[+Explanation] Multiple Peaks Fitting of XPS, Raman, PL spectroscopy │Curve Fitting using Origin

[+Explanation] Multiple Peaks Fitting of XPS, Raman, PL spectroscopy │Curve Fitting using Origin


This is a Raman spectrum. I have fitted with 14 peaks. I will show:
How did I do it? In a previous video,
I showed how to do it. Many viewers asked for a similar video with
explanation on several steps. Some of them also asked several questions. In this video,
I will try to answer those questions, and try to explain each step. Let’s start from the beginning. I am taking data to a new workbook. Plot the data in “Scatter” type. For better visibility, adjust the dimensions. Adjust the scale. Change the symbol to open circle. Now, inspect the spectrum. Count the possible minimum number of peaks. This quick inspection shows,
that the number of peaks cannot be less than 10. Let’s check it again. This time, concentrate on the humps. In the previous counting, we consider it as
1. But, when we see closely, it appears not to
be one. I am assuming, at least five peaks here. So, the count is now, 12. Let’s go to next hump. At least three peaks here. So, two more peaks to add. Total 14 peaks. If you are confused with the number, don’t
worry. Make a few trials between the minimum number
of possible peaks, and the maximum possible number of peaks. Start with the minimum number. and Follow the steps, I will show now. Go to analysis. Peaks and baseline. Multiple peak fit. Open dialog… There are 3 peak types available here. Gauss, Lorentz, and Voigt. Which one should you use? Let’s have some basic knowledge about these
peak types. This is how the Gaussian and Lorentzian peaks
differ from each other. In a very non-specialist language,
if in the spectrum number of peaks are spread more near the baseline,
and overlapping each other, then you can go for Lorentzian. If the peaks are very sharp, then go for Gaussian. The third tye is Voigt. Voigt is a combination of Lorentzian and Gaussian. Here, you can see how the peak shape is changed,
from Lorentzian to Gaussian through several Voigt types. The peaks can also be some other types. We don’t talk about those here. Please, keep this in mind,
there are no strict rules for finding the peak type before fitting. Which means you have to make a few trials
with different types. You will know your peak type ONLY when you
will get the best fit. I suggest you start with Voigt or Lorentz. Here, I will go with Lorentz first. The number of peaks: 14. Here, my suggestion is, go as slow as possible. If you think you did wrong at one of the peaks,
do it again for all. I will go very carefully. This fit is almost perfect! Don’t get worried, if you are not getting
at the first trial. As I said earlier,
If you are confused with the number, don’t worry. Make a few trials between the minimum number
of possible peaks and the maximum possible number of peaks. Start with the minimum number. Follow the steps, I have shown. Then repeat for the maximum number. Analyze and understand. Try for another number of peaks between minimum
and maximum. Repeat. And finalize, where you get the best peak
fit. Narrow down your range between minimum and
maximum number by several trials. The second fact was the peak type. Now, after the first trial,
you can see the individual peaks and their contribution to the envelope. See that difference between the envelope,
the red line, and the experimental data, the black points. If the difference is more near baseline, change
Lorentz to Gauss. In most of the cases, like Raman, PL, Lorentz
will fit better. But, you should try Gauss, if the fit is not
perfect with the other two. If you want to practice with the data file
which I used in this video, you can download it from our website: insilab.com
. Now I will show some useful functions for
very new learners. Double click to customize the individual peaks. Change the colors and width so that you can
make it like this. Find out a theoretical or computational published
work on your material or at least a material with similar structure. Compare the peak positions with those reported
computational values. Then, you can index the peaks like this. With this, I conclude today’s tips. As you have seen, this video is made on request
from some of you. So, keep commenting, asking an suggesting.

4 thoughts on “[+Explanation] Multiple Peaks Fitting of XPS, Raman, PL spectroscopy │Curve Fitting using Origin

  1. hello I am using origin 2016 and in the multiple peak fitting dialougue there is no box of " number of peaks" can u please help. I tried taging the peaks but it is not working too

  2. Amazing. This video is really helping. You explained it very well and in a very easy way. I'm using Origin Pro8 but unfortunately, Origin Pro8 doesn't have Voigt to fill multiple peaks. Could you please help me how to fit multiple peaks for Voigt using OriginPro8? Thank you in advance.

Leave a Reply

Your email address will not be published. Required fields are marked *