A Comprehensive Reanalysis of $\textit{Spitzer}$'s 4.5 $\mu$m Phase Curves, and the Phase Variations of the Ultra-hot Jupiters MASCARA-1b and KELT-16b

TL;DR

This study reanalyzed extit{Spitzer} phase curves of ultra-hot Jupiters MASCARA-1b and KELT-16b using a new open-source pipeline, revealing detailed atmospheric properties and model dependencies, and providing insights into their thermal and phase variation characteristics.

Contribution

The paper introduces a new open-source pipeline for analyzing extit{Spitzer} phase curves and applies it to reanalyze data of two ultra-hot Jupiters, offering a comprehensive comparison with previous results.

Findings

MASCARA-1b has a phase offset of approximately 6 degrees west.

KELT-16b has a phase offset of approximately 38 degrees west.

Strong correlation between dayside temperature and irradiation, with a weaker correlation for nightside temperature.

Abstract

We have developed an open-source pipeline for the analysis of \textit{Spitzer}/IRAC channel 1 and 2 time-series photometry, incorporating some of the most popular decorrelation methods. We applied this pipeline to new phase curve observations of ultra-hot Jupiters MASCARA-1b and KELT-16b, and we performed the first comprehensive reanalysis of 15 phase curves. We find that MASCARA-1b and KELT-16b have phase offsets of $6^{+11}_{-11}~^{\circ}$W and $38^{+16}_{-15}~^{\circ}$W, dayside temperatures of $2952^{+100}_{-97}$ K and $3070^{+160}_{-150}$ K, and nightside temperatures of $1300^{+340}_{-340}$ K and $1900^{+430}_{-440}$ K, respectively. We confirm a strong correlation between dayside and irradiation temperatures with a shallower dependency for nightside temperature. We also find evidence that the normalized phase curve amplitude (peak-to-trough divided by eclipse depth) is correlated with stellar effective temperature. In addition, while our different models often retrieve similar parameters, significant differences occasionally arise between them, as well as between our preferred model and the literature values. Nevertheless, our preferred models are consistent with published phase offsets to within $-8\pm21$ degrees ($-1.6\pm3.2$ sigma), and normalized phase curve amplitudes are on average reproduced to within $-0.01\pm0.24$ ($-0.1\pm1.6$ sigma). Finally, we find that BLISS performs best in most cases, but not all; we therefore recommend future analyses consider numerous detector models to ensure an optimal fit and to assess model dependencies.