When I do refinement of XRD with GSAS EXPGUI, the resulted CHI**2 is 3.6, while wRp and Rp are 0.2376 and 0.1684, respectively. Why? How could the wRp and Rp could be reduced to an acceptible number? Thank you.
Dear Yuan..
The resulted CHI**2 is 6.2. So , you must refine the data to reduce the value like 1 or lesser than 1. Then only, you may see the reduced wRp and Rp value.
To my knowledge CHI**2 value is 1 or lesser than 1 is a good refinement.
It seems from the residual (blue line) that the intensity of your peaks are not well fitted. Are you completely sure that your structural model (atomic positions and occupations) are correct? However, it is not easy to see from your picture whether the problem with the intensity originates from the magnitude of the peaks or their shape...
Dear Yuan,
you might have a problem with preferred orientation or the starting structural data are incorrect.
Best, Bernd
From the graphical comparison it seems that different problems are present: wrong intensities (check your structural model) and also wrong peak shape and broadening
Thanks to all of you for the kindful answers to my questions. In this measurement, heating devices were used to keep the temperature at a constant value. However, the device is covered with an aluminium foil to reduce thermal convection. I suggest that the covered aluminium foil absorbed some x-rays, which reduced the intensity of either incoming or diffraction that makes the peak shape and profile does not fitting well in the unit cell calculation functions. Have you ever knew the similar phenomenon?
Thanks again for concerning this question.
Typically, it would be astonishing to see such effect caused by a thin Al foil (if this is a thin foil that is commonly used in labs for thermal homogenization) but after all you are using a low brilliance Cu radiation from some lab source with a Bragg Brentano geometry. Uneven Al cover with varying thickness could somewhat mess up your diffracted signal. It is still very likely though that your model has some issues with atomic positions and incorrect instrumental profiles like suggested before. You should at least consider measuring some standard like LaB6 with and without foil to establish the profiles.
U introduced to many parameters or u had a noisy data (as Dr Ian said).
U use type 2 (pseudo-voigt standard), I sugest to use type 3 (with FCJ Assymetry) or 4 (with Stephens anysotropic strain broadening).
Dont refine zero because that for neutron and synchrotron, not for flat plate BB.
Try not to use damping (even 70% shift), u need more iteration. U need to come with better refinement strategy: refine lattice parameter, preferred orientation, occupancy, displacement etc
It wud be nice if u cud post ur raw data.
I also, I suggest you work refinement with the type 4 function, because with type 2 is not refined zero "SFHT"(one of the first to be refined), and see if you have peaks in the diffractogram not engage with the theoretical prediction because it can also be the presence of a second phase.
Refine GU, GV, GW, LX, cell, atomic positions, U, etc, there are lots of things you can still try.
- Badges
- Science topic
- Similar topics
- Filing