Dear RG members,
I analysed a samples twice in XRF analyzer with two different options first "Oxides-analysis" and second was "Elemental-analysis". With the first option, the % of SiO2 was 66, whereas in the second option the % of Si was 59. However, based on the atomic mass, the % of Si in SiO2 is 40. Therefore, the XRF should give 26.4% of Si, if SiO2 is 66%. Can anyone please justify this?
Thanks,
Mohit
Dear Mohit Somani
Your question is always asked by the beginners of XRF users and some users will give up using the XRF because XRF gives wrong answer. However your question is a very important question to become a skilled XRF user.
Usually the ED-XRF cannot detect oxygen, and the sample is often a mixture of metal oxides, in such a case, you can use "Oxides analysis". The XRF instrument usually detects Si, Fe, ..., but could not detect oxygen. And the sample is rock or something like that, the quantitative analysis program calculates the weight ratio of SiO2, Fe2O3, ...from the peak intensity of Si, Fe, ... without detecting oxygen peak. In this calculation, FeO or Fe3O4 or FeOOH are not considered. SiO or SiC are also not considered.
If your sample is an Fe-Si alloy, you must use alloy analysis. The computer program supposes that the samle is composed of Si and Fe. However if the sample contains boron, carbon, oxygen, nitrogen, ... which are not detectable by XRF, the accuracy becomes worse.
When you want to analyze a soil sample, oxide program will give better result than alloy program. But soil is usually not completely oxidized, containes organic components, and thus the result is not accurate enough.
The details of the computer programs (matrix effect correction) are quite successfully explained by Marek Bielewski.
You can not analyze the sample of which you do not know well.
Yours sincerely
Jun
The issue is not as easy as calculating the simple percentages.
If you declare sample made of oxides, the software "understand" the observed signal - peak areas or cps, depending on the algorithm - was modified by the non-fluorescent elements, like oxygen (the same physics would also work for carbon or hydrogen atoms). So, the absorption of fluorescence by oxygen atoms and scattering plays here a role. The most prominent effect is absorption. The software can also try to correct for enhancement effects - mainly additional excitation by fluorescence from other sample constituents. Intensity of this additional excitation also depends on sample absorption, so on the presence of additional atoms, like oxygen.
In case you declare sample made of only fluorescent elements the matrix effects (so the absorption and enhancement) are calculated only for the elements detected in the sample.
You will find a lots on XRF physics in:
https://www-pub.iaea.org/MTCD/Publications/PDF/te_1401_web.pdf
https://www-pub.iaea.org/MTCD/publications/PDF/IAEA-CMS-21_CD_web/PDF/CMS_21.pdf
https://www-pub.iaea.org/MTCD/Publications/PDF/te_0950_scr.pdf
Dear Mohit Somani
Your question is always asked by the beginners of XRF users and some users will give up using the XRF because XRF gives wrong answer. However your question is a very important question to become a skilled XRF user.
Usually the ED-XRF cannot detect oxygen, and the sample is often a mixture of metal oxides, in such a case, you can use "Oxides analysis". The XRF instrument usually detects Si, Fe, ..., but could not detect oxygen. And the sample is rock or something like that, the quantitative analysis program calculates the weight ratio of SiO2, Fe2O3, ...from the peak intensity of Si, Fe, ... without detecting oxygen peak. In this calculation, FeO or Fe3O4 or FeOOH are not considered. SiO or SiC are also not considered.
If your sample is an Fe-Si alloy, you must use alloy analysis. The computer program supposes that the samle is composed of Si and Fe. However if the sample contains boron, carbon, oxygen, nitrogen, ... which are not detectable by XRF, the accuracy becomes worse.
When you want to analyze a soil sample, oxide program will give better result than alloy program. But soil is usually not completely oxidized, containes organic components, and thus the result is not accurate enough.
The details of the computer programs (matrix effect correction) are quite successfully explained by Marek Bielewski.
You can not analyze the sample of which you do not know well.
Yours sincerely
Jun
- Badges
- Science topic