Photometric Response Functions of the SDSS Imager

TL;DR

The paper presents in situ measurements of the SDSS 2.5m imager response functions across the $u,g,r,i,z$ bands using a permanently installed monochromatic illumination system. It constructs a reference response curve and analyzes how seasonal, secular, and chip-to-chip variations affect photometry, finding a pronounced aging effect in the $u$ band (about a 30% throughput decrease from 2000 to 2004) but that calibration largely mitigates its impact. Across other bands, time variations are small and largely temperature-driven; differences among detectors and auxiliary telescopes are generally small and are effectively canceled by the SDSS photometric calibration, yielding final catalog residuals of order $0.01$ mag. The study provides critical insights for interpreting SDSS photometry, quantifies color-term variations (delta$(u-g)$ of about 0.01–0.02 mag for extreme colours), and offers guidance for the design of future photometric instrumentation and calibration schemes.

Abstract

The monochromatic illumination system is constructed to carry out in situ measurements of the response function of the mosaicked CCD imager used in the Sloan Digital Sky Survey (SDSS). The system is outlined and the results of the measurements, mostly during the first 6 years of the SDSS, are described. We present the reference response functions for the five colour passbands derived from these measurements, and discuss column to column variations and variations in time, and also their effects on photometry. We also discuss the effect arising from various, slightly different response functions of the associated detector systems that were used to give SDSS photometry. We show that the calibration procedures of SDSS remove these variations reasonably well with the resulting final errors from variant response functions being unlikely to be larger than 0.01 mag for g, r, i, and z bands over the entire duration of the survey. The considerable aging effect is uncovered in the u band, the response function showing a 30% decrease in the throughput in the short wavelength side during the survey years, which potentially causes a systematic error in photometry. The aging effect is consistent with variation of the instrumental sensitivity in u-band, which is calibrated out. The expected colour variation is consistent with measured colour variation in the catalog of repeated photometry. The colour variation is delta (u-g) ~ 0.01 for most stars, and at most delta (u-g) ~ 0.02 mag for those with extreme colours. We verified in the final catalogue that no systematic variations in excess of 0.01 mag are detected in the photometry which can be ascribed to aging and/or seasonal effects except for the secular u-g colour variation for stars with extreme colours.

Figures (43)

• Figure 2: Monochromatic illumination system overview. (1) lamp; (2) light condensation mirror system; (3) input slit; (4) monochrometer; (5) shutter; (6) mirror switch to change the output direction; (7) integration sphere; (8) photodiode.
• Figure 3: Response function (quantum efficiency) of the ultraviolet-enhancing coated thinned back-illuminated CCD used at the Photometric Telescope. The solid points are the measurement with the present system, and the dashed curve is the quantum efficiency for the same CCD measured at SITe at a room temperature, but tilted using a laboratory measurement (open points) at a cooled, operating temperature, used in F96.
• Figure 4: Response functions for the $u$ band. The thick solid curve is the reference response function for the 2.5 m telescope imager defined in the text, thin solid curve is the F96 standard, the dashed curve is for the USNO 1m telescope system, and the dotted curve is for the SDSS Photometric Telescope (after the replacement of filters: see text)
• Figure 5: Same as Figure 2, but for the $g$ band.
• Figure 6: Same as Figure 2, but for the $r$ band.