Hi Thomas, it is a Si single crystal, i.e. using a XRD diffractometer you can be happy to see a single reflex from Si. Obviously (001)-Si was very good aligned. Therefore, 111 must be GaAs...

BTW: Why do you think Miller-Bravais indexing is better than Miller indexing? I asked a similar question here several months ago and the result was that there is no specific advantage (expect of the fact that you can recognize symmetry-equivalent planes and directions by permutation). But please note, that unfortunately the reciprocal lattice definition for Miller-Bravais indexing is totally different to all other systems what finally puzzles people more than it helps. I guess that the 4-indexing system is still "preferred" since it shows that we can even deal with 4 numbers instead of 3 which are totally sufficient for 3D problems. For XRD - where you don't need the symmetry-equivalence - it is absolutely useless. I even do not know whether there is a Laue indexing defined considering 4 indices.

Samsul was correct to use hkl and not {hkl}: the latter is for planes and h,k,l are coprime...whereas for Laue indices hkl also 222 or 400 are possible since they describe lattice points (vectors) in reciprocal space, and not like {hkl} directions in reciprocal space.  

1. "theta-2theta (TTH) scan which exhibits poly-crystallinity as shown in the attached file" - How do you conclude this from the subject diffractogram Shamsul? Is it really possible to tell if a sample is epi or poly from a single equatorial scan conventional linear diffractogram?

2. From "Slit - 20 x 2 mm", how did you get a "10 x 1 mm (vertical line source)"? Did you image it? Where you using additional slits?

3. Monochromatic CuKα or CuKα1 = 1.5405 Å? Remember that in practice it is not a "delta function" but is a narrow distribution depending on the optics used.

4. I suggest you repeat the scan of the questionable region at higher resolution.

5. 50 degree scan at 0.025 degree per step and 10 sec dwell time would have to take you 33 min not counting the travel time from one data point to the next? Wow! You are patient and at best ambivalent about the Nano structure of your "epi"/"poly" film, correct?

6. I'll share more soon!

One of the greatest challenges with conventional diffractogram is its "spatial blindness".

Here are some factors to keep in mind in this case of GaAs film growth on a Si substrate:

1. May we assume that the Si substrate is mono-crystalline and of (001) surface orientation without any "off-cut"?

2. What is the expected growth direction for the GaAs film? If poly, do you expect preferred orientation to any extent? What is the "prior experience" in literature?

3. What method are you using for the film growth? What is the typical film thickness? What is the substrate thickness?

4. Well articulated question with good back-ground information! Good job! You should expect plenty of expert opinions from RG/LI to help you understand your paradigm better.

5. If you believe there are peaks from an additional phase, are you able to find other reflections for that phase? Do consider the fact that there may most likely be "preferred orientation" in the film even if not epitaxial. Is your intention to grow epitaxial or poly?

6. Below are examples of real time 2D XRD signal from poly, bi & mono crystalline HgMnTe grown using a modified Bridgman method. The distinctions are unequivocal:

Stats @15:43 EST 8/6/14 1 / 0  ·  2 Answers  ·  42 Views 3 followers.

You'll appreciate the "checks & balances" in RG. Especially under the keen acumen of the likes of Dr. Edward Andrew Payzant of ONL :-)

I'd also suggest you include up to 15 topics (used to be max 5 before) up top to get even greater exposure and hence enhanced participation by other experts as well. I'll share with my network ("followers") as I suggest you do too. The more the merrier! The stats will speak for themselves. Watch them.

I see splitted lines in XRD patterns of cubic structure. At first look, I would think about the structure distortions immediately.

GaAs (A3B5) has face-centered cubic structure. Most of lines of your XRD prove this statement: (111), (200), (220), (311). Certainly, I cannot understand why you have marked (002). It might be very important to be sure that in common case (200) is absent. It is valuable information for interpretation of distortion type.

 I would use…

 …more accurately recorded XRD patterns

… rotation option of your instrument to be sure that the splitted lines really exist.

