The characterization of thin films using the conventional symmetrical Bragg Brentano configuration (theta/2theta) generally produces weak signals from film itself and intense signal from substrate, which is why it is not suitable. Moreover, the depth of analysis varies during the symmetrical sweep theta/2theta (the analysis depth z is approximately equal to sin(theta)/2mu, where mu is the linear absorption coefficient). In these cases, it is convenient to use the technique of X-ray diffraction at grazing incidence to minimize the contribution related to the substrate.
In this geometry, the incidence angle (alpha) is fixed at a small angle (exceeding the critical angle of total reflection, typically between 1 and 3 °) and the angle between the incident beam and the diffracted beam (2theta) is varied, moving only the detector arm. Thus, the incident beam go over a long way on the film surface or interest, reinforcing its diffraction pattern, while the signal from the substrate is reduced due to the small angle of incidence (alpha). Under these conditions, it can be shown that the depth of analysis (z), when alpha is very small, does not depend on the 2theta angle and, in that case, the depth of analysis is controlled by the angle of incidence and does not vary during the sweeping. In this way, one can perform similar studies to the powder diffraction X-ray but at a controlled depth of analysis, either for phase identification, crystal structure study, quantitative analysis, etc.
Unlike conventional powder diffraction geometry, which observes planes parallel to the surface of the sample, in the case of grazing incidence crystal planes inclined with respect to the surface of the sample are observed, whose normal is the bisector of the angle formed by the incident and the diffracted beam. So, it is not suitable for studying orientation.
The instrumental contribution to the width of the diffraction peak is higher in GI configuration than in Bragg Brentano configuration. So, it is not suitable for studying crystallite size.
https://www.researchgate.net/post/Whats_the_difference_between_GAXRD_and_Gonio_mode_of_XRD
Depending on the type of fimls, this might or might not be the best technique. You work at low angle to maximise the signal from the film
The characterization of thin films using the conventional symmetrical Bragg Brentano configuration (theta/2theta) generally produces weak signals from film itself and intense signal from substrate, which is why it is not suitable. Moreover, the depth of analysis varies during the symmetrical sweep theta/2theta (the analysis depth z is approximately equal to sin(theta)/2mu, where mu is the linear absorption coefficient). In these cases, it is convenient to use the technique of X-ray diffraction at grazing incidence to minimize the contribution related to the substrate.
In this geometry, the incidence angle (alpha) is fixed at a small angle (exceeding the critical angle of total reflection, typically between 1 and 3 °) and the angle between the incident beam and the diffracted beam (2theta) is varied, moving only the detector arm. Thus, the incident beam go over a long way on the film surface or interest, reinforcing its diffraction pattern, while the signal from the substrate is reduced due to the small angle of incidence (alpha). Under these conditions, it can be shown that the depth of analysis (z), when alpha is very small, does not depend on the 2theta angle and, in that case, the depth of analysis is controlled by the angle of incidence and does not vary during the sweeping. In this way, one can perform similar studies to the powder diffraction X-ray but at a controlled depth of analysis, either for phase identification, crystal structure study, quantitative analysis, etc.
Unlike conventional powder diffraction geometry, which observes planes parallel to the surface of the sample, in the case of grazing incidence crystal planes inclined with respect to the surface of the sample are observed, whose normal is the bisector of the angle formed by the incident and the diffracted beam. So, it is not suitable for studying orientation.
The instrumental contribution to the width of the diffraction peak is higher in GI configuration than in Bragg Brentano configuration. So, it is not suitable for studying crystallite size.
in addition to what Claudia already has compiled (nice summary btw.), also grazing incidence XRD can be performed even below the critical angle of total reflection, resulting in a very small penetration depth (depending on the material 1-3 nm only). The peaks are again weaker in contrast to incidence angle above the critical angles, and furthermore their position is systematically shifted because of the influence of the refraction of the X-rays in the material. Reference papers are given e.g. by Michael Toney and Sean Brennan, http://prb.aps.org/abstract/PRB/v39/i11/p7963_1.
Best regards, Dirk
I'm interested from a fundamental perspective and will participate once I've understood the direction of the discussion better. I have not yet reviewed all the wonderful links posted. I promise I will. I need to! Thanks to all participants willing to share!
