Carlos, it all depends what you want to do with the values you obtain! If you want to play bingo, you can use any type of data.
This said, have a look at other similar threads on the same topic (domain size analysis) and the comments (a lot of them also mine) on Scherrer equation. In the end the intensity always count but believe me, I have measured domain sizes with a very low intensity. What is important is the signal/noise ratio but for a nanomaterials you can easily get a good one as you can collect data with a large 2theta step and count the points more.
@Venkat: my suggestion is that you have a look at the literature and be sure of what you write, to avoid people getting severely confused. Especially if you are not expert in the field. The Williamson-Hal approach is definitely a step forward with respect to Scherrer equation, but it bears the same limits. It cannot handle dislocations unless you modify it (so-called modified Williamson-Hall plot). However it can be demonstrated that also in that case reducing a whole peak profile to a single point is definitely not a great idea! And you can't get information on the size distribution that is usually an important information for people working with nano-sized powders, right?
'' if u want tell crystallite size only from low intensity then its not correct bcoz amount in that sample is very less.''
Thank you for your answer. you answered clearly to my question in the last part.
Carlos, it all depends what you want to do with the values you obtain! If you want to play bingo, you can use any type of data.
This said, have a look at other similar threads on the same topic (domain size analysis) and the comments (a lot of them also mine) on Scherrer equation. In the end the intensity always count but believe me, I have measured domain sizes with a very low intensity. What is important is the signal/noise ratio but for a nanomaterials you can easily get a good one as you can collect data with a large 2theta step and count the points more.
@Venkat: my suggestion is that you have a look at the literature and be sure of what you write, to avoid people getting severely confused. Especially if you are not expert in the field. The Williamson-Hal approach is definitely a step forward with respect to Scherrer equation, but it bears the same limits. It cannot handle dislocations unless you modify it (so-called modified Williamson-Hall plot). However it can be demonstrated that also in that case reducing a whole peak profile to a single point is definitely not a great idea! And you can't get information on the size distribution that is usually an important information for people working with nano-sized powders, right?
Dear Matteo Leoni, can you explain more. I did not catch your ideal completely.
It would be very kind of you.
Best regard
Carlos tell me what is it not clear:
- signal/noise ratio is the important parameter
- collect few points but collect them well (for a powder in the nanometers range you can collect pattern with 0.1 deg step size without losing information)
- be sure of the hypotheses behind a formula before using it (e.g. Scherrer formula is not always applicable, especially quantitatively!!!!!)
- check the relevant literature/conferences
Hi,
I have doubts with signal/noise ratio is the important parameter. How I can do this, and what is the appropriate ratio...these questions remain in my mind.
Thank you for answering!
For me the interesting question depends on the participants besides the contents. This one is! I'm still checking it out. I'll be back!
As usual Matteo is on target!
SNR - Signal-to-noise Ratio - Is one of the most critical parameters in most XRD measurements. This determines both contrast and resolution. I'll post some examples of XRD profiles (diffractograms & rocking curve profiles) illustrating a variety of SNR data. In general, as the SNR deteriorates so do both contrast and resolution. I'm sure there are better explanations on Wikipedia.
There are generally two types of back-ground signal (amended from "noise") in XRD (at least):
1, Dependent on 2-Theta such as absorption, ambient, Compton etc.
2. Independent of 2-Theta - Time dependent detector scintillation noise.
"Noise is the electronic one plus the spurious radiation entering the detector plus cosmic rays". (as correctly suggested by Matteo)
Both these factors need to be quantified and deconvoluted in order to correctly analyze X-ray diffractograms.
Don't forget, data present in the XRD pattern is a measure of the Nano structure of the sampled VOXEL. Other than those mentioned above. So, with the right detector sensitivity and numerical tools, it can be resolved. Inability is only temporary!
@Ravi, glad to see you here. be careful in what you write as words are important (my scalpel is always ready..)
