I have seen a lot of reports which show that the particle size observed in SEM is larger than the crystallite size observed in XRD, proving the agglomeration of particles. I was wondering if it is possible to observe an exact opposite phenomenon, vis-a-vis the particle size observed in SEM is smaller than the crystallite size obtained from XRD. If it can be observed, then what can be the possible reasons for the phenomenon.
Sure it can happen, operators errors are pretty common, but theoretically it's impossible. SEM gives real size (if there are no agglomeration), but XRD gives either a right one, or shows size smaller than it should be ( Scherrer equation is rather a rough approximation).
Sure it can happen, operators errors are pretty common, but theoretically it's impossible. SEM gives real size (if there are no agglomeration), but XRD gives either a right one, or shows size smaller than it should be ( Scherrer equation is rather a rough approximation).
Can you tell me, how much is difference in the crystallite and particle sizes? If you have recorded SEM images after gold coating, there is always possibility to get nanoparticles of gold in the order of 10-15 nm.
Dinesh Kumar Sorry sir, I was just asking about a hypothetical situation. I didn't actually receive any results in such manner. Thank you for your suggestion.
A scanning electron microscope often images objects in two dimensions. i.e. it is not always possible to accurately size a third axis. Crystals are three dimensional objects. So, it's possible that the true crystal size is greater than the SEM image suggests e.g. if the crystal presents its shortest dimensional axis.
Peter Stacey , do you really suggest to determine particle size by measuring only one particle?
SEM image shows particle size. But using XRD peak, Scherrer formula gives crystallite size. Particle size and crystallite size are not same.To the best of my knowledge, crystallite size should be smaller than that of particle.
Md. Moazzem Hossain , you are right, bigger particles usually consist of many crystals, and in that case SEM measured size will be unquestionably higher than XRD data. But for nanoparticles in most cases particles and crystals are the same thing.
Vladimir Dusevich Sir, I think that Peter Stacey sir is pointing out the situation in which columnarly grown cuboidal or cylindrical crystal structures are grown which have smaller height compared to the length and breadth or diameter, thus resulting in larger observed particle size. This is all still a speculation though.
Peter Stacey Thank you for sharing the possibility sir.
Can we call " columnarly grown cuboidal or cylindrical crystal structures" particles?
Vladimir Dusevich Sorry, sir. I was referring to "columnarly grown cuboidal or cylindrical particles", e.g., Carbon Nanotubes or other prismatic particles created through electrospinning, laser ablation, CVD etc.
CNT will not attach themselves to specimen stubs in "stand up" position. So, no problems with SEM measurements.
Vladimir Dusevich Yes sir, I agree that there is an absence of practical situations for this problem to be real. It took me some time to get the flaw in the explanation. I understand now that it is theoretically not possible for a sample to have a larger crystallite size than particle size.
I thank everyone for their enlightening points.
Dear all,
may I add two important remarks:
1. The crystallite "size" seen in SEM-pictures is a two-dimensional cut through a three-dimensional structure. So, the observed "particle / grain size" is not the real one, but a section through 3-dimensional objects. The real diameter in 3 dimensions can be calculated under the assumption that the particles / grains are approximately spherical.
However, you have to assume additionally that you imaged in the SEM a representative part of your sample. This may not always be the case.
2. XRD is on the contrast an integral method, i.e. it measures sample properties in a quite large volume and averages over this volume.
The determination of grain / particle sizes by XRD also involves some assumption and mathematical treatment of peak broadening causes by small (< 1micron) crystallite / particle sizes. To my knowledge, XRD does not distinguish between particle and grain sizes. So, basically the number you obtain is the particles size only if the particles are single crystals, which is often not the case.
Please, do take the numbers given in publication not as so accurate as they often are presented: there are a lot of sources of errors as described above.
With best regards
Torsten Staab
You can check the following publications on this topic:
@ARTICLE{staab:1998a,
AUTHOR = {T.E.M. Staab and R. {Krause-Rehberg} and B. Vetter
and B. Kieback},
TITLE = {The influence of Microstructure on the Sintering
Process in Crystalline Metal Powders
investigated by Positron Lifetime Spectroscopy:
{Part~I}: {Electrolytic} and Spherical Copper
Powder},
JOURNAL = JPCM,
YEAR = {1999},
VOLUME = {11},
NUMBER = {7},
PAGES = {1757--1786},
KEYWORDS = {} }
@ARTICLE{staab:1998d,
AUTHOR = {T.E.M. Staab and R. {Krause-Rehberg}
and B. Kieback},
TITLE = {Positron Annihilation in Fine-Grained Materials and Fine
Powders --- an Application to Sintering of Technically used Metal
Powders},
JOURNAL = {J.~Mater.~Sci.},
YEAR = {1999},
VOLUME = {34},
NUMBER = {},
PAGES = {3833-3851},
KEYWORDS = {} }
crystalite size refers to the grain boundary of any particle. thus in any particle so many grain boundary present. If the particle contain only one grain boundary then the size is equal to the crystalite size. but it is very rare case, always crystalite size is less the size of the particle obtained from SEM image
Torsten E. M. Staab , the question was about particles, not about bulk material
Ikbal Bahar Laskar , it is not " very rare case", since in most cases particle size and crystal size are the same thing for nanoparticles.
Dear Vladimir Dusevich,
it does not matter: you can embed powder particles and then make a metallographic picture by grinding and polishing, i.e. looking at a section - the same what you do with a "bulk material". In fact my publications are about powder metallurgy: how to come from a powder to a bulk material. There particle size where between 1 and 20 micron, while grain sizes were sometime smaller than a few 100nm.
