I have carried out XPS analysis of CNFs/ACF samples to know the nature of C1s, O1s spectrum. I have often seen that people use mixed Gaussian Lagrangian function to interpret the results. What is the necessity of doing this?
Thanks in advance.
Sorry Bhaskar, people use Gaussian - Lorentzian functions. The reason is that the profile of the measured XPS data is not only originating from the atomic level considered, but also from the instrument and its resolution. Thus it might be that XPS profiles are often best described by a mixed function. Further effects may arise if you have a metallic species. In this case, the ejected photoelectrons may interact with the conduction electrons, loosing very small portions of energy. This may be accounted for by using an additional tail function (i.e. the peaks are asymmetric. Hope this helps, kind regards, Dirk
Hi Bhaskar
Dirk made valid argument on use of mixed Gaussian-Lorentzian functions while analyzing XPS peaks. If you are dealing with carbon nanotube forests, you might be able to get away with using 100 % Gaussian peak profile. Metallic peaks are Lorentzian (having extended tails) and polymeric samples usually yield Gaussian peaks. Nevertheless, because of instrument factors, the peak profiles are often not ideal, therefore a mixed function. A careful peak fitting approach along with literature research will give you the ratios of Gaussian to Lorentzian character in the peak.
PLease be free to ask any follow-up questions.
Cheers!
Hello, Bhaskar! I usually use a ratio of 90:10 (Gaussian/Lorentzian) for C1s and O1s spectra. This ratio almost always gives a good result.
Dear Fetisov,
Could you please tell me which software you use to process the XPS data. Also, could you please send me the link of one paper which describes the C1s and O1s spectrum incorporating the Gaussian Lorentzian mixed functions... Thanks in advance...
Please, find several programs for XPS curve fitting as below,
http://www.uksaf.org/software.html#1
Hi,
I would like to recommand CasaXPS for XPS peak fitting and the Casa Cookbook for additionnal information.
http://www.casaxps.com/
regards.
In fact, the mixed Gauss-Lorentz function is an approximation of Voigt function that is a convolution of gaussian and Lorentzian components. The lorentzian come from inherent line shape, i.e. the photoemission process and the gaussian comes from spectrometer and broadening from slightly different local bonding. As it was mentioned earlier, the G/L ratio is instrument-dependent (and pass-energy dependent) because of gaussian dependency. In modern software packages for peak fitting you can choose to use convolution of Gauss and Lorentz functions (Voigt function) which is more precise. Moreover, the inherent line width given by lorentzian component with should be instrument independent. The recent computer have so high computing power that the approximation is no longer needed compared with the situation 20 years ago.
Bhupinder Singh wrote: If you are dealing with carbon nanotube forests, you might be able to get away with using 100 % Gaussian peak profile.
This is not true for the SWCNTs. In this case, the carbon configuration is graphitic, therefore the C1s line is strongly asymmetric, like in many metals.
You can read more about carbon spectroscopy in my paper:
https://www.researchgate.net/publication/263591416_Spectroscopy_of_carbon_from_diamond_to_nitride_films
Article Spectroscopy of carbon: From diamond to nitride films
Thank you very much Kaciulis sir for sharing those papers. I will definitely read those articles. Thanks...
Hello all,
I have a silly question to ask...... how can we get the ratio of gaussian/lorentzian in a viogt fit with origin software ?
- Badges
- Science topic