Kepler-1624b Has No Significant Transit Timing Variations

TL;DR

Kepler-1624b's TTVs were previously interpreted as evidence for a nearby companion, but subsequent analyses yielded conflicting results. The authors reanalyze archival Kepler data, extend the baseline with TESS observations, and incorporate three new ground-based transits, performing a joint fit and a Bayes-factor comparison between one-planet and two-planet models. They find the TTV amplitude to be significantly weaker and preferentially support a one-planet model. The study underscores the fragility of low-amplitude TTV detections in large catalogs and informs migration constraints for gas giants around M-dwarfs.

Abstract

It is relatively rare for gas giant planets to have resonant or near-resonant companions, but these systems are particularly useful for constraining planet formation and migration models. In this study, we examine Kepler-1624b, a sub-Saturn orbiting an M dwarf that was previously found to exhibit transit timing variations with an amplitude of approximately 2 minutes, suggesting the presence of a nearby non-transiting companion. We reanalyze the transits from archival Kepler data and extend the TTV baseline by 11 years by combining TESS data with three new ground-based transit observations from Palomar and Las Cumbres Observatories. We jointly fit these datasets and find that the TTV amplitude is significantly weaker in our updated analysis. We calculate the Bayes factor for a one-planet versus two-planet model and find that the one-planet model is preferred. Our results highlight the need for careful analysis of systems with relatively low amplitude TTV signals that are identified in large automated catalogs.

Figures (3)

• Figure 1: Orbital period versus planet radius for all planet pairs with period ratios within 2% of a first-order resonance. The colors of the points indicate the effective temperature of the host star, and each planet pair is connected by a solid black line. The location of Kepler-1624b in this space is marked by a star. This plot was generated using data from the NASA Exoplanet Archive Akeson2013.
• Figure 2: Kepler light curve before detrending (left) and after detrending (right). The red line shows the Gaussian kernel convolution trend described in §\ref{['Kepler']}. Blue points indicate masked data ($\pm 0.8$ transit durations around each transit midtime).
• Figure 3: Transit light curves for Kepler-1624b from multiple instruments. (a) Kepler, (b) Palomar/WIRC, (c) TESS, (d) LCOGT/SINISTRO $i$ band, and (e) LCOGT/MuSCAT3 multiband. For Kepler, the light curves of individual transits are phased up using the individual best-fit transit times. For TESS, we phase individual transits using transit times from the best-fit one-planet model. For (a), the gray points show unbinned phased data, where each point corresponds to a 30-minute integration. For (b), unbinned photometry is shown as gray circles, and 10-minute binned photometry is shown as black circles. For (c), we use 10-minute bins for the black points and 2-minute bins for the gray points. For (d), we show the unbinned data as gray points. For (a)--(d), the best-fit transit light curves are overplotted in red, with the 1$\sigma$ confidence range overplotted as light red shading. For (e), transits in $g$, $r$, $i$, and $z_s$ bands are plotted with different colors with an arbitrary vertical offset, and the gray points represent the observed data, while the colored lines show the best-fit transit model for each band with a 1$\sigma$ confidence range overplotted with the corresponding color of shading. For some of the panels, the 1$\sigma$ confidence shading is not visually noticeable due to very tight confidence ranges.