Contamination

[Chris McCowage]

Posted to CCD-world:
-+-+-+-
Contamination problems.

We are currently having contamination problems and wish to consult the
community. We have followed the various discussions regarding cryostat
contamination with great interest.

We operate two spectrograph cameras on the AAO 2dF system. The cameras
have embedded tek 1K detectors and are cooled by CTI closed cycle
cryodynes.  The volume of the vessels is approx 33 litres. 

The two cameras are virtually identical, however they exhibit different
behaviour. The problem we are experiencing is contamination of the field
flattener leading to halation. The contamination is visible to the naked eye
as a light frosting. Camera 1 generally has this at a quite low level and it
does not cause measurable halation in the spectrograph images. However the
problem is serious in camera 2. The problem is worst at high ambient
temperatures and requires frequent thermal cycling and re-pumping of the camera.

The contamination evaporates/sub-limes when the cryostat is thermal cycled and
re-pumped. We always reactivate the getter by electric heating. Unfortunately we
have no temperature sensing on the field flattener, but the field flattener
clears some time after the detector mounting block has got to around 273K. The
field flattener could still be colder.

We would be interested to hear of peoples experience of analysing cryostat
exhaust with residual gas analysers or mass spectrometers. We have had
difficulty in locating a facility in Australia prepared to analyse an "unknown"
material for fear of contaminating their instrument.

Both camera 1 & 2 have been subjected to careful leak testing and have very low
pressure rise rates. In fact, camera one can last a long run and require
dry air flushing to the corrector to prevent dewing due to the pressure rise in
the cryostat but not have contamination while camera 2 has no external
condensation problem, but suffers from contamination of the internal field
flattener. Camera one can remain servicable for over 21 days.

The materials used in these cameras are similar to the materials that the AAO
has used for the last 20 years in the instrumentation and detector development
programme. They include the following.

	Optics
	 
	  aspheric corrector ( cryostat window )
	  camera primary mirror
	  field flattener
	  
	Detector
	
	  Tek 1K
	  mounting block heated by strain gauge attached with epo-tek H70E
	  silver impregnated epoxy
	  1N914 diode used as temperature sensor attached with epo-tek
	  Teflon detector mounting block stack with nylon screws
	  Decoupling capacitors
	  Fibre glass board ( G10 ) used as detector spider vanes
	  Wiring uses thermocouple wire with teflon insulation, a couple of
	  points secured with a drop of cellulose cement.
	  
	Getter
	
	  Linde Type 5A Sieve ( Crystalline calcium alumino-silicate ) mounted
	  in own design tray with a single resistor for reactivation.
	  
	Internal finishes
	
	  Electro polished stainless steel
	  Machine finish aluminium. Originally included a significant area of
	  black anodising, which has been removed without improvement in
	  contamination. Now replaced by a water based paint of a type used in
	  previous cryostats.
	  
	Seals
	
	  "O" rings, dry or with minimum of vacuum grease apiezon class "M",
	  no vacuum grease used has a vapour pressure of 10e-6 or higher.
	  Three piercings for motor mic driven focus drive, "o" ring seals. (
	  no leaks )
	  
	Cleaning
	
	  Optics Teepol and water
	  Internal surfaces Gensolv 2000 1,1-Dichloro-1-fluoroethane CH3CCl2F
	  Or ethanol followed by isopropol alcohol

Trapped volumes have been avoided in the design of the internal 
construction.

Camera 1 has almost unmeasurable pressure rise of less than 0.001 lusec.
Camera 2 has no detectable leaks and a similar low pressure rise.

Leak rates are measured as a volumetric flow rate - a pressure rise of 10
milliTorr in a vessel of 100 litres is 10 times the leak rate as the same
pressure rise in a 10 litre vessel. The unit that we use is litre-microns/sec
also known as the lusec.

Leak rate (lusecs) = Volume(litres) X Pressure rise in microns (10-3 Torr)
                     ----------------------------------------------------
                                       seconds

Futhermore a plot of pressure rise (millitorr) versus time reveals a linear
relationship. Our advice is that a leak will cause a linear increase in
pressure while outgassing will cause the pressure to rise to a steady state
value determined by the vapour  pressures of the desorbed gases.

There was an unfortunate event early in the development of the system where a
vacuum pump failed and there may have been back streaming of vacuum oil from
the backing pump. The internals of the cryostat have been cleaned as above.

We now maintain a strict regime of pumping with oil free systems, reactivation
of getter material and management of detector temperature during cooling.

Any input, recommendations and strategies would be appreciated.

Regards,
Chris McCowage

-+-+-+-
For information about CCD-world, send email to owner-CCD-world at cfht.hawaii.edu.