Mani Renuka The techniques are all accurate (and can be proven so by the appropriate material and documentary standards - ISO & ASTM) but are measuring different particle properties and the raw data are processed in a different manner. XRD does not measure particle size - via the Scherrer equation crystallite size is approximated (as there are other causes of peak broadening - strain and instrumental). DLS measures particles in liquid suspension including the protective and stabilizing 'coating' indicating how the particles will interact with their environment. SEM is a 2D visualization of a 3D particle and looks at the electron rich areas. A good example is a Au colloid where the Au particles would appear be much smaller than the particle size determined by DLS as only the core is examined rather than the core + shell protective layer (essential for the particle to be stable). In the whole history of electron microscopy no more than a few hundred grammes of material has been imaged... Take a look at the NIST gold colloids nominally of 10, 30, and 60 nm. Note on the certificates that different techniques produce different values for the same material... What numbers you use in practice will depend on your end application. See the attached material. There are many questions and answers on RG that will help you if you search.

All the cited techniques are accurate, it depends on which parameter you want to determine. Some of them determine crystallite sizes, other grain or particle sizes, or aggregates of them

Dear colleague

This is not a trivial question and when you use XRD you access at the coherent diffracting domains of a material which provide you a measure on the crystal dimension under Bragg condition. By DLS you access at the coherent light diffusion of a particle. DLS supposed that your particle is hard spheric ball. Then DLS trend to oversize the particles because signal issied from large ones is more intense than the small ones. A consequence if you used the two methods is smaller values with XRD and larger with DLS. I suggest you to use SEM (beam reflection) or TEM (beam projection) which will provide you images of your particle. Then if you want to measure size from the SEM or TEM picture I recommand you to calculate the fractal dimension (Df) of the particles distribution. Df provide information about the complexity of agregates but also the necessary informations for sizes distribution.

I hope that this answer will help you. Best regards

Philippe

Each of the tests is for a specific application.

DLS does not give the diameter of a single particle, but shows the diameter of interconnected particles in the aqueous medium and indicates the amount of accumulation or stability of the particles in the aqueous medium.

The lower the Z-Average value, it means that the particles are better dispersed and have good stability in the water environment.

TEM indicates the diameter of the particle.

Mani Renuka The end application will dictate which number or numbers you would want or need to use, so good particle characterization laboratories have a range of techniques available to them that provide complementary information. Thus, we have to take great care.

Accuracy is easy to define – closeness to an agreed ‘truth’ or standard. Often this is a certified standard material, ideally from NIST. We also have to distinguish particle size from particle size distribution. And distribution is the bane of all particle sizing techniques.

Hence, we have to take care with a statement such as ‘DLS supposed that your particle is hard spheric ball. Then DLS trend to oversize the particles because signal issied from large ones is more intense than the small ones’. DLS is a very powerful technique allowing many billions of particles to be simultaneously and quickly assessed. It’s an ensemble technique and therefore of much lower resolution than single particle analyses. The accuracy of any DLS technique is easily verified with an appropriate standard such as the NIST 100 nm material (SRM 1963). As the movement of particles (diffusion coefficient) is measured in DLS then the numbers that arise relate to the transport properties of the particles in suspension (with aggregates and agglomerates and the core plus surrounding protective layers). These numbers will be more related to, say, the viscosity and rheological properties of the system than XRD crystallite size or electron microscopy primary particle size.

We may want sensitivity to the large material in a system (e.g. protein aggregates which may decrease the efficacy of a treatment) and DLS will provide that because of the way the data are manipulated (intensity rather than number based). However, some form of imaging is essential for morphological information, but we have to distinguish between qualitative and quantitative information. If one tries to measure the 100 nm SRM1963 material by x-ray techniques, one finds that the material cannot be measured or sized as it’s a latex polymer. Thus, another reason to be careful with generalizations.

Representative sampling of any system is vital, and it would need 10000 random particles to be images for statistical validity at the 1% standard error system. Such information is rarely obtained. Indeed, no more than several hundreds of grams in total have ever been imaged in the entire history of electron microscopy… If we have to prepare a sample for TEM, say by microtoming, then every image could be stated to be an artifact as it’s a 2D slice from a 3D particle. Consider slicing a cube or a human lengthways of sideways and trying to work out the proportion of the body that is the liver. The illustration from Claude Bragdon’s classic text from 1909 (A primer of higher space (the fourth dimension) Manas Press) illustrates some of these (slicing) difficulties. We can see everything from triangles to hexagons when a cube is sliced. We have a single shape (cube) but an apparent size and shape distribution from the images. How is that referred back to the actual starting cube?

XRD, zeta sizer