Multi-epoch ultraviolet observables for breaking the radius-albedo degeneracy in directly imaged exoplanets

Direct imaging measures how bright a planet appears in reflected starlight, but brightness alone cannot tell whether the planet is large and dark or small and bright. This radius-albedo degeneracy limits the characterization of non-transiting exoplanets, including those targeted by the Habitable Worlds Observatory (HWO). From multi-epoch observations we construct ultraviolet observables that are independent of planetary radius: the normalized 400 nm lightcurve shape, the ultraviolet intensity and polarization colors, and the degree of linear polarization. In a clear 360-400 nm spectral window where Rayleigh scattering is strong and bright ultraviolet surfaces are nearly colorless, these observables constrain the atmospheric column, the surface reflectivity, and, when relevant, the observed phases before radius is inferred. Using the GPU accelerated vector radiative transfer model vSmartMOM, we test how uniquely these radius-free observables determine the scattering state at signal-to-noise ratios (SNRs) of 5, 20, and 100. For the chosen six phase sequence, spectropolarimetry gives median radius consequences of 9.5%, 3.3%, and 0.45% when the phases are known. If the phases are unknown, six distinct epochs keep the penalty modest at low SNR and negligible at moderate-high SNR, with median consequences of 14%, 3.2%, and 0.47%, respectively. Removing polarimetry degrades the result, especially in the HWO simulations where phase coverage is restricted by inner working angle. For solar twin systems at 6 and 12 parsecs, the required six-phase campaigns fall in the allotted range of a few hundred hours, making this a clear observing path to planetary radii from HWO's reflected-light detections.