I'm trying to perform semi-quantitative analysis on multiple-phase sediment samples. I've succeeded in identifying the different phases, but I don't know how to adjust their peak heights optimally in order to obtain concentration values that are as accurate as possible.
Should the pattern peaks all be within the scan peaks?
Is it alright to have pattern peaks that surpass the scan peaks?
Should I only be looking at the tallest peaks?
Thank you very much in advance!
I assume you made sure that your samples didn't have a preferred orientation of the minerals - that is quite tricky if your materials contain platy minerals like mica. If you didn't succeed in 'randomising' your sample, the relative heights of the different peaks for a single mineral may be affected, and you will then have to be very careful which peak to choose for doing your quantification.
Top-of-the-range, award-winning software is available for unravelling your minerals' secrets. The DIFFRAC.SUITE and DIFFRACplus software provide a common look and feel for measurement and analysis software. Powerful tools are provided to analyze your diffraction data and to provide information on the phase content, the phase composition, crystal structure and microstructure, crystallinity or crystallite size.
yes you should look for tallest peaks
I assume you made sure that your samples didn't have a preferred orientation of the minerals - that is quite tricky if your materials contain platy minerals like mica. If you didn't succeed in 'randomising' your sample, the relative heights of the different peaks for a single mineral may be affected, and you will then have to be very careful which peak to choose for doing your quantification.
You need to scale the 100% peak of the pattern to your data. This is based on the I/Ic figure in the database, and these are all measured from the 100% peak. Be careful that the pattern does have a valid I/Ic value as a default of 1 can be used for absent data and this can affect the accuracy of the results. It is often worth searching the database for other patterns of the phase of interest to find the range of I/Ic values and peak hights to see what sort of error is likely.
As has been mentioned you MUST have removed any preferred orientation effect for this to work. Also you MUST have identified every phase as the results are normalised. Any missing phases will cause an overestimation of the remaining components.
Sediments can be tricky as they can contain clay minerals that do not have well constrained I/Ic values, or if they do they are for very well crystalline clays, that are unlike anything you find in the real world sediments!
If you can I would recommend using a Rietveld quant package such as TOPAS as this can successfully model any preferred orientation and make the appropriate corrections.
I think that you can successfully model any preferred orientation and make the appropriate corrections in TOPAS. However it is not so easy and you can do this only in a case when you have structures (unit cell and positions of atoms) of all phases, what is a case of sediment samples could be impossible to do. In other cases, this method will be not very good for quantitive analyses. Am I right, Ian J Slipper? Omar Choa could you write which solution did you apply, please?
- Badges
- Science method
- Similar topics
- Methodology
- Crystallography
- X-ray Diffraction
- XRD Analysis