This is my first time using XRD and the highscore plus software, so any help would be appreciated.
I want to execute a search and match to determine which minerals are present within my samples. (As search and match gives many minerals that could be in the data, I'm combining the candidates with the minerals I have determined the sample via SEM-EDS).
These are my steps so far, and I'm wondering if I should be making an improvements and if I am doing it right.
1) Background treatment (I accept, but don't subtract).
2) Search Peaks - I go through and add additional peaks.
3) Profile fit - I do an automatic profile fit, and then look at the difference plot to see if the differences are low enough. Is this step necessary to accurately perform a search and match?
4) Perform search and match.
Extra questions:
Do I need to conduct a rietveld refinement after determining the minerals (I am unsure what exactlya rietveld refinement is, but I have seen it being used).
Is there a way to show the percentages of the different minerals making up a sample? and also the XRD pattern of a certain mineral on the same sample pattern to show which peaks correspond to the mineral?
Any help would be greatly appreciated.
Dear Janki Patel ,
Your procedure is right. Nevertheless, you may always improve the response by considering some points:
1) If you got XRF, you can limit many options to only the interested ones, naturally taking into account the combination between them.
2) Not always, you end up with some peaks that initially can not be directly addressed. Some physical phenomena may explain those peaks, like solid-solution, satellite lines, stack fault, and some others.
3) Concerning your background, it is not strictly necessary to subtract; however, the subtraction only for the identification propose may help decrease the amorphous phase mismatch and small particles' contribution effect, helping in this step.
4) Rietveld is a diffraction fitting procedure that uses a reference (model) structure to fit the observable peaks to the phase's structure in consideration. Remember that increasing the phase types in your sample will decrease the facility exponentially to execute the fitting. All the Rietveld bases are developed for a single-phase fitting; nevertheless, they can describe a mixture of them.
5) To reach a more reliable phase content in your sample, a fitting is necessary. The relative amounts are calculated as an area contribution under the phase XRD pattern and added to all the different patterns in the same sample. The accumulation point to point in your angle range may describe the experimental XRD pattern.
Several drawbacks may occur in this procedure, like strong peak superposition, that will negatively reflect the phases' description, here straitly concerning the calculated area.
6) There are other ways to calculate the relative amount of each phase, not only by fitting the XRD but will lack in precision, only an idea of the whole present amount.
I hope this short answer helps you, somehow.
As a beginner in the field, practice with a known phase mixture can bring you confidence in interpreting real unknown samples.
Best regards,
WNM
In case you are working for clay minerals in soils and sediments then the methodology for their identification is different.
Dear Janki Patel,
Your steps are good for a Search/Match, the Profile Fit is not really required (but doesn't harm). Search/Match mainly acts on Peak Positions and these are very accurate already from a Peak Search. Profile Fitting can be employed to get more accurate values for Peak Area, Peak FWHM, and Peak Shape. A Rietveld fit is typically employed to get more accurate Phase Quantities after a PhaseID. Most Reference Patterns have included RIR (Reference Intensity Ratio) values. if that is the case you can get a rough Quantification directly after the PhaseID by just pressing the PieChart button. Here is a nice YouTube video that shows most of the required actions:
https://www.youtube.com/watch?v=dASaENblC_4&t=287s
Also as mentioned in another answer, whenever you know something about the chemical composition you should enter that information into your search. You can do that on the Restrictions page of the Search/Match dialog there you can click Restriction set [Edit..] and on the Chemistry page, you can Include and Exclude Elements from your Search.
Good luck with your analysis!
Hey Thomas Degen I have noticed that Rietveld fit won't work if the divergence slit was set to automatic, is there any way around that? and thank you for your answer
Hi Janki Patel, I am not sure which version you have, in V4.x you can also model/fit ADS Scans. If you have V3.x however you need to do a Divergence Slit correction first, from ADS to FSD. You find that action under Treatment => Corrections => Convert Divergence Slit...
Thomas Degen I think my university has the V3.x (I am not sure which one) in that case. I will try to do a Divergence Slit correction first. Thank you for all your help
Doriana Vinci , that's a pretty bold claim: "high score plus gives only semiquantitative results, even if a Rietveld refinements performed".
We, like others who write analytical software, make sure by many automated tests, using scans with known composition, that the analytical (QPA) results are correct and reliable.
Most problems arise long before the Rietveld fit is executed. Mainly in the sample preparation stage, like having too big or too small particles. Then leading on to the measurement stage, introducing sample displacement, over-irradiating the sample (violating the constant volume assumption) or having not enough counts (bad counting statistics) or too much background due to fluorescence, etc.
Many of these problems are not fixable anymore in the modeling/fitting stage.
Numerous Round Robins tests, like https://www.researchgate.net/publication/319416105_Summary_of_Results_of_an_International_Round_Robin_for_Mineralogical_Analysis_of_Iron_Ores
Show that QPA indeed can differ a lot, but completely independent of the fitting software that was used. Much has to do with the experience of the operator still, unfortunately. So, the best way to judge (and improve) your results is, as always, to prepare your own samples with known composition and feed them through the same Sample Preparation/Measuring/Analysis chain. To get an estimation of the true precision of the whole experiment.
- Badges
- Science method
- Similar topics
- Methodology
- Crystallography
- X-ray Diffraction