Dear Surender,

Yes, you could have a peak splitting indicating a formation of another crystal structure or due to the secondary phase,

If you want to confirm if this peak splitting is attibuted to a secondary phase you could use in situ XRD using different temperatures (as suggested by Georgiana) and this technique is called High Temperature X-Ray Diffraction (HTXRD) and associate the result with Thermal analysis (TG/DTA) to monitor the decomposition of the precursors until the formation of your ferrite. But, to be sure the presence or not of this another compound you can also analyse the precursor by IR and Raman Spectroscopies.

My co-workers and I studied one similar behaviour of a perovskite compound by HTXRD, TG/DTA and IR spectroscopy, so you can see in the paper attached below.

You can also anneal your sample at higher temperature if this overlapping disappear.

As I mentioned in previous comments, this peak splitting could be also assigned to a another crystal phase of your ferrite. So, I suggest you prepare different compositions analyse your sample by XRD, refine your samples by rietveld, You can monitor the evolution of this peak according to the composition.

Good luck!

Article SrSnO3:Nd obtained by the polymeric precursor method

Dear Surender,

Are you synthesizing a new spinel ferrite which is not published until the moment or does your compound already exist in the litterature? Did you analyse your sample by other techniques to know if it is monophasic? Because for my opinio this behaviour may be due to the present of another phase in your sample that presents a diffraction peak close to the ferrite ones, overlapping the main peak of your ferrite. That is why the profile peak present this shape.

All the best!

Thanks Andre, There are only two references till date for samples with almost similar my composition. But the temperature and method of synthesis is different. My samples are in nano range. Please check the next attachment of the same figure.

The Red dotted section completes the peak shape profile. Only one single peak is expected to be present at 62.5 degrees. The A section of the peak is almost full but B section is chopped off from the top. You may note that all peaks(except the one in question) in the spectra represents single phase cubic spinel structure.

Assuming that your point about presence of second phase is true and hence section A and B of the peak in question are assumed to represent independent phases. How do we explain the following:

If we calculate crystallite size using Scherrer formula (D=0.94 lambda/beta cos theta) . All Other peaks (except one at 62.5) will give crystallite size half as that of the size given by peak in question. the crystallite size is about 20 nm as calculated from all other peaks. However it will be about 40nm if calculated from section A(or section B) of the peak. Which is not possible. So I believe that sections A and B represents two parts of the same peak. Comments, suggestions and explanations are welcome.

Dear Surender,

Because of the XRD profile I have realized that your sample is nanoparticle (broad diffraction peaks), but it doesn't justify the assymetrical appearance of the peak at 62.5º in the XRD pattern.

As already I told you I think this behaviour is due to the presence of impurity or secondary phase, due to a possible coexistence of two different crystal structures of your sample or even more due to a possible crystallization of a new and/or different phase.

I would like to tell you that even it existes in the litterature samples with similar composition of yours, it doesn't mean that they will have the same crystal strucutre, do you agree? To help you understand, i give you a simple example: TiO2 can crystallize in three different kinds of structure (brookite, anatase and rutile) and depending on the type of synthesis method, temperature of synthesis, temperature of annealing or other parameters, you can have TiO2 with one kind of structure or with the coexistence of two of them or even more both of the three. Moreover, you can also observe a begining of a phase transition of the material. I may be wrong, but it can be happening with your sample. So, I would suggest you to use another characterization technique to confirm the presence of a monophase or multiple phases of your sample.

Have you ever tried Rietveld refinament to confirm if your sample is monophase and the kind of crystal structure it can present? If not, do it and i am sure that you will have success on that.

I hope it helps, if not I would like to know by other colleagues why this peak presented this profile.

All the best!

I see that a deep discussion is going on here. If you kindly let me to give my opinion just looking at the peak, it seems like there is convolution of 2 peaks between 62.75 to 64 degrees besides the intense peak at about 62 degrees. Could you also add an image of a reference spectrum from a similar study? And, is this the first time you did this characterization or did you observe this a few times?