"There are generally two types of noises in XRD (at least):
1, Dependent on 2-Theta such as absorption, ambient, Compton etc.
2. Independent of 2-Theta - Time dependent detector scintillation noise.
Both these factors need to be quantified and deconvoluted in order to correctly analyze X-ray diffractograms."
NO ! Noise is the electronic one plus the spurious radiation entering the detector plus cosmic rays (that give extra counts). Of course in some cases the noise from the detector can be virtually zero (it depends on the detector). Plus you have the counting statistics that in the end contributes to the noise as well.
Compton is inelastic scattering (so it is signal), absorption gives a decrease in intensity plus anisotropic broadening. Among them the only quantity that you deconvolve is absorption. Compton can be subtracted, whereas the noise.. remains there (so be careful about using the word "deconvolution")!
Background is for sure another issue, but I assume already that you are reasonably able to model the background or that you are using a zero background holder. With good microstructure models, the modelling of the background becomes less critical. Taking the background into account can be tricky in several cases, but I find it more critical to have a signal/noise ratio of 1:10 versus a signal/background of 1:10 (Ravi yes I mean it as I write it as I talk about critical)
Ready for your constructive surgery any time Matteo! I need the lobotomy for some of my other misconceptions anyway.
Good news is that I shall amend my errors. I completely agree with you. I should have used the term background signal rather than "noise" as I always do and recommend others to do as well.
Maybe I have the SNR definition different from you. I'll figure it out. Would you mind defining it please, as you have used it? Thanks!
SNR (Wiki): http://en.wikipedia.org/wiki/Signal-to-noise_ratio
Here are examples of:
1. Low signal to noise ratio diffractogram:
Figure 1A: http://www.flickr.com/photos/85210325@N04/7988760159/
Figure 1B: http://www.flickr.com/photos/85210325@N04/7988731098/in/photostream/
2. High SNR diffractogram:
Figure 2A: http://www.flickr.com/photos/85210325@N04/7944785794/
Figure 2B: http://www.flickr.com/photos/85210325@N04/7893352226/
Figure 2C: http://www.flickr.com/photos/85210325@N04/7975835813/
These images should be self explanatory. I have to run out now, but I will explain the images in greater detail later based on interest.
http://www.youtube.com/watch?v=IU0m4yI7D-k&list=PL7032E2DAF1F3941F
I agree with Matteo that the S/N ratio is an important parameter that should be considered when measuring crystallite size with Scherrer equation. For example, at the first time I used this equation to calculate the size of nanodiamond particles (with a low SNR), the result (~3nm) was 1nm smaller than the real size (~4nm) obtained by TEM observations. The second time, I increase the SNR by preparing a better sample for XRD analysis, the calculated result (~3.5nm) was just 0.5nm smaller than the real size.
Hi Ravi, thank you for your question.
I mean at the second time, I prepared a sample for XRD more carefully. I tried to cover the substrate (carbon tape) by nanopowder as much as possible but still kept the roughness of powder layer very limitted. By that way, I could reduce the SNR of XRD analysis.
I am sorry, not "reduce", I could " increase"the SNR of XRD analysis
SNR - Clarification! Here's how I use it for XRD data. Do I need to reconsider?
Figure 3A: http://www.flickr.com/photos/85210325@N04/9730898656/
The sample in this case (same as Figure 1A from earlier) is a 2mm thick composite ribbon of propellant formulation containing RDX, Al powder, CAB polymer (amorphous) binder and an Al foil calibrant. I've indicated what I consider as signal, noise and back-ground. That is how I always get a value larger than 1 for SNR.
In this case a conventional 1D diffractogram (using a 0D point/scintillation detector) in the reflection mode with Cu K Alpha (Ni filter) & Rigaku MiniFlex is contrasted with the 2D diffractogram which is in the transmission mode all else identical. Big difference!
http://www.flickr.com/photos/85210325@N04/9730898656/
What reflection were you using for the Scherrer Equation? What was the assumption for the grain size and shape?
