I have scanned a sample with X-rays for quantitative analysisof phases present. In the X-Ray diffraction plot (attached) peaks of phase with fcc and with bcc are easily identified.
But, a peak is still unidentified as shown in the attachment with question marks. The unknown phase should be carbide which is confirmed by optical micrography and SEM studies. To get the exact volume fraction of phases in the sample, lattice parameter of carbide is needed. Since no sister peak of unknown phase (that may be carbide) is obtained no matter how long the angular range of the scanning is employed, we are not able to compute the stiochiometry or lattice parameters of the inidentified phase. How can we move ahead? Can any other technique help us in identifying the unknown phase?
If carbides are visible in optics so size of it's grains is not very small. So it seems possible to obtaine electron diffraction from these inclusions by TEM or HREM, if you have these equipment. Another way consist in some kind of separation, for example magnetic, of carbide phase.
"No sister peak of unknown phase is obtained no matter how long the angular range is employed" is unclear - does it mean long exposure time, thus good statistics, or high angle range, or low angle range (your figure starts at 60°). The low-angle range is especially important for phase identification.
In any case, as you have an idea of the possible phases causing the peak, you can limit your search to the chemically possible metal elements + carbon. Either take a powder diffraction database (pdf) or calculate the pattern from the structures in a crystal structure database. Compare to the observed peak and select the phase that produces a peak most close. Then adjust lattice parameter. There is software to do this automatically, of course; it is probably installed at the diffractometer you took the data on. Of course, with only one peak the identification can never be 100% sure.
"Can any other technique help us in identifying the unknown phase?"
Absolutely. You should always consider other techniques. Microscopy (optical and electron), and EDS (or XRF) would be an obvious next step.
I haven't had time to check, and you really have not given enough detail to do it right, but is there any possibility that the peak at 62°2theta is due to contamination of your incident spectrum (either a W L-line reflection or a source K-beta line)? These sort of artifacts are common for cases when you have a very strong diffraction peak (the one at 68°2theta) present.
Dear Daniel, thankyou for your response. By "long angular range" I mean to say that i have tried 2theta values ranging from 40 degree to 130 degree but no sister peak was observed! In that case it is getting difficult to index this single peak. Moreover i have used Xpert high score there I am getting vague idea but i can't fully rely on that. As per your suggestion let me try with crystal structure data base. Thanks for the suggestion. Can you please tell me the software you were talking about?
Dear Edward, Thanks for your response. Sorry for not provinding the details. Relevant details about plot are as follows.
Target: chromium
current: 30mA
Voltage: 40KV
machine used: Bruker diffractometer.
There is little possibility of k-beta line as we have used vanadium filter which can easily chop off these lines.
It is an interesting coincidence that your "impurity phase" line is exactly in the right spot to be a K-beta version of your strongest Bragg peak. I just ran a quick calculation for Cr Ka and Cr Kb and it is an awfully good match for your situation.
beta filters work by preferentially attenuating the beta compared to the alpha, but for a very strong Bragg peak there is usually a small amount of beta visible in the pattern. Again, a beta filter does not eliminate beta lines from very strong Bragg peaks.
This could be good news for you, as it may mean that you do not have an unknown impurity phase present at all. Just the expected BCC and FCC phases.
FYI - if you would like to see this effect directly, put a single crystal plate of silicon on your instrument and scan from low angle to just below the first expected Bragg peak (which will depend on the crystal orientation). Note: don't scan the Bragg peak as the very strong intensity at that 2theta might damage your detector.
You will see lots of features in the diffraction data, corresponding to W L-lines, and Cr Kb, and the absorption edge from your V filter. You can usually ignore these artifacts in powder XRD scans, but not if your Bragg peaks are really strong.
