On the suitability of the virial equation for modeling the solubility of solids in supercritical fluids