The properties of charged, charged at the pzc and uncharged interfaces are studied by means of the semi-thermodynamic approach developed [Electrochim. Acta, 39 (1994) 865; Russ. J. Electrochem., 31 (1995) 747]. It is shown that an uncharged monolayer located between two conducting phases is mathematically equivalent to the corresponding charged one at the pzc. The free surface of solutions is equivalent to the previous ones, but only in the stable region. When a 2D phase transition occurs at an adsorbed layer, both the adsorption isotherms and the critical values of the interaction parameter an markedly different at the free surface from those at the uncharged metalelectrolyte solution interface. The origin of this difference is indicated and discussed. It is shown further that the conventional use of the canonical ensemble disregards the existence of two domains of different electrode charge density, both at the uncharged and the corresponding charged metalelectrolyte interface when the intermolecular interactions are treated approximately and the adsorbed layer undergoes a 2D phase transition. This is the real cause of the polarization catastrophe at charged interfaces and the reason why the canonical ensemble predicts totally artificial adsorption isotherms throughout the region of a 2D phase transition and higher and incorrect values for the critical interaction parameter at the uncharged metalelectrolyte interface. The existence of domains of different charge density is taken into account if we consider surface equilibria, and this eliminates these artifacts completely.