One more time all lines should be analyzed, they have structure or not. The summarized information give the distortion type. You can find even in Internet (distortion of cubic structure) how to do.

Thank you all for your comments.

@Ravi, referring to your questions:

1. May we assume that the Si substrate is mono-crystalline and of (001) surface orientation without any "off-cut"?

Yes- Si substrate is single crystalline and of (001) surface orientation without any offcut.

2. What is the expected growth direction for the GaAs film? If poly, do you expect preferred orientation to any extent? What is the "prior experience" in literature?

111 is the preferred orientation for the GaAs film although it exhibits poly.

3. What method are you using for the film growth? What is the typical film thickness? What is the substrate thickness?

MBE grown GaAs film by layer-by-layer growth approach. Film thickness is around 80-90 nm. Substrate thickness is 500 um. 

Providing more info soon.

"although it exhibits poly". How did you conclude this? And if so, how poly?

Is it still a hypothesis? My gut feeling is that because of the lattice mismatch the likelihood of a highly strained interface layer is increased. I'll read-up: http://iopscience.iop.org/1063-7869/51/5/R02

http://iopscience.iop.org/1674-4926/35/7/073002?fromSearchPage=true

What is the lattice mismatch between Si (001) and GaAs (111)?

Have you considered the use of buffer layers and or deliberate miss-cuts to better match the lattice parameters conducive to low defect epi growth?

But before that, you need to get "Back To The Future" in XRD and equip yourself with a serious 2D real time detector system. You are in the "stone ages" with the 0D Geiger counter that dates before the Braggs. That said, you may use dental film until you have the temerity to seek funding and acquire a 2D device. Since you have a Bede, I don't even know your 2D choices. Place a dental film for exposure at each of the questionable 2-theta positions by sticking to the front of your detector and exposing it. You'll need to measure the sample to detector distance. One such exposure will reveal the enigma once and for all. Let's quit putsing around and let's get serious.

Do you have the ability to take advantage of dental film? Find an old dentist around, hopefully they still have the stuff. Commandeer their entire stock including the developer and fixer solutions. Worth its weight in gold to you!

@Ravi,

Apologies. Sorry for not expressing myself clearly. In fact, I am not a XRD guy. So, sometimes its a bit difficult to follow your questions and answer to the point. Lets resolve this issue bit by bit. First of all, you asked how do I conclude the as-grown films exhibits poly. I concluded in the following way:As you see, the TTH scan shows many diffraction peaks, e.g. (111), (220), (311) and so on. Isn't it enough for claiming that it is poly? Do you think I would need to perform TEM or any other measurements to verify it. Please advise.

BTW, My intention is to grow single crystalline GaAs.

Sorry again.

No problem not being a "XRD guy".  We're learning together. So, don't worry, I'm not perfect.

Now, in my humble opinion your single linear equatorial diffractogram is insufficient to conclude poly or epi film. However, a single 2D image would be sufficient to confirm poly or epi as I demonstrated earlier with MMT (HgMnTe).

Did you have to optimize the sample position to maximize the Si reflection intensity? If you rotate the specimen about Phi (surface normal) then if poly the film reflection will be about the same. If epi then they will have maxima and minima for Phi=0 to 360 degrees depending on the symmetry involved. That'd be one quick test with the conventional linear diffractogram. Here are some examples of materials with strong preferred orientation in real time!

Well, the XRD pattern looks indeed very interesing :-). I don't want repeat all the recommendations given by Ravi. Only some short remarks to topics which are not yet discussed:

The 111 and 222 of the assumed hexagonal GaAs are  (in case it really exists) 100 and 200. This would be perpendicular to the hexagonal stacking. If you really assume a preferred orientation of GaAs (111) parallel to the substrate suface then I would first expect (001) of the hexagonal GaAs parallel to the substrate. On the other hand, the four-fold symmetry of the surface Si-atoms does not fit very well to the three-fold symmetry of GaAs. Indeed, the preferred orientation of the cubic GaAs is - from my point of view - not very strong. To see the required differences you really need statistically more reliable data. Theoretically the integral intensity for cubic GaAs should be 

111  82%

200 0.01%

220 100%

311  70%

222 0%

400 22%

and the peaks displayed in your dataset matches this more or less (if you consider the increasing peak width with increasing theta, or the influence of non-isometric crystal shapes).

