Oxygen vacancies in semiconducting BaTiO3 -based ferroelectrics: Electron doping, history dependence of TC, and domain wall pinning

We measured the complex Young modulus of BaTiO3-δ (BT), BaxSr1-xTiO3-δ (BST), and (Ba0.85Ca0.15)(Zr0.1Ti0.9)O3-δ (BCTZ) during heating and cooling runs at various O deficiencies [which dope electrons rendering the samples semiconducting/conducting ferroelectrics (FEs)] and aging times. The elastic energy loss has peaks due to the jumps of isolated O vacancies (VO's) and reorientations of pairs of VO's in the paraelectric phase, from which the respective rates and activation energies are measured. These rates control the mechanisms of domain clamping, pinning, fatigue, and anything related to the VO mobility. In the FE phase, the drop of the losses due to the domain wall (DW) motion upon introduction of VO's monitors the degree of pinning. In addition, large shifts of TC are observed at the same value of δ upon varying the permanence time in the FE state, up to ΔTC= 21 K in BST, while no aging effect is found in BCTZ. The phenomenology is explained by considering that TC is depressed mainly by the mobile electrons doped by VO's. Each isolated VO dopes two electrons as itinerant Ti3+ ions, but when it forms a stable linear VO-Ti2+-VO pair, the two electrons of the Ti2+ are subtracted from the mobile ones, halving doping. The rise of TC during the initial aging is then explained in terms of the progressive aggregation of the VO's. Prolonging aging for years leads to a decrease of TC, explained assuming that the most stable position of VO is at 90° DWs, whose geometry is incompatible with the pairs. Then after enough time, the VO initially aggregated within the domains dissociate to decorate the 90° walls, increasing doping and lowering TC. The absence of such effects in BCTZ is due to larger activation energy for pair reorientation and pair binding energy. Then at room temperature, practically all VO's are paired and static over a timescale of hundreds of years, explaining the superior resistance of BCTZ to fatigue.