It depends both from the deposition method and the type of substrate. The type of substrate (especially its structure) can influence both direction and smoothness especially when ultra thin films are prepared (monoatomic or monomolecular layers). In practice physical deposition methods such as PLD, sputtering, ion beamand etc..., can provide deposition of more homogeneous ultra thin films when compared with the chemical methods especially chemical bath depositions because in my opinion chemical methods require more experience (Chemical bath methods). When you perform deposition of thicker kind of thin films (>100 nm) the substrate won't influence the smoothness really significant, but its structure will influence the deposition direction if the adhesion is good.

Best regards

Probably one of the best method is to use CVD over the single crystalline copper substrate ( preferentially (111) orientation) with a controlled hydrogen injection by the pulse technique during high temperature (1010-1030 C ) carburizing annealing. The main reason for the hydrogen injection is to scrap up (clean) the oxide layer over the copper surface to obtain freshly formed copper surface, which acts as a catalysis for the Carbon deposition on copper surface. Adhesion energy for a freshly formed 111 copper surface is minimum with respect to all other orientations, and secondly; because of the hexagonal packing of 111 plane is the best match for the graphene or graphite formations ''mismatch free energy is almostly nil).

I am agree with the Sasho Stojkovikj that " It depends both from the deposition method and the type of substrate ".

Thank You! All