Condensation on AR coatings?

[Peter Sinclaire]

Roger Smith writes:

> 	After liquid Nitrogen runs out the flask and molecular sieve tend to
> warm faster than the massive and thermally isolated detector mount.  Of course
> radiative loading onn the detector helps initially but not when the
> temperature gets closer to ambient (goes as T^4). So the sieve which has
> soaked up all the contaminants dumps a fair fraction of it back into the
> vacuum and other material is liberated from the dewar walls.  All this then
> goes and condenses on our precious detector mount.  Of course when the
> detector warms up too, the contaminants will evaporate again into the vacuum.
> 
> 	On cooldown, the can and sieve cool first and do their job soaking up
> the contaminants, so the detector looks acceptably clean.  Thus CCDs which
> don't need UV flooding can survive for many months and many thermal cycles
> with little problem.
> 
> 	As a reminder that this process is going on we have an incident about
> once a year where someone is late refilling the dewar.  ...They find it half
> warmed up and hurriedly refill it.  This is a bad move since it results in
> freezing contaminants which have condensed onto the surface.  While this frost
> more than halves the QE and makes flat fields look grainy, complete recovery
> is acheived by warming completely and cooling again.  (Its good to pump the
> dewar out, but this can be omitted in emergencies.)
> 

This gives me the idea that we should have an inteligent Temp. controller,
which should keep a minimum (prefixed) current on the heater, even if the
CCD is warmer than the set temperature: thus, the CCD will always be
warmer than the other dewar components... Or even better, the prefixed
current should be switched on, whenever the temperature goes higher than
the let's say, 10 degrees above the set temperature (i.e when the LN2 has
evaporated). What do you think?

ABout the AR coatings being contaminated: we had a year ago an accident in
which a TEK512 AR coated (TEK#33) got frost on its surface due to a sudden
lost of vacuum. It produced an ugly stain which covered 1/5 of its imaging
surface. It was clear that the efficiency in this area droped
significantly when we tested it later. Baking the chip on a vacuum oven
did not help. Eventually, my colleagues (Peter Moore and Nicolas Haddad)
attempted to clean it with a very soft piece of optical tissue cycling
between freon and ethanol. It worked fine and the AR coating seems to have
survived! The stain can hardly be seen now, either by eye inspection or on
Flat field images. The only signature we see, is the border of the stain
which has a slightly lower sensitivity but disappears with flat fielding.
The CCD went back to work at the telescope a few months later.

Peter

 --

============================== E.S.O. =================================
Peter Sinclaire                     e-mail: psinclai at eso.org (Internet)
Optical Detectors, TRS, La Silla            psinclai at dgaeso51 (Bitnet)
European Southern Observatory       Tel:  +56 (2) 699-3425
P.O.Box 19001, Santiago 19, CHILE   Fax:  +56 (2) 695-4263
++++++++ Over 19 million pixels serving science every night! ++++++++++