Improving vacuum system efficiency?

[rkum99]

So I was looking through the internet on how one could improve the efficiency
of a vacuum system (with a diffusion pump). Lots of websites describe systems with,
cold traps to remove condensable vapors such as water. If I am not mistaken, cold
traps also help prevent back streaming, similar to a baffle. So, I suppose my first
question is - how does this all come together in quantitatively? Depending on the
pressure level the system is operating at, how do baffles and cold trap affect its
pumping speed and capability? If there isn't any easy explanation for that question,
would there be any purpose in investigating it?

Also, what are some methods people use to improve pumping speed or the pressure
levels a system can attain (other than improving the quality of the components in
the system). Taking cost into consideration - what would some options be?

I'm pretty much new to all of this - so If I'm speaking nonsense here, sorry :P.
My "scientific jargon" is probably messed up too. Thanks for any and all answers!

 

[M Quack]

IMHO the best way to improve the efficiency of a diffusion pump is to replace it by a turbo pump.

The problem with diffusion pumps is that the oil eventually breaks down and fragments get into your vacuum system. Cold traps and baffles are ways to reduce this contamination.

For a better answer you have to provide a bit more information about your vacuum system and what you are trying to do with it.

The vacuum system I work with usually consist of an oil-free roughing pump (membrane, scroll or roots) backing a turbo pump. This gets us to about 10^-7 to 10^-8 mbar. Beyond that we close a valve between the pump and the vacuum vessel and then use ion getter pumps installed on the vessel. With that we reach about 10^-9mbar.

Keeping the vacuum vessel clean is the most important part. Wear gloves when you work on vacuum components, clean stuff that goes into vacuum with ethanol to remove grease, etc. Avoid water in the vessel, e.g. vent to dry nitrogen and keep a small nitrogen flow while the vessel is open to avoid humidity from the air building up in the vessel.

 

[rkum99]

I guess I should describe my situation a little more - I don't actually have a vacuum system yet. As of
now, I'm pretty much trying to design a system, preferably as efficient as possible. My University's physics
lab has a bunch of stuff in the basement for people to take (some of it is junk, some is in good condition).
Some stuff is pretty dangerous (there is a block of sodium in oil that's been resting for decades, but no one
wants to touch it in case it explodes O_O). Point is, major components like the actual vacuum pump/roughing
pump will probably be scavenged from there.

Yeah, I suppose the easiest way to improve a vacuum system is to replace it with better, more efficient pumps?
But under the assumption I'm using a diffusion pump, what would be the best option? I will probably be using an
oil-free roughing pump - I feel like that would eliminate some extra work (If I remember correctly, don't people
leave oil-based mechanical pumps running continuously to keep the oil warm and dry? - the vacuum system
probably would not be operating continuously, so the pump could rust). So under the assumption of using a
diffusion pump and, say, a roots blower - how could I improve the system. I guess having a baffle is necessary
to prevent back streaming - but what are some other options. Unfortunately, I haven't really seen any additions
in books or online other than a cold trap, and naturally, a baffle.

Once again, hope I am not saying nonsense :).
Thanks for the response!

 

[M Quack]

Why do you need vacuum in the first place?

 

[rkum99]

Honestly, I have no idea currently. I assumed I could use it for other projects in
the future. If you think this is a lost cause, I would appreciate you telling
me - hopefully it isn't. If you had any particular application for which the system's
usefulness could be improved (sorry, weird wording), I'd appreciate the knowledge.

Like I previously mentioned, I'm using somewhat used components, so this isn't
for any real professional application - I would use new, better components in that
situation. I was hoping for some method to make the system more efficient/capable
so I could make the most out of what I make.

Thanks again!

 

[M Quack]

If the goal is to get some hands-on experience with vacuum equipment, then I'd suggest installing some diagnostics like pressure gauges and an RGA if you can get your hands on one, and then trying out different
configurations such as with and without baffles.

 

[monodisperse]

Sodium isn't sweating dynamite. Place it in tetrahydrofuran under an inert atmosphere and slowly add isopropanol.

Cold traps are generally used as a basic cryopump. They actually harm the pumping rate unless well designed because they're typically the lowest conductance component of the vacuum system.

If you've using a diffusion pump then don't worry about the oil in the roughing pump. It won't diffuse against the diff pump. It's the diff pump working fluid you have to worry about. Unbaffled the pump will never pump below the vapour pressure of the working fluid. However, a liquid nitrogen trap cuts the vapour pressure of these to essentially nothing.

 

[rkum99]

Yeah, I hardly know anything about that. Probably a little exaggerated, but I guess the superoxides that have formed over the years could be explosive. Anyway, I obviously don't know anything about this. So better stay away from it either way :).

So maybe I framed my question wrong. Efficiency normally applies to the pumping speed of my system, but I also was looking for ways to decrease the minimum pressure the system can attain. So a cold trap, by removing certain vapors, could assist the pumping process? Then again, like you said, conductance rate is slow, so it would detract from the pumping rate. It seemed like a standard in many of the designs I glanced at, but if I can't get my hands on a decent one, disregard one completely? I seem to be getting a wide range of feedback on cold traps...

If you've using a diffusion pump then don't worry about the oil in the roughing pump. It won't diffuse against the diff pump. It's the diff pump working fluid you have to worry about. Unbaffled the pump will never pump below the vapour pressure of the working fluid. However, a liquid nitrogen trap cuts the vapour pressure of these to essentially nothing.

I'm a little confused at what you're saying here - sorry if this is something I should know. So are you saying that when baffled, the pump will pump below the vapor pressure of the oil? I somewhat assumed that baffles were used to reduce the amount of backstreaming in the system.
Similarly, why would a cold trap cut the pressure to essentially nothing?

Thanks for the help!

