Grain boundary sliding in creep To prevent grain boundary sliding so that creep is less likely to occur, usually engineers would design components of larger grains or have columnar grain structure to prevent grain-boundary sliding. Why can these two methods prevent grain boundary sliding? For columnar grains, would they be more easy to slide against each other since the grains are parallel?
 A: The creep mechanism of grain boundary sliding is also sometimes referred to as superplastic or granular flow. The macroscopic deformation of the sample is caused by the rearrangement of grains as they slide past one another along grain boundaries. The microscopic displacements along grain boundaries is caused by the glide and climb motion of grain boundary dislocations. Dislocation glide motion may be conservative (require no diffusion of point-defects) until they reach obstacles that block glide motion. Dislocation climb motion (non-conservative motion requiring diffusion of point defects) may enable dislocations to move past the obstacles. The rate of diffusion of point defects to or from non-conservatively moving dislocations becomes the rate-limiting step in the entire creep mechanism. Steady-state creep due to grain boundary sliding is therefore a type of diffusion creep and the creep rate is inversely proportional to some power of the grain size.  As the characteristic length of the poly-crystal, grain diameter will be related to the average diffusion path length (at least at steady-state.)     
