In protein adsorption, the surrounding solvent has an important role in mediating protein-surface interactions. Therefore, it is of paramount importance that the solvent methods employed to model these kinds of processes are able to correctly capture the complex mechanisms occurring in the protein-water-surface interface. Here, we test the suitability of the two most popular implicit solvent methods based on the Generalized Born formalism to describe the adsorption process of the immunoglobulin G (IgG) on a hydrophobic graphene surface. Our results show that in both cases, IgG experiences an extreme and early (in less than 40 ns) unfolding as a result of the adsorption to the surface in contrast with previous experimental findings. A detailed energy decomposition analysis of explicit and implicit solvent simulations reveals that this discrepancy arises from the ill-characterization of two energy components in implicit solvent methods. These findings help to elucidate how implicit solvent models may be improved to accurately characterize the protein adsorption process.