C 1s peak of adventitious carbon (AdC), often used for charge referencing XPS spectra, shows markedly large shifts from the “recommended” value of 284.8 eV that basically disqualifies its reliability. In some earlier papers we attributed this spreading effect to the vacuum level (VL) alignment at the AdC/sample interface, which makes the measured position of C 1s peak EBF highly sensitive to the sample work function ϕSA. Recently, it was suggested [M.C. Biesinger, Appl. Surf. Sci. 597 (2022) 153681] that it is instead the differential charging in the native oxide layers that sometimes accounts for C 1s shifts and that electrically isolating samples from the spectrometer would solve the problem. To evaluate this hypothesis, we performed a series of experiments with Au and Al foils electrically isolated from the spectrometer, while varying the surface potential in a relatively wide range by adjusting the charge neutralizer settings. Markedly, the C 1s peak positions recorded from Au and Al foils are distinctly different when referred to their Fermi levels, at respectively 284.80 ± 0.05 eV and 286.31 ± 0.06 eV, independent of the surface potential. This confirms the interpretation presented in our previous papers (experiments performed in a conventional way with samples connected to spectrometer), that the binding energy of C 1s peaks from Au and Al foils differs significantly due to the corresponding difference in their work function values, such that the sum EBF+ϕSA is constant at ∼ 289.6 eV, as imposed by the VL alignment. In addition, the energy separation between metal and oxide peaks in Al 2p spectra from Al foil is independent of the surface potential (controlled by the charge neutralizer settings), the photoelectron current (varied by adjusting x-ray power) and the Al oxide thickness (in the range from 0.7 to 4.7 nm). These observations disprove differential charging as the general cause of C 1s peak shifts at least for the case of Al foils with thinner oxide layers. As many thicker oxides are well-known to develop charging, a similar type of analysis can be performed on the case-to-case bases to determine the reasons for C 1s peak shifts.
Read full abstract