see also; http://homepage.ntu.edu.tw/~kcyuan/form/For_Rigaku-Upload_powder.pdf

Dr. Sekar,

If you want to know several parameters from XRD analysis than the size of the domain particle must be less than 100 nm.

Hi, With larger particle size, all the signals are observed and it is easy to refine them accurately. If you have small particle sized materials, offcourse you can make the XRD patterns but you will see very broad peaks. Sometimes your two small peaks will be merged into a single peak. Still you shoudl be able to get good information if you are able to refine the patterns correctly. I have been working with particles sizes of 5-10 nm for almost half a decade, I have had no problems with small particle size. The most important thing is give very slow stem scan if your particles are small. You will succeed.

The in situ Nano structure of the rock sample contains an incredible wealth of information about the ambient conditions when the rock solidified and its thermo-mechanical history since. Why would you want of "crush", "grind" or "mill" this information into an unrecognizable "mush" :-)

Instead, go real time and 2D XRD imaging!

If all that does not matter, then "ball mill" it till the Debye-Scherrer arc/ring becomes continuous. My guess would be 1-40um size would work. See examples attached illustrating effect of powder size on the continuity of the Debye-Scherrer arc/ring. https://www.flickr.com/photos/85210325@N04/7920887956/in/album-72157632728981912/ Very small size (10's of nm) would be influencing the FWHM. Beware of the effect of sample prep, preferred orientation and process induced Nano structure (size & strain).

When you use the term "particle size", define it. The coherent diffracting domain (CDD) size  is determinable using XRD technique besides the in situ residual stress/strain.

https://www.flickr.com/photos/85210325@N04/7920887956/in/album-72157632728981912/

I believe most of the comments that mentioned are true. XRD test can show you a good results from micro to nano particle sizes, but as it was mentioned and considering Scherrer equation, The peak will get broaden as the size of the crystallites reduce. Besides that you should consider other parameters that can broaden your peaks to be able to have a good estimation of your average crystallites sizes. It will be easier for us to help if we know what are you looking for in XRD results

pulverizing to -200 mesh is often suitable for mineral characterization.

Dear all, thank you for your valuable comments. I am not working deep into nano as well as mineralogy. I need to characterise the nature of an unknown sample using XRD to get some preliminary idea as said earlier by Dr. Kenneth M Towe. Though I find some idea, I could not conclude to go with analysis.

Dear all, thank you for your valuable comments. I am not working deep into nano as well as mineralogy. I need to characterise the nature of an unknown sample using XRD to get some preliminary idea as said earlier by Dr. Kenneth M Towe. Though I find some idea, I could not conclude to go with analysis.

I suggest reading a basic book on powder XRD to get a better handle on the subject. The point behind all this "pontification" is that the XRD signal contains tremendous amount of Nano structural information. It all depends on your objective in terms of depth of analysis. The Debye-Scherrer photographic film method or the conventional diffractometer would work in your case as long as you are consistent (and random) with the powder size (large, small, whatever). This would give you some semblance of precision.

The best course of action in your case would be to just acquire the XRD data for several samples and orientations, then figure out what's up :-) Post the XRD data when possible for best feedback. Use an internal or external "known standard" to contrast the results with and calibrate the instrument. Reduce the potential contributions to the XRD signal from factors enumerated earlier by many.

I'd also suggest you include relevant topics up top like XRD, PXRD, etc. See attached discussion for additional relevant topics. You may include up to 15 such topics. This will help boost the circulation and participation rate from the "experts" on RG and yield superior feedback :-)

https://www.researchgate.net/post/Would_anyone_have_a_clue_as_to_the_origin_of_unexpected_spots_in_the_real_time_2D_XRR_signal_from_thin_film_samples_on_substrates_Yuneda_Spots

The ideal average particle size of the XRD is 10~50um (P.302, Fundamentals of Powder Diffraction and Structural Characterization of Materials). If you are going to do the Rietveld analysis, it is better to decrease the particle size further to 1~5um (Snellings, R., et al. (2014). "The existence of amorphous phase in Portland cements: Physical factors affecting Rietveld quantitative phase analysis." Cement And Concrete Research 59(0): 139-146.).

Good luck.

Minimum crystallite size as can be seen by XRD seems to be above or around 1 nm. (This is the case for Ni. But I am not sure that this is applicable to all the compounds/ materials.) This (1 nm size particle sample powder) will provide a very broad hump only. Below or smaller than 1 nm most probably, it would give a noisy background without any hump. Therefore, it appears that particles size above 1 nm determined from XRD by Scherrer formula will show XRD peaks: peaks shall be sharper bigger the particle size. (Scherrer size is generally smaller than those determined from AFM, STM or TEM, since latter show everything together such as not so periodic surface regions, ligands, etc which in turn most probably cannot be detected by XRD: XRD perhaps sees only the central cores of particles wherein the atoms are PERIODIC, not the others.)

