Flat-Field Color Dependence

IRAC will be using a "super flat" to correct for pixel-to-pixel gain variations. This flat is computed from dedicated "dark field" images near the ecliptic plane (high zody) and the ecliptic poles (low zody). Differencing high vs. low gives the gain variation while canceling any time-dependent zody variations. The ecliptic plane that crosses the Galactic plane is avoided. Using a large stack of differences, gathered over the lifetime of Spitzer, the median "flat" represents the "super flat" with supremely good accuracy, comparable to a tenth of a percent. But beware, this supreme flat comes from the zody, which is very cold (few hundred kelvin) and much redder than astrophysical sources. The flat that works for very red, cold objects is not the same as the flat that would work for blue (stars/galaxies) sources. This is because the effective bandpass changes with wavelength; see below.

For point sources, the dominant photometric uncertainty arises from the flat-field pixel-pixel difference between the zodiacal emission (from which the flat is derived) and the color of stars & galaxies.

"IRAC has significant scattering, as well as spatial distortion. As a result, the extended and point source effective gains are slightly different. In addition, there is a variation in the effective filter bandpass as a function of angle of incidence, which in turn depends on the exact position of an object on the array (Quijada et al. 2004, SPIE 5487, 244). As a result of this, while the flat-field perfectly corrects the extended zodical background (or any extended object) with a similar spectral slope, it is incorrect for many other objects.

This effect has been directly measured. Objects were sampled at many different locations on the array, and their flux measured from the BCD images. The systematic variations in their measured fluxes were used to derive the corrections. The amplitude of this effect is sizeable. It may reach 10% peak-to-peak, depending on the detector array. This is larger than any other source of uncertainty in the IRAC calibration."

For an example of how to use the location-dependent flat-field corrections, see Spitzer NEP Data Reduction; step 5.

What can we expect for WISE? Given the very fast optics, f/3.375, there should be a strong wavelength dependence on the angle of incidence; thus, a strong color dependence is expected. Ground tests and/or modeling (e.g., Code-V) should look for this dependence and give IPAC some idea of what to expect. In flight, the way to derive the correction per band is to track the same stars at the poles and measure their differences as they 'rotate' about the focal plane during the survey.


Last update - [19 Dec 2007]
[Tom Jarrett - IPAC]