Pump Descriptions Vacuum Results The Cleanroom VUV Testing and Integration
Also known as extreme ultra-violet, light whose frequency lies on the border between ultra-violet and x-rays is a thing virtually unknown on Earth's surface. Readily absorbed by matter, VUV light passes through our atmosphere the way you can shine a flashlight through clam chowder. However, since this light is intensely emitted by the sun's atmosphere, we want to look at it. The only way to do that, of course, is to look from space. MOSES is one of many experiments designed for just such a purpose. Since any sane investigator would like to have their equipment tested before flight, there has to be a way of using VUV light on Earth. This is done by reproducing the space environment, or close to it, in a vacuum chamber. To this end, I have been working on converting the Space Science and Engineering Lab into a new testing environment complete with a 100 cubic foot monster of a vacuum chamber kindly donated by Goddard Space Center.
Liquid Nitrogen ->
<- Chamber
Sorption Pump ->
^ Vane Pump
Pump Descriptions
The chamber, old as it is, has been the least of my troubles. Our goal is to reach 10^-6 torr (atmospheric pressure being around 760 torr), and to do so in our case requires three stages of pumping. These are mechanical, sorption, and cryogenic, which I'll explain as we go. The first two are housed in the small cart to the left of the chamber. Inside the perforated steel in the bottom is a rotary carbon vane pump. This is the primary stage, and operates essentially as an oil-less mechanical pump. Reliable to a tee, it consistently brings the chamber to under 100 torr. Beyond this point, it becomes necessary to utilize the sorption pumps. These are amazingly simple devices, with no moving parts aside from the valves to open them. They operate on a temperature based pressure differential, in which the pumps are submerged in liquid nitrogen. The 'pumps' are actually stainless steel bottles which contain Zeolite, a highly porous ceramic bead. Each gram of Zeolite (also called molecular sieve) has about 1000 square meters of surface area, and hence provide an excellent condensation structure on which the evacuated gases are trapped. These can usually get a container to < 5 x 10^-2 torr.
A Typical Sorption Pump
The final stage of pumping, which we have yet to implement, involves the cryogenic pumps. There are two that attach to the far side of the chamber. These work on an almost identical principal to the sorption pumps, except that they use liquid helium (provided by He compressor/refrigerators) and a type of charcoal instead of the Zeolite. Dave Linard, the man who originally ran the chamber for NASA will be visiting later this summer to help get them running. They should eventually push us to 10^-6 torr.
One of our Cryogenic Pumps
Vacuum Results
Over the past two months, our vacuum capacity inside the chamber has gone from zero (you could hear air rushing in, and the gauge wouldn't budge) to our current record of 9 x 10^-3 torr. Getting there has been an enormously tedious process. Ninety percent of this improvement was made by leak testing. For gross leaks, pressurizing the system with nitrogen gas to about 1500 torr and then spraying joints looking for bubbles and/or sputtering noises worked well. For finer (the most numerous) leaks, we would spray suspect joints with alcohol while in the micron (10^-3) range and look for spikes in pressure on our gauges. Many joints required extra soldering, and many more required several rebuilding attempts before they sealed. Some figures for our initial two pumpdowns are below. The gaps are normal - there are blind spots in our pressure gauges.
The Cleanroom
Even if the pumps end up working perfectly and the chamber becomes utterly airtight, there will always be another source of leaks - people. Whenever any kind of contamination gets into a vacuum, it outgasses. This is merely evaporation or sublimation of its particles and is known as a virtual leak. These seem like leaks because they increase the pressure by adding to the gas inside, but cannot be detected externally. To MOSES, the most dangerous of these are hydrocarbons (any molecule composed solely of carbon and hydrogen). This is due to their nasty tendency to polymerize upon exposure to VUV light. What you end up with is a lot of gunk on your optics. To combat this, and other forms of contamination, a cleanroom will be constructed around the loading end of the chamber. The current design for this yet-unconstructed apparatus is a stainless steel floor surrounded by low-outgassing vinyl curtains, suspended from a tile ceiling interspersed with HEPA filters (Kinda like what the scientists kept E.T. in). Since the curtains won't quite touch the floor, the positive pressure inside will make the room into a kind of inverse fume hood. There will be a gowning room for people to put on protective clothing, and enough clean work supplies inside to not have to leave until you're done. Another precaution, since we will be dealing with CCD's (essentially very sensitive digital cameras), is electro-static dissipation, or ESD. This basically involves grounding out everyone and everything in the room, and maybe even making the air slightly conductive with an ionizer to help keep charges down. Much will also depend on procedural matters, but remains a very important matter since even the kind of shock you get from the doorknob is enough to fry hundreds of CCD's.
