CO2 snow cleaning procedure
A. How it works
We are using the equipment sold by the R.Zito company. It consists of a 99.99% pure liquid CO2 tank (with syphon) connected to an adjustable valve and a long nozzle blowing CO2 snow at high speed. The liquid CO2 (-56.6C), when released out through a fine nozzle, will suddenly expands without exchanging heat (adiabatic expansion), thus becoming a mixture of gaseous CO2 and dry ice. Two principles govern the cleaning technique: 1/ When colliding with dust particles, the snow flakes transmit their momentum which makes the dust move away, and 2/the frozen dust particles contract and break free from the surface. Blowing gas is much less efficient because the dust particles are much heavier than the molecules of gas blown. In theory (and it has been demonstrated in practice by many experiments), the flakes don't scratch the coating because they slide over the surface on a cushion of CO2 gas which sublimes off the flakes.
Remember that when fresh Aluminium is exposed to air, a thin layer (about 30 Angstroms) of aluminium oxide Al2O3 will form on the surface. That layer is a tremendous protection as its Knoop hardness is 2100 kg.mm^-2 (compared to only 140 kg.mm^-2 for bulk Aluminium).
We also use a filter at the output of the tank to prevent oil residuals from contaminating the mirror (this oil could have been introduced accidentally in the tank when filling up at the factory). The filter is supposed to clean up to 100 tanks of CO2.
A copy of the Spanish instruction booklet ("Instrucciones para el aparato de limpieza que utiliza la nieve de dioxido de carbono", 1995) is kept in the old console room of the 4m telescope.
See the following papers for more information:
Zito R., Cleaning large optics with CO2 snow, SPIE vol. 1236, Advanced Technology Optical Telescopes IV (1990), p 952
Zito R., Removal of adsorbed gases with CO2 snow, SPIE vol. 1494, Space Astronomical Telescopes and Instruments (1991), p 427
Kimura W. and Kim G., Comparison of laser and CO2 snow cleaning of astronomical mirror samples, SPIE vol. 2199, Advanced Technology Optical Telescopes V (1994), p1165
Magrath B. and Nahrstedt D., A cleaning process for the CFHT primary mirror, PASP 108, July 1996, p620
Torii Y., Hayashi S. and Toda M., In-situ cleaning of the primary mirror of Subaru telescope, SPIE vol. 3352, Advanced Technology optical/IR telescopes VI (1998), p 808
The entire process (from taking over the telescope and cleaning the mirror to bringing it back to the zenith) takes about 25 minutes at the 4m. At the 4m, bring the telescope down toward the Cassegrain access stairs (North). Set up the ladder on the edge of the mirror cover petals, lay the wooden board between the ladder and the inside wall of the mirror cell. Don't forget to use shoe protectors when you enter the mirror area. Unroll the flexible high pressure line from the tank. Open the tank.
C. Cleaning session
Before cleaning, take a measurement of the reflected light and scattered light of the surface in a few spots. Hang the ground cable of the nozzle to the hook on the chimney (to prevent static electricity discharge to the operator). Open gently the nozzle valve and let the CO2 snow blow away from the mirror for 15 seconds, the jet should be between 1 and 2 meters long. Then start flushing the mirror from the top to the bottom, successively on each side of the chimney. Try to maintain the nozzle extremity 30 to 40 cm above he mirror with the jet incidence angle at 45 degrees and move the nozzle in a regular pattern at a speed equivalent to 1m^2 in 20 seconds. If ever the nozzle gets obstructed (by a "cork" of snow), close the nozzle valve, close the tank, open the nozzle valve to empty the line. Warm up the valve with your hands or wait a few minutes, shaking gently the nozzle.
Repeat the same steps to terminate the process (i.e. make sure the line is emptied). Don't forget to measure the reflectivity/scattering when you are done.