I think everytime we use the Scherrer equation, we must assume that the shape of particle is sphere or cube or something, then we assign a corresponding value for the shape factor K. For example, in my experiment, I assumed that the shape of nanodiamond particles is quasi sphere, then I assigne the value of 0.9 for shape factor K.
http://www.physicsforums.com/showthread.php?t=327576
After obtaining the size by Scherrer equation, I confirmed these results (size and shape of particles) by TEM observation.
In my opinion, I think Scherrer equation is just a tool for guessing the size of particle. If we want to confirm the size, we should use high-resolution microscopy techniques like TEM or something.
@Son: have a double check and a deep thinking about what you are doing and what you wrote (especially in the last sentence). You take a formula that gives you a "size" based on the analysys of a large fraction of your specimen and then you want to confirm the result by analysing just a few domains with the TEM... In my opinion the opposite should be done i.e. get the information of a few domains using TEM and then validate it on a larger scale using apropriate analysis techniques.. With TEM is quite likely that you see what you want to see (you just look for isolated or not highyl packed domains).. whereas with diffraction is hard to cheat. Or am I missing something?
Scherrer equation is a great tool to compare specimens (especially if they are within an homogeneous class). But beware, if you find the same result with TEM and with Scherrer equation then (and this is what people keep ignoring) you just demonstrate that your result is WRONG! The Scherrer value and the TEM mean size should not be coincident. It can in fact be demonstrated that in a specimen with a size distribution (i.e. in every specimen you can possibly analyse) Scherrer equation gives you a ratio of high order moments of the distribution (4th over 3rd). What you usually extract from TEM data is the mean i.e. the first moment of the distribution. Unless your distribution is a gaussian (impossible as it would give a non-physical probability of
Matteo! Another awesome explanation indeed. Thanks! Seems like "deja vu" all over again. I'm re-learning. Your students are particularly fortunate to have your tutelage just as we were with the late Prof. Sigmund Weissmann @ RU in the 1980's.
Subjective assumptions, as we are discussing here, are at the "root" of the inconsistencies with the XRD results from various investigators. Once we as experts introduce some consistency in our method, we will be able to compare our results better. As Matteo points out, the dramatic difference (orders of magnitude) in the sampling volumes between TEM & XRD cannot be ignored. Besides, the destructive sample prep in the case of TEM has its own contribution.
The next basic issue is the understanding of the difference between "integral breadth" and FWHM pertaining to the Scherrer formula. In the simple case of a Gaussian distribution assumption (seldom is the case, as Matteo rightly opines), the integral breadth, β, is related to the FWHM peak width, H, by β = 0.5 H (π / log e 2)^1/2
http://pd.chem.ucl.ac.uk/pdnn/peaks/gauss.htm
This link has computations for other possible distributions as well.
"kept the roughness of powder layer very limited" - How did you accomplish that? Pressing? Have you eliminated the effect of any possible preferred orientation? Or at least tested for its presence? In the reflection mode, you may do this by repeating the scan after rotating the sample surface w.r.t. its surface normal. I follow this step routinely for powder samples using the reflection mode. In the transmission mode, it is trivial to detect "preferred orientation" with a photographic film or 2D area detector. A 2D detector would reveal such preferred orientation in the reflection mode as well. In the transmission mode one is able to capture the entire 2D diffraction pattern and thus have a more complete depiction of the effects of any preferred orientation as we are capturing the entire diffraction "cone" in forward diffraction. However, the sample thickness will be the limiting issue in the transmission mode.
Have you considered using an internal standard? Known standard!
