It is not a problem to use XRD for quantitative phase analyses. If you know the phases you even don't need any other technique like EDS. If you do NOT know the phases you have to identify them first. This might be a bit tricky sometimes if they are totally unknown and you do not have any idea about the chemistry, but XRD is a very powerful technique which is very sensitive to the chemistry as well (peak shift). If you finally identified the phase you have to test how much background you have since this gives you the amount of amorphous material. You are a bit lost if the background is not only formed by one phase. Then you can only make estimations. If the background is more or less of one chemistry there are procedures to find out how much amorphous phase is inside your mixture (mainly comparison with self-made mixtures). The quantification of each phase is commonly dome by a full pattern refinement software (unfortunately often called Rietveld program which is actually only for crystal structure refinement and not for phase quantification, i.e. it only works with the diffraction pattern of ONE phase). This full pattern refinement considers many things like the scaling factor which you are interested in, peak broadening (size and strain...but here you should not expect that much, especially when you work with 10 phases or more), but also texture. At least BGMN I worked with has powerful approximations for texture phenomena. Therefore, it is possibly the most successful software regarding phase analysis of clay minerals, cf. Reynolds-Cup (link)). Finally, the quantification is well done if the crystal structures are completely described,i.e. with all H, H2O or OH groups, i.e. also elements which are weak but contribute to the signal if their amount is "big", cf. ettringite where these element contribute more than 5% to entire signal). The result is the fraction in at% which you now need to recalculate into w%. Software should do this for you, but you should make sure that the values given are really in w% and not in at%. Since we are calculating with atoms all calculations are related to a%.  

Summary: except of the problem with the background which can be tricky (but not unsolvable in case of a single phase) you can definitely determine the phase faction quantitatively. As far as I know it is the only technique! Preconditions are of course an excellent homogenization which is however always better than EDS in REM which only determines the chemical composition but cannot distinguish between magnetite and hematite, aragonite and calcite, tridymite, cristobalite and quartz etc. (the list is ENDLESS). And for good software texture is not a big problem. You have to keep in mind the minimum amount of a phase. An average approximation is 2% but this is only the case if all phases have comparable diffraction power.

You already see: as any other technique full pattern refinement has its request on the knowledge of the user. But it gives you a lot back. You only need to know hoe it works. And for this you will find an endless number of publications and several excellent books. And the number of software (free like Fullprof or GSAS as well as commercial like TOPAS which is used in cement industry for QUANTITATIVE analysis) is sufficient but has sometimes very special applications. 

http://www.clays.org/Reynolds.html

You can determine the relative fraction of your constituents in a semi-quantitative manner by performing a Rietveld refinement of your XRD pattern, assuming that all constituents are reasonably well crystallized, not nano-sized, and that your pattern does not exhibit a significant preferential orientation. In your target compound is the only well crystallized one, you can add a (precisely) known amount of a standard compound (typically LaB6 or Si) and perform the Rietveld refinement of the obtained pattern.

Hi,

You can use Bruker Diffrac EVA, an XRD analysis software, or even other softwares to analyze your data semi-quantitatively. In a software that I've mentioned before you can see on the tutorial tab how to use the function. It's quite practical and easy to use. However, I haven't tried the Rietveld refinement yet with the software, guess it'll be more reliable. I'm not sure.

Dear Friend,

I am sure it is not possible to measure wt.% with XRD analysis. For sure it is a phase characterization technique and gives you the quantitative values, You can be used EDX ord  XRF for recognizing the wt.% of you additive or sth else.

Phase percentages can be made by XRD by proper peaks deconvolutions and taking into account the matrix effect. Intensity and area of the peak is indicative of the phase content. Normally software Rietveld analysis helps can give the qualitative/quatitaive values of each phase content in the material, provided deconvolution, peak broadening due to stress if any and matrix effect is taken care very carefully

The best is to find a software like MAUD, GSASI or II is free but you need to learn how to use them). If you can, please contact me so i will see what we can do here. (Ron Jenkins and R. Snyder book explains all the details how to do the calculation.

