Invited Paper AS-MoM5
ASSD 30th Anniversary Lecture: Sensitivity Factors in XPS: Where Do They Come From and How Accurate Are They?

Monday, October 19, 2015, 9:40 am, Room 212D

The most common method for quantitative analysis in X-ray Photoelectron Spectroscopy (XPS) incorporates relative sensitivity factors. In the 1970s and 1980s, several research groups studied the relative intensities of photoelectron lines and attempted to generate tables of relative sensitivity factors. Calculations were also made of ionization cross-sections using Mg and Al Kα x-rays, and experimental measurements were often compared to theory. The seminal measurements were made by Wagner and co-workers, who used two different types of electron energy analyzers from two different manufacturers and compared the results with theory and with other measurements. They found that the agreement in data from the two instruments on the same compounds was good, and they generated a table of empirically derived sensitivity factors. Today, many XPS analysts still use the relative sensitivity factors from Wagner’s work, particularly if they use an instrument where the sensitivity factors were not provided. An early ASTM round-robin on XPS measurements on catalysts by 12 laboratories found a large spread in measured peak intensity ratios, even for instruments of the same manufacture, concluding that a calibration procedure for the intensity response of instruments was needed. A later ASTM round-robin on pure Au and Cu concluded that the spread in intensity ratios was typically a factor of ten. Based on these results, Seah spearheaded studies on instrument transmission functions and developed a method so that the relative instrument response function between two different analyzers was better than 5%. In 1990, he published standard reference spectra for XPS and claimed all instruments may be calibrated absolutely to an accuracy of ± 2%. The National Physical Laboratory (NPL) developed and sold a procedure for generating an instrument transmission function, but it was not widely adopted. At least one manufacturer developed a model for measuring the transmission function in their instruments and measures it on each instrument before shipping it to the purchaser. A digital database of spectra was also developed at NPL, and could have provided a set of relative sensitivity factors for XPS, but it was never sold. It is important to know the origin of the sensitivity factors supplied by manufacturers and any limits placed on their applicability. By measuring spectra from a number of reference compounds one can quickly learn how reliable they are. For the most accurate measurements, one should measure relative sensitivity factors from standard compounds that match as closely as possible the compositions of unknowns.