To explore the origin of the Fermi level pinning in germanium, we investigate the Ge(001) and Ge(001):H surfaces. The absence of relevant surface states in the case of Ge(001):H should unpin the surfaceFermi level. This is not observed. For samples with donors as majority dopants, the surfaceFermi level appears close to the top of the valence band regardless of the surface structure. Surprisingly, for the passivated surface, it is located below the top of the valence band allowing scanning tunneling microscopy imaging within the band gap. We argue that the well known electronic mechanism behind band bending does not apply and a more complicated scenario involving ionic degrees of freedom is therefore necessary. Experimental techniques involve four point probe electric current measurements, scanning tunneling microscopy, and spectroscopy.