My interest is, to clarify my own conceptions of XRD and the data obtainable in the reciprocal space. In other words, what in the material Nano structure can we measure by probing the reciprocal space using the intersection of the Ewald surface (detector plane) and the corresponding diffraction vector (hkl), for instance. This is the ultimate effect of all XRD diffractometery. The difference usually is the "resolution" achievable.
In general my conception is, various periodicities in the Nano structure cause diffraction effects in different portions of the reciprocal space (diffractogram) or the Fourier space. The smaller periodicities may be measured at larger angles and the larger one at the smaller angles. This is consistent with the reciprocal relationship in Fourier space. Now in the "Grazing Angle" method the sampling VOXEL is exceptionally "thin" in depth (due to the limited penetration depth as a consequence of the geometry). Therefore the diffraction pattern would display the "sync" function which is the result of the Fourier transform of the "square" function in real space represented by the "film", for example. By monitoring the characteristics of the "sync" function in reciprocal space one is able to determine the characteristics of the corresponding "square" function in real space, namely the "thickness of film". In general this effect is predominant very close to the incident beam vector just as in the case of SAXS for "Grazing" angle XRD.
Due to the nature of the X-ray detector (0D Point/Scintillation) used in conventional diffractometery the values one measures are the average over relatively "large" sampling surface areas compared with the scale of variations present in the sampling area. Meaning, non existent spatial resolution in data. Unless of course, one uses a micro-focus source and is willing to turn "Rip Van Winkle" collecting XRD data.
I know I virtually have my head on the "chopping block" this time around. So, "basturi" or "sword", please feel free to help me edit the above generalization to perfection. Thanks! I shall gladly recant in the interest of self-improvement!
Rip Van Winkle (http://en.wikipedia.org/wiki/Rip_Van_Winkle)!
Is how I felt @ Rutgers Grad School in the 1980's collecting 3D Reciprocal Space data with a 1D LPSD (Linear Position Sensitive Detector - innert gas filled Braun) from a poly-crystalline AL2024 fatigue specimen for days and weeks at a stretch. I can't even imagine how it feels for those still handicaped with the 0D conventional, century old, "point counter"!
http://www.flickr.com/photos/85210325@N04/11343736043/
Claudia! "conventional symmetrical Bragg Brentano configuration (theta/2theta) generally produces weak signals from film itself and intense signal from substrate, which is why it is not suitable"
That would depend on the sensitivity, dynamic range and spatial resolution of the detector system you use and the geometry/optics you would choose. Your statement is generally true for the conventional 0D point/scintillation counter based XRD systems with limited capabilities.
1. Large angle symmetric Bragg reflection: Here is an example of "film thickness mapping" using real time Bragg XRD Microscopy for a 0.5um MBE epi film of super lattice InAs/InAsSb 25nm layer x10 periods on (001) GaSb substrate with high penetration depth of about 15-20um:
Video:http://www.youtube.com/watch?v=De_Nh7iN6y4&list=PL7032E2DAF1F3941F
MBE epi:http://www.flickr.com/photos/85210325@N04/10653618416/
2. Topographically mapping relative crystallographic orientaion using Grazing angle or ACT- Asymmetric /Crystal Topography ZnSe wafer (224) reflection with low penetration depth of about 1.1um:
Video:http://www.youtube.com/watch?v=dFCQS8oUyT0&list=PL7032E2DAF1F3941F
ZnSe (224) ACT - http://www.flickr.com/photos/85210325@N04/8068646477/in/set-72157632729045818
I look forward to demonstrating the dramatic advantage of this real time XRD technology for GIXRD, SAXS & XRR soon. I'll post findings then.
http://www.youtube.com/watch?v=De_Nh7iN6y4&list=PL7032E2DAF1F3941F
Dear Ravi, always online (?!) - also on Sunday evenings :-D ?
I think what Claudia mentioned are really thin films, of say some few nm, but not microns. In that sense - if you are investigating a really "thin"film, then asymmetric Bragg diffraction with grazing incidence will help!
Best regards from the ESRF - Dirk
Thanks for staying on target. These are test balloons. Go ahead shoot them down!
