This question sounds quite generic, therefore I can give you a generic answer:

- X-ray diffraction. "Conventional" to get a bulk value for "crystallite" (size of coherently scattering domains) size from the analysis of the peak shape functions of your diffractogram. Or you can use small beam sizes to do profiles, meshes (~maps), etc. in your sample. From your question this might be too "coarse" a method...

- Electron Back-Scattering Diffraction imaging. You can map a surface using the electron diffraction signal: you'll get an image with the different crystallites and you can get a size distribution plot with some simple image processing. In addition you'll have info about orientation and general crystallographic info.

If you can segment the different crystals in your sample (es. because you have evident grain borders) you'd probably be good just using SEM back-scattering or secondary electron imaging, tho. Or any kind of imaging able to separate the different grains (optical microscopy with selective etching, etc). Then you can just separate the single crystal in your image and do some more or less fancy image processing, from counting the size and number of the different crystals to some stereological analysis...

You can also have a look at this: http://www.sciencedirect.com/science/article/pii/S0263436808000760

- If you need extremely detailed 3D info on a small sample and you are ready to spend some significant time in the data analysis (not to mention the complicated experiment) you can use box beam diffraction tomography. Es:

http://www.sciencedirect.com/science/article/pii/S0921509309004511

and works "derived" from the one above.

Cheers

Marco

-

Hi uday

As suggested by Marco, this question seems fairly general; again as proposed by Marco, you can use SEM equiped with EBSD which allow to obtain different maps (i.e. grain and subgrain boundary (GSB) as well as the called IPF Images (for determining grain orientations with respect to the loading direction [if it is the case]). A softaware, called Channel 5 (HKL technology) allowed to provide misorientation angle distribution relative to the fraction that characterize what you refer the level of non-uniformity. You can found many explanatory articles by introducing into a search engine (as an example), the following keywords: "evolution of microstructures and textures upon deformation" .

Grain size distribution before deformation can be easily obtained through the use of quantitative optical metallography. Measuring the grain size after cold deformation i.e grain subdivision looks as not trivial task because of high dislocation density. TEM has  small field of view to get true size distribution. SEM will not help too because of too many grain boundaries both high and low angle. EBSD will not be helpful because of bad Kikuchi  pattern from the deformed metal. It seems reasonable to carry out recrystallization and form equiaxed grains free of dislocations.  In this case any method will do, while the most still will metallography.