First, understand the difference between particle size, grain size and crystallite size

Particle size could be either an agglomerate of many grains/crystallite or single grain/crystallite. For example, most of the case, the powders produced by mechanical alloying route will have particles consist of many grain whereas powders produced by chemical route will have an individual particles made up of single grain/crystallite. If you are sure that your individual powder is not an agglomerate of many grains, than you could use results from XRD and support it with SEM/TEM, provided you don't have any other planar defects (read the below paragraph to know why).

Most of us, use the term crystallite size and grain size interchangeably; but there is a slight difference between crystallite size and grains. For example, crystallite size from XRD and grain size from TEM/SEM will be almost similar only if the material of your interest don't have subgrains, stacking fault , twins, or any such planar surface defects.

Secondly, to be very accurate, what we calculate from XRD is "Coherently Scattering Domain Size" which includes crystallite size, stacking fault, twin and so on; and most of the time, we assume that domain size is equal to crystallite size by ignoring the other contribution to domain size. This will be wrong, if you have other planar defects in your material.

I should also emphasis, all the methods used to calculate Size or Strain from XRD pattern are based on analytical models except Warren Averbach method (fourier and stokes deconvilution method) and double voigt method. By using analytical model, we are biased to one particular profile (Pseudo-voigt/Pearson and so on) and results from these methods will never be accurate . . [ Read this article "Voigt-function model in diffraction line-broadening analysis -Davor Balzar"

www.du.edu/~balzar/Documents/IUCRbook.pdf ]

In my opinion, to find out the particle size, use particle size analyser / SEM /TEM than XRD technique.

Hi shrikrishna,

XRD can only measure particle size if and only if each particles are monocrystalline. This is by far not case with most of the inorganic materials.

SEM and TEM are both very accurate techniques if used carefully and rationally. However, I want to rise one point. A paricle is a 3-dimentional entity. whereas a SEM or a TEM micrograph is a 2-D representation of that. If the particle has some orientation preference (i.e. Not oriented randomly) analyzing a cross-sectional image can give wrong reprenstation about the 3rd direction.

I hope I could confuse you a lot. Anyways best luck dude....

Well Shrikrishna, I think that the size distribution obtained from TEM and SEM images tends to be more realistic and accurate, but they also present some limitations (as pointed above by Biplab). With respect the other techniques, you should consider that the particle size indirectly obtained from measurements of several size dependent physical properties tends to be smaller than real size of particles... For example, very often, the particle size distribution of a magnetic nanoparticles system obtained by magnetic resonance or magnetization measurements (FCM and ZFCM) present particle sizes about 1 - 3 nm smaller than the real. This occurs due to surface anisotropy of particles (among other surface effects) which tend to align antiferromagnetically the atomic or molecular magnetic momenta near particle´s surface.

I would say, it deoends upon your particles. if you want to characterize micron size particles then SEM would be quick technique, also you could use particle size analyzer and it gives you average particel size. if you are working with nanosized particels , then go for TEM, one thing, in TEM you might be dealing with agglomerates if you are not careful enough with sample preparation.

I hope this helps as this is a pretty basic question

i also consider SEM and TEM the best choice; moreover SEM/ EDX could offer you even more information. theoretically, the measurements take place in vacuum and your particles should be extremely dry, so first make sure your particles are not from a material that may collapse during drying.

Thanks to all for useful discussion. I tried to analyze data using SEM N XRD. The calculated particle size from log normal distribution function fitted to sem data is almost double than calculated from xrd using Williams Hall plot. Now which one I should trust? From SEM I can see surface morphology as well dispersed particles.

Use STM or TEM if and only if your system is not orientation dependent. Both STM and TEM are 2D measuring techniques.

Dear Shrikrishna,

I still have the doubt. The and Scherrer formula relates XRD peak position and FWHM with crystallite size whereas the Williamson-Hall equation adds another parameter (which is strain). How are you sure that the particles what are you seeing in SEM are single crystalline?

Can you please tell us the material under investigation? Just curious in how you are getting a linear correlation (as you mention it is double) between the crystallite (calculated from XRD) and particle size (seen in SEM) can exist.....

I will wait for you to enlighten me in this matter.....

meanwhile enjoy..

Using Scherrer formula you can assume the particle size for only the spherical particle for other structure you can't find out the size using scherrer formula the assumption for the other than spherical structure will be wrong.

First, understand the difference between particle size, grain size and crystallite size

Particle size could be either an agglomerate of many grains/crystallite or single grain/crystallite. For example, most of the case, the powders produced by mechanical alloying route will have particles consist of many grain whereas powders produced by chemical route will have an individual particles made up of single grain/crystallite. If you are sure that your individual powder is not an agglomerate of many grains, than you could use results from XRD and support it with SEM/TEM, provided you don't have any other planar defects (read the below paragraph to know why).

Most of us, use the term crystallite size and grain size interchangeably; but there is a slight difference between crystallite size and grains. For example, crystallite size from XRD and grain size from TEM/SEM will be almost similar only if the material of your interest don't have subgrains, stacking fault , twins, or any such planar surface defects.

