The geometric albedo of the hot Jupiter HD 189733b measured with CHEOPS

Context. Measurements of the occultation of an exoplanet at visible wavelengths allow us to determine the reflective properties of a planetary atmosphere. The observed occultation depth can be translated into a geometric albedo. This in turn aids in characterising the structure and composition of an atmosphere by providing additional information on the wavelength-dependent reflective qualities of the aerosols in the atmosphere. Aims. Our aim is to provide a precise measurement of the geometric albedo of the gas giant HD 189733b by measuring the occultation depth in the broad optical bandpass of CHEOPS (350-1100 nm). Methods. We analysed 13 observations of the occultation of HD 189733b performed by CHEOPS utilising the Python package PyCHEOPS. The resulting occultation depth is then used to infer the geometric albedo accounting for the contribution of thermal emission from the planet. We also aid the analysis by refining the transit parameters combining observations made by the TESS and CHEOPS space telescopes. Results. We report the detection of an 24.7 \ub1 4.5 ppm occultation in the CHEOPS observations. This occultation depth corresponds to a geometric albedo of 0.076 \ub1 0.016. Our measurement is consistent with models assuming the atmosphere of the planet to be cloud-free at the scattering level and absorption in the CHEOPS band to be dominated by the resonant Na doublet. Taking into account previous optical-light occultation observations obtained with the Hubble Space Telescope, both measurements combined are consistent with a super-stellar Na elemental abundance in the dayside atmosphere of HD 189733b. We further constrain the planetary Bond albedo to between 0.013 and 0.42 at 3σconfidence. Conclusions. We find that the reflective properties of the HD 189733b dayside atmosphere are consistent with a cloud-free atmosphere having a super-stellar metal content. When compared to an analogous CHEOPS measurement for HD 209458b, our data hint at a slightly lower geometric albedo for HD 189733b (0.076 \ub1 0.016) than for HD 209458b (0.096 \ub1 0.016), or a higher atmospheric Na content in the same modelling framework. While our constraint on the Bond albedo is consistent with previously published values, we note that the higher-end values of ∼0.4, as derived previously from infrared phase curves, would also require peculiarly high reflectance in the infrared, which again would make it more difficult to disentangle reflected and emitted light in the total observed flux, and therefore to correctly account for reflected light in the interpretation of those phase curves. Lower reported values for the Bond albedos are less affected by this ambiguity