do that? Maybe you can put a divider in and start scooping oil over to one side. Ok, but eventually you’ll be left with a surface that’s still coated with oil and no way really to completely remove every single atom of it, plus your divider is not a perfect seal. That’s kind of how vacuum is, it’s easy pumping the air out at first but as you get down to lower pressures it becomes almost impossible.
So that’s one issue, another issue is leaks. All seals leak, the only question is how much. Copper seals (conflats) are basically the best we have and while they do work very well they still leak a little bit. Even if you have a magic pump that can create a perfect vacuum your seals will leak anyway.
Other major issue is outgassing. Materials, even metals, will outgas, meaning when placed under vacuum they will give off gasses that are embedded in the materials which will contaminate the vacuum. You can reduce this by using vacuum compatible materials but it’s usually not perfect. Another issue is off gassing of dirt and residue that’s on your materials. Things like grease, oil, water, etc will simply vaporize under vacuum. Even fingerprints will contaminate a vacuum. If you’re trying to do things as good as possible you would clean all your parts multiple ways (acetone, IPA, ultrasonic) and then assemble them in a cleanroom. Then you bake it at several hundred degrees, under vacuum, for several hours to try and drive off as much residue as you can. That’s not perfect either though, there’s always going to be some residue left behind.
I am sitting at work monitoring a TVAC (thermal vacuum) test for space-bound hardware right now, and have about 2 decades experience building, operating and maintaining high vacuum, ultra-high-vacuum and Extra-high vacuum systems for use in scientific research, aerospace testing, and nano-materials fabrication (like your sputtering setup, PVD etc).
One of the main concepts that a lot of people don’t grasp that makes reaching deep vacuum levels hard is the difference between what is called viscous flow and molecular flow. At pressures we are used to, gasses are said to be in the viscous flow regime. This means that the molecules of gas are densely packed enough that they interact with (impact and bounce off of) each other a LOOOOT more than they interact with other stuff (surfaces they would exert pressure on). This gives them the property known as viscosity. This also means that if you manage to pull or push on this ‘chunk’ of air over here, it will have an impact on the chunk of air adjacent to it, which will have a slight impact on the chunk of air next to IT, which will have a slight impact on the chunk of air next to… well you get the point. The air is so closely packed that the molecules sort of ‘drag’ each other along when they are pushed or pulled. Kind of like being in a closely packed crowd. If some people start moving one direction, you could get pulled along with them. This makes it relatively easy to evacuate (pump down) a chamber when in the viscous flow regime. You can just use a pump to start pulling air out of the side, and it will drag a lot more other air with it, which will drag other air with it, etc. And you get a steady flow out of the chamber.
However as you remover more and more air the pressure inside reduces more and more, and the molecules get less and less densely packed. At some point they will transition into the ‘molecular flow’ regime (usually somewhere around 1×10-3 Torr). This means that the molecules are more likely to interact with another surface (like the walls of a chamber) than with each other. The technical definition is that the ‘mean free path’ (the average distance a molecule will travel before it hits another molecule) is larger than than the dimensions of the chamber the gas is in). Instead of people in a densely packed crowd, the molecules are now billiard balls bouncing randomly around a poool table the size of a ballroom. They just go around bouncing off the walls, and very rarely hitting each other. Now you no longer have any ‘viscous drag’ to help you out. If you want to get those billiards balls of the tables, it’s really tricky. You can basically make a big door and wait for them to randomly bounce out through the door… but then theres a good chance they bounce off a wall in the hallway and just end up back in the ballroom again. So you make the back wall of the hallway suuuuuper cold (just 10-12 degrees above absolute zero) so that when the billiard balls hit it they freeze and stick to the wall (cryocapture pump array), or maybe you have a sort of big fan blade at the end of the hallway and when the ball reaches it, it gets whacked by the angled backside of the blade and gets knocked into another room (turbomolecular pump) or maybe a machine that shoots a huge waterfall of billiards balls down the hall into a giant pit so your billiard ball gets caught up in them and pulled out with them (diffusion pumps). Either way, it gets REAL tricky to get those last few billiards balls out and keep them out. And you find out pretty quickly that geometry (shape of your chamber, how big the ‘door’ for escaping billiards balls is, etc) becomes a lot more important to how quickly the evacuation goes than how big your capture/pump device is. And theres not much you can do to make it go faster, it basically becomes a matter of statistics (you’re just waiting for those last few molecules of gas to randomly bounce their way down the throat of your pumping system and interact with your UHV/EHV pump)
And then the LAST piece of the puzzle is: EVERYTHING LEAKS. At temperatures above absolute zero, gasses will diffuse through solid metal. Much less any sort of seal. Even UHV seals like the copper ones you describe (Conflat seals) will leak SOME amount. If you’re only pulling down to the 10-6 Torr range? you can get by fine with KF/ISO style elastomer rubber seals. They’ll leak a little, but your pump will be able to keep up to maintain your pressure level. Want to get down into the 10-9 range? Gonna need single-use copper Conflats. They’ll reduce the leaking enough that your same pumping system can get you a little deeper. Wanna go deeper than that? How much money and time you got?
Latest Answers