@Amin - I note that the paper you have linked misspells Scherrer as "Scherer" throughout the manuscript. Also the conclusions regarding lattice parameters are made without providing an estimate of the errors, and are therefore not terribly convincing. The Scherrer crystallite size error estimates in Table 3 (e.g., ±0.2nm) are absurdly small., though the W-H plots and error estimates are better. And the conclusion begins: "Well crystallite of AA 6061(100 − x) − x wt.% Al2O3 (x = 0.4, 8, and 12 wt. %) nanocomposite powders was fruitfully fabricated by MA," which is not very clear.
But never mind all that, I simply don't understand why you think that specific paper provides any insight to Abhinav Varshney's question "How do we find lattice parameters with x-ray diffraction", or how it relates to identifying a possible second phase in his pattern. Can you elaborate?
Dear Edward, Thanks for such an overwhelming response and suggestions. I will definitly try with single crystal silicon.
If carbides are visible in optics so size of it's grains is not very small. So it seems possible to obtaine electron diffraction from these inclusions by TEM or HREM, if you have these equipment. Another way consist in some kind of separation, for example magnetic, of carbide phase.
Can you send all information about your sample?
The phases that you have identified and how do you prepare the sample?
I was just thinking of the extracting the carbides by acid dissolution.
People regularly does this.
Abhinav! Here are some missing pieces to this puzzle and my observations.
1. "Point" source or "line" source? It is most likely "line". How about dimensions? Not critical, just for general knowledge and survey.
2. Please display the entire diffractogram from 5 degrees to 130 degrees. Send me the data as Excel spread sheet with relative intensity alone in a column if possible.
3. Describe the sample and sample prep better. What is the chemistry?
4. Can you make a thin section of your sample and conduct XRD in the transmission mode? Here is an example of transmission Laue XRD with a composite sample and the possibilities:
https://www.flickr.com/photos/85210325@N04/7988731098/in/set-72157632728981912
5. Do you have photographic film? I presume you do not have any modern real time XRD imaging tools yet. You could digitize film with a modern ubiquitous scanner.
6. Do you have access to Guinier?
7. If there were K alpha 1 or 2 (Cu) or K beta or other components in the incident beam then they would be repeated at each reciprocal spot. Using the K beta wavelength compute the expected 2-theta to verify if the suspect peak coincides. At higher angles they would resolve better. Here is an example using Cu radiation without a Ni K beta filter:
i. https://www.flickr.com/photos/85210325@N04/13109812555/
ii. https://www.flickr.com/photos/85210325@N04/13125116403/
iii. https://www.flickr.com/photos/85210325@N04/13166377625/
8. If you are operating "spatially blind" with a conventional zero dimensional point or scintillation counter, then you are going to need the help of other methods such as suggested by others to fully characterize your sample Nano structure.
9. What about a "standard" diffractogram to calibrate your instrument? Please post.
10. Have you eliminated any effects due to potential "preferred orientation" by repeating your diffractogram after rotating the sample about its surface normal? Example of preferred orientation in aluminum foil and how it could be missed in a conventional equatorial scan diffractogram:
https://www.flickr.com/photos/85210325@N04/7944785794/in/set-72157632728981912
11. Cr wavelength at 2+ Angstroms or >0.2nm would reduce the reciprocal space observable compared with say Cu at about 0.154nm. Any reason you prefer the Cr wavelength? The penetration depth would be shallower as well. Have you estimated the sampling depth or extinction depth in your sample material yet?
12. If the suspect peak were a "carbide" peak then you would also see other reflections from that phase. It is highly unlikely that only one of the carbide peaks would be observable in a conventional diffractogram.
https://www.flickr.com/photos/85210325@N04/10515372183/
Abhinav - here is an easy experiment to try that does not require anything new: remove the vanadium filter and scan again. You should get the same pattern, plus the Cr K-beta peaks. If I am correct, you will see a very large K-beta peak at the same location where you now see your weak "carbide" peak. I believe that what you are seeing is just the Cr K-beta peak that has been attenuated by the filter (but not completely eliminated).
