Mirror Cleaning Experiment: I. Interim report

Sloan Digital Sky Survey Telescope Technical Note 19960113

• Introduction
• The experiment
• Results
• Discussion
• Acknowledgements

Introduction

Contamination on optical surfaces degrades the throughput of an optical system. Equally troublesome is the increase in scattered light due to contamination. Contamination is particularly pernicious on the unprotected aluminum that is commonly used on the large first-surface mirrors of reflecting telescopes. Unprotected aluminum is so soft that vigorous cleaning damages the coating. As the aluminum coating corrodes, the short wavelength scattering and reflective properties degrade especially rapidly. Because of their size, such mirrors are costly to realuminize. Furthermore, risk to the optic from the handling necessary for realuminization can be minimized, but not eliminated.

Unprotected aluminum mirrors are often realuminized at one to two year intervals. Significant degradation of performance often occurs prior to aluminization. It is believed that much of the irreversible portion of degradation is due to corrosion of the aluminum coating.

We can speculate as follows regarding the mechanism for the degradation of the aluminum surface. As long as surface contamination is dry, it is likely that little degradation occurs. However, when exposed to the night sky, the particles radiate more energy in the thermal infrared than they absorb from the sky and cool below the ambient air temperature. If they cool below the dew point, water vapor will condense on the particles. Moreover, many contaminants are hygroscopic and will absorb water at temperatures well above the dew point. Occasionally, despite the best efforts of the telescope operators, precipitation may fall directly on the mirror thereby wetting the contamination.

Once contamination is wet, soluble substances will be dissolved. Often, the result will be an alkaline solution that will rapidly attack the aluminum coating. Upon drying, deposited salts and the parent contamination are likely to be strongly attached to the surface and cleaning is not likely to dislodge them.

It is obvious that the longer contamination resides on the mirror surface, the more likely it is that it will become wet and degrade the surface. (We imagine that wetting events are episodic, e.g., associated with high humidity or precipitation.) Consequently, frequent cleaning should minimize degradation.

The experiment

Six test mirrors are exposed to contamination and cleaned on a nominal biweekly schedule (Table 1). Each mirror is a ø102 mm polished silicon wafer coated with 100 nm of aluminum. The mirrors are mounted on the 3.5-m telescope for a week at a time. During that period, the covers of their protective cases is opened whenever the 3.5-m telescope primary mirror cover is opened for nighttime observing.

Table 1: Schedule of mirror exposure and cleaning.
Week #1
      Monday
               Frequent Flyer (FF) case arrives at Apache Point Observatory (APO).
      Tuesday AM
               FF case reinstalled in 3.5 telescope.
               FF case covers and Sessile (SS) case covers removed.
               Measure FF, SS, and control mirrors with scatterometer.
               Install control mirror cover.
Week #2
      Tuesday AM
               Remove control mirror cover.
               Measure FF, SS, and control mirrors with scatterometer.
               Install FF case covers, dismount, and ship to STI.
               Install control mirror cover.
               CO2 clean exposed mirrors in SS case.
               Install SS case cover.
      Thursday
               FF arrives at STI, Bellevue, Washington.
      Friday
               FF#1 mirrors are CO2 cleaned.
               FF#2 mirrors are laser cleaned.
               Light scattering measurements performed on cleaned mirrors.
               Ship FF case to APO.

Scattering measurements at APO were made using a µScan Scatterometer manufactured by TMA Technologies, Inc., P.O. Box 3118, Bozeman, MT 59715, (406)586-7684. To describe this instrument, we use zenith angle and azimuth angle with the surface assumed horizontal. The reflected beam azimuth angle is zero. This instrument uses a 670 nm laser diode light source located at 25° from the surface normal (25°,180°). One scatter detector is located on the surface normal (0°,0°). The other is 50° from the surface normal (50°,180°). The reflectance detector is mounted in a light trap (25°,0°).

Measurements were made at the center of the test mirrors. The specular reflectance and bidirectional scattering distribution function (BSDF) at the angles of the two scatter detectors were recorded. The BSDF is a generic term that is identical to the bidirectional reflectance distribution function (BRDF) for a reflective surface.

where

• = measured scattered power
• , solid angle given by receiver aperture (A) and sample to receiver radius (R)
• = measurement position from surface normal
• = power incident on the sample

For the special case of an ideal Lambertian surface, the BSDF is and is independent of angle.

A calibration mirror is measured each time scattering is measured to monitor the performance of the µScan instrument. This mirror has a highly reflective, protected and enhanced silver coating.

The laser and CO2 cleaning techniques have been described elsewhere ("Comparison of Laser and CO2 Snow for Cleaning Large Astronomical Mirrors", Kimura, W.D., Kim, G.H., and Balick, B., PASP 107: 888-895, 1995 September). Results from the STI Optronics measurements will be reported later.

Results

Data were taken beginning July 3, 1995 except for a period between October 17 and November 21 when the telescope was shut down for maintenance. On November 28 and December 12, the sessile mirrors were not cleaned due to a lack of CO2.

A plot of all the scattering measurements shows a zig-zag behavior (Figure 1). The scattering performance is degraded by exposure. Cleaning tends to restore the surface to approximately its pre-exposure level although slow secular degradation occurs. This trend can be seen clearly in those data obtained after cleaning (Figure 2).

Measurements of the calibration mirror (cal in the Figures) are quite consistent and small. This we interpret as indicating proper operation of the instrument. This leaves unexplained the puzzling behavior of the control mirror that shows more degradation than the exposed mirrors.

Fig. 1: All scattering measurements.

Fig. 2: Scattering after cleaning.

Discussion

Presumably the amount of degradation that occurs during exposure is proportional to the integral of the airborne dust concentration during the interval that the mirrors are exposed, i.e., during telescope operation. Thus the degradation during exposure is modulated by both dust concentration and weather. Without examining observing logs to get exposure time, it is not possible to separate the two effects. However, independent measurements provide evidence for time variable dust concentrations at the Very Large Telescope (VLT) site ("Survey of airborne particle density and the aging of mirror coatings in the open air at the VLT Observatory", P. Giordano, M.S. Sarazin, Proc. of S.P.I.E., 2199, 1994, p.977-985). In the next month or so we plan to implement a dust concentration monitoring program at Apache Point Observatory to help understand these effects. In an effort to understand the control mirror measurements, we are adding a second control mirror.

Acknowledgements

It is a pleasure to thank Eddie Bergeron, Karen Gloria, Dan Long, and Bruce Gillespie of Apache Point Observatory for their assistance.

Date created: 01/13/96
Last modified: 03/17/96
Copyright © 1996, Walter A. Siegmund
Walter A. Siegmund