IRTF Near-Sun Scattered Light

Go straight to the solution(?)

Observers at the NASA Infrared Telescope Facility (IRTF), on Mauna Kea, Hawaii, sometimes need to take daytime images at small solar elongations (time-critical monitoring of Jupiter or Io, or comets, for instance). When looking sufficiently close to the sun, (less than about 35 degrees), NSFCAM image quality and photometric accuracy are severely compromised by the appearance of donut-like diffraction patterns on the chip, which do not subtract out in object/sky pairs. This page documents the appearance of these "donuts" and describes our attempts to eliminate them.

Test images, 96/11/06

Images of daytime sky in the L filter, 0.15 second integrations, 0.15 arcsec/pixel.

Daytime Sky 40 degrees from the Sun
The sun is on most of the secondary spider, and most of the secondary mirror housing, but not on the lowermost part of the housing or the secondary mirror itself.

Daytime Sky 30 degrees from the Sun
Note the "donuts" that aren't seen further from the sun. The sun is now on the whole secondary mirror housing, including the secondary mirror itself.

Object-Sky frame, 30 degrees from the Sun
Beamswitch 60 N. Note that the "donuts" don't subtract out very well, making accurate photometry difficult. The donuts are, however, fairly reproducible at a given pointing position and don't appear in the difference between two consecutive object frames taken without beamswitching between them. Varying the beamswitch direction seems to help with subtraction (see "The Solution?")
The fact that the donuts appear when the secondary mirror is illuminated sounds like an important clue, but wrapping the lower secondary mirror housing in black paper, including the circumference of the mirror itself, does not eliminate the donuts.

We also tried getting object/sky pairs by moving the chopping secondary instead of the whole telescope, but this did not help significantly.

Attempt to remove the "donuts"
On 96/11/08 we used black paper and black cloth to cover up the silvered surfaces of the secondary spider and the secondary mirror housing, in case reflection off these surfaces was the cause of the "donuts". We weren't able to get test images that day, but on 96/11/29 John Stansberry repeated the experiment and found that it gave at most a modest improvement in scattered light. The recent silvering of the IRTF secondary support structure thus does not seem to be the source of the problem.

The Solution?

On 96/11/29, John Stansberry discovered that subtraction of the donuts in object/sky pairs appeared to be improved by changing the orientation of the beamswitch. Taking the sky 60 arcsec W of the object frame, instead of 60 arcsec N, for instance, improved the donut subtraction significantly. Thus, experimenting with the beamswitch direction seems like the most promising approach at present.

Return to John Spencer's Home Page