Chemical control of the polymorphic phase boundaries in doped barium titanate

Abstract

Barium titanate (BaTiO₃), a known ferroelectric material, is of great interest as a future lead-free piezoelectric if appropriately doped to tailor the electric properties for specific applications. This work focuses on the study and rationalisation of the dielectric properties of a series of A-site/B-site co-doped compositions: SrZrO₃-BaTiO₃, CaZrO₃-BaTiO₃, LaScO₃-BaTiO₃ and GdScO₃-BaTiO₃. The effect of sintering conditions and microstructure on ceramics is shown to have a significant impact on the physical properties of these materials. Pellet inhomogeneity, air sensitivity in pre-calcined powders and the presence of parasitic grain boundary capacitances are all shown to have adverse effects on properties, including the magnitudes of relative permittivity and T[subscript(C)] calculated from total capacitance data. These can be overcome by careful control of synthesis conditions. Dielectric spectroscopy measurements on the optimised materials show that increasing addition of SrZrO₃, CaZrO₃ or LaScO₃ causes the phase transitions between the various polymorphs of BaTiO₃ to coalesce. In each case T[subscript(C)] is reduced whilst each of the other phase transitions is shifted to higher temperatures, until the coalescence temperature is reached. When doped with GdScO₃ T[subscript(C)] is observed to fall, but so too are the rhombohedral/orthorhombic and orthorhombic/tetragonal transitions, resulting in a stabilisation of the tetragonal polymorphic phase. This is suggested to result from an antipolar displacement of small Gd₃₊ ions, resulting in 8-coordinate ion and stabilisation of the tetragonal polymorph. The addition of dopant species is shown to result in two different high temperature conduction regimes. Both mechanisms are observed within single compositions over different temperature ranges. It is suggested that this is due to a change between n- and p-type electronic conduction processes or mixed ionic/electronic processes. Finally, it is shown that trends observed in changes to T[subscript(C)] cannot be accounted for by simple and widely used size-based arguments alone, but requires consideration of cation size variance and charge dilution effects in order to fully understand the impact on T[subscript(C)] of dopant addition.