BTW What are you doing this early on Monday morning there? It is 4:45PM here. I'm enjoying NY Jets "football" on TV with my 19yr. old son while blogging. I just heard "Santa" riding around in a "fire engine" (fire truck) in our neighborhood. Happy Holidays!
Here is an example of a 50nm MBE monolayer of LaSrCuO4 grown on (001) LaSrAlO4 substrates:
Video: http://www.youtube.com/watch?v=LLoSxmZIQnQ&list=PL7032E2DAF1F3941F
LSCO/LSAO (004) Symmetric Reflection: http://www.flickr.com/photos/85210325@N04/8036412739/in/photostream/lightbox/
This method works well even if the film is real thin. I just haven't measured the limit yet. The SNR from the 50nm film is phenomenal at
Looks convincing. An yes, with the new 2D-detector stuff things can be done that were impossible a few years ago with point detectors (and they still are!). In fact I have beamtime at the ESRF for a project - SR-beamtime is rare so you have to take what you get, even if it is the week before X-mas. Enjoy the Jets - I prefer european soccer and the NBA ...
I'm open to both soccer & football, Dirk. I watch them both. What's your favorite NBA team? We're rooting for the NY Knicks! I used to play "field hockey" in my early youth. My son played football and wrestled in High School despite my trepidations. I never discouraged him. That game helped my son eliminate "fear" from his lexicon. We as parents were always on "pins & needles" due to the possibility of injuries, concussion etc. I went to his games not only to see who won but to make sure if injured proper care given. Like little gladiators! Except, the guys hitting my son didn't look so little at all. Rough stuff!
What sort of detectors do you use at ESRF? Why not a real time 2D detector? I’d think, specifically because "beam time is rare" and limited. That was the precise reason that motivated me to move to real time 2D imaging. I'm still analyzing video data from 2 years ago and finding more and more details. It is a veritable "Pandora's Box" of information. I have not yet got to the "diffused scattering component" in the videos. I'm learning more and applying to analyze "old recorded data". As you point out with beam time, we shoot (video) first as much as we can and then ask as well as answer all questions later. Exciting stuff!
Check out this LI discussion and jump in if you like: http://www.linkedin.com/groupAnswers?viewQuestionAndAnswers=&discussionID=5815615304121749504&gid=2683600&commentID=5817985212927221760&goback=%2Egmr_2683600&trk=NUS_DISC_Q-subject#commentID_5817985212927221760
http://www.flickr.com/photos/85210325@N04/10659886546/
We are using the Pilatus 100 K from DECTRIS (https://www.dectris.com/pilatus_overview.html#main_head_navigation) currently - we also have them in Dortmund at our beamlines, where the photon flux is much smaller. Here at the ESRF (even at a bending magnet) we have so much flux that we still have to use some absorber foils between sample and detector, even if you are investigating the weak diffuse scattering. The detector is so sensitive, and each of the 100K pixels of 170 x 170 micron may count up to 2x10^6 photons per second with 20 bit dynamic range. It is phantastic to have this equipment. In our home lab I was happy about 10 years ago to have the Philip (panalytical) X celerator, a linear array of Si diodes that substantially speeded up scanning diffractometry. And nowadays we are again one step further with the new area detectors, in the home lab as well as in the synchrotron labs. I may post some diffuse scattering stuff later ...
I fully agree with you, the old recorded data often contain many information that awaits to be discovered and analysed.
NBA? Dallas with Dirkules - or how do you call him in the US?
Enjoy the evening Dirk
Dirk Nevitski? Dallas! Get to bed Dirk!
Later on! What is the typical frame rate or image frame download rate for that detector @ 20 bit DR? Are you able to store video type data files? Please post some typical data. What video formats are possible?
guys, it is true that with 2D detectors you can have more information, but I would not get over excited. Prices are still the issue and the "obsolete" 0D detector (as Ravi call it) is still leading in most laboratories. There are applications where 2D can be good and others where it is not. In the lab it is true that you can use the 2D information and reconstruct a 1D pattern, but if you need a clean profile (e.g. for line profile analysis), then the old 0D detector with a monochromator is still the winner: I want to characterize the specimen, not the effects of the filtering including the edge cut on the white emission of the tube.
