Ultra‐thin SiO 2 on Si VIII. Accuracy of method, linearity and attenuation lengths for XPS

In the study of ultra‐thin films (<10‐nm thick), there is a range of methods that can provide accurate measurements of differences in thickness. However, in a pilot study under the auspices of the Consultative Committee for Amount of Substance (CCQM), results for the archetypal system of SiO 2 on Si show that the methods have different offsets such that, at all thicknesses, positive or negative amounts in the range up to 1 nm may be observed between methods. All the methods studied give thicknesses that are greater than those measured by X‐ray photoelectron spectroscopy (XPS) by amounts between 0.2 nm and approximately 1 nm. Significant parts of these offsets, of other methods with respect to XPS, may be attributed to contaminations which increase the apparent thickness but that do not affect XPS. However, not all of these offsets can yet be explained at the 0.2 nm level. The remaining part of the offsets could be thought to have arisen either from the XPS or from the other methods. In this study, by measuring SiO 2 deposited in situ on amorphous Si by XPS it is shown that the XPS linearity is consistent with the previous estimate of ± 0.025 nm, down to fractions of a monolayer, with no significant offset and that, therefore, it is the offsets seen using other methods that need further study. Recent calculations of the film thickness dependence of the attenuation lengths (ALs) for this system, using NIST SESSA software, are not consistent with these data although earlier calculations are. This work shows that XPS, with the AL calibrated by one or more other methods that are valid for differences in thickness, can provide a traceable measurement of thickness in all laboratories.