It is not a problem to use XRD for quantitative phase analyses. If you know the phases you even don't need any other technique like EDS. If you do NOT know the phases you have to identify them first. This might be a bit tricky sometimes if they are totally unknown and you do not have any idea about the chemistry, but XRD is a very powerful technique which is very sensitive to the chemistry as well (peak shift). If you finally identified the phase you have to test how much background you have since this gives you the amount of amorphous material. You are a bit lost if the background is not only formed by one phase. Then you can only make estimations. If the background is more or less of one chemistry there are procedures to find out how much amorphous phase is inside your mixture (mainly comparison with self-made mixtures). The quantification of each phase is commonly dome by a full pattern refinement software (unfortunately often called Rietveld program which is actually only for crystal structure refinement and not for phase quantification, i.e. it only works with the diffraction pattern of ONE phase). This full pattern refinement considers many things like the scaling factor which you are interested in, peak broadening (size and strain...but here you should not expect that much, especially when you work with 10 phases or more), but also texture. At least BGMN I worked with has powerful approximations for texture phenomena. Therefore, it is possibly the most successful software regarding phase analysis of clay minerals, cf. Reynolds-Cup (link)). Finally, the quantification is well done if the crystal structures are completely described,i.e. with all H, H2O or OH groups, i.e. also elements which are weak but contribute to the signal if their amount is "big", cf. ettringite where these element contribute more than 5% to entire signal). The result is the fraction in at% which you now need to recalculate into w%. Software should do this for you, but you should make sure that the values given are really in w% and not in at%. Since we are calculating with atoms all calculations are related to a%.  

Summary: except of the problem with the background which can be tricky (but not unsolvable in case of a single phase) you can definitely determine the phase faction quantitatively. As far as I know it is the only technique! Preconditions are of course an excellent homogenization which is however always better than EDS in REM which only determines the chemical composition but cannot distinguish between magnetite and hematite, aragonite and calcite, tridymite, cristobalite and quartz etc. (the list is ENDLESS). And for good software texture is not a big problem. You have to keep in mind the minimum amount of a phase. An average approximation is 2% but this is only the case if all phases have comparable diffraction power.

You already see: as any other technique full pattern refinement has its request on the knowledge of the user. But it gives you a lot back. You only need to know hoe it works. And for this you will find an endless number of publications and several excellent books. And the number of software (free like Fullprof or GSAS as well as commercial like TOPAS which is used in cement industry for QUANTITATIVE analysis) is sufficient but has sometimes very special applications. 

http://www.clays.org/Reynolds.html

Dear Azimi,

you can determine the absolute weight fraction  of your major constituents using Rietveld method that implemented in many softwares, but you have to use  known amount of internal standard (10%) such as (Lab6, Al2O3, Si) and also you have to insert the unit cell parameters (CIF) for each constituent and also for internal standard. you can find the Cif file using the COD website.

best regards

I can see different response to the questions, some thinks it is possible and some not.

Can Azimi tell us which one works for him? if he managed to quantify it.

Regards,

Elsadig

Elsadig Eltai : The determination of the phase fraction in mixtures consisting of crystalline phases with known crystal structure and complete description (inclusing H atoms as well for a correct normalization) is a common tool. Like all other techniques it is not a technique for unexperienced people, i.e. you need always to know what you are doing. However, as far as I know, the output is given in fractions which need transfer into mass % since it is caclulated concidering the relative amount of unit cells. in other words, to get the mass fraction you need to correct this by the mass of each unit cell. Thats all.

Contributions which claim that this is impossible are incorrect (since they do not know the technique?). It is more complicated to apply it for mixtures containing unknown or amorphous phases. Then you need to use standards in order to get some kind of reference in your mixture. How accurate this becomes is also another question since for an unknown phase you do not know how "heavy" this phase, however, you can at least determine the relatice fraction of all known phases. With the content of an amorphous phase it is even more complicated. But he did not talk about these special cases...

Thanks alot Gert Nolzefor the reply