The actual divergence of the source should also be considered when integrating the 2D information to obtain a 1D pattern, as the breadth of the profile might change with the integration angle. A 1D detector in that case could be a better choice. And again which type? There is no 1D detector able to perform well in the general case. And if you have a bulk(y) specimen? that's different from a foil or a capillary in transmission that would be the ideal specimen.
I still believe that detector, source and application behave as one, so the three should be optimised/considered simultaneously.
This said, I still believe the information given here is inaccurate/partial. I have previously pointed out to https://www.researchgate.net/post/Whats_the_difference_between_GAXRD_and_Gonio_mode_of_XRD
What Claudia describes is one of the possible grazing incidence modes and, as I said here and as I was saying in my previous answer to this issue, it is not necessarily the best one. You can have at least two more operational modes: if you do work in grazing psi angle you can have a traditional Bragg Brentano scan at fixed penetration.
The applicability of line profile analysis in grazing incidence mode depends again on the specimen. If you take some powder, you spread it over a zero-background holder and you collect data in grazing incidence mode, you can definitely do an accurate line profile analysis. The same therefore applies for a randomly oriented polycrystalline film. You collect the pattern of a standard in the same conditions and use that. You are not in the ideal case, but "being not suitable" is definitely not correct. You just see the divergence more (that's why you expect the instrumental profiles to be more odd), and therefore need to spend more efforts in modelling your instrument (I always don't advise so). If you spend more time in data collection (different incidence angles, different rotation of the specimen or you use a 2D detector) you can get information also on an oriented film.
"Obsolete", not yet. "Archaic" - perhaps? The ubiquitous 0D detector can duplicate any observation that a 2D detector can provide but with significantly higher time investment. It worked for the Braggs, it's fine for me. The key advantage of the 2D device is its sensitivity to any existence of "preferred orientation". Of course, for random powder type samples, this would be irrelevant. The data from the 0D and 2D device would be identical. For XRR (X-ray Reflectivity) type of measurement the effect of any preferred orientation may not be observable with a 0D or 2D device as this phenomenon may be insensitive to changes in structure factor due to specimen morphology (random versus epi). Due to a lack of my experience in XRR type of measurements, I better stop sticking my big foot in my mouth any further. I shall read more. I shall perform such observations with the 2D detector and then comment further. In the examples I've used, the epi films are highly oriented and the Bragg reflections have been optimized for the to approximately coincide with the diffractometer axis (rocking axis) while examining the (hkl) set of planes. Coupled Omega-2Theta scan was used for these examples.
BTW What is the cost of a typical conventional 0D X-ray scintillation detector system these days? (U$D?) Including all the "bells & whistles" in the electronics.
Matteo! Please post or forward links to study "different incidence angles, different rotation of the specimen" in the diffraction mode. I'm very much interested in understanding this aspect a lot better. I'm reviewing the details on the other post as well. Thanx!
Ravi, be careful!!!!! this is not XRR.. if you talk about grazin incidence diffraction you are well above the total reflection angle as you talk about diffraction effects and not of total reflection!!!
Oh! Yes, Matteo. Thanks! This discussion helped me realize the idea of using shallow incidence angle for WAXS versus XRR. I have not tried either experimentally yet. I understand now the possibility of the "Whole Pattern" method and shallow angle of incidence. It is for WAXS, correct? There is some cross-over between XRR & SAXS, yes? I do concede to the "total reflection" versus "diffraction effects" as you have stated. Depending on the periodicities in the thin film the total reflection effects may extend beyond the 1-5 degrees mentioned earlier. I see some interesting and complex hetero-epitaxial structures. Some of this signal is "drowned" in the diffused scattering signal and more often than not neglected. I'm still learning.
Dear all,
The choice of detector always depends on the your experiment.
For instance, experiments performed on organic molecular films need long acquisition time but can be damage, therefore the detector choice is important.
The best way to obtain fast information of orientation is a quick collection using 2D. Of course, PILATUS is the best, but it does not have high resolution.
Moreover, the footprint increases the peak width in the 2D image.
Time and resolution can be improved by scanning the GIXRD using 1D detector with appropriated slits. You do not necessarily need synchrotron to do it.
However, I think that 2D detector can be used as point detector: by placing the detector at a distance such that the footprint is negligible, you can use the pixel resolution.
