Ever since Herbert Gieiter, in 1981, attracted sudden attention to the merits of ultrafine-grained solids, which he called 'nanostructured', investigators have been concerned about the problem of keeping the grain size ultrafine even when the material is heated. The problem arises because nanostructured solids, in particular ceramic ones, can be superplastically shaped, exploiting vacancy flow between adjacent grain boundaries. The admissible imposed strain rate for superplasticity increases rapidly as the grain diameter, d, diminishes, typically as the third inverse power of d; so keeping grains fine is of paramount importance for practical application of such superplasticity.
One favored way of keeping grains fine even at the fairly high temperatures needed for superptastic behavior is to disperse a fine second phase which will drag, and slow down, migrating grain boundaries. An example of this strategy, applied to a zirconia-based ceramic, was described by B.-N. Kim et al. [Nature (2001) 413, 288-291; see page 14 for further details]. Another, rather mysterious two-step approach can be applied during sintering of the nanostructured powder: a high-temperature stage followed by continued sintering at a slightly lower temperature. This was convincingly demonstrated by I.-Wei Chen and X.-H. Wang [Nature (2000) 404, 168-171] with yttria, Y203, either nominally pure or doped with other oxides. They provisionally attribute this'frozen microstructure' phenomenon to a change in the pore size distribution caused by the two-stage annealing, but their argument is distinctly fuzzy.
All the efforts to prevent grain coarsening in nanostructured polycrystals have hitherto depended on ways of impeding grain boundary motion by one means or another. Meanwhile the driving force for such motion, which leads to grain coarsening, is simply the reduction in the total area of grain boundary that unavoidably accompanies coarsening. However, there is now a potential strategy that is based on a completely different idea. If the effective specific energy of the grain boundaries can be made zero or even negative, then the driving force for grain growth vanishes and there will be no growth at all. This intriguing idea was first advanced in quantitative form by J. Weissmuller [Nanostructured Materials (1993) 4, 261]. The idea depends on the very well-documented process of grain boundary segregation of a minor constituent (solute). Putting the matter at its simplest, such segregation will occur if the free energy of the grain boundary is lowered. As the grain size grows during coarsening, less of the segregated solute is needed because the area of grain boundary per unit volume diminishes, and so the excess solute returns to the grain interior. This return increases the free energy of the grain, and may indeed increase it so much that the global free energy change is positive. If that happens, then the driving force for grain growth vanishes. This idea is outlined again by Gleiter - who was Weissmuller's collaborator - in his major review of nanostructured materials [Acta Materialia (2000) 48, 1-29]. This kind of 'segregation-stabilized grain structure', as we may call it, is an important new concept in the venerable topic of grain growth theory.
This novel idea has just been examined further, in quantitative detail, by R. Kirchheim [Acta Materialia, in press], and linked to the process of precipitation of excess segregated solute. He points out that if such precipitation is kinetica[ly hindered, then the overall free energy can rise sharply and at a certain grain size, the driving force for grain growth will disappear. Recent measurements on fine grained ruthenium aluminide containing some iron impurity [K.W. Liu and F. Mi~cklich, Acta Materialia (2001) 49, 395-403] are analyzed in quantitative support of this model.
It is tempting to suggest that the striking and puzzling findings by Chen and Wang with two-stage sintering of yttria, outlined above, might be linked to a 'segregation-stabilized grain structure'.