I am agree with Dr. Li.  Usually micron particle of this range can produce minimum size broadening.

I am agree with Dr. Li. 1~5um is better.  the particle size distribution must be uniform by filtering. the annealing is  necessary to  get rid of the mechanical stress.

micrometers or below. based on the crystallinity of your rock samples

Dear Dr. Sekar

I'm not an expert on qualitative x-ray diffraction analysis. But I think that -200 mesh (-74 micronmeter) is suitable for a qualitative XRD analysis, at least. Also, I heard from my friends in materials science that XRD peaks of a material can not be clearly distinguished  if it is %2 by wt. in the whole sample. 

Let us try to make this a pleasant learning experience for all of us :-)

Interesting to note the following recommendations:

• Rinkesh thinks "the size of the domain particle must be less than 100 nm".
• Mahmut thinks "200 mesh (-74 micronmeter) is suitable" and "friends in materials science" know better :-). I hope we are included among them :-)
• Xuerun Li suggests "10~50um" & "1~5um".
• Yanmin thinks "1~5um is better".
• Biswajit agrees with 1-50um (I think?)
• Ning thinks "micrometers or below".
• Gunadhor thinks "Minimum crystallite size as can be seen by XRD seems to be above or around 1 nm."
• Suman has been "working with particles sizes of 5-10 nm for almost half a decade".
• Forget about what Ken or Ravi thinks just for the time being :-) 
• What should Pragadeesh in Karnataka think of all this information?

Let's consider the paradigm here for a moment. As far as I can observe that would be more than 2 to 3 centuries of combined experience in XRD and its practical application among the participants in this active discussion. Would you agree? Then why such dramatic ambivalence? Anyone want to volunteer to explain?

My initial guess would be the ubiquitous "spatial blindness" factor of conventional diffractometry. What do you think? :-)

Great thanks to Mr. Ravi for making these conversations under scrutiny, giving everyone an opportunity to learn the correct perspective towards crystallographic studies!!! Too much of information lead to confusion. I look forward for the expert comments for Ravi's points, to make this discussion a more productive one.

Mr. Ananth, the conflict is due to fact that we are working in different fields. However, I am working on mineral processing and my prediction is suitable for Dr.Sekar. Reducing the size to nanometer may results in mechanial activationn which could cause the change in peak pattern. 

Dear Kenneth,

Let me know if there is anyone that can reduce the dimensions of a rock sample from cm to nm scale, and also match the peak pattern of that so-called ultra-pulverized rock to the previously-determined crystallographic data  of minerals. Then, I will accept mineral characterization by XRD can be done with any method as long as I am consistent with particle size. Otherwise, it is better to make some realistic suggestions to Dr.Sekar.

The point in contention,"to make some realistic suggestions to Dr.Sekar", is a technical one not an emotional one. The size of the grain/particle I must presume to be much larger than the CDD (coherent diffracting domain) size in powders. Would that be reasonable Mahmut?

The 75um size you are recommending will lead to large spots (diffraction topographs) on the Debye-Scherrer arc. This will lead to inconsistency in the powder diffractograms using conventional diffractometry, is my opinion based on my experience. Besides, large particle sizes will also invariably lead to some amount of "preferred orientation" just by the process of manual loading of the powder. See example of Si powder standard from a Bruker B-B arrangement attached showing spottiness and "preferred orientation"https://www.flickr.com/photos/85210325@N04/15210921656/in/album-72157647414035470/ . If you positioned your detector to any of your powder peaks and then hung a dental film in front of the detector for exposure, you'd see a similar image with 75um powder. Please examine images attached of several sizes below as well. These images are at a tiny SDD (sample to detector distance,

Ravi, I was only talking about that size was suitable for powder diffraction, I did not know whether such powders are suitable for Debye-Scherrer method since I don't know this method.. Still, I guess that sizes as 10 micron 1 nm size alter crystal structure of mineral phases in rock, even in Debye-Scherrer. The minerals if not gold or rare-earths, have the grain sizes in the range of 100 micronmeter or so.

I understand. The powder diffraction technique is just an extension of the Debye-Scherrer method but quantitative without the need for "photo-densitometers". Ever heard of this? :-) Ancient stuff!

In the Debye-Scherrer method a photographic film is generally used to capture the Ewald surface interaction with the diffraction vectors simultaneously over the entire equatorial plane. Sort of like with a curved linear position sensitive detector (Inel). The specimen is generally rotated about the diffractometer axis to catch several topographs and then the result is some sort of a convolution (integration) known as the Debye-Scherrer rings/arcs. These have to be carefully identified, analyzed and deconvoluted. Many conventional methods exist today for this sort of deconvolution based on the optics and geometry chosen.

