Hello Hu,

Yes both GBs and twin boundaries are "boundary" type (2-dimensional) defects, they have very different nature accross the boundary itself. Twins are mostly coherent boundaries accross which the number of bonds are just like in the perfect lattice, with the difference that the bond angles are now highly different compared to the perfect lattice. Whereas accross a grain boundary, with some difference depending on whether it is a low angle or high angle boundary, there is hardly any coherency at all. This difference makes the grain boundaries a "fast track" for diffusion. Creep being inseperable from the diffusion phenomenon, grain boundaries acts as an accellerating factor for creep. Twin boundaries on the other hand do not posses the same potential as the ordinary grain boundaries.

One can furher elaborate on the effects of GBs on creep. For example, under certain stress-temperature conditions another phenomenon called grain boundary sliding that takes place.

Now, we can safely comment with "thought experiment" on the behaviour of twins in relation to this particular deformation mechanism, i.e. grain boundary sliding. Let's ask if the twin boundaries contribute to creep via sliding. Since by their birth the twin boundaries can contain dislocations they can slide too. However, in practice there can hardly be a hypotetical case like this. And to my best knowledge there is no such study that reports on an isolated situation of this type. However, let's allow ourselves to answer this hypotetical case. Twin boundaries, naturally assuming that there exist more than one inside any grain, they actually acts as dividers of a grain into  "several portions" . In the picture twins actually create a strengthening effect and hinders each others movement (twins cannot cut accross each other as each one reorients the grain by a large angle accross the twin bounary). So we may safely say that as they exist in some population in an individual grain twins cannot contribute to creep deformation via sliding along twin boundaries. However, if there exist only one twing boundary in an individual grain we may envisage a set of temperature and stress conditions under which that twin boundary can show sliding.

I hope this helps.

Hello Hu,

Yes both GBs and twin boundaries are "boundary" type (2-dimensional) defects, they have very different nature accross the boundary itself. Twins are mostly coherent boundaries accross which the number of bonds are just like in the perfect lattice, with the difference that the bond angles are now highly different compared to the perfect lattice. Whereas accross a grain boundary, with some difference depending on whether it is a low angle or high angle boundary, there is hardly any coherency at all. This difference makes the grain boundaries a "fast track" for diffusion. Creep being inseperable from the diffusion phenomenon, grain boundaries acts as an accellerating factor for creep. Twin boundaries on the other hand do not posses the same potential as the ordinary grain boundaries.

One can furher elaborate on the effects of GBs on creep. For example, under certain stress-temperature conditions another phenomenon called grain boundary sliding that takes place.

Now, we can safely comment with "thought experiment" on the behaviour of twins in relation to this particular deformation mechanism, i.e. grain boundary sliding. Let's ask if the twin boundaries contribute to creep via sliding. Since by their birth the twin boundaries can contain dislocations they can slide too. However, in practice there can hardly be a hypotetical case like this. And to my best knowledge there is no such study that reports on an isolated situation of this type. However, let's allow ourselves to answer this hypotetical case. Twin boundaries, naturally assuming that there exist more than one inside any grain, they actually acts as dividers of a grain into  "several portions" . In the picture twins actually create a strengthening effect and hinders each others movement (twins cannot cut accross each other as each one reorients the grain by a large angle accross the twin bounary). So we may safely say that as they exist in some population in an individual grain twins cannot contribute to creep deformation via sliding along twin boundaries. However, if there exist only one twing boundary in an individual grain we may envisage a set of temperature and stress conditions under which that twin boundary can show sliding.

I hope this helps.

twin boundaries resist sliding motion as compared to grain boundaries hence they have high resistance to creep

At very high temperature where the  Herring&Nabarra creep dominates the whole scenerio , T/TM  GRT 0.75 ,  the dislocation mechanisms don't play significant role but the vacancy flow driven by  the stress gradient controls the whole creep phenomena. Then it is clear than twin and or the low angle grain boundaries can't play any role by hindering that vacancy and/or interstitials  flow but the large angle grain boundaries especially those having sub-microscopic sizes can contribute to the high temperature plastic deformation by sliding and rotating, and in addition allowing high mobility paths for the vacancy flow due to low activation of motion for the diffusion compared to bulk. The large angle grain boundaries are also perfect sink and source for the vacancies. 

I agree with the above remarks, however there is a delicate question as follows. If twins are just ADDED to regular granular structure (e.g. within austenite grains by annealing), how can they hinder the mentioned phenomena (diffusion and sliding) taking place at regular grain boundaries ?   The only option is the structure where some 90% of HABs are twin ones, however I cannot imagine this in reality. A more realistic reason is the contribution of lattice dislocation slip to the creep also actual to some extent. Indeed, twins will make a real barriers in this case.