Publ. Astron. Soc. Pac., 131, part no 12, 4502-124502 (2019/December-0)
Photometric precision of a Si:As impurity band conduction mid-infrared detector and application to Transit spectroscopy.
MATSUO T., GREENE T.P., JOHNSON R.R., McMURRAY R.E., ROELLIG T.L. and ENNICO K.
Abstract (from CDS):
Transit spectroscopy is the most promising path toward characterizing nearby terrestrial planets at mid-infrared (3-30 µm) wavelengths in the next 20 yr. The Spitzer Space telescope has achieved moderately good mid-infrared photometric precision in observations of transiting planets, but the intrinsic photometric stability of mid-IR detectors themselves has not been reported in the scientific or technical literature. Here, we evaluated the photometric precision of a James Webb Space Telescope Mid-Infrared Instrument prototype mid-infrared Si:As impurity band conduction detector, using time-series data taken under flood illumination. These measurements of photometric precision were conducted over periods of ∼10 hr, representative of the time required to observe an exoplanet transit. After selecting multiple sub-regions with a size of 10 x 10 pixels and compensating for a gain change caused by our warm detector control electronics for the selected sub-regions, we found that the photometric precision was limited to 26.3 ppm at high co-added signal levels due to a gain variation caused by our warm detector control electronics. The photometric precision was improved up to 12.8 ppm after correcting for the gain drift. We also translated the photometric precision to the expected spectro-photometric precision (i.e., relative photometric precision between wavelengths), assuming that an optimized densified pupil spectrograph is used in transit observations. We found that the spectro-photometric precision of an optimized densified pupil spectrograph when used in transit observations is expected to be improved by the square root of the number of pixels per a spectral resolution element. At the high co-added signal levels, the total noise could be reduced down to 7 ppm, which was larger by a factor of 1.3 than the ideal performance that was limited by the Poisson noise and readout noise. The systematic noise hidden behind the simulated transit spectroscopy was 1.7 ppm.