Secondly, to be very accurate, what we calculate from XRD is "Coherently Scattering Domain Size" which includes crystallite size, stacking fault, twin and so on; and most of the time, we assume that domain size is equal to crystallite size by ignoring the other contribution to domain size. This will be wrong, if you have other planar defects in your material.

I should also emphasis, all the methods used to calculate Size or Strain from XRD pattern are based on analytical models except Warren Averbach method (fourier and stokes deconvilution method) and double voigt method. By using analytical model, we are biased to one particular profile (Pseudo-voigt/Pearson and so on) and results from these methods will never be accurate . . [ Read this article "Voigt-function model in diffraction line-broadening analysis -Davor Balzar"

www.du.edu/~balzar/Documents/IUCRbook.pdf ]

In my opinion, to find out the particle size, use particle size analyser / SEM /TEM than XRD technique.

Use Schrrer formula to estimate the particle size not williamson hall plot!! you can trust SEM more than the XRD, because XRD gives you the average but you can clearly size single particle( depending upon particle size) in SEM

Hope it helps

please consult other existing and extensive threads here on RG. You may search for "Scherrer", you may also look for the discussion contributions of Matteo Leoni, who has given excellent advice elsewhere.

I don't trust particle size measurement by TEM or SEM, because it is a very small (extremely small!) selection of the whole volume you can only look at.

I would always prepare a colloidal dispersion in a suitable liquid (in fact: in several different liquids) and measure and compare particle size distribution (using Laser Doppler method or else). In addition, you can prepare samples via freeze-drying and check in SEM / TEM whether you had dispersed well or you were still measuring aggregates of your nanoparticles.

By using the Scherrer formula you can calculate the crystallites structure.If you got a peak broadening in your XRD spectrum,then you can calculate the FWHM from that peak and the corresponding 2theta value also.By using these values you can easily calculate the size.

It is very hard to resolute particles using SEM, TEM is the one.

I agree with the response Dr. Praveen

what is the size of your particles? And are they monodisperse? With XRD you always get an average size of a large entity while looking at individual particles with SEM and TEM. SEM is only suitable for relatively large particles (depending on the resolution of your SEM) and for non-conductive materials charging effects may falsify your measurements.

As mentioned by others, especially TEM is a 2D projection of the particles. But anyhow, you should think about the definition of a particle size (supposing you do not have an agglomerate). Most often, particles are not spherical. If you e.g. have rod-shaped particles with edge lengths of 5x5x20nm, what is then your particle size? 5nm or 20nm? As far as I know, with XRD you get a value between 5nm and 20nm, which will of course not fit with TEM-measurements. Furthermore, even particles as small as 5nm in diameter may consist of several subgrains (eg. due to twinning). How this influences your XRD-result, I do not exactly know.

Often there are nanoparticles of very different sizes present in a sample. A problem are then the extremely small nanoparticles. For these it may be helpful to not only use conventional bright field TEM, but high angle annular dark field STEM (so called z-contrast imaging). With this technique you may see very small particles (1-3nm length) which are easily overlooked in conventional TEM images. If this technique is not available, at least try to also take images at comparably low acceleration voltage (80-100kV) and not only at a 200-300kV, if you expect really small particles. Thanks to the stronger interaction of the beam with the material at lower voltages, you get some more contrast for the particles.

When preparing the nanoparticles from a suspension on a TEM grid, keep in mind that they might distribute according to their size / weight on the grid...

Thank you all of you to have valuable discussion. As I mentioned earlier there is lot difference in the size calculated from XRD and SEM. @Elisabeth I have nanoparticles from 30 nm to 135 nm in size. As i have calculated size from SEM, it seems to be good enough to trust because I can see well dispersed nano particles with no agglomeration. As i got very small size from XRD, does it mean that it has core-shell like structure? Because from XRD we can get crystallite size. If suppose XRD size is 4 - nm smaller than the usual one, then why SEM shows so much higher values?

@ Shrikrishna Shivaji Gaikwad.. The answer to your last question "why SEM shows so much higher values than XRD? " must be obvious from the explanation of Mr Praveen, that XRD gives the coherent domain size which could be less than the SEM because X rays visualize a part of a grain separated from the other part of it  from a defect as a coherent domain and treat it as a single crystallite.. whereas in SEM what you visualize is the complete grain wherein you will end up with a higher size.

Dear colleague,

Crystallite Size is Different than Particle Size. A particle may be made up of several different crystallites or just one crystallite so in this case (particle size = crystallite size)

Crystallite size often matches grain size, but there are exceptions

Crystallites are coherent diffraction domains in X-ray diffraction.

Particles are chunks/pieces (usually very small, below 1 mm) of solid matter, ensembles of atoms. Particles can be as small as two atoms (the nitrogen particle for example, N2)

Grains are volumes, inside crystalline materials, with a specific orientation.

Particles can be polycrystalline, single crystal or amorphous. A 100 nanometer particle of gold, for instance, can be made of:

a single gold crystal,

many grains with a grain size