2014ApJ...796..100T

We analyze all existing secondary eclipse time series spectroscopy of hot Jupiter HD 189733b acquired with the now defunct Spitzer/Infrared Spectrograph (IRS) instrument. We describe the novel approaches we develop to remove the systematic effects and extract accurate secondary eclipse depths as a function of wavelength in order to construct the emission spectrum of the exoplanet. We compare our results with a previous study by Grillmair et al. that did not examine all data sets available to us. We are able to confirm the detection of a water feature near 6 µm claimed by Grillmair et al. We compare the planetary emission spectrum to three model families–based on isothermal atmosphere, gray atmosphere, and two realizations of the complex radiative transfer model by Burrows et al., adopted in Grillmair et al.'s study. While we are able to reject the simple isothermal and gray models based on the data at the 97% level just from the IRS data, these rejections hinge on eclipses measured within a relatively narrow wavelength range, between 5.5 and 7 µm. This underscores the need for observational studies with broad wavelength coverage and high spectral resolution, in order to obtain robust information on exoplanet atmospheres.