For a rough fit of the hexagonal phase I had to reduce the lattice parameter for a_o from  3.912 to 3.88 Angstroem. On the first view this looks quite much but the lattice distortion would be below 1%.  In so far it would fit, but in contrast to the cubic phase this looks like a very sharp texture since the strongest interference 013 at 50° is not visible at all.

Regarding your 002 reflection of Si which is commonly absent: could it be Umweganregung? This effect is actually known since the end of the 30th of the last century and occurs only for very specific directions of the primary beam.

International Tables for Crystallography Vol C, page  453

sorry...some old terms havn't been translated into English like Bremsstrahlung

try this link

http://www1.uni-hamburg.de/mpi/rossmanith/Reprints/jac_03b.pdf

Hi, what Gert means you can find in a similar system here:

http://onlinelibrary.wiley.com/doi/10.1002/pssa.200409030/pdf

However, I don't think your indexing is clear:  not only the reflections marked by * but all black ones should come from cubic Si. if you index a hexagonal system you should use Miller-Bravais indices instead of Miller indices. If your GaAs is relaxed, then the lattice constant of cubic GaAs is ~4% larger than that of Si, so your red {111} could be those of cubic, relaxed GaAs.  The rest of the XRD scan is really too noisy to decide whether there are also other, weaker peaks from this. Another strange thing in your XRD scan is that the red {111} reflection is also very narrow, implying large crystallites.

If poly, then the likelihood of the peak split (1 degree +-) being due to miss-oriented grains is higher than a different phase. Especially, if other strong reflections of that suspect phase are not also present. This is the typical paradigm of ambivalence when analyzing a conventional linear equatorial scan diffractogram. In your case let us assume a sampling size defined as 1mmX10mm. In order to analyze epitaxial film it is quintessential to eliminate the handicap of spatial blindness. Go 2D! Use film if you have to! You'll learn and understand faster than this "grope in the dark" method now in practice for over a century. Otherwise, you still can't escape determining the 3D reciprocal space with the conventional method in order to be precise and unequivocal with XRD. An oblique attempt is what is lovingly called the "RSM" (Reciprocal Space Map") with the conventional 0D detector and a gigantic sampling area/volume (1X10mm^2). A time prohibitive exercise in futility! :-)

BUT YES! In order to confirm the existence of any new phase you will eventually need to use the TEM to CORROBORATE! 2D real time Bragg XRD Microscopy can help pre-screen and identify the sampling area for TEM!

Hi Thomas, it is a Si single crystal, i.e. using a XRD diffractometer you can be happy to see a single reflex from Si. Obviously (001)-Si was very good aligned. Therefore, 111 must be GaAs...

BTW: Why do you think Miller-Bravais indexing is better than Miller indexing? I asked a similar question here several months ago and the result was that there is no specific advantage (expect of the fact that you can recognize symmetry-equivalent planes and directions by permutation). But please note, that unfortunately the reciprocal lattice definition for Miller-Bravais indexing is totally different to all other systems what finally puzzles people more than it helps. I guess that the 4-indexing system is still "preferred" since it shows that we can even deal with 4 numbers instead of 3 which are totally sufficient for 3D problems. For XRD - where you don't need the symmetry-equivalence - it is absolutely useless. I even do not know whether there is a Laue indexing defined considering 4 indices.

Samsul was correct to use hkl and not {hkl}: the latter is for planes and h,k,l are coprime...whereas for Laue indices hkl also 222 or 400 are possible since they describe lattice points (vectors) in reciprocal space, and not like {hkl} directions in reciprocal space.  