Dear Matteo! Thank you very much for your valuable comment, I have learned alot from that. Maybe you misunderstand my opinion. We know that XRD is a technique more simple than TEM. When we have a new unknown material, or powder, I think the first thing is to check it with XRD. With XRD for 5 or 10 min we can indicate what the powder is, and at the same time we can measure the size of particle using Scherrer equation. With TEM, we need longer time to indicate the powder (by EDX), and much longer time to get the distribution of particle size as you said. So, here is exactly what I did. I checked the powder with XRD first, and calculate its particle size by Scherrer equation. After that, I observed it with TEM, not only for measuring particle size, but also to confirm their shape and the size of their agglomerates. I repeated the observation many times to get the most accurate size distribution of particles. Then I compared it with the result obtained by Scherrer equation. Is there any problem with my procedure?
Dear Ravi! I "kept the roughness of powder layer" just by using a special silicon holder for XRD. The holder has a small pit that user can fill up with powder. I tried to fill up the pit with powder as much as possible, but kept the surface of the filling-up powder not higher than the surface of silicon holder. I hope that you can understand my explanation. In addition, in my research I just used XRD to know the idetification of the powder so I did not think alot about preferred orientation of the sample. Thank you for your advice.
I think Matteo's point was that the statistical certitude of XRD measurements is much higher than the TEM, when properly used, due to the large XRD relative sampling volume in comparison to TEM.
That leads me to the next question Son Nguyen. How did you prepare the TEM sample? (Just curious, since I have no clue how it is done with loose Nano powders). Did you have to use a matrix/binder of some sort? What material system are you studying?
I am confuse. I will read it slow, and then respond calmly.
Thank you for ansering!
Carlos! Please post your low intensity XRD data and details of material morphology as you know i.e., expected size distribution, expected shape of particles/grains. Any TEM or SEM data? Even with low intensity you should still be able to detect relative differences between samples if such differences exist. However, the absolute value calculation might need more rigor.
Hi Ravi! My TEM sample was prepared just by diluting powder with acetone (using ultra sonication) and then dropping onto the microgrid. I think that people usually prepare their TEM sample by this way, is that right?
I plead ignorance Son! Please post a link and photos (when possible) so I may learn more. What is the "microgrid" made of? Can you use that sample on the microgrid (in situ) for XRD? Positioning it in the X-ray beam may be the biggest challenge. It would be a lot easier in the transmission XRD mode.
Carlos! Is this the material you are working with? Interesting stuff! Certainly send me some samples if available and when convenient. I'll be glad to examine it for you.
Structural and magnetic properties of ferromagnetic/ferroelectric multilayers.
M. Sirena, E. Kaul, M. B. Pedreros, C. A. Rodriguez, J. Guimpel, L. B. Steren
ABSTRACT: The La0.75Sr0.25MnO3 (LSMO)/Ba0.7Sr0.3TiO3 (BSTO) superlattices and bilayers, where LSMO is ferromagnetic and BSTO is ferroelectric, were grown by dc sputtering. X-ray diffraction indicates that the samples present a textured growth with the c axis perpendicular to the substrate. Magnetization measurements show a decrease of the sample’s magnetization for decreasing ferromagnetic thickness. This effect could be related to the presence of biaxial strain and a magnetic dead layer in the samples. Conductive atomic force microscopy indicates that the samples present a total covering of the ferromagnetic layer for a ferroelectric thickness higher than four unit cells. Transport tunneling of the carriers seems to be the preferred conduction mechanism through the ferroelectric layer. These are promising results for the development of multiferroic tunnel junctions
If so, you would have to be interested in the preferred orientation characteristics of these "poly-crystalline" not "epitaxial" thin films. I've requested the full text of this paper to examine the XRD data from it. Please send me the PDF version to: [email protected]
Here is an example of a similar 50nm super-conducting MBE epi film of LaSrCuO4 on LaSrAlO4 (001) substrate from BNL (Condensed Matter Group - Dr. Ivan Bozovic):
http://www.youtube.com/watch?v=FZpsgNXGHsc&list=PL7032E2DAF1F3941F
I suggest you should be acquiring the 2D XRD pattern instead of the archaic 1D Linear Diffractogram with a 0D scintillation/point counter in order to evaluate the "pole figures" (preferred orientation). Here is an example of a highly oriented rolled Aluminum foil (0.2mm thick) displaying real time pole figure acquisition capabilities using the transmission geometry:
http://www.flickr.com/photos/85210325@N04/7944785794/
http://www.youtube.com/watch?v=FZpsgNXGHsc&list=PL7032E2DAF1F3941F
Hi Ravi. Here I attach an optical micrograph of the micro-grid I used to prepare TEM sample. It is made from Cu, and unfortunately it can not be used for XRD. Regards
Why not? The Cu would be a great internal reference. What are the dimensions? Show the whole sample as it goes in to the TEM please (grid & powder). Put something in there for scale when you take the photo. How does the Nano powder remain in place? Do you etch the sample?
