Have you ever made these basic errors in XPS data analysis?

X-ray Photoelectron Spectroscopy (XPS), also known as Electron Spectroscopy for Chemical Analysis (ESCA), is a conventional technique for surface composition analysis. Besides characterizing the composition of materials, it can also characterize the chemical states of each component and quantitatively represent the relative content of each component. Therefore, XPS is widely used in various fields of materials research.

With the expanding application scope, XPS characterization is basically a fundamental part of articles. However, it is important for everyone to pay attention to the standardization and correctness of data analysis. In this sharing, we will discuss basic errors in XPS data analysis and share them with everyone.

(This sharing is not targeting anyone; it is simply an analysis of basic errors in data analysis.)

Case 1

In XPS data analysis, a crucial step is background subtraction. In the four graphs above, the background subtraction is incorrect, with the baseline positioned above the curve. This can have subsequent effects on peak fitting and qualitative and quantitative chemical state analysis. The correct approach involves selecting the baseline at the bottom of the corresponding element peaks (the region with relatively flat signals compared to the lowest signals), and choosing an appropriate background subtraction method. Common methods include Linear, Shirley, Tougaard, and Smart. It is advisable to use the same background subtraction method for the analysis of the same dataset.

Errors in the data analysis shown in the graphs include: incorrect baseline subtraction, arbitrary assignment of oxidation states for Ni elements, and incorrect ratio of areas for Mn elements' 2p3/2 and 2p1/2 peaks.

Case 2

The concept of orbital spin-splitting peaks was not correctly identified in the graphs. The figure below shows the data manual for the element Ca. From the data, it can be seen that the p orbitals of Ca exhibit orbital spin-splitting peaks, where the 2p3/2 and 2p1/2 pair represents a specific chemical state. In Figure 1, there are four peaks labeled as different chemical states for 2p3/2, which is incorrect. Additionally, there is only one chemical state labeled for 2p1/2, which is also incorrect. Furthermore, the area ratio of 2p3/2 to 2p1/2 should be 2:1.Figure 2 indicates a misunderstanding of the peaks, which according to the data manual, represent plasmons (loss peaks).

The errors in data analysis shown in the figures include: misunderstanding of the peaks and violation of basic principles of data analysis.

Case 3

The errors in the figures include a failure to consider overlapping peaks. The binding energy position of O1s falls between 528-535 eV, which overlaps with the range of Sb 3d5/2. Therefore, when analyzing the peaks of the Sb element, one needs to consider the O1s peak as well. However, in the graph, all peaks are attributed to Sb, leading to an unreasonable area ratio.

Error in the figure: Failure to consider overlapping peaks.