This geometry is suitable for specular and RC scans since you can obtain a 3D reciprocal map. Also XRR would benefit from this geometry: you can collected at the same time reflectivity and diffuse scattering, useful for in-situ and real time experiments.
Fabiola! "PILATUS is the best, but it does not have high resolution". Why? Are you able to use slit sizes smaller than the pixel dimension of the 2D detector. I'd like to understand the paradigm better. Thanx!
Does anyone know the "full frame" download rate for the PILATUS 2D detector?
In my opinion, with a 2D detector present and its ability to perform the 0D function, the conventional 0D point counter should have been obsolete. Why is it not? What drives its predominance still in present day XRD?
1. Cost?
2. Existing research?
3. Temporary ignorance?
4. Higher Sensitivity?
5. Other?
http://www.flickr.com/photos/85210325@N04/7890468142/in/set-72157632724090465
http://www.youtube.com/watch?v=IU0m4yI7D-k&list=PL7032E2DAF1F3941F
@ Ravi & Fabiola - the resolution of the Pilatus is good in terms of space (0.17 mm)^2 per Pixel, but the energy resolution is poor (you are only able to define a lower energy, and the filter is rather braod (around 500 eV FWHM from 0% acceptance to 100%), so you are only able to discriminte some lower energy background, but if you work slightly above some fluorescence line, then you are not able to reject the FY beackground ...
- In fact cost is an important point - you pay around 1 USD per pixel for a Pilatus ....
So 100 K means 100 000 bucks ...
Best regards to everyone
I could count on you to come up with something like that Dirk! 1 U$D/pixel, LOL!
Full frame download time frame? Please!
@ Ravi:
PILATUS is the best for collecting quickly a 2D-GIXRD image with high dynamical range, with respect to CCD.
However PILATUS has bigger pixel size (0.172 mm)^2 than that of CCD which is (0.05 mm)^2; therefore, using the same experimental set-up CCD has a higher spatial resolution.
NB: In the case of GIXRD you need to take into account of the footprint which decreases spatial resolution. You can overcome this problem by increasing the distance between sample and detector.
Makes sense Fabiola. What are the typical "full frame" download rates for your CCD and the PILATUS devices? I'm trying to estimate how fast one is able to grab a full XRD frame for your dynamic measurements.
"PILATUS has bigger pixel size (0.172 mm)^2" about 172um. Active area? $100k. 20 Bit.
"CCD which is (0.05 mm)^2" about 55um. Active area? Cost? DR?
"footprint" = FoV (angular detector Field of View of the reciprocal space) ....?
Do these devices allow you to create video data for time lapse imaging? What sort of software support is available for such adaptations?
Dirk! I notice you bringing up the "energy discrimination" issue. Even in monocromated and conditioned incident beams there is always an inherent "instrumental" spread. This is normally calibrated or deconvoluted using "known" standards. XRD measurements are almost always "relative". So, may I presume you are referring to energy dispersive applications.
Ravi, the Pilatus is a grid of single photon counters.. so expect quite good performance when used at its best.
With the toy you get a pattern in 2.7 ms @ 20 bit/pixel and 1M photons/pixel/sec. The frame rate is not super high (300 Hz), but well it's not bad either!
Wow! Does 300 Hz mean 300 fps? Can you use it in a real time video mode? Anyone has sample video data?
Is it 2.7ms (0.027sec) per frame? What frame size max?
Hello all,
I recently found this thread and it has been very helpful. I was wondering about the grains that GI probes. Some sources I've found say that GI sees both the preferentially oriented grains as well as the un-oriented grains. I wanted to get another opinion though, since I don't see why the oriented grains would be at a proper angle to diffract into the detector. As mentioned above, I think a GIXRD scan will only pull a signal from the crystal planes/grains that are at an angle bisecting the source and detector arms.
Context: I was doing GIXRD on an inorganic multi-layer film I believe to be textured, saw all of the material peaks I wanted including the peaks not present in BB geometry, but at very low counts regardless of my attempts to optimize the alignment/signals and I was wondering if the texturing is the cause. (Low signal is compared to a polymer film of similar thickness which I think should have a much lower scattering strength compared to the inorganics.) If possible, I'd like to use BB and GI to get as much information about the microstructure as I can.
