Uniform Reanalysis of JWST MIRI 15μm Exoplanet Eclipse Observations using Frame-Normalized Principal Component Analysis

Abstract: JWST MIRI 15 micron time-series eclipse photometry presents a powerful way to probe for the presence of atmospheres on low-temperature rocky exoplanets orbiting nearby stars. Here, we introduce a novel technique, frame-normalized principal component analysis (FN-PCA) to analyze and detrend these MIRI time-series observations. Using the FN-PCA technique, we perform a uniform reanalysis of the published MIRI 15 micron observations of LHS 1478 b, TOI-1468 b, LHS 1140 c, TRAPPIST-1 b, and TRAPPIST-1 c using our new data reduction pipeline (Erebus) and compare them to different potential atmospheric and surface compositions. We also investigate additional public data sets with the sole purpose of understanding the instrument systematics affecting MIRI. We identify and categorize important detector-level systematics in the observations that are generally present across all 17 analyzed eclipse observations, which we illustrate as eigenimage/eigenvalue pairs in the FN-PCA. One of these eigenimage/eigenvalue pairs corresponds to the prominent ramp effect at the beginning of the time-series observations which has widely been reported for JWST and Spitzer photometry. For JWST/MIRI, we show that the detector settling time scales exponentially with the apparent magnitude of the target star $T_\mathrm{set} \mathrm{[hours]} = 0.063\exp^{0.427\cdot m_K} -0.657$. This uniform reanalysis and investigation of JWST/MIRI systematics is done in preparation for the 500 hour Rocky Worlds DDT survey, to demonstrate a data-driven systematic model usable across all MIRI 15 micron datasets.