Curve fitting of Raman Spectrum

[mmus...@laurentian.ca]

Hello everyone,

We are two Master's Student working first time working on Fityk software. We are trying curve fitting of Raman Spectrum using this software so far we have done that using the auto-add ( Lorentzian 15% & Guassian 15%) starting fitting option in the menu bar, and have reached . We have got okay results comparing to the published articles. I have attached one of the sample spectrum.  So,

1) does the fitting look right? We are not too sure.

2) We are trying to adjust the height the highest peak on spectrum , lower the height down so that the smaller peaks can be easily. So how can we go about doing that?

3) How can we shift x-axis data points by some units for example we would like to shift one of the sample spectrum by 6cm-1

4) Is there any useful resources you could recommend beside the manual.

5) Should I give try to other curve fitting software such as matlab, orginlab or magicfplot.

Thank you

[christian weikusat]

Hi there,

I usually use the Voigt shape for fitting raman bands, you may want to
try it. Also it would help to either make a background subtraction first
or only activate the parts of the spectrum with actual peaks.
So to answer the questions:
1) You have way too many functions, just use one for the main peak and
one or two for the second order (if you really need that at all). It
makes no sense to fit that many functions to a single peak.
2) I don't understand what you are trying to do here.
3) Very easy: Just type in the command line X=x-6
5) It's totally up to you and depends on the purpose you want to
achieve. I have used fityk exclusively for many years and am still very
happy with it.

Best regards,
Christian

[aa...@laurentian.ca]

Hi Christian,

Thank you very much for your kind input. We are looking for peak in the range of 120-135 cm-1 and can you help us to produce result. What are the importance of Lorentzian and Gaussian. Is it necessary to have a good background of these functions to deconvolute Raman peaks/spectrums  using fityk software.Is it necessary to cancel background noise to have good peaks. Can you take look at the Raman Peaks Indentification.png to see our deconvolution . If you think it is incorrect do you mind deconvoluing the S5 R3 sapphire.text attached herewith . We have been badly stuck and we couldnt get any help on campus so your help will very highly appreciated.

King Regards.

 Ali,

Grad student

Laurentian Uni. ON.

Susbury Canada