As Edward astutely points out, remove the V filter and try. Or, use the K beta wavelength and compute expected peak position. If the K beta is coming through then it must appear similarly and with approptiate relative intensity besides all other high intensity reflections as well. In majority of our work we use a hybrid Bartel (220) beam conditioner to eliminate the equivocation.
XRD is a precise science. There should be no room for so much ambiguity in the analyses. However, the use of the antiquated "spatially blind" linear diffractogram is at the root of the challenge. You fellows are working with tools that are older than the Braggs. I mean the use of the point counter in 2014. That dinosaur ought to be "moth balled". It amazes me that premier institutions like the IIT's are not equipped with the latest 2-dimensional area detectors. Many of the 2D detector users today are still reducing the wonderful spatially resolved XRD pattern back into a conventional linear "smudge" to interpret. Many have no clue how to handle the 2D XRD data yet. Crying shame!
If I need a microscope to verify what I observe with XRD, then I have a lot to learn. Throw the point counter out. You'd be better off going to the extinct photographic film and digitizing it with a scanner. I see a lot of pontification with the "spatially blind" 0D detector diffractograms both in RG and LI. The conclusions and interpretations are so nebulous that it all seems futile. At best, it is worth re-verifying with optical microscope, TEM, SEM etc. where the sample prep is a night mare. That totally defeats the purpose of XRD. Study and learn Guinier besides the basic XRD books to understand what happens to the reciprocal space image if multiple high intensity components are present in the incident beam spectrum of wavelengths. Get the fundamentals right!
If one were to observe the entire 2D XRD pattern, the presence of K Alpha & K Beta components would be instantaneously obvious as duplicate rings or spots as would effects of any “preferred orientation”. In fact the duplicate images (K alpha & K beta) may be used to calibrate your instrument precisely. Better than any external or internal standard. Then you wouldn’t need to ask RG experts to explain one peak on a diffractogram. A diffractogram from a known standard would alleviate such quandary as well. So my point was that this experimental data is lacking in many respects in order to come up with an unequivocal interpretation. Work on the basics!
Abhinav! My idea here is not to answer your question because you are the one with the best vantage point (data wise) to get to the correct answer to your question. My idea is to use the fundamentals of XRD and materials characterization to get us to that right answer.
Some may call this shameless "self promotion". But I wish someone would have boldly promoted it to me in 1979. I had no clue until 1985. Never looked back since. I've used the 0D, 1D and 2D methods since then and know & understand the differences clearly now.
Here is another reason to throw the antique 0D detector out, "don't scan the Bragg peak as the very strong intensity at that 2theta might damage your detector". How then would a XRD experimentalist measure incident beam characteristics, (Io)? Have you any clue what your incident beam looks like? Use film and find out.
P.S: I notice that the image below is missing dimensions. The largest beam is about 2mm X 20mm with 76um pixel size & 8 Bit DR.
https://www.flickr.com/photos/85210325@N04/9529858827/
"How then would a XRD experimentalist measure incident beam characteristics, (Io)?"
A rather obvious solution is to simply attenuate the beam in the region where it is too intense. Does not require a 2D detector or film, just some common sense.
Sure Edward! That is what most folk still using point counters are employing, just plain old “common sense”. To switch to the more modern 2D detectors would require "uncommon sense" indeed.
BTW I've never ever seen any of you wizards share the details of the "Io" for each conventional linear diffractogram. Would that be too much for the 0D detector? Most seem to subscribe to the view that it is pointless, characterizing the incident beam. Out of sight, out of mind! The argument for not using 2D yet in 2014 is still just "common sense". I've seen similar justification since the 1980's Edward! It is embarrassing to see other folks pontificate about XRD rocking curve analyses with the ubiquitous 0D detector still. This is a super sensitive method that is rendered analytically ambiguous by integrating the intensity data over a ridiculously large sample area, generally. Then they are touting the phenomenal "dynamic range" of the conventional 0D data which is typically integrated over a “huge” area 0.5mm X 10mm (effective "pixel size" of 5mm^2) receiving slit dimension. This would be equivalent to about 7H X 132V (10^3) 2D detector pixels at about 76um per pixel.
