Deconvolution is quite an art. It depends on the type of spectrum you have (raman, xrd, FTIR etc), the number of peaks below the curve, their respective weights and widths, the background etc. In Origin you have the NonLinear Curve Fit where in the Advanced Options you can select the number of peaks and the type of peaks you want to adjust. There, you give the initial values of the peaks (position, width, area), if data are with fixed y0 and so on. Then you begin to move all the parameters at the time until the function converges or you move some parameters while let some fixed until you get satisfactory results. It is not easy, I can tell you. If you want I could do an example for you of some of your data and try to build a detailed procedure, although, as I mentioned before, each type of data and even each spectrum need time to get it.
Angel, I watched that video before I asked this question, but it is not for exactly for deconvolution that I wanted. Anyway thanks for your link and answer.
Hi Sergio, the software, FitYK, you specified is new to me. Let me try to do deconvolution in FitYK. But If you could provide the steps to be done, for deconvolution will be much appreciated. Thanks in advance
There is a free version of the software that can be downloaded. For you to save the spectrum you have to go to DATA--> Export and then choose the options that you need, this will create a file with .dat format that can be opened in origin!
Deconvolution is quite an art. It depends on the type of spectrum you have (raman, xrd, FTIR etc), the number of peaks below the curve, their respective weights and widths, the background etc. In Origin you have the NonLinear Curve Fit where in the Advanced Options you can select the number of peaks and the type of peaks you want to adjust. There, you give the initial values of the peaks (position, width, area), if data are with fixed y0 and so on. Then you begin to move all the parameters at the time until the function converges or you move some parameters while let some fixed until you get satisfactory results. It is not easy, I can tell you. If you want I could do an example for you of some of your data and try to build a detailed procedure, although, as I mentioned before, each type of data and even each spectrum need time to get it.
Thank you Sergio for your reply. Yes...I can't save spectrum of fitting in fitTY, but anyway, for saving a spectrum of muti-curves, we must use Origin.
I was facing problem with originpro8 software. I contacted origin team , they advised me to upgrade.now with origin9 decovolute is as simple as ABC.go to analysis......peaks and baseline----multiple fit peaks fit.....and after assigning peaks positions and fitting you will get a very nice spectra.