Dear,
thank you very much for discuss this question, it is very interesting to me and I can get knowledge.
Until now, I have obtained information from your answers in order to continue working. Really, thank you very much.
@Ravi Ananth, Structural and magnetic properties of ferromagnetic/ferroelectric multilayers, was just a colaboration. I work in semiconductors thin films to solar cells applications. Currently, I am working in ZnS (zinc sulphide) grown by chemical bath deposition. I have XRD patterns and TEM images (including electron diffraction).
In this moment I can not send to you any data, due to I am in vacations (I do not bring nothing academic material with me!), I will come back 23 sept. Please, wait for me!
PD: From your answer, I know that I can do (or can not do!)
Dear Ravi, in my laboratory, XRD and TEM samples are contained by different holder with different preparing methods. For TEM, we dilute powder in solvent and drop them onto the grid that I attached before. Then the grid is put into the holder and kept from falling out of the holder by a lid. In case of XRD, we just fill up a small pit on the holder (that I am gonna attach a photo here). Then we press the powder by a glass slide or something to make them attach to the pit. So, I think I can not use TEM sample for XRD measurement, though the dimension of TEM grid and XRD pit are the same (around 30 mm in diameter). Regards.
Dear, really I am sorry for the delay. I am very busy, but I Will update de XRD patters.
Best regards
Son! Post photos of the XRD diffractometer and sample stage (sample holder in there) to get the best suggestions. Think out-of-the-box!
http://www.flickr.com/photos/85210325@N04/10659886546/
http://www.flickr.com/photos/85210325@N04/10659886546/
This is the XRD patterns of my thin films. Really I want to know what i can say with this information.
Thank you
Carlos! I, as several others have, suggest you get a firm understanding of the principles of XRD in order to "know what i can say with this information".
One of my favorite prescription for this ailment, among other books, is "X-ray Diffraction In Crystals, Imperfect Crystals and Amorphous Bodies" by Andre Guinier. 1962 - Great book! Worth acquiring for your XRD collection.
(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."
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 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 0D XRD data/results.
I'd also recommend you review this "Friday Evening Discourse at the Royal Institution", Professor Stephen Curry - See more at: http://richannel.org/seeing-things-in-a-different-light#!
I'll be adding more comments regarding the data that you have posted soon as well.
http://www.flickr.com/photos/85210325@N04/10515372183/
Thank you very much for your comments and suggestions, were very useful to me. I will buy a book of XRD, some friends recommended me 'Cullity' too.
Moving on, I show you the XRD patterns of my samples due to I think that the peak intensity is very low and I am trying to identify the crystalline structure. Maybe this information must be complemented with other measurements?. What do you think?
Comments & suggestion regarding your data:
1. Your challenge is not just "low peak intensity" but poor SNR (signal to noise ratio) as well. Use max kW (kV & mA) allowable. Increase dwell time. Maybe only in the "regions of interest" (ROI) on the diffractogram. Examine the cause and presence of the apparent higher noise in the 8h sample. Improve angular resolution in ROI.
2. Post the XRD scan for the as-received (0h) sample. Important to identify the various components in the XRD scans. The uniquely distinguishable shapes of the various components in the XRD event are obvious from your scans. The key is in correctly normalizing the data and being able to effectively compare the various XRD scans.
