On the surface analysis of copper oxides: the difficulty in detecting Cu3O2