Incorporation of hydrogen in cubic and uniaxial oxidic crystals deduced from polarized Raman scattering

Polarized Raman scattering investigations of the ir OH stretch mode vibration have been performed in a variety of prototype oxidic crystals, KTaO₃, SrTiO₃, K_1-x{Li}ₓTaO₃, TiO₂, LiTaO₃, with the aim of providing information on the sites of the incorporation of hydrogen. The traditionally used polarized ir absorption is limited by ensemble averages and its lacking sensitivity to crystallographic orientations of defects. Contrarily Raman scattering is sensitive to the symmetry of the dynamical mode. The problem of extracting information on a single defect from an ensemble is treated in the behavior-type theory, which has been developed recently for defects in cubic crystals and was extended to tetragonal and trigonal crystals in the course of these investigations. The principles of the method and the extensive tables are summarized. We describe an experimental method for determining the precise optical alignment of the scattering geometry. In the cubic phase of the perovskites SrTiO₃ and KTaO₃ with isotropic ir absorption, the Raman data allow us to reject one out of three models, where the protons vibrate between the two next-nearest oxygen ions along the cubic axes, by excluding the related C₂[100] symmetry.In tetragonal SrTiO₃ the absorption band splits into three components. In the second model the protons vibrate between the O^2- and the Sr²⁺ ions on the face diagonal of the cube, in the last along the edges of the oxygen octahedron between the two nearest-neighbor oxygen ions. The second model can be rejected, because the claimed vertical mirror plane {σ}(010) is not observed in the Raman spectra. The third model claims no mirror symmetry because of the tilting of the octahedra. In K_1-x{Li}ₓTaO₃, x=0.023, a polar tetragonal phase occurs due to the off-center freezing of the LiK ions along the polar axis. Again three subbands of the OH mode occur. But the bands, which reflected the breakdown of the {σ}(010) symmetry in SrTiO₃, do not occur in K_1-x{Li}ₓTaO₃. The observed vertical mirror plane {σ}(010) reflects the nontilted arrangement of the octahedra in K_1-x{Li}ₓTaO₃. In TiO₂ (rutile) the Raman data reveal a C₁ₕ(001) symmetry for the prominent OH stretch mode. This firmly establishes the model, where the protons vibrate in the center of the open channels between the next-nearest O^2- ions in the basal plane. In trigonal LiTaO₃ the crystal symmetry is so low that hardly any models can be discriminated on the basis of their symmetry, because most of them deal with C₁ symmetry. This is detected in the experiment and thus any orientation not along the trigonal axis or the vertical mirror plane is compatible with this result. Still one model proposed for LiNbO₃ can be excluded.