On the Technical Information Service Page a frequently asked question is: “Pumping from atmosphere to 1 × 10-6 Torr is taking over five hours. What's the problem?” Obviously, there is no hope of answering without a long assessment that could easily be made by the questioner.
- Given the chamber's volume, cleanliness, pumps used, conductance from chamber to pumps, is a pump-down time of < 5 hours reasonable?
- Is this the first time the chamber has been pumped down?
- Does it take five hours for every pump down?
- Has the time slowly increased?
- Has the time suddenly increased?
Let's examine each of these questions.
- Is the chamber volume very large and the rough pump speed very small?
- Does the chamber contain very large surface areas and is the high vacuum pump small?
- Is the high vacuum pump's base pressure close to 1 × 10-6 Torr?
- Are the chamber walls clean or dirty, oily, pitted, or corroded?
- Does the high vacuum pumping port have the same ID as the high vacuum pump inlet?
- Is the high vacuum pumping port's length more than three times its ID?
- Patience! The initial pump-down removes the loosely bound vapor layers absorbed on every surface. Regard it as “vacuum conditioning”. Pump the system for several hours, let up to atmosphere with dry nitrogen and pump-down again. Then, if the pump-down still takes five hours, consider B and C.
- Examine the simple things that can cause long pump-down: check for leaks with a leak detector; regenerate the foreline trap; check that gas inlet valves are fully closed; consider the outgassing characteristics of the construction materials; check that the cross-over pressure is appropriate for both rough and high vacuum pumps; check if the foreline pressure is at an acceptable value for the high vacuum pump.
- Did someone goof in the original design? Check using a computer modeling program such as VacTran®. When you allow for typical outgassing rates, are the results consistent with the measured pump-down time?
Five hours every time
Convince yourself the system does not leak and then, using a computer program like VacTran, model the system. Check that the calculated pump-down time is similar to the measured one. If not, add outgassing sources (roughly modeling the real parts in the chamber) until the pump-down time equals five hours. Now you have two options:
Reduce gas load by either:
- modeling the removal of non-essential components or reducing surface areas
- modeling a reduced outgassing rate achieved by baking or plasma cleaning; this is usually the least expensive option to improving pump-down times
- If little can be done about the total gas load, play “what if” games with the model by changing pumps and conductances to see what must be done to improve the pump-down time. This is always and expensive option.
Slowly gotten worse
Time to consider maintenance issued:
- Is the fluid in any oil-sealed pump contaminated with a vapor?
- Are the foreline or system traps overloaded?
- Is something inside the chamber thermally decomposing?
- Have the fill-full sensors of an LN2 trap changed position?
- Are the chamber walls contaminated will oil from the pumps?
- Are O-rings aging due to high temperatures?
- Does the cryo-pump need regeneration?
More potential issues can be added but the real question is, can the main cause be detected? Fortunately, in most cases the answer is yes, so attach an RGA. Of course, there are drawbacks: RGAs are not cheap and you must learn to interpret spectra. But, as a vacuum diagnostics tool, the RGA has no equal.
Suddenly gotten worse
- Check your vacuum system's logbook. What did you last do? Change a flange or gasket? Add a component? Change the pump fluid? Then check that the change did not cause a leak, increase the outgassing rate, or reduce the effective pumping speed.
- If the sudden increase in base pressure occurs after the first chamber bakeout or after three or four pump-downs following system commissioning (when no changes have been made), then make another leak check of the whole system. Real leaks are easily blocked by ice (the effect of vacuum on water trapped in the leak during the final chamber cleaning) or combination of machining oils and “residues”. When the blockage evaporates or disperses, suddenly the chamber has a leak that was previously not there.