I was surprised at their comment on turbos pumping water sluggishly. It IS a bit lighter than either N2 or O2, but the turbo's blades are way above sonic speeds, and that normally scuffs them clean of any adsorbed water. Adsorbed water elsewhere will only get pumped if it gets loose from the surface.. hence either heating, or glow discharge scrubbing or a third way... super dry N2 flush at atmospheric pressure. In any case... the lines to the pumps have to be hot or the water will just move there and sit.
Incidentally, the boil-off from an LN2 dewar is reasonably dry, but the lines might not be. And std compressed N2 is usually very wet, unless you paid a lot extra for super dry. Any kind of non-metallic transfer line should be considered a prime source of water. An atmospheric N2 flush through a large dessicant pack or molecular sieve will usally result in a much reduced pumpdown time and much shortened desorption times. A hot dry N2 flush works best ( other than the 250 C backeout) of all the other non bakeout methods.
A final note... don't forget the roughing pump oil as a major resorvoir of water in your vacuum system. Use of the gas ballast and dry gas flushes will keep the pump rate up.
Dave Cooper
Created on Monday, February 26, 2001 6:25 PM EDT by David Cooper