by M. Biesinger (UWO)
Insulating samples pose a unique challenge for XPS analysis. As photoelectrons are lost during the photoemission process a positive charge will build up on the sample. As this occurs, the kinetic energy of the emitted photoelectrons will decrease resulting in a shift to higher binding energy of the observed peaks in the spectrum. A number of schemes have been developed to compensate for this. Most commonly a low energy electron flood gun is used to replace emitted electrons. A slight overcompensation is normally used, setting up an equilibrium state with peaks in the spectrum shifted a few eV to lower binding energy. In post processing the peaks are then shifted back by reference to a set internal standard such as adventitious carbon or a known species in that particular sample.
In some cases only a portion of the sample is insulating with some discrete areas or layers of the sample being conducting or semi-conducting. In these cases a phenomenon known as differential charging can occur. It is possible to imagine a sample with insulating domains or island structures that have varying thicknesses on a conducting surface. The spectral features from these areas will be shifted to lower binding energies by the action of the charge neutralizer whereas the spectral features of the underlying material will not be shifted. This may be complicated if the islands or domain structures behave as semi-conductors and their conductivity varies with thickness. The resulting spectra may be broadened significantly as a result. Layered structures may also charge differentially. The thin oxide layer on a metallic (conducting) material can behave as a conductor, as a semi-conductor, or as an insulator as the film grows in thickness. A multilayered system may include alternating layers of conducting, semi-conducting and insulating species in a variety of combinations. All of this can result in changes in selected peak positions that must be understood or erroneous assignment of chemical states may result.
One excellent way to combat differential charging effects is to electrically isolate (or “float”) the entire sample from the specimen holder. This “specimen isolation” technique is similar to that used in secondary ion mass spectrometry  and effectively makes the entire sample (both insulating, semi-conductive and conductive areas) behave non-conductively. Methods employed with good success here at Surface Science Western include mounting samples on non-conductive double sided tape or mounting on glass slides.
The Kratos AXIS Ultra and Nova employ both a low electron flood gun as well as a set of sub-sample magnetic fields that return over-focused and under-focused photoelectrons (that are not being admitted through to the spectrometer) back to the surface of the sample. This system has had excellent success in analyzing a wide range of insulating samples.
 M.A. Kelly, Analysing Insulators with XPS and AES, in: D. Briggs, J.T. Grant (Eds.) Surface Analysis by Auger and X-ray Photoelectron Spectroscopy, IM Publications, Chichester, 2003, pp. 191-210.
 J.B. Metson, G.M. Bancroft, N.S. McIntyre, W.J. Chauvin, Surf. Interface Anal. 5 (1983) 181-185.