Additionally, would it be useful for GI to first do an XRR and position the omega in one of the XRR troughs to take advantage of the "guided" mode of the thin films I want in order to get larger GIXRD peaks from that film layer? Or, another thought I had, would it be useful to include layers of materials with complementary refractive index to help confine the x-rays to the interesting material (provided the additional layers don't affect the subsequent layer microstructures of course)?
Dear Claudia ,
as you know similar to GIXRD ,in the XRR the incident angle is small and exit angle is large , but It is not possible observe x-ray peaks in higher angle in the XRR while in the GIXRD it is possible observe x-ray peaks in higher angle .why?
Good link!
Check out the PPT attachment as well.
http://www.iisc.ernet.in/currsci/jun252000/DUTTA.PDF
Abbas, no your statement "in the XRR the incident angle is small and exit angle is large , but It is not possible observe x-ray peaks in higher angle in the XRR while in the GIXRD it is possible observe x-ray peaks in higher angle" is wrong. In XRR both incident and exit angle are small, in GIXRD incidence angle is small but exit angle my be large!
Here is an example of XRR and "total internal reflection" for the NIST2000SRM that we recently worked on at MIT. We used a Bragg-Brentano parallel beam geometry. For the first time at least in my experience did I actually see the "totally reflected beam" along side the "main beam". It was an eye opening experience for me :-)
My question is, have any of you tried Rocking Curve Methods (Omega-2Theta coupled or Omega only decoupled rocking curve scans) with the grazing incidence angles? Any references in literature? For polycrystalline materials with random orientation, you could always find grains in Bragg condition for a multitude of incidence angles. However, with monocrystals this becomes a challenge. The grazing incidence conditions are restricted to specific (hkl)'s. So, detector and system sensitivity become paramount!
https://www.flickr.com/photos/85210325@N04/15563390403/
https://www.researchgate.net/post/What_is_the_best_way_to_interpret_3D_reciprocal_space_data_from_XRD_rocking_curve_analysis2
http://www.iaea.org/inis/collection/NCLCollectionStore/_Public/28/044/28044496.pdf
Great resource to include in your XRD colleaction! Cut/paste into Google for best results.
Reviewing this thread after 4 years is interesting. I've learnt a lot more since. Thanks to all contributors!
Check this out! You cannot do this with a point counter over the next decade:
https://www.researchgate.net/post/How_do_I_figure_out_the_spatial_thickness_of_the_individual_layers_from_these_well_resolved_diffractograms?
"The choice of detector always depends on the your experiment." Why should it be? Given the choice, I'd always choose 2D, best spatial resolution and best dynamic range.
"
Dear Alwin,
Thanks a lot for this interesting question. The proposed reply of Claudia is almost complete. I agree with Claudia.
With my best wishes!
Dr. Adel OUESLATI
"What is glancing angle incidence XRD, how is it used to study thin films? What are the properties that can be studied using G-XRD"
The principal reasons for using small angles of incidence is shallow penetration depth and expanded incident beam "foot print" (larger area coverage). In the case of thin film XRD, that would increase the signal strength from the "film" relative to the entire penetration depth (diffracting volume).
If after all that trouble you then in the name of "cost" decide to "smudge out" all the advantage by using an "archaic" 0D fancy Geiger counter, then I'd encourage you to think of the alternative 2D spatially resolved detectors even for your dinosaur 2kW sealed tube unit.
See this example for a hetero epitaxial sample of GaN-AlxGa1-xN-AlN on Si:
https://www.researchgate.net/post/How_do_I_figure_out_the_spatial_thickness_of_the_individual_layers_from_these_well_resolved_diffractograms?
Glad you're still following with interest Adel . I'll be in Alwin's town in February 2018 showing a bunch of "eager to learn" folks the dramatic advantages of high resolution real time 2D XRD imaging and analyses for thin film Nano structural characterization. Including OAT, Oblique Angle Topography:-)
Alwin! I see no XRD facilities in your university. How come?
http://www.siet.ac.in/index.php
There is such a preponderence of information here from so many experts that I've decided to save this discussion as a PDF file. Use as you wish!
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science