Could 12,000 Journal Papers a Year Be Wrong?

To identify the chemical composition of different materials, a technique called X-ray photoelectron spectroscopy, or XPS, is often used. It has existed since the 1960s and is today a standard method in materials science. But researchers at Linköping University have in an article in Scientific Reports been able to show that the method often gives incorrect results.

Method criticism is a virtue in research, but here a couple of generations of researchers seem to miss being critical of a method with shortcomings that have actually been known from the beginning. For a long time, it also included our own research group. But now we hope that X-ray photoelectron spectroscopy will start to be used with greater thought, says Grzegorz Greczynski, senior lecturer at the Department of Physics, Chemistry and Biology at Linköping University.

Together with Lars Hultman, professor at the same department, Grzegorz Greczynski has been able to state that the technology X-ray photoelectron spectroscopy, also called XPS, can give misleading analyzes due to an incorrect calibration method. The results are published in the journal Scientific Reports.

Nobel Prize-winning measurement technology

XPS is used to analyze the chemical composition of different materials. It is a standard method in materials science and is used in over 12,000 scientific articles annually. The technology was developed into a useful and powerful method for chemical analysis in the 1960s by Kai Siegbahn, professor at Uppsala University, which later awarded him the 1981 Nobel Prize in Physics.

The pioneering effort behind the Nobel Prize should not be questioned. When the technology was first developed, the error was comparatively small due to the more roughly calibrated spectrometers of the time.

But due to the rapid development of spectroscopy and spread to several disciplines of science, the instruments have been improved to such an extent that the underlying error has become of decisive importance for future development, says Lars Hultman.

According to the Linköping researchers, the original XPS method has been misused by a fatal assumption in a calibration method that uses the element carbon to calibrate the experiments. But it turns out that carbon-based chemical compounds, which naturally condense on most samples and are used for calibration, give rise to measurement signals that depend on the environment and the substrate to which they adhere. Therefore, there is no unique signal from carbon and large errors occur when a more or less arbitrary value is used as a reference to calibrate the measuring instrument.

Perfect storm

Already in the 1970s and 80s there was criticism of the calibration method, but for some reason it fell into oblivion and the method became standard. According to Lars Hultman, there are a number of different factors that interact for the errors to have gone unnoticed for almost 40 years. He believes, among other things, that the dramatic increase in magazines that came with the digital publishing opportunities comes into play as well as a lack of review.

In addition to a rapidly growing number of researchers failing to be critical, there also seems to be a form of laxity among editors and reviewers of the scientific journals who agree to publish obviously unphysical interpretations of data. View from inside an XPS spectrometer.

It can be likened to a perfect storm. There are probably the same type of problems with a lack of method criticism in several scientific disciplines, and it risks damaging the credibility of science in the long run, says Lars Hultman.

Other methods

Grzegorz Greczynski hopes that their discovery will partly lead to the further development of XPS technology and partly contribute to a more critical way of thinking in the research world.

 Our experiments show that the calibration method, which is enormously popular, leads to contradictory results and should therefore be discontinued. There are various options that can provide more reliable measurement calibration, but they require more of the user and some need to be further refined. At the same time, it is important to encourage method-critical conversations in the lab, development department, or seminar, concludes Grzegorz Greczynski.

Reference
Greczynski G, Hultman L. The same chemical state of carbon gives rise to two peaks in X-ray photoelectron spectroscopy. Sci Rep 11, 11195 (2021). doi: 10.1038/s41598-021-90780-9

This article has been republished from the following materials. Note: material may have been edited for length and content. For further information, please contact the cited source.