Secondary eclipses of WASP-18b -- Near Infrared observations with the Anglo Australian Telescope, the Magellan Clay Telescope and the LCOGT network

We present new eclipse observations for one of the hottest "hot Jupiters" WASP-18b, for which previously published data from HST WFC3 and Spitzer have led to radically conflicting conclusions about the composition of this planet's atmosphere. We measure eclipse depths of $0.15\pm0.02\%$ at $Ks$ and $0.07\pm0.01\%$ at $z'$ bands. Using the VSTAR line-by-line radiative transfer code and both these new observations with previously published data, we derive a new model of the planetary atmosphere. We have varied both the metallicity and C/O ratio in our modelling, and find no need for the extreme metallicity suggested by Sheppard et al.(2017). Our best fitting models slightly underestimate the emission at $z'$ band and overestimate the observed flux at $Ks$-band. To explain these discrepancies, we examine the impact on the planetary emission spectrum of the presence of several types of hazes which could form on the night-side of the planet. Our $Ks$ band eclipse flux measurement is lower than expected from clear atmosphere models and this could be explained by a haze particles larger than 0.2 $\mu$m with the optical properties of Al$_{2}$O$_{3}$, CaTiO$_{3}$ or MgSiO$_{3}$. We find that $z'$ band measurements are important for understanding the contribution of photochemical hazes with particles smaller than 0.1 $\mu$m at the top of the atmosphere.