Dear Andre and Dilek, Thanks for the response. A file is attached. Rietveld analysis of pattern in question was carried out using GSAS. Stopped at chi square = 2.2 . but the trend indicate that the peak at 62.5 degrees represents single set of parallel planes corresponding to (440) peak . Cation redistribution will improve the intensity of (440) and other peaks. Not a single peak representing second phase appears in the entire diffraction pattern obtained in the 2theta range 10 to 100 degrees. For convinience, Second file contains diffraction pattern of a sample representing single phase spinel ferrite as well as some peaks representing 2nd phase. So there are plenty of additional peaks in sample containing 2nd phase. Another image of a reference specctrum is attached as 3rd file. I observed this behavior with 2 other samples also. Opinions are invited.

2nd file

3rd file

It really looks like convolution of two peaks. You said that you observed it more than once. Sometimes it is better to trust your data.

Looking at the 2nd phase peaks you are right, they do not appear in your case, but are these reference peaks from a powder too and are they from a sample with the same processing as yours? Good luck.

Thanks Dilek. Reference peaks are of the powder prepared by different technique as I could not find any reference with exactly same composition and method as mine. Could this be peak splitting? If I am correct, GSAS cannot handle peak splitting?

You could also try in situ XRD using temperatures in the range ambient to above the calcination temperature already employed by you (for example, 25-900ºC). In this way, you can see the transformation by decomposition of your precursor until the spinel ferrite (alone or in combination with other XRD phases) is obtained.

Another useful technique will be infrared or better, in situ infrared. It will provide you information on the vibration bands spectrum of your sample indicative of the different species at the surface.

Hope it helps!

Dear Surender,

Yes, you could have a peak splitting indicating a formation of another crystal structure or due to the secondary phase,

If you want to confirm if this peak splitting is attibuted to a secondary phase you could use in situ XRD using different temperatures (as suggested by Georgiana) and this technique is called High Temperature X-Ray Diffraction (HTXRD) and associate the result with Thermal analysis (TG/DTA) to monitor the decomposition of the precursors until the formation of your ferrite. But, to be sure the presence or not of this another compound you can also analyse the precursor by IR and Raman Spectroscopies.

My co-workers and I studied one similar behaviour of a perovskite compound by HTXRD, TG/DTA and IR spectroscopy, so you can see in the paper attached below.

You can also anneal your sample at higher temperature if this overlapping disappear.

As I mentioned in previous comments, this peak splitting could be also assigned to a another crystal phase of your ferrite. So, I suggest you prepare different compositions analyse your sample by XRD, refine your samples by rietveld, You can monitor the evolution of this peak according to the composition.

Good luck!

Article SrSnO3:Nd obtained by the polymeric precursor method

Dear Andre, Thanks for the paper and suggestions. I appreciate.

You're welcome!

We are here for helping each other.

I wish you all the best in your scientific research!

Great discussion! I'm not done reading all the coments in detail.

Come join us and share your knowledge as well:

http://www.linkedin.com/groupItem?view=&gid=2683600&item=218925637&type=member&commentID=123681577&trk=hb_ntf_COMMENTED_ON_GROUP_DISCUSSION_YOU_CREATED#commentID_123681577

We are currently discussing conventional diffractograms (1D - Relative Intensity & 2Theta). I see you folks are using conventional diffractograms exclusively. We would love get hear words of wisdom for us.

Once I've familiarized myself with this discussion thoroughly, then I could comment or post examples.

Just click! You are all automatically pre-approved to join the "X-ray Diffraction Imaging for Materials Microstructural QC" group:

http://www.linkedin.com/groups?gid=2683600&goback=%2Egde_2683600_member_218925637

I think this type of pattern is possibly due to the presence of impurity or secondary phases. One can simply anneal the sample and see the XRD pattern. By SEM and EDS also you can check the morphology for the same.

Figure 1: http://www.flickr.com/photos/85210325@N04/8538203093/in/photostream

Why aren't you doing the obvious? Running higher resolution scans of the peaks in question.