There are several choices  based on comments in this discussion. It would be nice to understand the "driving forces" for these choices. You obviously have had success with your choice for your specific case. Have you conducted any systematic studies regarding the choice of other sizes or distributions?

No, because when you go to the laboratory of an academic center, they provide you powder diffractometer since it is more practical than Debye-Scherrer. Just you do is to obtain diffractogram and match the 2Q peaks with some certain minerals.

I understand that you are quite a way dealing with the science of diffraction but sometimes engineers require fast and accurate solutions. In Dr.Sekar's case, he just wanted to see which minerals are present in his rock sample. So, why does he require some specific and possibly tedious techniques ?

In short, scientists generate methods and engineers get results. :) 

From latest answers, I can see that reducing the particles in microns to nanometers could lead to phase change. From my side, I like to make my points very clear. I am going to use the particles in microns range only and nm level particles will not be used as I need to fluidise the particles in a setup. 

So, I wish to know whether I need to use particular (consistent) sieve range or I can use any of the particles ranging sizes from 100 um to 420 um (which I have). What I am looking the XRD results are the elements present in the samples and their phase as a rough estimate qualitatively but semi-quantitatively. The word I meant "good XRD results" are distinct results with clearly observable peaks and their analysis for understanding.

Please bear with me if I have made any error.

For particle size above 100 micrometers, XRD peaks shall be generally sharp. No need to worry about nanosize or hump. However, the peaks for the samples you make have to match with the original/ parent compound without any extra peak. If there are extra peaks, you have to find out that they are related to your compound or not. Otherwise, it should be due to impurities arisen from the unreacted part of the initial compounds. Then, impurity content is quite high. Otherwise, within the capability of XRD (say,

Mahmut! You are totally off the mark in the empirical use of scientific data. You certainly don't realize that "real time" and 2D are less tedious than "blind" and "groping in reciprocal space". I suggest you read some of the classics like Andre Guinier's book that I posted earlier and look over some of the original works of the Braggs, Laue, Ewald, Weissmann and others. This is known since the turn of the century but ignored by folks in favor of expediency.

Engineering may be the practical extension of science! But arbitrary use of XRD is not advisable in Engineering or Science. I strongly recommend against the "black box" (voodoo, magic) approach to XRD analyses.

Now if this was as simple and pragmatic as you seem to indicate, then why not quickly try all of the potential permutations and combinations queried and share the experimental results to analyze and determine the correct answer. I guess because the conventional methods being used are " some specific and possibly tedious techniques"?

So here is a suggestion for  Pragadeesh who "wish to know whether I need to use particular (consistent) sieve range or I can use any of the particles ranging sizes from 100 um to 420 um (which I have)". The answer obviously from this cacophony is, "ALL OF THE ABOVE". Try all permutations and check for consistency. Don't look for an answer that you like or that is only expedient :-) Burn & learn on your own! Pain from others' experience cannot be felt!

Isn't is flabbergasting to have this fundamental difference of opinion after nearly a century of conventional diffractometery? Any of you geniuses have a  good explanation other than my diagnosis of "temporary blindness"? It is also astounding that the Debye-Scherrer technique would be so ill understood as a "specific and possibly tedious technique". It would be futile to even be using conventional X-ray diffractometery if the connection and relevance between the two is not comprehended :-)

"reducing the particles in microns to nanometers could lead to phase change", I missed that one somewhere!

To change phase composition by mechanical deformation is not normal unless the extreme strain is responsible for such phase alterations. Some references to such behavior in mineral powders as a result of XRD sample prep processes would be educative :-) I'm aware of the creation of defects by the process of "pulverizing" or "grinding". This will only modify the CDD (coherent diffracting domain) size, in my opinion. I'd like to know if different.

It should also be emphasised that the reason for the "hump" in Nano size "particles" is because now you have many more smaller particles nearly satisfying the Bragg condition populating the Debye-Scherrer arc. Do remeber the effect of the incident beam size & characteristics as well. In fact, the reason behind rotating the sample in the Debye-Scherrer method (100+ years old) is precisely to overcome the challenge of "grain"/"particle"/"agglomerate" size and bring particles into Bragg condition. Conventional diffractometry using a 0D point counter does not afford this luxury, as far as I know. Instead one needs to resort to Bragg-Brentano type geometry (and other methods) to include additional signal.

Ken and I seem like "harsh teachers", don't we? In fact, I've never seen Ken as excited as in this discussion :-) Glad it is not theological! Go Ken!

