Microstructural analysis using MAUD is easy and quick. However, I have some doubts as no instruction manual is available online for microstructural analysis.
1. But how to be sure about the calculated values like crystallite size and micro-strain? How do I cross-verify these values?
2. How to correct sample displacement error? I am working with steel samples. The data have been taken from a Rigaku xrd machine. From the machine software (smart-lab studio) I got the peak positions. Using that I calculated the lattice parameter by Bragg formula. The lattice parameter calculated by the Bragg formula differs from that obtained by Rietveld refinement using MAUD(Bragg formula: 2.867Å MAUD: 2.90Å). So, I am doing something wrong in Rietveld refinement using MAUD. I guess that there may be some sample displacement error that I am not refining. But as I do not have any manual for MAUD I am not sure how to do that. Can anyone help me with that?
Dear friend Nilabja Kanti Sarkar
Performing microstructural analysis using MAUD software involves the following steps:
Regarding your doubts:
In summary, microstructural analysis using MAUD software involves data preparation, Rietveld refinement, and microstructural analysis. To ensure the accuracy of the results obtained, it is important to cross-verify them with other techniques and compare them with material properties and processing history. Regarding the correction of sample displacement error, you can refine the sample displacement parameter during the Rietveld refinement process using the MAUD manual as a guide
Thank you so much Kaushik Shandilya for such detailed reply. I really appreciate it.
The link you have provided is not working (https://maud.radiographema.eu/manual/). Can you help me here?
This is how I do microstructural analysis now (and probably doing somewhere wrong).
I first get an instrumental broadening file. I generally take XRD of a NIST-Si powder sample as standard and refine it in MAUD. Fortunately, there is an instruction document available for that. Here I get the following parameter values
__riet_par_2-theta_offset0
_riet_par_asymmetry_value0
_riet_par_asymmetry_value1
_riet_par_caglioti_value0
_riet_par_caglioti_value1
_riet_par_caglioti_value2
_riet_par_gaussian_value0
_riet_par_gaussian_value1
I keep these values constant for my actual sample (steel) analysis.
I load my actual data file in .xy format. I load .cif file for pure iron (space group number 229, lattice parameter 2.8667Å). I click on the quantitative analysis option in the Refinement wizard. (I did not do sample displacement error refinement.). The software refines and gives a result where crystallite size and microstrain are calculated using the Delf-Isotropic model. (I guess this is the default setting). The software also calculates background etc.
Then I change my model to Popa-Isotropic or Popa-Popa to get a better fit.
Let me ask some specific questions if I may,
1) In which step I should refine the sample displacement error?
2) Should I keep the _riet_par_asymmetry_value and _riet_par_gaussian_value constant for actual sample analysis (here steel samples)? Caglioti values have to be kept constant of course.
3) There can be more than one 2theta offset parameter be refined, like __riet_par_2-theta_offset0, __riet_par_2-theta_offset1, etc. How to decide how many I should add and refine?
4) In the sample displacement error, there is an x,y,z error. Should I refine all 3?
5) How to cross-verify them by TEM or SEM? I guess EBSD can be done in SEM but what about TEM? How to do it in TEM?
0.1) So how is the Si standard refinement? Is it good, are peak profiles and positions acceptable? If there are systematic peak displacements in opposite directions in low and high scattering angles, then you may want to refine the sample displacement here too. That would be the _riet_par_spec_displac_z, if you are in Bragg-Brentano geometry. Check that all the orientation angles for the sample and the dataset are = 0 (just in case, since I think that they may influence the correction). Note that it is scaled in mm, so if you have it larger than about +-0.1, it indicates that something may be wrong.
0.2) I think that using Wizard->Quantitative analysis would (by default) also refine the _riet_par_2-theta_offset0, which you in fact would prefer to have fixed.
1) It does not matter, in which step you refine the sample displacement error, as long as the structure model is close to the experimental pattern (e.g., all peaks are in more or less correct positions, have "good" intensities, shapes, etc.). It should converge to reasonable value.
2) _riet_par_asymmetry_value and _riet_par_gaussian_value are designed to reflect the instrument contribution to the resolution function. All other changes in peak profiles and asymmetry are considered to be due to microstructure of the material, and should be treated accordingly.
3) As far as I am aware, with these parameters you may build a polynomial in scattering angle to correst for profile positions. Reasonably, you need only _riet_par_2-theta_offset0, but in some specific cases (e.g., the pattern is registered by a film, which is somehow distorted) you may put any number of terms. Of course, depending on your experiment setup and sample, there may be other (more or less significant) corrections for diffraction peak positions, due to sample transparency, refraction, dispersion, flat surface, etc...
4) Depends on experiment geometry.
5) How to cross-verify what? Remember, that with MAUD models you may get the size of spherical (or in shape of the "composite" domains in the Nicolae Popa model) coherent scattering domains (or the distribution, if you use the corresponding model), and some root-mean-square microstrain value. It may be possible, e.g., to assume that coherent scattering domain = grain (most likely valid if their size is below 0.1 mkm), and that all microstrain comes from dislocations and use some extra models to deduce dislocation types and densities, and then do TEM/SEM measurements, but it looks like a lot of work with an uncertain outcome considering all the simplifications in MAUD models...
Dear friend Nilabja Kanti Sarkar I feel Roman Nikolayevich Vasin has provided some deeper you may want to consider
- Badges
- Science topic