Oxide Ion Conduction in A-site Bi-containing Perovskite-type Ceramics

Oxide ion conductors have drawn significant attention due to their important technical applications in electrochemical devices. This project is based on a new oxide ion conductor Na0.5Bi0.5TiO3 (NBT) which indicates (undoped) conducting NBT is a potential electrolyte material that possesses a high level of nearly pure oxide ion conduction. Na non-stoichiometry in the starting composition (Na-series), acceptor doping (Mg2+ → Ti4+, Na0.5Bi0.5Ti1-xMgxO3-x) and donor doping (Nb5+ → Ti4+, Na0.5Bi0.5Ti1-xNbxO3+x/2) in NBT have been investigated. It has been shown that similar to a previously reported Bi non-stoichiometric series (Bi-series), the electrical properties of NBT are highly sensitive to low levels of Na non-stoichiometry. However, the defect mechanisms for Na and Bi non-stoichiometry are different and leads to a contrasting influence on the properties of NBT ceramics. Na-rich samples from the Na-series behave like Bi-deficient samples from the Bi-series whereas Na-deficient samples from the Na-series behave like Bi-rich samples from the Bi-series. Generally speaking, the bulk conductivity (oxide ion conduction) of NBT is dependent on the Na/Bi ratio in the nominal starting composition. Samples with a Na/Bi ratio ≥ 1 exhibit high, nearly pure oxygen ion conduction with a low activation energy (< 0.9 eV) for bulk conduction whereas samples with a Na/Bi ratio < 1 are electronic insulators with a high activation energy (~ 1.6 eV) for bulk conduction. Mg B-site acceptor doping, (Na,Bi)Ti1-xMgxO3-x, can further enhance the bulk conductivity and produces oxide ion transport numbers, tion, close to unity. This doping also stabilises NBT ceramics to reducing atmospheres (eg 5%H2/95%N2 at 500 oC) to demonstrate their potential as an electrolyte material for Intermediate Temperature Solid Oxide Fuel Cells. In contrast, Nb donor doping, (Na,Bi)Ti1-xNbxO3+x/2, systematically suppresses the oxide-ion conductivity; very low levels of Nb doping (0.002 ~ 0.003) leads to a mixed oxide ion and n-type conduction and an intermediate tion (~ 0.5). A further increase of Nb doping level (≥ 0.005) suppresses the oxide ion conduction and results in dielectric materials with predominant n-type electronic bulk conduction with tion ≤ 0.07 at elevated temperature (eg > 600 oC). It is worth noting that, extremely Bi-rich (undoped) NBT (Bi ≥ 0.52) also induces mixed ionic/electronic behaviour by reintroducing higher oxide-ion conductivity with tion ~ 0.4–0.6. The ferroelectric Aurivillius phase Bi4Ti3O12 (BiT) has also been determined to exhibit high levels of oxide ion conduction. Un-doped BiT shows mixed p-type and oxide ion conduction at low temperature; however, tion approaches near unity close to the Curie Temperature, TC ~ 675 oC. As BiT contains both extrinsic and intrinsic defects, the Bi nonstoichiometry has limited influence on its electrical properties. Isovalent doping (La3+ → Bi3+; Bi4-xLaxTi3O12) acceptor doping (Sr2+ → Bi3+; Bi4-xSrxTi3O12-x/2) and donor doping (Nb5+ → Ti4+; Bi4Ti3-xNbxO12+x/2) are all investigated. La-doping (x ≤ 2) can shift TC to lower temperature and makes BiT a potentially good oxygen ion conductor at ~ 600 oC, but the bulk conductivity gradually reduces with increasing x. Sr-doping has a rather limited solid solution limit (x ≤ 0.12) compared to La doping but can maintain the bulk conductivity while lowering the TC. Nb donor doping on Ti-site can compensate oxygen vacancies and suppress the oxide ion conduction. K0.5Bi0.5TiO3 (KBT) has been determined to be a mixed conductor where the ionic contribution can be oxide ions and/or protons. The proton conduction in KBT is controlled by the K/Bi ratio in the nominal starting composition. Samples with a starting K/Bi ratio > 1 exhibit substantial proton conductivity whereas samples with a starting K/Bi ratio ≤ 1 exhibit lower proton conduction. Compared to NBT, the oxide ion conduction in KBT is significantly lower and relatively independent of the starting A-site non-stoichiometry.