Atomic structural features of dopant segregated grain boundary complexions in alumina by EXAFS

The primary objective of this undertaking was to characterize the atomic structural features of dopant-segregated interfaces in a (pseudo) single phase microstructure and relate the same to atomic diffusion in the grain boundaries. Alumina was chosen as a model host system based on prior observations of grain boundary complexions in this system by electron microscopy. Two types of dopant chemistry were selected that are known to produce dramatically different microstructural behavior in alumina. These were, (i) rare earth element doping (Y) and (ii) Y-Si co-doping in alumina. In Y-doped alumina microstructures, different Gibbsian excess of the segregated dopant has been known to produce two distinct types of interface complexions. On the other hand, three distinct types of disordered grain boundary complexions have been observed in Y-Si codoped alumina. A quantitative grain growth study was performed in dense microstructures of these materials and different kinetic regimes of boundary mobility were identified. Subsequently, samples annealed at various temperatures were quenched to preserve the grain boundary structure and characterized using synchrotron extended X-ray absorption fine structure spectroscopy (EXAFS) at the Y K-edge. Distinct local structural features of the dopant segregation induced interface phases were observed and were used to distinguish between each of the complexion types. Computation of EXAFS spectra of theoretical clusters by ab initio methods and fitting the same with experimental data identified several types of interface complexions including: (i) sub-monolayer adsorption, where oversized isovalent dopants (Y) occupy substitutional cation sites at the grain boundary core and reduce the interface energy, (ii) saturation of dopants at the interface leading to bilayer adsorption, where dopants (Y) substitute host cations on both sides of the boundary interpolating into the crystals, (iii) multilayered adsorption, where a pseudo-amorphous phase of high specific volume appears at a particular combination of temperature and chemistry, (iv) thin intergranular amorphous films in the Y-Si-O system, and (v) thick intergranular wetting films with a structure similar to that of yttrium aluminosilicate bulk glass. The structure of these interface phases correlated directly with the experimental boundary mobility, indicating a progressive increase in the diffusion rate across the boundary with increased disorder in the interface phase. Additionally, results of three other doped alumina systems including La-doped, Zr-doped and Cu-Ti co-doped alumina, are presented in the appendices that provide substantial evidence for the existence of various complexion types. A first order transition in boundary mobility was observed in Zr-doped alumina that may be related to the transition from a sub-monolayer adsorption to a bilayer configuration. Metallic copper of constrained dimension was found to exist in dense Cu-Ti co-doped alumina that was postulated to exist as an intergranular layer in the microstructure. These findings provide validation to the concept of interface kinetic engineering exploiting grain boundary complexions, wherein by engineering the chemistry of a polycrystalline interface dramatically different microstructures can be processed.