Dear Jingyang Zhang ,

I am not an XPS specialist, however, how could you be sure that Ar+ treatment did not affect the surface in the same way for another element?

Would you still probe the pristine surface of interest?

It may not change significantly the surface, therefore a reasonable uncertain remain.

If your material has carbon, it may be used as a calibration target, as P, O and so on.

Best regards,

WNM

What kind of sample are you investigating?

If it's metallic or a small-bandgap semiconductor, the C1s method wouldn't improve your result quality, anyway, and even applying no calibration at all would be a better way to do it. See:

https://www.researchgate.net/project/Why-you-should-not-use-the-C-1s-peak-for-referencing-XPS-spectra

If you look at XPS discussions here on RG you will find that the adventitious carbon method is not as unanimously accepted as some people claim it is.

Yes ..as carbon is more common and available it is taken as reference...but in absence of that any other element can be considered...as per my knowledge...the things are going like this...measurement of energy is carried out to know binding state of elements...that is with reference to floating potential...hence device will create a spectrum but that needs reference to analyse the positions...carbon peak position is the pin point for spectrum...but same way any other element with specific dominating peak can be considered as reference.

If possible dropcast your sample on Au foil and simultaneoulsy record the XPS of Au 4f spectra. Most people use carbon or other noble metal like Au and Ag to do charaterization as these values are well known and easily identified. Before doing the measurment just do Ar ion sputtering to remove the oxide layer from the surface, this is also a standard procedure for calibrating the XPS instruments. Thank you.

Thank you all for your attention and constructive suggestions.

After I posted this question I subsequently took other internal elements as the benchmarks to correct my raw spectrum and got some tips. Here provides some feedback for any future discussion on this processing method.

My sample is a certain kind of high entropy alloy (HEA), containing five metal elements and also mixed with C and O on the very top surface. I'm concerned about the chemistry difference between the surface and inner bulk. When carbon is no longer sufficient for calibration, for example, I set the Fe dominating peak as a reference, the validity of this operation was considered by comparing the peak position of other metal elements with its reference value in the NIST database. The mismatch between the other four elements and its corresponding reference value is lower than 0.5 eV, which I think is a trustable way.

Of course, if one wants to improve the credibility, I also suggest conducting more than one calibrations based on different elements for those weak-carbon spectra. Specifically, Fe element for the first time and Cu element for the second time... The mismatch may get reduced.

Welcome for any further discussion on this point. Thx

Since your substance is an alloy, you should have states at the Fermi level. Therefore the Fermi cutoff is probably the best place to check how your measurement needs to be recalibrated or whether it requires a calibration at all; I wouldn't be surprised if, for a well-conducting alloy, you could just use the data as it is.

I agree with Jürgen and recommend considering the Fermi level, which is becoming more reliable in the community than C1s signal from your sample. You may find useful discussion in this review:

Article X-ray photoelectron spectroscopy: Towards reliable binding e...

The challenge goes for samples which are not conductive and it is not possible to get the Fermi cutoff.

Babak Bakhit Jürgen Weippert , in theory, I fully agree that it might be a good idea to use the Fermi level for energy calibration. However, in practice, have you ever measured low binding energy electrons with XPS? The fermi level is broadened even if you use a monochromatic x-ray source so it is really hard to find the correct value… and the intensities of those fast electrons are small because the transmission function of the spectrometer decreases with the kinetic energy of the electrons. The question in my mind is, why is there any charging observed in this particular case? I feel that the sample is not properly connected to the spectrometer, or the sample holder, because a shift of the peaks should not occur for a metallic sample.

Furthermore, Jingyang Zhang did you make use of a monochromatic source, or a dual anode with an aluminum window? In the latter case, photoelectrons excited in the window will flood your sample and equilibrate / compensate charging, so you may keep the shifts from the first C1s spectra.

good luck, Dirk

Indeed I have measured the Fermi edge with XPS (it was in a Synchrotron session, though).

As you can see from both of my answers, I stated that in case of doubt applying no calibration at all would be quite a viable option.

Jürgen Weippert , sorry, indeed better to avoid any wrong calibration in any case! With regard to the SR-XPS, can you post a Fermi-edge spectrum? What was the excitation energy? I did some SR-XPS with moderately low excitation energies aound 300 - 500 eV, there the Fermi-edge was o.k., but from my lab. experience, puh, hard to see any sharp decay at large kinetic electron energies ... Best regards, Dirk

The excitation energy back then was ca. 500-515 eV; I don't have the data here anymore, though; that session was part of my PhD thesis work at KIT. If you're really interested, I can try to organize the files. In the final publications on the investigated Cs fullerides we only showed surveys plus details of C1s and Cs4d.

However, in the surveys (fig. 4B), the cutoff can also be seen rather decently:

Article C68 : a non-IPR Fullerene Capable of Binding Extraordinary A...