You fellows look like youngsters and should be a lot more open-minded than some of the dunderheads (my contemporaries and older) that I've run into on RG and other public comment sites (including some venerated, premier R&D institutions :-). The "blindfolded" approach is defended by so many XRD users that it is pathetic. It is like trying to reach our moon on a hot air balloon :-)

If learning is not an objective, then let's have some fun at least!

 "If there are extra peaks, you have to find out that they are related to your compound or not. Otherwise, it should be due to impurities arisen from the unreacted part of the initial compounds"

In a mineralogy sample the only "extra peaks" would be from impurities accumulated during the XRD sample prep process.

"To know better about better or good quality nature of the sample it is made, it will require other experimental techniques", only if you are married to the conventional diffractogram without recourse :-) Otherwise, XRD may be used with similar precision and results as any microscopy tool. One may use 2D XRD imaging devices all the way from XRR (reflectivity), GIXRD, µGIXRD, SAXS (small angle), WAXS (wide angle), Diffuse X-ray Scattering to XRD RSM (reciprocal space maps) just as folks have used photographic film for over a century now.

If it was I who had such a mineralogical rock sample, I'd do the following:

Ken and Ravi,

Thank you very much for advising me the reference works you suggested. I will definitely read them. I also suggest Dr. Sekar to review these books for his upcoming mineral identification.

I feel more enthusiastic and finding good interest in learning new topics from the cited sources while seeing experts' eagerness in explaining the insights of the analysis. 

I would also like to post one of my XRD results for the experts' critics and for that, please post the file format that will be convenient for discussion.

Dear Dr.Kenneth, please suggest me a format to post a diffractogram.

LOL! Nice cartoon! It was a figure of speech Ken. Would potentially generate less "gibberish" than all this pontification here.

BTW if the incident beam is a "tiny point source" like a micro-focus source, then even a boulder would appear flat. Just got to make sure that the surface being examined coincides with the diffractometer rocking axis. Good luck achieving that without a real time 2D imaging system :-)

Ken the point you have made is an excellent diversion to the next practical application. That would be a portable diffractometer which could be mounted (using suction cups or other means) on one of the noses at Mount Rushmore for precise XRD analysis. Sort of like the Mars Rover! This is the design we would create in order to perform on-site XRD rocking curve analyses on an aircraft structural member or the supporting cables of the Golden Gate bridge or Titanium alloyed turbine blades to detect accrued fatigue damage. When you can't rock the sample (like a boulder) then  the detector and source combination may be rocked about a common vertex, yes? "Taking Mo to the Mountain rather than bringing the Mountain back to the lab". Good new in mineralogy is that samples may be chipped off the old block without significantly changing the in situ Nano structure of the sample.

Parallel beam optics with a micro-focus source would still work in the case of a non-flat sample surface, wouldn't it? Back in grad school we were able to generate "non gibberish" XRD rocking curve data from polycrystalline AL2024 cylindrical tensile test samples and predict location of fatigue failure in the "neck" region. We accomplished all that without real time XRD imaging. We were using the "tedious" film method then.

Lights out, phosphor screen and binoculars was the modus operandi for beam alignment in those days! Concern for radiation exposure was dominant then. Contrast it with directly imaging the transmitted beam along with the XRR signal today in real time without burning the detector!

https://www.youtube.com/watch?v=TabCUq0oaqQ

https://www.flickr.com/photos/85210325@N04/19248953711/in/dateposted-public/

When one is able to generate the right diffractogram with good SNR for all peaks, then there would be no need to read any more books except "User Manuals". There is sufficient "canned" software provided by equipment manufacturers to perform precise peak and constituent identification. In order to quantify constituent fractions one would need a lot of calibration standards or reference standards. This would have to be accomplished with "known mixtures" of individual constituents produced in the lab. Not trivial!

Post the diffractogram as a graph but include the data as an Excel file for others to download and analyze as well. Or, a text file with data. Not the machine specific format!

I wouldn't try argue with you Ken. No sneeze zone:-)

This video recommends 45um size for most polycrystalline materials. Not even sure what that means? Agglomerate size?

https://www.youtube.com/watch?v=lwV5WCBh9a0

BTW Pragadeesh, one of these 25mm/27um real time 2D Bragg XRD Microscopes is available for demos in Bengaluru India, besides NJ USA. It was designed, developed and tested at the Indian Institute of Science, Bengaluru! Can't be that far from you in Karnataka, India :-)

There seems to be a significant amount of XRD based publications from India, I notice. We haven't yet used the 2D imaging device for conventional powder diffraction work. But the need for such activities is clear from many of these RG discussions involving powder XRD (PXRD) :-)

I think this link can help

http://www.instanano.com/characterization/theoretical/xrd-size-calculation/