We also have a non-technical discussion of the film killer in our Introduction to Cryogenics section.
The problem was a leak which threatened to deplete the superfluid helium supply well before the end of the mission. The planned lifetime is 2 years, and the superfluid helium volume only 18 to 20 liters. To meet the planned lifetime, we must hold the total helium loss rate to 40 micrograms/sec. According to our thermal calculations, our helium evaporation rate will be at or slightly below this rate. Therefore, if helium were leaking out (as well as evaporating) we might run out of helium too soon and have to end the mission earlier than we'd planned.
We discovered that superfluid helium was leaking out of the XRS cryostat through the porous plug. The superfluid helium formed a film and "escaped" by flowing away along the walls of the vent tube. The rate at which helium was escaping was about 40 micrograms/sec. That leak rate is the same as the evaporation loss rate that we were expecting. We had to stop that leak, or we'd run out of helium in half the time we'd planned.
The leak through the porous plug is not new with XRS. However, in previous satellites, the leak was such a small proportion of the total helium loss rate that it could safely be ignored. The effect showed up, for example, in tests of the porous plug for COBE, the Cosmic Background Explorer. The rate of helium evaporation in COBE was so much larger than the leak through the porous plug that there was no need to deal with the leak.
The solution is two-fold: a narrow vent tube restricts the amount of of film flow while a heat exchanger evaporates the film and cools the helium tank. (As a backup, there is a series of knife-edges downstream of the heat exchanger.) The vent tube, where it runs from the porous plug to the heat exchanger, has a 0.15 cm I.D. This diameter restricts the film flow to about 9 micrograms/sec. The film and vapor leaving the porous plug are 50 mK higher in temperature than the superfliud helium bath. The copper heat exchanger uses this temperature difference to evaporate the film. The cooling produced by this evaporation then cools the helium tank. We chose the sizes of the porous plug, vent tube, and heat exchanger so that they would work together to give us the cooling rate we needed.
Downstream of the heat exchanger is a knife-edge device, a common method of reducing film flow in dilution refrigerators. In our tests, the knife-edge did not perform as well as the heat exchanger, and is therefore only used as a backup. The principle behind the knife-edge film killer is that a superfluid film is reduced when it flows around a sharp bend. If the bend is sharp enough, the film does not flow past it. Unfortunately, there is a catch: the film evaporates where it stops, cooling the material of the knife-edge. As the knife-edge cools, some of the helium vapor condenses on the downstream side. This newly condensed, somewhat reduced, film then continues to flow downstream. Our film killer uses a series of knife-edges etched into 2 silicon wafers.
"Suppression of Superfluid Film Flow in the XRS Helium Dewar", P.J. Shirron and M.J. DiPirro, Advances in Cryogenic Engineering 43, ed. P. Kittel, p. 949-956 (1998)Abstract
Reference and Other Links
General Links
Cryogenics & Fluids Home Page