I have to disagree slightly :-). The RIR method is actually coming from a time we did not have computers and algorithms to predict "absolute" intensities as comparison between different phases. Therefore it was required to make defined mixtures in order to see how the intensities change. But also in those days major problems exists (which are already forgotten) e.g. mainly preparation problems (homogeneisation, grain distribution, textures for different compositions etc.): Vegard's rule is a rule and only fits ideally if the conditions matches in the same way. Nothing against this approach, but it is hard to generalize, although the other way - full pattern analysis - is not the ideal one since it is "only" a mathematical approach and only considers parameters we are implementing, activating etc. And this might be wrong as well, even if R-values become dramatically better.

Regarding Murats question I have to add that the major problem is the missing information about the scattering power of your phase. If you don't know the chemistry you cannot say anything. Imagine two structure with the same unit cell, e.g. Cu and Al. The diffracted intensity of both pure phases is quite different so that also a mixture of 50-50 would show much more intensity coming from Cu than from Al. If you now imagine you wouldn't know this percentage you would certainly underestimate the Al fraction and overestimate the Cu content. But already during preparation the more heavy Cu particles could be more placed in deeper regions so that your sample actually does not shows a sufficient homogenisation since you perhaps have more Al in the top. But even in case of an ideal homogenisation the different absorption would reduce the Al content and overestimate the Cu. AND: don't underestimate the influence of H. There are many phases containing crystal water or OH groups which are known but not described in the crystal structure. E.g., for ettringite 11% of the entire scattering power is not considered during missing description of H. For the phase identification it doesn't mean anything, but for the phase fraction is is dramatical.

Therefore, it is a hard job to find reliable quantitative phase descriptions of mixtures. The annual Reynolds cup shows you how "wrong" even experienced scientist in this field can be. The majority of us even wouldn't find the correct number of phases mixed together. And: They will never work with RIR!   

Hi,

First you have to identify the phase by matching the spectra with the data in the crystallographic database. Then in the Rietvield fitting, you have to put that material as second or third phase (as per the composition of your material) and have to get the best possible fitting. Once the fitting is done, it tells you the correct weight percentage of that phase.

Thanks for your response. I explor XRD spectrum in MDI Jade software. can it help in analysis for finding the Phase %age?

Yes it can - you have to to add all known phases and then refine them.

There is an option (not sure however how it is exactly called) which allows you to identify the ratio of the phases.

I want to add something else - please don't call it spectra! It is a diffraction pattern - spectra are usually in some form energy dependant while diffraction pattern are obtained diffraction of X-rays

Personally, I would use an internal standard approach (e.g. corundum) and a Reference Intensity Ratio (RIR) method. Rietveld refinement is for sure a fantastic way to determine phase quantification, but, in my opinion, it is not user friendly as a RIR method. For a better explanation I am attaching a brief presentation of ICCD.

Hope that it could help.

What is the criteria for FullProf? It usually gives an output called Chi2. Can I calculate the percentage of a phase by this way?

My personal experience also supports the Francesco's approach, in which several power mixtures of two or three phases are prepared with known mass ratios, then a calibration curve is plotted based on your X-ray diffraction results. You read the phase ratio of a particular phase using this calibration curves. 

You can do that for a particular phase as you said in the last sentence , Dr. Tsai.

Excuse me fro writing here but, my problem is how can I find the weight of a phase that is unrecognized? I have no calibration curve.

Regards.

To answer Murat's problem, you have to do it in three steps. If the unknown phase is in a polycrystalline sample, and mixed with other phases, You have to find a way to prepare a sample of single phase, single crystal is even better. Then analyze and characterize it like a crystallographer. That's a lot work involved, depending how much details you would like to have. After you have characterized the unknown phase, you proceed to mix with the internal standard and plot the calibration curve.

I have to disagree slightly :-). The RIR method is actually coming from a time we did not have computers and algorithms to predict "absolute" intensities as comparison between different phases. Therefore it was required to make defined mixtures in order to see how the intensities change. But also in those days major problems exists (which are already forgotten) e.g. mainly preparation problems (homogeneisation, grain distribution, textures for different compositions etc.): Vegard's rule is a rule and only fits ideally if the conditions matches in the same way. Nothing against this approach, but it is hard to generalize, although the other way - full pattern analysis - is not the ideal one since it is "only" a mathematical approach and only considers parameters we are implementing, activating etc. And this might be wrong as well, even if R-values become dramatically better.

Regarding Murats question I have to add that the major problem is the missing information about the scattering power of your phase. If you don't know the chemistry you cannot say anything. Imagine two structure with the same unit cell, e.g. Cu and Al. The diffracted intensity of both pure phases is quite different so that also a mixture of 50-50 would show much more intensity coming from Cu than from Al. If you now imagine you wouldn't know this percentage you would certainly underestimate the Al fraction and overestimate the Cu content. But already during preparation the more heavy Cu particles could be more placed in deeper regions so that your sample actually does not shows a sufficient homogenisation since you perhaps have more Al in the top. But even in case of an ideal homogenisation the different absorption would reduce the Al content and overestimate the Cu. AND: don't underestimate the influence of H. There are many phases containing crystal water or OH groups which are known but not described in the crystal structure. E.g., for ettringite 11% of the entire scattering power is not considered during missing description of H. For the phase identification it doesn't mean anything, but for the phase fraction is is dramatical.

Therefore, it is a hard job to find reliable quantitative phase descriptions of mixtures. The annual Reynolds cup shows you how "wrong" even experienced scientist in this field can be. The majority of us even wouldn't find the correct number of phases mixed together. And: They will never work with RIR!   

 Huang Wang, Did you find the way to estimate the phase percentage using JADE software? Any body who knows how to do that?

 For Jade software try this: 

You have to add all known phases and then refine them. Once the fitting is done, find the option ''easy quantitative'' which allows you the weight percentage of that phases if its PDF cards have RIR.

you should refer The International Centre for Diffraction Data (ICDD) to phase Identification of crystalline substance and crystalline phases in a specimen is achieved by comparing the specimen diffraction spectrum with spectra of known crystalline substances. Xray diffraction data from a known substance are recorded as a powder diffraction file (PDF). Most PDFs are obtained with CuKα radiation. Standard diffraction data have been published by the ICDD, and they are updated and expanded from time to time. When we need to identify the crystal structure of a specimen that cannot be prepared as powder, matches of peak positions and relative intensities might be less than perfect. In this case, other information about the specimen such as chemical composition should be used to make a judgment.

can we use xpert-highscore quantification method??

Any body who can shar a pdf file