Title: A Combined VLT and Gemini Study of the Atmosphere of the Directly-Imaged Planet, beta Pictoris b

Abstract: We analyze new/archival VLT/NaCo and Gemini/NICI high-contrast imaging of the young, self-luminous planet $\beta$ Pictoris b in seven near-to-mid IR photometric filters, using advanced image processing methods to achieve high signal-to-noise, high precision measurements. While $\beta$ Pic b's near-IR colors mimick that of a standard, cloudy early-to-mid L dwarf, it is overluminous in the mid-infrared compared to the field L/T dwarf sequence. Few substellar/planet-mass objects -- i.e. $\kappa$ And b and 1RXJ 1609B -- match $\beta$ Pic b's $JHK_{s}L^\prime$ photometry, and its 3.1 $\mu m$ and 5 $\mu m$ photometry are particularly difficult to reproduce. Atmosphere models adopting cloud prescriptions and large ($\sim$ 60 $\mu m$) dust grains fail to reproduce the $\beta$ Pic b spectrum. However, models incorporating thick clouds similar to those found for HR 8799 bcde but also with small (a few microns) modal particle sizes yield fits consistent with the data within uncertainties. Assuming solar abundance models, thick clouds, and small dust particles ($<a>$ = 4 $\mu m$) we derive atmosphere parameters of log(g) = 3.8 $\pm$ 0.2 and $T_{eff}$ = 1575--1650 $K$, an inferred mass of 7$^{+4}_{-3}$ $M_{J}$, and a luminosity of log(L/L$_{\odot}$) $\sim$ -3.80 $\pm$ 0.02. The best-estimated planet radius, $\approx$ 1.65 $\pm$ 0.06 $R_{J}$, is near the upper end of allowable planet radii for hot-start models given the host star's age and likely reflects challenges with constructing accurate atmospheric models. Alternatively, these radii are comfortably consistent with hot-start model predictions if $\beta$ Pic b is younger than $\approx$ 7 Myr, consistent with a late formation, well after its host star's birth $\sim$ 12$^{+8}_{-4}$ Myr ago.