Grain boundary is a transition region in which some atoms are not exactly aligned with either grain. Grain boundaries are actually a finite thickness, perhaps few atomic distances. The mismatch of the orientation of neighboring grains leads to a less efficient atomic packing within the grain boundary and hence the atoms in the boundary have a less ordered structure and a slightly higher internal energy. Doping of certain elements may increase the width slightly, but still in nm range. If you consider any precipitate along the grain boundaries, then the situation is different. You can treat the precipitate as a part of grain boundary due to the lattice mismatch to the host grains. In this case you are not increasing the grain boundary thickness, but you are inserting an additional layer of different crystal structure. 

Hi, are you talking abou the thickness of grain boundary or thedendrites? If the dendrites, the dendrite arm spacing is greatly related to the cooling rate. But for the grain boundary, it forms when grains tough together and the thickness is hard to change in my opinion. 

I agree with Artur. A grain boundary has no thickness. You need to use another term. In case you don't mean grain boundary but phase boundary you might talk about a diffusion thickness...?

Grain boundary is a transition region in which some atoms are not exactly aligned with either grain. Grain boundaries are actually a finite thickness, perhaps few atomic distances. The mismatch of the orientation of neighboring grains leads to a less efficient atomic packing within the grain boundary and hence the atoms in the boundary have a less ordered structure and a slightly higher internal energy. Doping of certain elements may increase the width slightly, but still in nm range. If you consider any precipitate along the grain boundaries, then the situation is different. You can treat the precipitate as a part of grain boundary due to the lattice mismatch to the host grains. In this case you are not increasing the grain boundary thickness, but you are inserting an additional layer of different crystal structure. 

@ S.T. Abraham: you are certainly right. Nobody would doubt your comment, but we need to keep in mind about what we are talking. In your "scale" you can even forget the definition of cystals since they are also full of defects and never perfect (except perhaps some semiconductors) or consider any relaxation along the surface. Therefore I would really urgently recomment  Shivaprasad S to clarify his question. It is still not clear to me where his problem is. It makes no sense to mix all definition of grain boundaries if the scale is not clear. Geomoetrically it is a surface and has no thickness. Experimentally in very high resolution the boundary is a strong crystal defect with a  thickness.

I agree with Gert. We cannot discuss the matter if it is not precisely defined.

Another point: WHY would you want to increase the grain boundary thickness???? As S.T. Abraham pointed, you would then have insertion at "original" boundaries and then you would have two new boundaries: between the original material and the inserted one and the next on the other side. As I said, what is the point in increasing something that is actually a defect? Why wouid you need to separate grains?

Dear Shivaprasad S,

A grain boundary can be defined as the region where two incoherent grains meet each other. It is imaginable that the thickness of this region taken along the normal on the grain boundary depends on the incoherency angle. However, I would expect that this region thickness is always smaller than the dimensions of the metal's unit cell. The consequence is that from a practical point of view it seems impossible to increase significantly this thickness. In case the idea is to increase the volume fraction of the region around grain boundaries in order to increase strength, then grain size reduction seems more appropriated (Hall-Petch equation). The same approach may be useful in case it would be desirable to increase the amount of alloying elements in the direct vicinity of grain boundaries. 

You might want to check out complexions at GBs. Here's a useful reference http://www.sciencedirect.com/science/article/pii/S1359645413005570

Grain boundary thickness increases when teh radius of the particles approaches nanometer size. See our paper from 1990:

https://www.researchgate.net/publication/233858912_HREM-Studies_of_the_microstructure_of_nano-crystalline_Palladium

Article HREM-Studies of the microstructure of nano-crystalline Palladium

A grain boundary can be defined as the region where two incoherent grains meet each other that is a  grain boundary has no thickness  and The smaller grain size have more grain boundaries per unit area . 

Interface between adjacent grains may be modified by using larger ions but that was not reproducible for all structure. It can also modified by heating and cooling rate and temperature during the grain growth process.