From siegmund@astro.washington.edu Wed May  1 17:53:45 1996

John Varsik has nearly a month of particle density data now for Apache
Point Observatory. It seems likely, however, that we are not measuring the
density of particles that are most responsible for the degradation of
telescope mirrors.

John's results indicate that significant departures occur from the Giordano
and Sarazin (1994) results that the number density of particles in the air
is proportional to d^-4. Also, it is apparent that the measured particle
density in the >0.3 and >0.5 micron channels does not necessarily correlate
well with visual appearance, e.g., the visibility of a flashlight beam.

Our concern is the rate of degradation of the scattering and reflective
properties of optical surfaces, particularly aluminized mirrors. This
depends on the scattering efficiency, deposition efficiency, cleaning
efficiency, chemistry, and optical properties as well as initial
distribution of particles with size.

Since we have no information on deposition efficiency and cleaning
efficiency, early this week, Patrick, Russell and I examined some
contaminated surfaces using an optical microscope.

1) Microscope slide exposed near the entrance to the UW Physics-Astronomy
building. The slide was indoors and was exposed about a week.
Biggest particle in 4 mm^2      45 microns
Number > 1 microns              88 mm^2
Number > 3 microns              6 mm^2

2) Microscope slide in Physics-Astronomy building rm B330 (our lab). The
slide was exposed about a week.
Biggest particle in 4 mm^2      20 microns
Number > 1 microns              52 mm^2
Number > 3 microns              1 mm^2

3) Frequent flyer, CO2 cleaned

Biggest particle in 4 mm^2      30 microns
Number > 1 microns              1200 mm^2
Number > 3 microns              20 mm^2 (5 particles, 15 corrosion spots)

4) Frequent flyer, laser cleaned

Biggest particle in 4 mm^2      30 microns
Number > 1 microns              2000 mm^2
Number > 3 microns              ? mm^2

I didn't record the density of particles larger than 3 microns for case 4.
My recollection was that the density was comparable to case 3. The surfaces
of cases 3 and 4 were examined after cleaning.

My subjective impression is that particles larger than 3 microns dominated
scattering except perhaps for case 4 where the small particles were very
dense. In the absence of size sorting and with an initial distribution of
d^-4, one would expect 0.5 to 5 micron particles to dominate scattering.
The cross-section for big particles (>>lambda) goes as d^2. For small
particles (<<lambda), it goes as d^6. Consequently, large particles are too
sparse and small particles scatter too poorly to be significant.

John finds exponents between -1 and -5 for the size power law (over a 14
day period). In the absence of size sorting, flatter power laws will cause
larger particles to dominate scattering. I think this is what I'm seeing on
the test mirrors.

1) We should change the threshhold of the 0.5 micron channel of the
particle counter to 2 microns to get a better measure of the density of
particles that are the putative contaminates.

2) I think we need to do a more quantitative job of determining the size
distribution of contamination on the mirror. Wayne is increasing the
resolution of his test equipment, although it still may not quite be
adequate for this task. Alternatively, Wayne has identified a lab at the
Univ. of Wash. that can measure the optical properties of samples and they
may be able to help.

3) It would be useful to monitor the contamination of the same 4 mm^2 of
surface over several months. This was suggested by Don Brownlee. If we
marked a square on the test mirrors with a Sharpie, we could photograph it
at one or two week intervals, we could determine the size dependent
deposition efficiency and the size dependent cleaning efficiency, and
monitor the growth of each area of corrosion, at least in principle. This
would complement the  experiment that is currently underway.