 

[monodisperse]

Yeah, I hardly know anything about that. Probably a little exaggerated, but I guess the superoxides that have formed over the years could be explosive. Anyway, I obviously don't know anything about this. So better stay away from it either way :).

I quenched a bunch of decades old sodium a few months back. You just get a lot move sodium hydroxide on the surface.

So maybe I framed my question wrong. Efficiency normally applies to the pumping speed of my system, but I also was looking for ways to decrease the minimum pressure the system can attain. So a cold trap, by removing certain vapors, could assist the pumping process? Then again, like you said, conductance rate is slow, so it would detract from the pumping rate. It seemed like a standard in many of the designs I glanced at, but if I can't get my hands on a decent one, disregard one completely? I seem to be getting a wide range of feedback on cold traps...

Depends. You normally want to protect the pumps, since they're expensive, but unless you have a wide bore cold trap the pumping speed will suffer. However, the conductance of the whole system is less than the conductance of the lowest component of the vacuum line. Hence a 4 mm stopcock will cut pumping down to ca. 0.5 l/s regardless of the pump. So it may not actually matter.

The cold trap however acts as a cryopump. Condensible gases (mainly water) will rapidly condense at liquid nitrogen temperatures. If you can't use liquid nitrogen and pumping rate isn't a problem you can pack a trap with P2O5, molecular sieves, activated charcoal and cool with a dry ice slush which will also work. Pumping speed is throttled, but ultimate vacuum is better.

I'm a little confused at what you're saying here - sorry if this is something I should know. So are you saying that when baffled, the pump will pump below the vapor pressure of the oil? I somewhat assumed that baffles were used to reduce the amount of backstreaming in the system.
Similarly, why would a cold trap cut the pressure to essentially nothing?

Without a baffle/ trap the working fluid will diffuse ("backstream") throughout the system at its vapour pressure. With a properly working trap this isn't a problem. It was more of a problem with mercury (2x 10^-3 torr @ 20oC) than with modern fluids (DC-704 = 3x 10^-8 torr). This will be the limiting vacuum, if nothing else limits it at a lower vacuum.

 

[rkum99]

The cold trap however acts as a cryopump. Condensible gases (mainly water) will rapidly condense at liquid nitrogen temperatures. If you can't use liquid nitrogen and pumping rate isn't a problem you can pack a trap with P2O5, molecular sieves, activated charcoal and cool with a dry ice slush which will also work. Pumping speed is throttled, but ultimate vacuum is better.

Yeah, diffusion pumps aren't to great at pumping water (I think), which a cyropump does
great. I think I could use liquid nitrogen, but it supposedly could create liquid oxygen. Sounds
like it would be a relatively rare occurrence, but in the event that the liquid forms, careful and
detailed procedures need to be undertaken. However, the systems being constructed in a
lab, so in such an event, I could always request assistance.

I am guessing that I could reach a pressure of something around 10^(-5), but I honestly
have no idea since I don't have all my equipment yet. So could I disregard a baffle in my
system if I use oil with a very low vapor pressure? If I am not mistaken, the trap also acts
as a baffle, preventing backstreaming as well. If I use a cold trap, would it sufficiently
prevent back streaming?

I'm not sure if I'll have more questions. As of now, it seems like I should collect my equipment
and see what I get. Based on that, it will probably be easier to formulate an idea of a system.
Probably would be easier for people to help me too. :)

Thanks!

 

[monodisperse]

Yeah, diffusion pumps aren't to great at pumping water (I think), which a cyropump does
great. I think I could use liquid nitrogen, but it supposedly could create liquid oxygen. Sounds
like it would be a relatively rare occurrence, but in the event that the liquid forms, careful and
detailed procedures need to be undertaken. However, the systems being constructed in a
lab, so in such an event, I could always request assistance.

Liquid oxygen never condenses in a vacuum, even a relatively low one. It only occurs when the line is let down to air without first removing the LN2 dewar.

I am guessing that I could reach a pressure of something around 10^(-5), but I honestly
have no idea since I don't have all my equipment yet. So could I disregard a baffle in my
system if I use oil with a very low vapor pressure? If I am not mistaken, the trap also acts
as a baffle, preventing backstreaming as well. If I use a cold trap, would it sufficiently
prevent back streaming?

Ultimate vacuum is limited by a number of factors. One is effective pumping speed. 10^-6 torr is not really achievable through narrow tubing. Again, the system rather than the pump is usually the limiting factor. Quoted pump ultimate vacuums are directly at the high vacuum connection. The other major factor (ignore leaks) is outgassing of the interior surfaces. This can be minimised by baking out, but even so the system often takes a significant period of time to pump down to base vacuum.

The trap will, of course, condense any vapours backstreaming from the pump. This means upstream from the trap you indeed can pump below the vapour pressure of the working fluid. Indeed this is why they achieved 10^-6 in the 1920's with mercury.

If interested, Review of Scientific Instruments has recently dropped the subscription walls on old articles, where a lot of the original articles were published. - http://scitation.aip.org/content/aip/journal/rsi

A good summary of good technique is in Strong's 1938 book, which is online at: https://archive.org/details/ProceduresInExperimentalPhysics

 

[rkum99]

Hate to bring back a semi-old thread, but I suppose this is a future update of sorts.

I ended up using a system you could see for leak detection - simply a mechanical pump,
a diffusion pump, and a cold trap in series. On its own, it could not pump out a chamber,
and requires an auxilliary mechanical pump to pump it down to medium vacuum. I've been
getting decent pressures on it, considering that the thing is 30+ years old - low 10^(-5) range.

Here is an oddity though. Using very crude measurements, the cold trap didn't seem to
reduce conductance, but actually improvement. With the use of the cold trap, I got a higher
vacuum and was capable of pumping slightly faster. Any ideas why?