Much more frequently than might be expected, observant vacuum system operators notice the reading on the high vacuum gauge changing either in a regular or irregular manner. Typically, the verbs used to describe the behavior are cycling, wandering, and bouncing. And so that we are all talking about the same effects, I'll define those actions:
- Pressure Cycling: indicates that the chamber's pressure rises and falls with some regularity over periods of minutes or hours, between fairly constant upper and lower levels.
- Pressure Wandering: means just that. The indicated pressure moves slowly but randomly throughout a range of pressures with no predictable cycle.
- Pressure Bouncing: covers circumstances where the pressure (intermittently and irregularly) jumps to a higher pressure, then decays slowly back to the original pressure.
Pressure Cycling
This is often the result of some outgassing material being temperature cycled. For example, if an internal heater uses on/off switching (rather than PID) for temperature control, heated objects may undergo quite large temperature swings. Since every 10°C rise increases the outgassing rate by a factor of ~2, even modest temperature changes are easily detected on an ion gauge. An effect as tame as the room temperature in a non-air-conditioned lab can cause noticeable pressure changes with a 24-hour cycle time.
Diurnal effects on partial pressures are also known. For example, one manufacturer of process mass spectrometers was conducting burn-in tests using the atmosphere as sample. One of the components monitored was CO2 which showed remarkably regular increases after dark and decreases throughout the day except for two spikes: one between 7:00-8:30 a.m. and the other between 4:00-5:30 p.m. The spikes, of course, were related to the local rush-hours, but the nightly increase was a puzzle until someone recognized it was summer and the trees and plant life surrounding the building were “breathing”.
One of the more bizarre reasons ever advanced for total pressure cycling was the switching of heavy-duty electric equipment unrelated to the vacuum. The power drawn reduced the line voltage at the vacuum system's diffusion pump, which, presumably, decreased the boiling rate and influenced the pumping speed enough to affect chamber pressure. Hmmmmm!
Pressure Wandering
Wandering has sometimes been connected with pumping hydrogen using diffusion pumps but usually has been tracked to faulty gauging, unstable voltages or thoroughly dirty pressure sensing elements. Presumably, if the line voltage wandered, for whatever reason, around its nominal value, any heated surfaces or diffusion pump might respond in a way that caused pressure wandering.
Pressure Bouncing
People are often thoroughly confused when they first see this effect. Since ion gauge pressures usually vary only slowly, it's no surprise that gauge controllers are built with long response times. It is a surprise, therefore, to see the gauge needle kick up, or the digits flutter to a higher value, and then slowly settle down. Although the gauge kicks are small, it's the time constant that limits their amplitude. The actual pressure burst may be quite large. We can't tell.
What we do know is, in all cases where the cause has been identified, it was associated with O-rings that are greased or (accidentally) oiled. For example, if your remedy for a leaking O-ring is a heavier grease layer and the O-ring seal happens to be warm, gas permeating through the elastomer forms a bubble of gas beneath the grease on the vacuum side. When its pressure exceeds whatever it is that holds the grease in a skin (surface tension? viscosity?), it bursts, releasing that pocket of gas into the system. Presumably the grease layer reforms and another bubble starts because bouncing will go on for days or weeks.
The most common site for this behavior is a diffusion pump's inlet flange that is O-ring sealed. Backstreaming pump oil covers the O-rings inner surface giving the right conditions for bouncing. The source of gas might be: permeation through the O-ring; leakage past the O-ring; or outgassing of the O-ring.