Nilakantha! Good basic question. Good answers! You may increase exposure of your question to a larger cross section of RG experts by increasing the number of "Topics" up top to as many as 15.

Summarizing all the possible reasons for variations in peak intensities of XRD powder patterns in comparison to those in reference files:

1. Grain size, particle size, crystallite size - Coherently Diffracting Domain (CDD) Size

2. Sample preparation and loading methods.

3. Data normalization method. For flat samples, one may use the total integrated intensity of the entire profile to normalize. Assuming constant material density and therefore constant diffracting volume per sample surface VOXEL. This would not be the same in the transmission mode such as in the Debye-Scherrer method with a capillary tube.

4. Peak intensity versus integrated intensity.

5. Stress/strain state within the CDD and that due to the surrounding CDD's.

6. Sample shape and alignment w.r.t. the incident beam. Flat, capillary, film, etc. 

7. Geometry and XRD optics used including incident beam conditioning and slits used.

8. Incident beam shape and size. "point source" versus "line source"

9. Presence of preferred orientation. Inevitable in most practical cases.

10. Use of internal/external standards would help ease the challenge and more????

Intensities appearing in X-Ray Diffraction data are Relative. Peak with highest intensity willl correspond to 100 in PCPDF database. Don't worry about the intensity just match the angle and proceed.

One of the main reasons for variation of relative intensities in PXRD patterns taken with a Bragg-Brentano geometry is preferred orientation of crystallites, which is especially significant when dealing with layered or fibrous compounds. 

This typically happens when pressure is exerted on the powder in order to make the surface of the sample flat. It is sufficient to avoid pressing the powder to get rid of most of the effect of preferred orientation.

Could you show us the comparison? Sometimes it only appears that the intensity is different since it is the integral intensity of the peak and not the peak height. Did you used a variable slit or another diffractometer geometry? The typical reason already Marco explained. The best is always to post the original data since words are misunderstandable...and also recommendations.  

As was said, check first the scale of the intensities and be sure that both, experimental and reference, are normalized. If you still have mismatching between them, then check if the differences have hkl indexes in common (all h00 peaks) and that will give you indication of preferred orientation. If it seems that it is not the case, then probably there will be interstitial and/or occupancies factors. That all will depend on your sample preparation and measurement device. good luck!

It should be because of prefered orientation. What is the texture and the phase of your sample? Is it big crystals or small powder like? what is the instrumentation?

Answer given by Rodrigo Castillo coversall most all the question possibilities, but still there is point which is the XRD method used, if the reference data is from powder which is mostly from that, then the comparison gives the same diffracted angle as well as relative intensity, that’s off course if the diffracted angles covers to a highest value or approaches , if your data is from single crystal diffraction, then you may miss the actual higher intensity value from diffraction and then the calculation system. For this case the slandered reference will work only for diffracted angle values

It depends also if all the reflections intensities are different from those of the reference pattern or only those of a particular hkl family. In this case  a preferred orientation is probably present. Otherwise many other factors could be the responsible: a low amount of sample in the sample holder, a sample loosely packed...It is important that the intensity ratios are correct and not their absolute values. 

As other colleagues have said mismatch in the intensity scale may be due to different factors (sample preparation, preferred orientation, fading tube) that can act individually or together. Better to use internal reference material and compare.

Nilakantha! Good basic question. Good answers! You may increase exposure of your question to a larger cross section of RG experts by increasing the number of "Topics" up top to as many as 15.

Summarizing all the possible reasons for variations in peak intensities of XRD powder patterns in comparison to those in reference files:

1. Grain size, particle size, crystallite size - Coherently Diffracting Domain (CDD) Size

2. Sample preparation and loading methods.

3. Data normalization method. For flat samples, one may use the total integrated intensity of the entire profile to normalize. Assuming constant material density and therefore constant diffracting volume per sample surface VOXEL. This would not be the same in the transmission mode such as in the Debye-Scherrer method with a capillary tube.

4. Peak intensity versus integrated intensity.

5. Stress/strain state within the CDD and that due to the surrounding CDD's.

6. Sample shape and alignment w.r.t. the incident beam. Flat, capillary, film, etc. 

7. Geometry and XRD optics used including incident beam conditioning and slits used.

8. Incident beam shape and size. "point source" versus "line source"

9. Presence of preferred orientation. Inevitable in most practical cases.

10. Use of internal/external standards would help ease the challenge and more????

Read a book on X-ray diffraction

Answers of Ravi Ananth and Daniel Hoeche (especially) may help you a lot. But even in the best books about XRD people usually forgot to mention that most PDF files contain a lot of errors (or at least mismatch from theoretical values) in both peak positions and intensity ratios. Or the data given is for special conditions (pressure, temperature, etc.). Check you instrument with your calibration standard and check the data in the PDF file you are using (there always is some description and references). This is usually a pretty good start before opening new frontiers in research... And don't trust unconditionally any of the databases.