Dear Bansi Lai,
"crystallite" and "grain" are used in a synonymous way, since the interface between one crystallite and the neighboring crystallite is typically called a "grain boundary".
However, you have to distinguish between "particle" and "crystallite/grain". Particles may be polycrystals (mostly metals - NOT nano particles in the range of 100nm or below) or single crystals (mostly ceramics).
Consider strain in your calculations as it is very much important and necessary for nano particles.
In other words don't just fall for a simple scherrer's formula but use Williamson Hall method to find your crystallite size.
A small error can be justified as an operational error so check out the relative error
Torsten E. M. Staab Thank you for your remarks.
Sachin Kumar I always try to use Williamson Hall method as much as possible. For me it provides the best result for crystallite size. Although, we are talking about a made up situation here with no practical data to support. Thank you anyway for your suggestion.
One of the Williamson-Hall methods for determining the crystallite size and induced strain could give a crystallite size similar to that of the particle size exhibited by micrographs (e.g. TEM).
Anish Bhattacharya Ideally it is not possible. in that case deal with the errors. Mention the error bar for crystallite size (take standard deviation of your linear fit of Williamson Hall plot) and note down the SD of your SEM particle sizes. If they overlap then you are pretty safe and can report this.
If this doesn't happen then check out other possible errors from your XRD data. recheck the shift in the positions of all peaks relative to your JCPDS card. XRD scan rate should not be too fast.
On the basis of general principles，xrd crystallite size can not be larger than sem particle size.
Sachin Kumar Thank you for the suggestion. Although, it seems like a desperate attempt to make sense out of misleading data.
Anish Bhattacharya What else can be done other than putting charge on experimental errors if you get something ideally impossible stuff. Have a good day
Instead of particle the term (primary) grain should be used for SEM images. The normal case in evaluating SEM images is measuring the diameters of the grains and averaging all values. This means calculation of a number weighted mean diameter. A big particle has the same weight / influence to the mean value as a small grain. For example: if 100 grains exist with diameters of 1 (99 grains) and 10 (nm, mm, ..) (1 grain) then the mean value is (99*1+1*10)/(99+1)=1.09.
If a grain consists of crystalline matter then the crystallites (coherently scattering domain) may be of the same size or of a smaller size. A more detailed discussion of the different particles (domains, crystallites, crystal, grain ...) may be found here (https://www.researchgate.net/post/What_is_the_difference_between_crystallite_size_grain_size_and_particle_size). The crystallite size usually will be determined as the volume weighted mean column height Lvol. The mean column height of a single sphere of the Diamater D is Lvol=3/4*D. There are two reasons (L
Robert Haberkorn Thank you sir for your valuable input. It cleared a lot of my confusion.
Robert Haberkorn
"Particle" is a usual term for SEM. Grain - it's not for particles, but for metals and such (but for these grains "grain" and "crystllite" are mostly the same thing). Just to clear some confusion.
Dear Robert Haberkorn in volume average crystallite size (Lvol) how this factor of 3/4 is coming and its importance. ( The mean column height of a single sphere of the Diameter D is Lvol=3/4*D. ) What about surface weighted mean column height (Lsurface) and sphere diameter D. Thanks.
Dear Ashok Bhakar there is an extensive literature in this area from textbooks by B.E. Warren to publications from authors including Langford, Scardi, Haberkorn, Balzar etc that can be consulted.
Dear Anish
Have you observed crystallite size greater than particle size in your experiment ? If yes, this is not expected. Experimental errors may be responsible for such observation. Minimum size of a particle can not be less than size of one crystallite.
Dear Anish,
please pay attention 1) at the definition of the length measured (volume-based? area-based? weigth-based? giration? etc.) and 2) the uncertainty that must/should be associated to the XRD estimate.
Regards
Dear Anish,
Particle >= crystallite;
crystallite could not be bigger then particles.
Best regards!
Grain and Grain boundary can be appeared clearly in details by TEM but you can not differentiate between the agglomerated particles or the grown one without any deformation .
Just to clear some confusion.
SEM images shows real particle size with compare to the XRD. Some times the instrumental error plays major role in XRD to calculate particle size through XRD. But some researchers shows the XRD particle size is smaller than SEM.
Practically and theoretically, it cannot be. SEM gives the agglomerate size by visualization but XRD gives the atomic orientation as grains which is called "crystallite size". If you calculate or see SEM smaller than XRD, that means you have put wrong scale to SEM or XRD data was interpreted completely wrong since SEM gives the agglomeration but not the grains in the same particle.
Hi
In general the crystallite size will be smaller than particle size. If it is a single crystalline then the particle size may be equal to crystallite size other than this the particle size will always higher since we have most probably poly grain particles.
Best
T. H. Al-Noor The link is not working. Please send an updated version.
I agree with Ali Özerw and Venkata Ramana Mudinepalli.
particle size by SEM should always be bigger than crystallite size by XRD.
The XRD technique requires several corrections and deconvolutions to reach qualitative particle size values: the presence of Kα2 radiation and instrumental line broadening have to be taken into account. The Williamson-Hall approach takes the strain effects explicitly into account works better for the largest sample Domain size. To separate strain and size broadening effects, a Gaussian deconvolution gives the best results.
Grains as far as material science is concerned refers to that volume of material within which the crystal structure and the orientation of the crystals are same. At the grain boundaries, you have a change in orientation of the crystal.
Particle size, on the other hand, is a macro phenomenon. It is the size of the particles which may have 1 or several grains within it. It is impossible to have a grain having multiple particles because the smallest thing we could have would always represent one single grain.
Ref: Article The Big Problem of Small Particles: A Comparison of Methods ...
- Badges
- Science method
- Similar topics
- Methodology
- Crystallography
- X-ray Diffraction