Binding energies (BE) from non-conductive samples and native oxides are commonly charge referenced to the C 1s BE of the hydrocarbon peak of adventitious carbon that appears on the surface of all materials exposed to air. Various studies of the usefulness of this method have been carried out (r.f. P. Swift, J. Surf. Interfac. Analysis, p.47, 1982) and none have revealed any significant problem with this everyday practice.

In this talk, C 1s, O 1s and metal spectra and BEs from native and pure binary metal oxides (e.g. MgO or Y₂O₃), and a series of residual gas capture studies, will show that there is indeed an obvious and significant problem when this practice is applied to binary metals oxides known to have highly polarized bonds or states (i.e. Al, Cd, Ga, Hf, Mg, Sc and Y). The hydrocarbon C 1s BEs for these native metal oxides routinely appear above 286.0 eV, which is significantly greater than the 284.6-284.8 eV that is commonly used. Another set of native metal oxides (i.e. Ag, Be, Co, Ge, Pb, Si, Zn and Zr) have C 1s BEs: 285.5 eV < C 1s < 286.0 eV.

This higher than expected BE problem is sometimes found for carbon captured by freshly cleaned pure metals exposed (>10 hr) to residual UHV gases and the surface contaminants that adhere to nearby, unclean regions of the sample. The cause of these high BEs might be related to strong surface dipole moments (image potentials) that protrude outward from the highly polarized metal-oxide bonds or states of the metal oxide layer that lie just under the adventitious carbon over-layer. Strong surface dipole moments are predicted (by A. Zangwill, Physics at Surfaces, Cambridge Press, 1988) to extend several tens to hundreds of angstroms above the surface of metal oxides into the carbon over-layer and the vacuum, whereas dipoles from pure clean metals are predicted to extend only a few angstroms above the surface.

To demonstrate this problem, spectra from grounded native oxides are compared to native oxides that were floated and irradiated with a flood gun set to OFF, 2 eV and 15 eV (using Mono-XPS). An overlay of five Al 2p spectra (from a 1978 synchrotron study) shows the progressive growth of Al₂O₃ where the Al 2p BE for Al₂O₃ = 75.7 eV, not 74.3 eV as determined by averaging 42 BEs from pure Al₂O₃ BE values that are stored in the on-line NIST XPS database of BEs (SRD20). The mean NIST value is ~1.4 eV lower than the 75.7 eV found for native Al₂O₃ on kitchen foil that gives a 286.3 eV C 1s BE, and also the 75.5 eV found in the 1978 synchrotron study. The experimentally measured 286.3 eV value is 1.5 eV larger than the 284.8 eV value recommended by the NIST database, handbooks and instrument makers.