Most of what I've done for the cleanroom so far is just planning. Assembly shouldn't take place until right around when the fall semester begins. Facility Services is doing most of the prep to the room, but some things (like a rack to mount the chamber electronics on) were built by Corey and I.
VUV Testing and Integration
There's only one missing piece to creating a space-like testing environment.. the sun! MOSES requires an emission source that operates at 304 Angstroms. A continuous source would of course cover that wavelength, but since we only need the two He II lines, an line emission source will probably prove more economical. The other concern is that the light needs to be collimated, or in other words, the light rays need to be parallel - like the sun's. A typical collimation system works like a telescope in reverse, with a pinhole point source at the focal length of a lens. Filters can be added as necessary.
Source ->
^ Filter ^ Pinhole ^ Lens
The picture above is slightly misleading, because VUV light cannot pass through a lens. It can, however, be moderately reflected - given the right expensive materials (gold, platinum..) And since reflecting telescopes are quite common (Cassegrain, Maksutov, etc.) then just using one in reverse should be relatively simple. The donation from Goddard that included the vacuum chamber also included a collimator designed to fit inside and ride along the railings on the walls. Upon initial inspection, the primary mirror seemed a bit dull, suggesting a substantial amount of scattered light. If we stick with this mirror, it will have to be recoated.
The Collimator
Further inspection of this unit included some rudimentary focal length tests. For this, a Foucault knife-edge test is normally the most efficient. But since a large collimated source is required to accomplish this, we opted to use a form of auto-collimation. Auto-collimation generally works by placing a flat mirror perpendicular to the plane of focus just outside the exit end of a collimator. If the light source is placed at the focus, then the light that bounces back from the mirror should come back to the same place. We reflected with a cube's corner (retroreflector) instead of a flat mirror to keep things easy and moved the laser/pinhole until the return image was lined up with it. The image below shows the laser and the spatial filter (pinhole) acting as a source and aligned towards the collimator.
Auto-Collimation Setup
Since we certainly can't use the helium-neon laser in the picture to illuminate the payload, the next big step is to find a VUV source. As I mentioned before, we can probably use an emission line source, and the easiest and most common is a hollow cathode tube. These operate by running an anode and a hollow ended cathode in a rare gas of your choice. The discharge ionizes gas atoms, which then speed into the cathode and "sputter" metal atoms into the gas phase. Upon decaying, the metal atoms emit a line spectrum. During the process, many of the rare gas atoms are also excited, lending their emission lines to the mix. If the fill gas happened to be Helium, then we would get the He II lines we desire. Using a less participative metal and imposing filters before the chamber could limit the final emission to just the lines we want.
Assuming the hollow cathode choice sticks, I'll have several important things to do in order to get it running. First, the cathode needs to be bought or built. Marilyn Bruner, an expert guest MOSES had for a week this summer didn't believe that it would be terribly difficult to make one from scratch, but if there isn't time for or faith in a homemade, we can easily order one. Next, since there has to be gas in the tube, and not in the chamber, some differential pumping will be necessary. To manage this, the tube will be mounted outside the chamber and a turbomolecular or similar pump will take care of excess gas. A high-voltage power supply will need to be installed to run the cathode. The final concern will be filters, which for VUV must be very thin. This means they will be on the order of a couple thousand atoms thick, and very sensitive to pressure differentials (a torr difference will pop one). Because both the tube housing and the chamber will be at pressures well under a torr, however, this should reduce to a merely procedural issue.
VUV Source Setup
About Me
Conclusion & Thank You's
Home