Hi Gert, you are certainly right about the correct use of the brackets (sorry) and the remark about the 111 reflection not being to Si if the substrate is an [001] oriented single crystal. The splitting of the 111 peak seems to be ~3%, so between that of Si and relaxed GaAs. Cross-sectional lattice imaging by TEM should show whether the grains are cubic, hexagonal or full of twins wherein the local 111 stacking sequence changes from cubic to hexagonal.

Since we know that the substrate is a Si (001) wafer the, discovered double peak shouldn't contain a 111 of Si.  So we still have the option of hexagonal GaAs, but I am out of the business for many years so that I don't feel myself able to judge how probable this is. At least it fits to a hypothetical 100 interference.

Nevertheless you are right, Thomas, that the best way would be to check this by TEM.

Thank you all for your valuable comments. 

@Ravi, referring to your questions: . For the powder scans  the Si reflection was optimized so that it was at a maximum intensity, and used that peak as the reference point for all the other peaks.

The tube is designed to output a 10 x 1 mm vertical line, the beam then passes through another physical 10 x 1 mm line slit. The 20 x 2 mm slits come after diffraction from the sample itself. Both are in front of the detector (one directly in front, the other spaced ~10 cm away). 

Rotate the sample 90 degrees (Phi, about the sample surface normal) and repeat the scan. Go further out to the shoulder of the Si substrate peak (80-90 degrees 2Theta).

preferred orientation of GaAs (111) - this can be proved by GaAs (111) PF.

Apparently there is a center pole, indicating that the film has a preferred orientation of GaAs (111) along surface normal (SN) direction.

There are 4 sharp peaks around circle of alpha=35deg. I think these peaks come from the Silicon substrate, which surface normal (004) so that its (111) pole is 35 deg off from SN. Ideally, the silicon 111 peak (28.4deg 2theta) would not be measured on this GaAs 111 PF measurement (27.3deg). But the Si single crystal peak is so strong and have long peak tails that can be caught.

I am also attaching a TEM image.

1. How long did it take for you to generate PF_GaAs.jpg?

2. What part of the sample did you get the TEM sample out of?

3. Did you take a transverse sample or a lateral sample w.r.t. the specimen surface?

Of course, the {111} pole figure can be proved by a measurement of the intensity around a theta equal to 111, 222, 333 etc. Unfortunately I cannot see how big strong the texture is. Sometimes you only measure a bigger GaAs crystal...(although the paek should be clearly sharper...except you used a very wide aperture). In order to make this clear you should measure down to alpha

The ratio of epi to substrate diffracting volumes are vastly different in this case. What is the estimated penetration depth in Si? Around 13-15um or so? The GaAs film is at most 90nm (?).

Is the pole figure for poly a synonym for RSM (Reciprocal Space Map) with mono/epi?

Here's a helpful table from the past:

Shamsul! The pseudo color image of the PF is what I want to focus on. In order to understand how this PF (Pole Figure) changes topographically for your specimen, you must be prepared to raster the beam on your sample surface and record the 3D pattern for each topographic "Pixel/VOXEL". Ideally, you must minimize the sampling area (1X10mm^2, pixel size) by using appropriate slits in order to improve resolution.

How long did it take you to create the current PF image for one topographic location of your specimen?

This means you are using a slit focus of your tube and not the point focus? (You can do this but you are loosing a lot of intensity using a point aperture.) If you don't use a point aperture or better a focusing optics (Multicappilar), you will have "blurring" since the sample is no more in focus conditions. Or do I not correctly understand your measurement regime? 

Very interesting topic, the question, as well as the answers - it was really a pleasure reading all this.

I find rather interesting to get what seems to be a rather strong 111- texture for a cubic film grown on Si(001)

Concerning the pole fig, for visualisation, it might be clearer to show it as a 3d projected image (polar coordinates) with the intensity (z axis) being color-coded; that would also make clearer the 35 or 55 deg. positioning (question raised by Gert) of the 4-fold symmetry peaks attributed to the Si - most probably this is indeed the case, the Si peaks should be close enough and their tails extending enough to be present in the PF.

apart extending as much as possible the range (psi) of the PF, or going furtherer in the Rec. space (peaks like (333), as suggested), it could also be useful to perform a PF on at least another GaAs characteristic peak, not belonging to the 111-family.

