Periodic table of most intense/most used XPS core lines[1]. |
A printable pdf version of this chart can be downloaded here.1. Most elements produce more than one peak in XPS. Thus, it is nice to have a convenient tabulation of the peaks that are most intense/most used in XPS analyses – this information provides a good starting point for an analysis.
2. An obvious, but still quite important, effect shown in the table is that the binding energy for a peak from a given orbital always increases as Z increases. This result would be expected from Coulomb’s law – the more protons in the nucleus, the more strongly the electrons are held in the atom.
3. As an important practical consideration, all of the peak energies listed in the table have binding energies less than 1200 eV. This choice allows the table to be used by analysts using both Al Kα and Mg Kα X-rays, which have energies of 1487 and 1254 eV, respectively. That is, this table is specifically targeted to those doing conventional XPS.
4. The table helps organize our thinking about XPS analysis. That is, one might regularly analyze a subset of the elements by XPS and memorize the best peaks for analyzing each of them without seeing how the preferred peaks for the different elements are related. This periodic table shows that the peaks that are most often used for XPS analysis are not randomly chosen among the possible signals produced by the atoms. Indeed, elements with about the same Z value generally have the same recommended/preferred peak. For example, beginning with lithium, which is the first element that gives an XPS signal, the 1s orbital produces the best signal for analyzing all of the subsequent elements up to Na. Then, from Mg to Si, the 2p orbital peak is preferred. After that, from P to Ga, the 2p3/2 peak is the preferred one, and so on. One of the reasons for the shifts we see here from one orbital to another as Z increases is that at some point, the binding energy of the orbital exceeds the energy of the probing X-ray, so a different orbital has to be considered.
5. We emphasize that the table lists ‘nominal’ binding energies, not exact ones. Clearly, it would not be possible for the table to list the exact positions of the peaks one may find in one’s XPS analysis because (i) elements in XPS undergo chemical shifts in response to the chemical environments they are in, and (ii) insulating samples often require charge compensation, which, in practice, can move peaks to higher or lower binding energies by a few eV.
6. As a follow up to the previous comment, we note that even though the table gives the best peak for an element for an analysis, e.g., for quantitative work, this does not mean that one should ignore the other peaks from that element in the spectrum. It is always a good idea to confirm the presence of an element in an analysis by making sure that the multiple signals expected from it are present in about the ratios expected for them.7
7. The recommended peak for some of the elements is an entire orbital that undergoes spin-orbit splitting, e.g., the recommended orbital for silicon is the 2p orbital, while for other elements the recommended peak is one of the two spin-orbit peaks, e.g., for the element after silicon, phosphorus, the recommended peak is the 2p3/2 signal. These recommendations are simply based on the energy difference between the spin-orbit peaks for a given element – when there is a large energy difference between the spin-orbit components, only one of them is listed, but when the peaks are too close to resolve well, the pair of them is recommended. For example, the recommended signal for Mg is the 2p peak, which only has a separation of 0.28 eV between its 2p3/2 and 2p1/2 components – these two peaks will show significant overlap by conventional XPS. Similarly, the 2p peak is recommended for Al, which has closely spaced spin-orbit components (0.44 eV energy difference), and the 3d peak is recommended for Ge (0.58 eV difference between its spin-orbit components).
8. In every case where one of the two spin-orbit peaks for an element is recommended, the peak with the higher j value is listed, i.e., the j = 3/2 state is recommended instead of the j = 1/2 state for p orbitals, the j = 5/2 state is recommended instead of the j = 3/2 state for d orbitals, and the j = 7/2 state is recommended instead of the j = 5/2 state for f orbitals. In each of these cases, the recommended peak corresponds to the larger of the two spin-orbit peaks, which should give a better signal-to-noise ratio for the measurement. In addition, the peak with the higher j value always comes at lower binding energy, which means that its background can overlap with the peak with the smaller j value, potentially complicating its analysis.
9. As a final note, we emphasize that the table here contains recommendations, not requirements. While XPS is convenient because most elements only show a few peaks, so overlaps between them are not extremely common, overlaps between peaks do occur. If the peak recommended in this table overlaps with a signal from another element, it may be best to consider a different peak from the element in question.
Reference: