Tuesday, December 11, 2007

UHV Part 2, the Bake-Out

Last month, I wrote about how we seal flanges using copper gaskets on our Ultra-High-Vacuum (UHV) system. Just sealing a system up, however, does not get rid of all the air and hooking a standard Hoover up to it won't do the job either. In fact, just hooking up all of our more sophisticated pumps won't do the job either. They work to keep us at UHV once the low pressures are achieved, but they won't do the job just by being turned on at atmosphere. In order to get to UHV pressures after venting the system to atmosphere, we have to do what is called a "bake-out."

The reason that UHV pumps cannot evacuate the chambers to UHV pressures from atmosphere is due to the fact that lots of atmospheric molecules (water is a particularly nasty one) stick to all of the surfaces. While the pumps can evacuate the gas molecules, they will be constantly replenished by molecules desorbing from the walls of the chambers to enter the gas phase. The idea of a bake-out is simple: heat the heck out of the chambers while pumping on them to encourage almost all of the molecules to come off of the walls to get pumped out over the course of a few days.

In practice, the basics of the bake-out are shown in these pictures. Once the system is completely sealed and leak tested, the entire system is wrapped with heating tapes and we place thermocouples everywhere on the system that we want to monitor. These heating tapes are phenomenally expensive and consist of nothing more than a piece of wire in some fiberglass that you plug into the wall; whoever manufactures these things must make a killing, particularly because they burn out all the time. Once the heating tapes are in place, we wrap the entire system in several layers of aluminum foil. This helps distribute the heat from the tapes over the entire chamber since the chambers themselves are made of the very poorly thermally conducting stainless steel. The tapes are then plugged into variacs, which let us apply however much voltage we want to them (to heat whatever parts of the chamber are wrapped in them to the desired temperatures) rather than just the 120 V from the wall.

For a few days, we slowly ramp the temperature up on the system without letting the pressure get too high. The goal is to get each part of the system as hot as we can without damaging any parts of the system (in this case, our microscope, which has a heat tolerance of about 150 C) or any of the seals, which vary from 150 C to 250 C. Once we have the system nice and hot, we say that it is at "full bake." We monitor the temperature at various places in and on the system with the thermocouples that we put on the system before wrapping it, plugging our thermocouple reader into each. We let the system sit at full bake for several days while we monitor the pressure, the presence of various atmospheric gases, and the temperature. Additionally, we do several procedures such as degassing filaments (turning any filaments on in the system to hot to desorb gases from them), regenerating our non-evaporable getters, and "burping" our ion pumps. (These terms are for another future UHV post on the five different types of pumps we use.) Finally, we switch the UHV side of the system over to the ion pumps and cool everything down. When we cool the system down, we cool the core of the system down first and the extremities last because we don't want gas molecules readsorbing at the extremities before the bake-out is complete. If all has gone well, when the system reaches room temperature, the pressures inside will be classified as ultra-high-vacuum!

Posted by Picasa


Ed & Ann said...

This is like watching Mr. Wizard only better. Thanks for trying to educate this old brain.


Anonymous said...

Greg, We have a mechanical pumped furnace that does'nt pump lower than 1mTorr. How high do you think I should heat the furnace to remove almost all water from this ? I need to prep a standard op procedure so we can get rid of the water every single time we open and close the furnace. We can go upto 300C (Stainless steel tube but has VCR gaskets) but no more. I have access to N2 and Ar as well and can purge and pump the system as many times as I can.
This is short of having a load lock onto the furnace flange which would be the correct way of getting stuff off the wall and keeping it that way.

Greg said...

I'm certainly no expert on bake-outs in general. Some of the parts of our system that only pump down to a millitorr are often not even involved in the bake, but they never come in contact with the UHV part of the system. To get to UHV pressures we heat almost nothing to higher than 250C, but even our turbo pump (which will pull 10^-9 torr) takes a couple of days to pump all the water out. Depending on what your tolerance for water is, I'd make sure that you're at least above 100C (the hotter the better, of course, within the constraints of your gaskets), but I have no idea how long you'd need to bake. Do you have any way to monitor partial pressures? We use an RGA, but that can't be turned on until 10^-5 torr or lower pressures.

Best of luck.

Anonymous said...

Never go above 250C as the copper gaskets are only rated to that temperature. If you are using Viton anywhere in your chamber, you have to stay below 150C. You can never go too hot as long as your chamber can support it.

Anonymous said...

Dear Greg, I am not expert in vacuum technology, however I have a question .While the system is bake-out,the mechanical pump(rough pump,or turbo molecular) must work during procedure?