Also, have in mind the resolution of your instrument during such measurements; what is the resulting angular acceptance of the detector with your slit setup ? Do not forget that with so assymetric slits (1 x 10 mm), the (angular) resolution function while you perfrom PF can drastically change vs. the sample elevation angle (psi). The illumination function with an assymetric beam as well? Maybe try to use some more 'symetric' values (ex. 2x2)

one of the suggestions I found really very intersting was given by Ravi - if possible, go 2D for the detector - the gain in the ammount of information and the acq. time are tremendous. even without doing complete PF, hints of presence of textures or poly can be rather easily found (was already pointed out in a post from Ravi), which could then guide the future investigations

Regards

cristian

I'm going to invite Matteo Leoni, Tapan Chatterji, Lawrence Margulies and a few others whose opinion I value immensely to join us and share their wisdom as well regarding the classic paradigm of 2D versus 0D data acquisition in modern XRD analyses. That will improve our quorum to discuss this fundamental XRD question astutely posed by Shamsul (one with the brilliance of the sun) who also wrote "I am not a XRD guy". I shall plagiarize him and say "I am not a XRD guy" compared to some of the expertise I've had the pleasure of encountering in RG and LI. With that attribute of humility, I'm excited about how much more there is for me to know! Thanks for facilitating it, all! It is time to shake that good old "knowledge tree" for lose fruit!

Here is a link (for those interested)  to the LinkedIn discussion regarding the pros & cons of 2D versus 0D in modern XRD: https://www.linkedin.com/groupItem?view=&gid=2683600&type=member&item=5884025088390500354&trk=groups_search_item_list-0-b-ttl&goback=%2Egmr_2683600%2Egde_2683600_member_5902554584810795012%2Egmr_2683600%2Egna_2683600 I was virtually booed out for my temerity to claim the incontravertible advantage of 2D image acquisition in XRD just as I was in the 1980's. Love it! Under-dog is here!

There is insufficient information in a single linear conventional diffractogram (using "point" or 0D source or "line" 1D/2D source) to characterize any epitaxial film/substrate sample processed using MBE or otherwise. I intend to tackle this subject with 4 explicit examples. I will demonstrate that the manner in which Shamsul and many others are using  XRD presently is a minutia of what is possible just by switching to 2D real time data acquisition. This may be achieved in any XRD lab (no mater from which Jurassic era) without a synchrotron.

Example 1. Mono & Poly Nano structure - ZnSe (224) Asymmetric Reflection using Panalytical PIXcel 3D - 5.5 hrs. data acquisition time and 0.01 degree resolution per image frame. 55um detector pixel size. 256x256 https://www.youtube.com/watch?v=dFCQS8oUyT0&index=38&list=PL7032E2DAF1F3941F

Example 2. Theory versus Experiment - Substrate GaAs (004) Symmetric Reflection - Real Time &

Thank you all for your valuable comments.

@Ravi and Cristian,

Referring to your questions:

1. How long did it take for you to generate PF_GaAs.jpg?

It took ~10hrs to collect this PF data. This is slow scan PF data for good quality. Rough PF could take only couples of hours. The software 3D Explorer was used to plot the data pattern.

2. The pseudo color image of the PF is what needs to be focused on. In order to understand how this PF (Pole Figure) changes topographically for your specimen, you must prepare to raster the beam on your sample surface and record the 3D pattern for each topographic "Pixel/VOXEL". Ideally, you must minimize the sampling area (1X10mm^2, pixel size) by using appropriate slits in order to improve resolution. How long did it take you to create the current PF image for one topographic location of your specimen?

The beam footprint on the sample was ~1/sin(13.6deg)mm x 3mm. The scan took ~10hrs.