1. What is the step size (reduce it) and dwell time (increase it) you used for this conventional diffractogram (Figure 1)?

2. What is your sample morphology? Powder wire, flat powder, powder on film, loose powder, pressed powder...?

3. What diffractometer and type of detector? (Just curious)

I also notice the neighboring peaks in Figure 1 are not perfect either. I suggest a higher resolution scan locally around the peaks. All the other suggestions are excellent as well. I’m learning a lot more about computing “particle size” using the Scherrer formula from you folks. Thanx!

The quick solution is the good old Laue transmission method using 2D film. This will give you the advantage of reviewing the diffraction event in 2D. I'm assuming your sample is thin enough for transmission. Items like sample preparation and other reasons for causing morphological changes (such as preferred orientation etc.) would become obvious. The standard deviation (scintillation noise) compared with peak amplitude seems high in figure 1. Meaning the SNR (signal to noise ratio) is low. This appears to be about 13 to 1 for the largest of the three peaks and 7 for the smallest. I would think that you do not have enough data to resolve the peak splitting at this SNR. With film you could use larger exposure times to integrate the weak reflections. Ideally, you need a 2D X-ray detector. Film will do for now.

Scherrer formula:

1. Can the grain size/particle size/diffracting domain size (whatever) measured using the Scherrer formula be different for individual (hkl)?

2. What if the domains in question were lamellar or “rod like” along a certain (hkl)? How is this aspect handled?

Surinder! "So I believe that sections A and B represents two parts of the same peak."

This seems unlikely for the domain sizes you are working with (

Thanks Ravi, I will answer remaining questions later. What is interesting is that my primary question still remains unanswered. I quote it again.

"If we calculate crystallite size using Scherrer formula (D=0.94 lambda/beta cos theta) . All Other peaks (except one at 62.5) will give crystallite size, half as that of the size given by peak in question. The crystallite size is about 20 nm as calculated from all other peaks. However it will be about 40nm if calculated from section A(or section B) of the peak. Which is not possible (if Scherrer formula is assumed to be correct). So I believe that sections A and B represents two parts of the same peak. Comments, suggestions and explanations are welcome."

I can draw only Two conclusions:

(1) Either Scherrer formula does not apply to my samples (which is almost impossible). OR

(2) A and B parts represent same peak, however part B is chopped from the top.

Please correct me if I am wrong.

Why I am ruling out 2nd phase?

I have gone through various possible phases that could be associated with my sample. All those additional phases have multiple peaks(minimum of 3 peaks). However in this sample there are no additional peaks that could possibly correspond to incorporation of 2nd phase. Just because One peak has strangeness associated with it, an automatic conclusion that the peak represents 2nd phase is little too harsh.This explanation, in addition with the failure of Scherrer formula in this case allows me to reach at conclusion (2).

Here is the header of the uxd file. I think this will help you form opinion.

; (content of file E:\SAIF User Data\2012\Surender Kumar PARSDC Kangra HP\110428A\1104-03(FEBC).raw)

_FILEVERSION = 3

_SAMPLE = FEBC

_+SAMPLE =

_SITE = "PARSDC"

_USER = "PARSDC"

