# Extracting elemental ratios from X-ray photoelectron spectroscopy (XPS)

Is it possible to extract information on elemental ratios from X-ray photoelectron spectroscopy (XPS)? For example, I am reading a paper and in it they say that the carbon to oxygen (C:O) ratios of the four samples are 2.8, 2.9, 5.1, and 23.3. However, the authors do not provide the actual XPS spectra, not even in a Supporting Information-type addendum.

I am a computational chemist and have no experience with XPS. I do know that elements, and even particular hybridizations, give characteristic peaks at characteristic binding energies (in units of eV, for example). This, I know, enables qualitative analysis of compounds. But is there a way to extract quantitative information from XPS? Suppose I want to determine the C:O ratio in a compound containing only C, O, and H. Do I take the ratio of the C and O peak heights, or do I take the ratio of the area under the peaks corresponding to C and O?

Do you know of any good review articles that would help me to understand the basics of XPS?

• ""and even particular hybridizations, give characteristic peaks at characteristic binding energies (in units of eV, for example)"" You mix up photoelectron spectroscopy and X-Ray-PES – Georg Jul 17 '12 at 10:03
• @Georg Thanks. What is the difference between photoelectron spectroscopy and X-ray-PES? – Andrew Jul 17 '12 at 14:37
• The latter uses Röntgen-rays, the other vacuum-UV to set free those electrons. That means that only the highest level electrons are expelled (carrying information on chemical bonds) or electrons "deep" in the atoms are expelled by Röntgen-rays, giving information on elemetary composition. – Georg Jul 17 '12 at 14:41
• @Georg XPS can also give chemical information (through less sensitive than UPS). – Greg Apr 29 '15 at 11:04

Quantitative Applications. Once, XPS was not considered to be a very useful quantitative technique. However, there has been increasing use of XPS for determining the chemical composition of the surface region of solids. If the solid is homogenous to a depth of several electron mean free paths, we can express the number of photoelectrons detected each second $I$ as $$I = n \phi \sigma \epsilon \eta A T \ell$$ where $n$ is the number density of atoms ($\text{atoms cm}^{-3}$) of the sample, $\phi$ is the flux of the incident X-ray beam ($\text{photons cm}^{-2} \text{ s}^{-1}$), $\sigma$ is the photoelectric cross section for the transition ($\text{cm}^{2}/\text{atom}$), $\epsilon$ is the angular efficiency factor for the instrument, $\eta$ is the efficiency of producing photo electrons ($\text{photoelectrons/photon}$), $A$ is the area of the sample from which photoelectrons are detected ($\text{cm}^{2}$), $T$ is the efficiency of detection of the photoelectrons, and $\ell$ is the mean free path of the photoelectrons in the sample ($\text{cm}$).
For a given transition, the last six terms are constant, and we can write the atomic sensitivity factor $S$ as $$S = \sigma \epsilon \eta A T \ell$$ For a given spectrometer, a set of relative values of $S$ can be developed for the elements of interest. Note that the $I/S$ is directly porportional to the concentration $n$ on the surface. The quantity $I$ is usually taken as the peak area, although peak heights are also used. Often, for quantitative work, internal standards are used. Relative precisions of about 5% are typical. For the analysis of solids and liquids, it is necessary to assume that the surface composition of the sample is the same as its bulk composition. For many applications this assumption can lead to significant errors. Detection of an element by XPS requires that it be present at a level of at least 0.1%. Quantitative analysis can usually be performed if 5% of the element is present.