3. Also, the resolution of PF instrument during such measurements; what is the resulting angular acceptance of the detector with your slit setup ? One should not forget that with so assymetric slits (1 x 10 mm), the (angular) resolution function while one perfroms PF can drastically change vs. the sample elevation angle (psi). The illumination functions with an assymetric beam as well? Maybe one should try to use some more 'symetric' values (ex.> 2x2)

The detector slit was 2mm.

Did you mean that sample spacial resolution? or scanning axis resolution? For quality PF measurement on a uniform sample, it may not matter. For structure orientation distribution function measurement, the points you mentioned do matter. XRD PF measurement by XRay is different from EBSD PF measurement by electron beam.

4. The 4 peaks at 35° I would expect at 55° since the distance between {001} and {111} is 55° not 35°. Moreover, the 2theta distance appears quite bit so that I won't expect that you will measure {111} for GaAs and Si at the same time. Your argument (tail of the Si) I can follow but wouldn't the peaks then that sharp?

The alpha angle depends on how you define it. If you say alpha at the surface normal is 90 degree, then the 4 peaks is at 55 deg.

@Ravi, 

Referring to your questions on TEM:

2. What part of the sample did you get the TEM sample out of?

it is from the region slightly above the substrate. 

3. Did you take a transverse sample or a lateral sample w.r.t. the specimen surface?

transverse sample w.r.t. the sample surface.

If these are 55° from the center to the 4 peaks I agree that these are obviously the {111} Si planes. Sorry for the misinterpretation but I've only seen pole figures which start with zero in the center.

Did you use point collimator?

Example 1. Mono & Poly Nano structure - ZnSe (224) Asymmetric Reflection using Panalytical PIXcel 3D - 5.5 hrs. data acquisition time and 0.01 degree resolution per image frame. 55um detector pixel size. 256x256 https://www.youtube.com/watch?v=dFCQS8oUyT0&index=38&list=PL7032E2DAF1F3941F

This sample exemplifies the XRD signal from a wafer containing both mono crystal and some near poly crystals at the "fracture edge". I make an attempt to simulate what one would see in a conventional diffractogram using an integrating window size of 1X10mm as Shamsul has presented and is typical in the industry. In the case of the ZnSe (224) data with PIXcel 3D detector (55um pixel) equates to roughly a window of 20x200 pixel array of data. Here's an example of two such windows (2x100 pixel array=0.1x5.5mm2) drawn on the topograph towards the center of the mono crystal: https://www.flickr.com/photos/85210325@N04/14387313558/ 

So, the narrow windows would just restrict the image (XRD signal) entering the active detector area. The integrated rocking curve of this entire region indicates two peaks separated by about 0.2-0.25 degrees (small compared to poly). If this integrating window was placed on the reflections from the "fracture edge" as shown here: https://www.flickr.com/photos/85210325@N04/14889281565/ The resulting XRD signal is displayed below: https://www.flickr.com/photos/85210325@N04/14702658058/in/photostream/ where the difference between first peak and last peak is nearly 0.5 degree.

@Gert, thanks!

Yes, you are right. Unlike TTH scan, the PF measurement is performed on the sample with Si(111). So, you are right. Those are obviously the {111} Si planes.

@Gert, 

the Parallel beam with incident slit size: 1mm X 2mm was used.

It seems that the conventional PF is measured by Bragg Brenteno focusing optics with a point x-ray source. They both have advantage and disadvantage. If you use conventional PF measurement and install a diffraction beam monochromator, the Si substrate peak may be avoided.

"conventional PF is measured by Bragg-Brento focusing optics with a point x-ray source". This method is "spatially blind". In the case of your epitaxial film the relative PF VOXEL-to-VOXEL spatially is really what one would want to establish to characterize the epitaxial film or the mono-crystalline substrate. The conventional PF is also called the RSM in the case of rocking curve analyses for mono-crystalline/epitaxial materials. A time prohibitive exercise in futility is to use this conventional method to characterize 2D surfaces of wafers.

