I am dealing with the in-situ XRD tensile test data of T91 steels from APS. In oder to find the change trend of dislocation density and coherent scattering volume. I used modified William-Hall to plot FWHM vs gC1/2 at certain strain stage, where g is the diffraction vector length and C is the dislocation contrast factor. Somehow the in linear fitting the y-intercept is always negative I wonder why? I used CeO2 as standard sample to calculate the instrumental broadening and used Pseudo_Voigt to fit the profile.By the way, which formula I should use to reduce the instrumental FWHM from the raw FWHM of the true sample.
in the second picture, if I used Pesudo_Voigt and know the fraction of gaussian
1-x and Lorentz x, then what n should I use as my formula?
thanks very much
If you are using pseudo-voigt. It would be better if you remove gaussian and lorenzian components of the instrumental broadening seperately and then recalulate the width of the peak again.
If you are using pseudo-voigt. It would be better if you remove gaussian and lorenzian components of the instrumental broadening seperately and then recalulate the width of the peak again.
Hello Mr. J. Wang, I want to ask you about calculation of FWHM. I also want to plot WH graph to calculate strain and size parameters. But I get stuck on calculation of FWHM. How to find out FWHM for XRD pattern, as there are many peaks in XRD pattern. So is it to calculate FWHM for individual peak manually using origin or you have some other line profile analysis software?? As FWHM should increase with increase of 2 theta. But if i do peak fitting using origin, then FWHM is not accordingly in trend with 2theta. There by, WH plot is not a linear graph and I am unable to find out strain and size parameters. Please help me to calculate FWHM for XRD peaks. Is there need of de-convolution of peaks??
Hello Mr. Gupta
There is a Matlab function that you can use to fit each peak to find FWHM using Pseudo-Voigt and you can just Google it.
As for de-convolution, that is what I am working on now because I used PV to fit peaks, then it would be kind of tricky to deduct the FWHM_instrument from FWHM_sample. If you are not getting right values of FWHM then your linear fitting of WH plot will not give you a good result.
Hello, Mr Loveleen K Brar,
can you explain in detail to me?
I have the FWHM_instr fitted by PV and got the FWHM_instr with fraction of Lorentz η1, and the same way got the FWHM_samp with fraction of Lorentz η2, how should I manipulate the data to find the real FWHM of my sample.
Thanks very much.
Dear Jinsheng,
in it's original formulation,e.g.
beta*costheta = lambda/t +2 (delta(d)/d)*sintheta,
where beta^n = FWHM(obs)^n - FWHM(instr)^n with 1
Jinsheng Wang,
I generally fit my peaks using origin "Pseudo -voigt fn 2" which gives me the individual gaussian and lorentzian widths. I believe matlab also has a simillar function available.
remove the gaussian and the lorentzian instrumental components seperately to get the final lorenzian and gaussian FWHMs for the sample.
Gaussian: sample^2 = total^2-inst^2
Lorentzian: sample = total - inst
Now combine the two, or use them seperately.
You should also look into using integral breadths rather than just FWHMs.
As suggested by Dr. Miguel Gregorkiewitz, Rietveld will also be a good idea for taking a closer look at your data.
Hello, Mr Loveleen K Brar,
What do you mean by saying "combine the two or use them seperately"? "combine" means literally "add" or if there is some formula I can use? How can I use them seperately to interpret my sample properties? By the way, I just want to find the grain size and microstrain. Forgive me for asking question that may seem too easy for you since I am new to XRD, thanks.
Jinsheng Wang,
One cannot simply add the gaussian and lorentzian widths. what i meant was that you can use either single of double-voigt analysis.
i think you will find the book chapter by D. Balzar quite informative and helpful. Some of the other resfs are also there in my paper (TaC from Ta-ethoxide). Please read the references there in. I think most of your confusions will be cleared.
Loveleen mam,
As you said you use Pseudo-voigt fn 2 for peak fittings, so I am trying to fit XRD peaks using the same function. Please help me more in this. My queries are as:
1. Should profile shape factor be equal to/between 0 and 1.2? If I put this boundary condition then software automatically fixes the value to 0 or 1.2. So I keep the value in between 0 and 1.2.
2. Is there any criteria to know whether gaussian factor should be more or lorentzian? I mean how to know the approximate value of profile shape factor?
2. Then based upon profile shape factor, individual widths for gaussian and lorentzian peak vary a lot. Even sometimes error becomes larger than the value of width. Is there any upper limit for widths?
Mam actually as i know, FWHM should increase with theta. So of-course then FWHM for both gaussian and lorentzian factors should follow the trend individually. Am I right? Please help me in this.
- Badges
- Science method
- Similar topics
- Methodology
- Crystallography
- X-ray Diffraction