3.6 AND 4.5 μm PHASE CURVES AND EVIDENCE FOR NON-EQUILIBRIUM CHEMISTRY IN THE ATMOSPHERE OF EXTRASOLAR PLANET HD 189733b

We present new, full-orbit observations of the infrared phase variations of the canonical hot Jupiter HD 189733b obtained in the 3.6 and 4.5 μm bands using the Spitzer Space Telescope. When combined with previous phase curve observations at 8.0 and 24 μm, these data allow us to characterize the exoplanet's emission spectrum as a function of planetary longitude and to search for local variations in its vertical thermal profile and atmospheric composition. We utilize an improved method for removing the effects of intrapixel sensitivity variations and robustly extracting phase curve signals from these data, and we calculate our best-fit parameters and uncertainties using a wavelet-based Markov Chain Monte Carlo analysis that accounts for the presence of time-correlated noise in our data. We measure a phase curve amplitude of 0.1242% ± 0.0061% in the 3.6 μm band and 0.0982% ± 0.0089% in the 4.5 μm band, corresponding to brightness temperature contrasts of 503 ± 21 K and 264 ± 24 K, respectively. We find that the times of minimum and maximum flux occur several hours earlier than predicted for an atmosphere in radiative equilibrium, consistent with the eastward advection of gas by an equatorial super-rotating jet. The locations of the flux minima in our new data differ from our previous observations at 8 μm, and we present new evidence indicating that the flux minimum observed in the 8 μm is likely caused by an overshooting effect in the 8 μm array. We obtain improved estimates for HD 189733b's dayside planet-star flux ratio of 0.1466% ± 0.0040% in the 3.6 μm band and 0.1787% ± 0.0038% in the 4.5 μm band, corresponding to brightness temperatures of 1328 ± 11 K and 1192 ± 9 K, respectively; these are the most accurate secondary eclipse depths obtained to date for an extrasolar planet. We compare our new dayside and nightside spectra for HD 189733b to the predictions of one-dimensional radiative transfer models from Burrows et al. and conclude that fits to this planet's dayside spectrum provide a reasonably accurate estimate of the amount of energy transported to the night side. Our 3.6 and 4.5 μm phase curves are generally in good agreement with the predictions of general circulation models for this planet from Showman et al., although we require either excess drag or slower rotation rates in order to match the locations of the measured maxima and minima in the 4.5, 8.0, and 24 μm bands. We find that HD 189733b's 4.5 μm nightside flux is 3.3σ smaller than predicted by these models, which assume that the chemistry is in local thermal equilibrium. We conclude that this discrepancy is best explained by vertical mixing, which should lead to an excess of CO and correspondingly enhanced 4.5 μm absorption in this region. This result is consistent with our constraints on the planet's transmission spectrum, which also suggest excess absorption in the 4.5 μm band at the day-night terminator.