The focusing geometry is typically a "cheap trick" to enhance signal amplitude by convoluting over a larger sampling area. This method was pertinent in the "pre-Noah" ages of our XRD fore-fathers who had to make do with integrating film or Geiger counters (0D detectors). There are no such limitations presently other than those self-imposed :-)

The fact that you are unable to unequivocally establish the poly-crystallinity or epitaxy of your film is symptomatic of the limitation of conventional linear diffractograms using a 0D detector.......

I personally prefer the parallel beam line source because it makes the "optimization" process almost trivial. This is so because practically it is "a piece of cake" to orient the diffracted beam (line) with respect to the incident beam (line) in comparison to doing so with a "point source" (without asymmetry).

@Gert, yes, you are right. 002 reflection of Si is nothing but Umweganregung (multiple diffraction) peak. 

good to know :-). Did you rotated your sample? How did you find it? Anyway, well done! There is still the major question: is there hexagonal GaAs? (I've never heard about it, but I am out of this business for many years, so that I wouldn't exclude anything :-). ) 

Ravi, you are "unfair" :-)... There are many and very excellent labs which still work exclusively with 0D detectors. It is simply the easiest way to prevent any kind of corrections for asymmetric beam conditions. The major problem is: TIME. It is simply very slow...so that we like to use 1D or 2D detectors.

@Gert, referring to your question on whether there is hexagonal GaAs: 

I would recommend to look at the paper I am attaching here.

Gert! "There are many and very excellent labs which still work exclusively with 0D detectors", why? The results of the conventional linear diffractograms seem as ambiguous as in the 1980's. What could be the advantage? Cost? Certainly not resolution. Gain? What? You concede, "not time"!

If only those poor suckers using 0D thought out-of-the-box and just "tasted" 2D detector once! It would all be history. The 2D detector could always be dumbed down to a 0D detector by integrating pixel arrays. (Just in case one is unable to tackle the 3D reciprocal space yet). I'm sure you've had the opportunity to  work with a 2D (real time or not) device. Which one?

My only suggestion to Shamsul is that before embarking on a search for any additional phases, eliminate the presence of poly-crystals leading to anomalous effects on the XRD signal. To check the presence of trace phases, you may produce a "well ground" conventional powder sample and use the Bragg-Bento arrangement to examine the sample to near forensic precision using the century old 0D Geiger counter. "Splitting" of Bragg peaks is a common enough phenomenon to identify. The lack of other reflections corresponding to any new phase would be another clue. Eventually, no matter how much we postulate TEM would have to be employed in order to be unequivocal.

Here is an example of a region on the ZnSe wafer with intense relative as grown in situ Nano structural perturbations: https://www.flickr.com/photos/85210325@N04/14756597361/

Detailed XRD RCP's (rocking curve profiles) in the same region reveal the possibility of large Bragg Peak position variations from 0.75 to 1 degree in some locations.This shift may be higher in the case of polycrystalline materials with large monocrystalline constituents. https://www.flickr.com/photos/85210325@N04/14891811366/

Ravi, this is a good question and certainly not easy to explain. My assumption is that since we have only one primary beam direction there is only one diffraction direction (in reflection, not transmission)  which does not need  complicateded corrections. I am really not sure whether high-resolved Rietveld analysis is done by 1D or 2D detectors. As I already mentioned: I am out of the business since more than 10 years and this is a long time. Nevertheless, you could also do in those days Rietveld with diffractograms collected by 1D detectors. 

@Ravi, Thanks for the info.

@Gert, yes, the sample was rotated. In fact, I have given a wrong info. The scan I provided here is 2theta-omega scan (NOT theta -2theta). Does it explain why 002 reflection of Si was obtained?

Although 2theta/omega and theta/2 theta scans are same with a minute difference in implementation. Detector and sample moves in 2theta/omega scan, where as source and detector moves theta/2 theta scan. Am I right?