_GONIOMETER_CODE = 20

; Goniometer : D8 theta/2theta, stage : Unknown

_SAMPLE_CHANGER_CODE = 0

_ATTACHMENTS_CODE = 0

_GONIOMETER_RADIUS = 217.5

_FIXED_DIVSLIT = 0.6

_FIXED_SAMPLESLIT = 1

_SOLLER_SLITS = 2

_FIXED_DETSLIT = 10.5

_MONOCHROMATOR = 0

; Incident beam monochromator : None

_SOLLER_SLITS_2 = 2

_THIN_FILM = N

_BETA_FILTER = Y

_FIXED_ANTISLIT = 6.76

_ANALYZER_CODE = 0

; Received beam analyzer : None

_WL_UNIT = A

_WL1 = 1.5406

_WL2 = 1.54439

_WL3 = 1.39222

_WLRATIO = 0.5

_ANODE = Cu

_V4_INF_OSUSER = Surender Kumar

_V4_INF_USER = PARSDC

_V4_INF_SITE = PARSDC

_V4_INF_CREATOR = XRD Commander

_V4_INF_SAMPLEID = C

; Data for range 1

_DRIVE = COUPLED

_STEPTIME = 60.45

_STEPSIZE = 0.0197335

_STEPMODE = C

_START = 10

_THETA = 5

_2THETA = 10

_KHI = 0

_PHI = 0

_X = 0

_Y = 0

_Z = 0

_DETECTOR = 5

; Detector type : Unknown

_DETECTORSLIT = unkn

_AUX1 = 0

_AUX2 = 0

_AUX3 = 0

_TIMESTARTED = 17

_TEMP_RATE = -1

_TEMP_DELAY = -1

_KV = 40

_MA = 30

_RANGE_WL = 1.5406

_3DPLANE = 0

_V4_ADDITIONALDETECTOR = 257

_V4_D8_DIBNUM = 1

_V4_DETECTORNAME = "Detector 1"

_V4_TTLDETECTOR =

_V4_DETECTORTYPE = 1

_V4_DETECTORFLAGS = 0

_V4_DETECTORHV = 630

_V4_DETECTORGAIN = 80

_V4_DETECTORLLD1 = 0.79

_V4_DETECTORWW1 = 1.568

_V4_DETECTORLLD2 = 0

_V4_DETECTORWW2 = 0

_V4_DETECTORPS = 9999

_V4_DETECTORDT = 0

_V4_DETECTORPILEUPTIME = 0

_V4_PSD2THETA = 8.4734

_V4_PSDCHANNEL1 = 0

_V4_PSDAPERTURE = 11.48

_V4_PSDTYPE = 5

_V4_PSDFIXED = 0

_V4_COUNTERS_MASK = 4096

_V4_DRIVES_MASK = 0

_V4_ENCODERS_MASK = 0

_2THETACOUNTS = 1

; 2THETA PSD

10.0000 301

10.0197 303

10.0395 291

............ .......

............ ........

It is correct that science and technology of X-ray diffractometry is well established. But, I think there should always be scope for deviation from well established thinking pattern.

"B is chopped from the top." - Why do you think so?

STEPTIME = 60.45 sec ?

STEPSIZE = 0.0197335 degree ?

WL1 = 1.5406 Angstrom Cu K

GONIOMETER_RADIUS = 217.5 mm

_FIXED_DIVSLIT = 0.6 mm or degrees?

_FIXED_SAMPLESLIT = 1 mm???

_SOLLER_SLITS = 2 mm????

_FIXED_DETSLIT = 10.5 mm ?????

_MONOCHROMATOR = 0

; Incident beam monochromator : None

_SOLLER_SLITS_2 = 2 mm????

_THIN_FILM = N

_BETA_FILTER = Y

_FIXED_ANTISLIT = 6.76 mm????

_ANALYZER_CODE = 0

; Received beam analyzer : None

With the Ni filter, you'd have to contend with K alpha 2, K beta. K alpha 2 should be 60% of K alpha 1 and absent should be K Beta around 20% of K alpa 1. But then that should show on all peaks. As 2Theta increases the separation increases.

I suggest try a higher resolution scan of the three peaks. Rotate sample and see if the profile still persists. i.e., Choose a different orientation with respect to the incident beam.

Running higher resolution scan is possible. The problem is I have run out of sample. I will prepare again once I am through with thesis.

"(D=0.94 lambda/beta cos theta) " - The Beta (FWHM) for the "split peak" is nearly twice that for the regular. I'd expect the computed particle size to change accordingly.

The correct way to use the Scherrer formula would be to fit the correct shape (Gaussian or whatever) and then extract the FWHM. The shape of the shoulder and tail of the peak from the left indicates consistent FWHM. A higher resolution scan would be essential to making that determination.