Doesn't it sound ridiculous to have to guess if the K beta is still present in the diffractogram or not? It seems quite obvious that such inadequacies stem from the blindness created by the use of the century old 0D scintillation detector in the XRD lab. Besides, one would go "old & grey" trying to generate 2D XRD data using the conventional 0D detector. Kind of what is done for pole figure measurements in modern times using the 0D detector. The repeated silly questions regarding these linear conventional diffractograms on RG & LI along with the ambiguity they create for analyses is quite obvious. It is symptomatic of the "spatial blindness" introduced by the 0D point counter. Majority of the folks using the 0D point counter method have never even bothered to characterize their incident beams. In this case for example, the wavelength composition of the incident beam is still an enigma. Besides, I'd bet these XRD "specialists" have never visualized the 2D XRD pattern or their incident beam yet due to the extinction of the photographic film from the XRD lab in modern times. The lack of "known standard" diffractogram with each experiment is another side-effect of the "spatial blindness". These are clearly due to a lack of the ability to "visualize" XRD phenomena. Once you begin to see and visualize XRD, the paradigm is quite different indeed.
One may always turn the 2D detector into a dumb 0D detector, as many do today, any time by choosing the appropriate integration window any way. The integrating window size may be changed for the analyses “on the fly” since the 2D XRD video data is stored in 3D to examine “on demand”. Why on earth would one want to suffer the pain of “spatial blindness” in a modern XRD Lab? There is really no justification for continuing to use the 0D detector in 2014 other than lack of funds or IQ or both. Except to calibrate the XRD system from time to time.
https://www.flickr.com/photos/85210325@N04/11343736043/
Incidentally, if any detector has the "risk of damage" due to a secondary diffracted beam from a single crystal on a sealed tube XRD system, it ought to be "moth balled". Haven't they figured out "safety trip switches" yet? That must be the XRD equipment manufacturer’s pabulum manifesting. Let's get "back to the future"!
0D XRD detectors aren't cheap either. I'd say they'd run you $5-$10k with all the electronics and software support. Am I out of touch?
My favorite Guinier quote:
X-ray Diffraction In Crystals, Imperfect Crystals and Amorphous Bodies by Andre Guinier. 1962 - Great book! Worth acquiring for your XRD collection. http://www.flickr.com/photos/85210325@N04/10515372183/
(pp. 162-163) 6.3.5. "Comparison between the Photographic and Counter Methods
It is obvious that the Geiger Counter and the ionization chamber have the advantage of precision and sensitivity. They alone permit a quantitative study of scattering. However, point by point measurements in reciprocal space are lengthy, while the photographic method gives directly a picture of the scattering in a single experiment over a large surface of reciprocal space. The resolving power of the film is excellent. Thus, sharply defined scattering regions, for example on rows or on planes of reciprocal lattice, show up very clearly on photographic film but are difficult to detect with a counter.
Let us note also that the counter is "blind". In many cases inaccuracies in the setup produce stray scattering which is easy to detect on the photographic film but which can spoil a series of experiments made with a counter.
The two methods are therefore complimentary and it would not be advisable to neglect one in favor of the other. The photographic method is admirably suited to the qualitative exploration of an unknown pattern, since one can then find totally unexpected phenomena. The counter is necessary for quantitative measurements on a pattern which is already known qualitatively." (end quote Guinier)
This was printed in 1962. This has to be considered in light of the present day 2D detectors that combine the benefits of the counter & the film in one. These writings indicate some causes for the present day inertia of moving from the 0D to the modern 2D detectors. The spatial "blindness" of the 0D scintillation counter/detector is the principal reason even experienced experts using XRD are ambivalent about the interpretations of 0D XRD data/results.
In your case where you can't confirm the exact phase of your sample from XRD, I think of Raman or FTIR that can help you out in the identification of the phase in your sample. Good luck!
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Annealing
- Metallurgy