Reitveld should be implementable on 1D or 2D XRD data. I'm not sure how that would be relevant in this particular case. I kind of understand its use in powder diffraction profiles but not for mono crystalline rocking curve analysis. Never-the-less, in this case the mono crystallinity of the substrate is not in question but that of the MBE film is. Unless Shamsul helps us with the potential grain size and Nano structural morphology of the film (poly), we would have to "wild guess" the cause of the multiple peaks in Shamsul's XRD profile. Chances are that the poly-crystallinity of the MBE film has dissipated the strain field in the film through creation of high defect density boundary layers (interface layers) some affectionately called grain boundaries. A vastly different scenario than a highly strained hetero-epitaxial layer. The use of buffer layers to minimize lattice strains is well known.

1. Omega-2Theta coupled scan - this is the conventional diffractometery. Some call this Theta-2Theta as well.

2. Omega only de-coupled scan - The detector 2-Theta is fixed while the sample alone is rocked through Omega. Generally used exclusively for XRD rocking curve analyses.

Our experimental observations do not show any difference in the RCP generated by the two scans noted above for a given (hkl)

Also remember that one could either move (rock) sample/detector combination to achieve the above modes or use the combination of detector/source (no available for Sychrotron) to achieve the same objective. This depends on the diffractometer make/model.

Rietveld I only mentioned because of the "side-discussion" of 0D...2d detectors. I only want point out that the focussing conditions are exclusively fulfilled for one point (position of the 0D-detector. For all other detectors some corrections are required...and a linearity are assumed which si perhaps not that exact as assumed. For common mesurements it is OK but for Rietveld it is not (except you summarizs over a 1D and move it as a 0D). But maybe thare are already solutions available. And maybe tha guys are know are oldfashined of they are still working with 0D detectors.

Regarding Omega-2Theta and Theta-2Theta I agree with Ravi. The principle is the same so that the results should look identical.

For the pole figure you used another XRD facility, right? Or is it equipped with another axis as well?

Shamsul! "2theta-omega scan (NOT theta -2theta)", No difference in rocking curve! Omega is the angle between sample surface and incident beam while "theta" is the angle between (hkl) and incident beam. Correct?

Yes, Ravi. Agreed. 

Shamsul! Do you have photographic film in your XRD lab by any chance? One short exposure in front of the detector where you believe the MBE thin film diffraction peak to be, would disclose volumes to you. Is such extinct technology still around @UCLA? Try an old dentist :-)

Back in 1984, I had similar equipment as yours but more crude without the beam conditioner just a Ni filter, that's all. A then 30yr old GE Diano model! That was about when Brian Tanner and Keith Bowen were getting their BEDE going at the 1986 Denver X-ray Conference. They were the "lead horse". They used a combination of 0D Geiger counter and 2D photographic film for their work then.

Here is a modern version of that awesome combination photon counter plus 2D array:

Here is one way to probe the 3D reciprocal space topographically using the Panalytical MRD & PIXcel 3D combination

Good work Shamsul! Now all you need to do is repeat the XRD scan for the entire 2D surface area of your sample to understand the sample Nano structure topographically.

https://www.researchgate.net/publication/265034540_Towards_van_der_Waals_Epitaxial_Growth_of_GaAs_on_Si_using_a_Graphene_Buffer_Layer?ev=pubfeed_overview

Article Towards van der Waals Epitaxial Growth of GaAs on Si using a...

Concerning the pole figure scan I shared in my previous post, I would like to know the following info:

1) How strong the texture is in my as-grown films? In other words, can I measure the degree of texture from the pole figure scan?

2)  What is the FWHM of the surface normal peak?  GaAs(111) is in this case.

3) Are (1) and (2) important figure of merits for evaluating the material quality ? In other words, can these info tell something conclusive?

Please note that the alpha axis peak: starting at 88 deg and its max is at 90 deg in my pole figure scan.  FWHM is measured to be about 1.5 deg.

The degree of preferred orientation could be calculated as follows:

spread of diffraction intensity/90=2/90=2.2%

What conclusion can we draw out of these info? Please help.