As mentioned by many, preferred orientation may be the main reason for intensity mismatch. One should also consider instrument settings and sample properties/preparation. If the instrument settings/sample properties with which the reference PDF is created differ from your parameters this type of disagreement can occur. It is always better to compare normalised peak intensities.  

The answers provided, especially by Ravi, had summarized the possibilities nicely. In my analysis, I do find some differences in peak intensities as compared to the standard pattern, due to one or more of the factors presented by Ravi. What bothers me in fitting the pattern is, however, intensity misfit around the tail of the peak more than misfit at the peak position. Also, asymmetry of the peak is a problem that could be present (even after deconvolution) which could lead to misfit at one tail of the peak.

Sami! " What bothers me in fitting the pattern is, however, intensity misfit around the tail of the peak more than misfit at the peak position".

I had the same challenge, until I decided to measure and quantify relative "misfit around the tail", peak shape and asymmetry :-) Here are some examples for several materials. We utilized identical XRD instrumentation for all samples and nearly identical XRD optics to evaluate precision. We'd appreciate any feedback from the many experienced and erudite in XRD among this audience. My interest is in understanding the Nano structural origins of the deviation from IDEAL Bragg condition. Especially, around the "shoulders" & "tails" of the Bragg profiles. Here are some terms I intend to use and have been using while discussing Bragg profiles:

• Canopy - Area above the FWHM
• Shoulder - Below FWHM before the "tail".
• Tail - nebulous region :-) Around 12.5% of maximum?
• Background - diffused scattering region? below 2% of maximum?

Any constructive suggestions to improve the definitions would be much appreciated. So if there are any objectionistas among us, please articulate your objections so that we may amend any errors or obvious oversights. Please don't object because you have no clue.  This is the leading edge of XRD. Surprise us all and tell us you have done this before and share your data/analyses :-)

YouTube Video with narration (monotonous:-) and links to real time raw 2D video data from BNL and AFRL:

https://www.youtube.com/watch?v=VSQ-eh-RDx8&list=PLmpho52CoOEiFLvpb3vxXH-jfAScTgN1H

Literature references for this sort of work from the present would be deeply appreciated. I'd guess only those with access to a synchrotron  are probably dabbling with this sort of experimental information and not those relegated to a

In most powder diffraction analyses that I see, the shoulder and tail fits are, to put it mildly, horrendous. One of the causes, I believe is, a universal peak shape assumption for the entire profile when fitting. We know from observations that even in monocrystalline material systems, the Bragg profile shape for any (hkl) may vary depending on the stress/strain state at the interrogated VOXEL. Besides, the "shoulders" & "tails" are highly sensitive to the defect distribution and type. I would clearly expect the "fit" to be "off" in these regions of the Bragg profile.

Since the sampling surface area (diffracting volume) may be assumed to be nearly identical for the two samples, any change in "shoulders" & "tails" would affect the peak intensity as well. I'd think, the integrated intensity of the diffractograms would be expected to be identical.

Back to the topic! So my suggestion to Nilakantha is, to normalize the two diffractograms w.r.t. the integrated intensity below each individual profile (complete) and then compare relative intensities. If you send me the original XRD data, from the archaic 0D "point counter" utilising the Excel format, I'll be delighted to show you how I'd accomplish it! :-)

It would also be helpful if you could display a "standard" powder profile for the same instrument with identical optics to estimate the "instrumental factors". I'm sure it is a bit much to ask for the size & shape characteristics of the incident beam, but that ought to be a part of your data as well. I notice, you haven't yet mentioned the beam conditioning apparatus you've uses. Please disclose for the best feedback!

I'd also suggest rotating the sample about its surface normal and repeating the diffractogram just to evaluate any potential effects of "preferred orientation" on the relative intensities. Just sample prep and loading may introduce some element of the unavoidable ubiquitous "preferential orientation". Normally assumed to be non-existent. Of course, one would be clueless without a 2D XRD detector :-)

 Доверяй, но проверяй, Doveryay, no proveryay, Trust but verify! Ronaldo Reaganov :-)

"Read a book on X-ray diffraction" - Excellent & secure grasp of the obvious!

I'd challenge any of you erudite to post a reference among many, that discusses the specifics as we have in this discussion. I don't mean this rhetorically but I'm genuinely interested in such a publication. Post it!

I was able to summarize my recommendations based on real time XRD imaging experience with powders and monocrystals spanning nearly 3 decades. I'm yet to find these in any book.

My book recommendations would include:

1. Compton & Alison

2. Andre Guinier

A static displacements of intrinsic atoms can be the main reason of the effect obtained. 

Background of that you can find in books.

What is a "static displacement" and what are "intrinsic" atoms?

There are two types of atoms' displacements in any crystal lattice: statical and dynamical.  They coexist. In distorted crystal lattice.

Intrinsic atoms are own atoms of crystal lattice.

Statical displacement are  independent on temperature vibrations. It means that atoms are really displaced from their normal positions in crystal lattice.

Dynamical displacements  mean atom vibrations (their normal modes).

with regards

Thank You all researchers for valuable suggestions and advices.

With regards.