Evidence for disequilibrium chemistry from vertical mixing in hot Jupiter atmospheres

Abstract

Aims. We present a large atmospheric study of 49 gas giant exoplanets using infrared transmission photometry with Spitzer/IRAC at 3.6 and 4.5 μm. Methods. We uniformly analyze 70 photometric light curves of 33 transiting planets using our custom pipeline, which implements pixel level decorrelation. Augmenting our sample with 16 previously published exoplanets leads to a total of 49. We use this survey to understand how infrared photometry traces changes in atmospheric chemical properties as a function of planetary temperature. We compare our measurements to a grid of 1D radiative-convective equilibrium forward atmospheric models which include disequilibrium chemistry. We explore various strengths of vertical mixing (Kzz = 0 1012 cm2/s) as well as two chemical compositions (1x and 30x solar). Results. We find that, on average, Spitzer probes a difference of 0.5 atmospheric scale heights between 3.6 and 4.5 μm, which is measured at 7.5 σ level of significance. Changes in the opacities in the two Spitzer bandpasses are expected with increasing temperature due to the transition from methane-dominated to carbon-monoxide-dominated atmospheres at chemical equilibrium. Comparing the data with our model grids, we find that the coolest planets show a lack of methane compared to expectations, which has also been reported by previous studies of individual objects. We show that the sample of coolest planets rule out 1x solar composition with >3 σ confidence while supporting low vertical mixing (Kzz = 108 cm2/s). On the other hand, we find that the hot planets are best explained by models with 1x solar metallicity and high vertical mixing (Kzz = 1012 cm2/s). We interpret this as the lofting of CH4 to the upper atmospheric layers. Changing the interior temperature changes the expectation for equilibrium chemistry in deep layers, hence the expectation of disequilibrium chemistry higher up. We also find a significant scatter in the transmission signatures of the mid-temperate and ultra-hot planets, likely due to increased atmospheric diversity, without the need to invoke higher metallicities. Additionally, we compare Spitzer transmission with emission in the same bandpasses for the same planets and find no evidence for any correlation. Although more advanced modelling would test our conclusions further, our simple generic model grid points towards different amounts of vertical mixing occurring across the temperature range of hot Jupiters. This finding also agrees with the observed scatter with increasing planetary magnitude seen in Spitzer/IRAC color-magnitude diagrams for planets and brown dwarfs.