3. Post the raw data in Excel format for the best advise. I may then be able to show you how to handle it in the future autonomously. Or, send it to me at [email protected]
4. Post the diffractogram without the vertical shift as you did before. Magnify the regions of interest for easier viewing & ease in comparing.
5. Examine the potential effects of the presence of "preferred orientation" in your sample. You may do this by rotating the sample about its surface normal w.r.t. the incident beam direction and repeating the diffractograms. If invariant then RANDOM orientation. I also suggest you obtain a transmission Laue image from your sample in forward reflection to clearly see any presence of preferred orientation. You may use a photographic film for this purpose.
6. I suggest you review the attached image of conventional diffractograms. This is an example of the use of "internal/external" standards to help compare various diffractograms effectively. The samples include ones similar to yours with amorphous and crystalline components. Here we used the Aluminum sample holder as our invariant standard in each diffractogram: http://www.flickr.com/photos/85210325@N04/8434970459/
7. Forget the Scherrer equation for your case for now. You'll be able to compare but not quantify (absolute) yet. Only relative values. You would need known standards to calibrate.
http://www.flickr.com/photos/85210325@N04/8434970459/
BTW Son! Where are your Aluminum holder lines? Was you beam smaller than the sample diameter? If so, open it up and get the Al lines for reference if possible. You may also consider mixing a known inert material in your powder for reference.
XRD is fun & precise, enjoy! Learn how! Femtometer resolution capability in reciprocal space using the rocking curve method! http://www.flickr.com/photos/85210325@N04/10221987094/
http://www.flickr.com/photos/85210325@N04/10221987094/
I realize the expertise of RG members interested in this discussion is exceptionally high by your scores, writings and profiles. So I'm going to dare to elicit your help with the analyses of some of our real time experimental findings from BNL and AFRL. Please feel free to jump in help me out.
1.https://www.researchgate.net/post/What_are_the_causes_for_ASYMMETRY_in_the_Bragg_X-ray_Rocking_Curve_Profile_RCP_for_a_symmetric_004_GaAs_reflection
2.https://www.researchgate.net/post/X-ray_Rocking_Curve_Analysis_of_Super-lattice_SL_Epitaxial_Structures_What_are_some_of_your_practical_experiences_with_this_method
You can knock me down a couple of notches if needed, no problem. But, I'm finding near perfect match with theory. I'm unable to find similar results elsewhere in literature yet. Am I the only one? There's got to be someone else besides me in the world that has thought of this. I'll post this in a few other discussions with different set of experts. Hopefully we could bring together a "Master Mind Group"!
Theory meets Experiment!
http://www.flickr.com/photos/85210325@N04/9430820747/in/set-72157635172219571
Dear Ravi Ananth,
Thank you, thank you, thank you very much for your help. It was very usefull and I will send to you the data. However, carry out new measurements is really complicated to me due to in my country is complicated to do somethings like that you suggests, because of the equipments are available to use only in one conditions and those that are available for more precise investigations are busy ... would have an appointment to 1 more year.
Related to your last comments, I don't understand! my language native is Spanish so sometimes I can´t get the ideas expressed in English.
Best regards
Carlos! You are welcome. I have done nothing spectacular yet. Join the group and share. Click and find out more. The rest of it was "fluff"! Don't worry.
My daughter speaks and writes perfect Spain Spanish (book Spanish) as well. I'll have her translate exactly and post it later, when possible.
Dear Dr. Carlos Rodríguez
Have a nice day
In short answer, yes you can. This article may help you in your study.
https://www.researchgate.net/publication/277952808_Final_-_Microstructure_and_crystal_imperfections_of_nanosized_CdSxSe1-x_thermally_evaporated_thin_films_-_final_form?ev=prf_pub
Thank you and good luck
Data Final - Microstructure and crystal imperfections of nanosize...
- Badges
- Science method
- Similar topics
- Methodology
- Crystallography
- X-ray Diffraction