There appear to be 2, maybe 3 Gaussians convoluted. (A & B)

Thanks Ravi, As far as the header of the file is concerned I reproduced what I got it from the operator, 2000 miles away. Your analysis is very informative. FWHM should not vary substantially no matter what the peak shape(Guassian or anything) is. A variation of 100% can not be due to non-fitting of peak shape! Yes I agree, higher resolution scan can allow deeper insight.

Next question. Is this profile reproducable for this sample?

If you fit mutiple Gaussians for example, their respective FWHM will be a lot lower than the convolution.

Yes I agree.

Surender: ""If we calculate crystallite size using Scherrer formula (D=0.94 lambda/beta cos theta) . All Other peaks (except one at 62.5) will give crystallite size, half as that of the size given by peak in question. The crystallite size is about 20 nm as calculated from all other peaks. However it will be about 40nm if calculated from section A(or section B) of the peak. Which is not possible (if Scherrer formula is assumed to be correct). So I believe that sections A and B represents two parts of the same peak. Comments, suggestions and explanations are welcome."

I can draw only Two conclusions:

(1) Either Scherrer formula does not apply to my samples (which is almost impossible). OR

(2) A and B parts represent same peak, however part B is chopped from the top."

If you would take some time to review other submissions to this forum regarding crystallite size estimates based on the Scherrer eqn, you will quickly see that this analysis is fraught with many assumptions which are rarely realized in real materials (which is why its use is strongly discouraged). Particularly look to comments by Professor Matteo Leoni on this subject.

Note that one or two larger grains, coincidentally oriented to meet the Bragg condition, can skew a diffractogram in precisely the way you have shown. Also any non-spherical grain shape can skew the results (since the size parameter will be a function of the hkl in that case). Note that sample preparation and mounting can be very important here.

So your conclusion 1: "The Scherrer formula does not apply to my samples", far from being "almost impossible", is in fact generally true for nearly all samples.

Also, if the peak appears split every time you run the sample after refilling the sample holder each time then it is likely real, and your structure model is therefore inadequate to fit the data properly - either the structure is wrong, or there is another phase present. If you are basing this on just a single scan, then all bets are off, but since you said you saw this with 2 other samples then I think you have to deal with this truth - either the structure is wrong, or there is another phase present.

Good luck!

Maybe it's more than one phase but still same in structure. If look closely we can see the differences. It also create asymmetric in 400 (>30 degree) cause the second phase has small lattice parameters.

I try to convolute 2 phase with same structure but with different in lattice parameters, as u can see small changes in lattice parameter can create another peak. If I can just make it right, which is difficult, i probably can create the same thing.

I think, the peak shape is not a strange. From the figure, it is clear that the sample contain more than 1 phase . I agree what Michel explained. U can use Highscore plus software or GSAS to analyse the profile .

Good luck

You should not rely on the Scherrer formula in ferromagnetic samples, from my work on FeGa the Scherrer formula crystalline calculation was hugely incorrect and I verified this by TEM. The magnetic moments present result in peak broadening meaning the FWHM is usually increased due to the internal magnetic field.

The shouldered peak looks like it could be a splitting due to either magnetic interactions or, as I have seen before a part of the structure is forming a superlattice, akin to D03 forming in a parent bcc lattice. May I ask what material you are using as spinels form some interesting lattices and chances are I have probably seen it before?

Many of the here above answers are meaningful. However before applying any model, or simply to establish the definitive presence of either additional phases or structure distortion of the unit cell (symmetry lowering), it is absolutely necessary to improve markedly the quality (statistic of counts) of the data collection by e.g. 5 to 10! Parallel it is important to calibrate the XRD instrument by recording a reference materials in terms of diffraction peak width (establish the instrumental diffractionn profile e.g. using a reference garnet powder).

You have more than 1 peak superpoused at this diffraction angle. Probably more than one phase. You should check any other possible phase.