Insights on the origin of the structural phase transition in BaV 10O 15 from electronic structure calculations and the effect of Ti-doping on its structure and electrical transport properties

Abstract

Band structure calculations at the level of LMTO-ASA provide insight into the electronic structure of BaV 10O 15 and the origin of the structural phase transition. A crystal orbital Hamiltonian population/integrated crystal orbital Hamiltonian population analysis provides evidence that the crystallographic phase transition is driven by V–V bond formation. As well, the energy bands near the Fermi level are very narrow, <1 eV, consistent with the fact that the observed insulating behavior can be due to electron localization via either Mott-Hubbard correlation and/or Anderson disorder. The partial solid solution, BaV 10− x Ti x O 15, was examined to study the effect of Ti-doping at the V sites on the structure and electronic transport properties. In spite of the non-existence of “BaTi 10O 15”, the limiting x = 8 , as indicated by a monotonic increase in the cell volume and systematic changes in properties. This limit may be due to the difficulty of stabilizing Ti 2+ in this structure. For x = 0.5 both the first order structural phase transition and the magnetic transition at 40 K are quenched. The samples obey the Curie–Weiss law to x = 3 with nearly spin only effective moments along with θ values which range from −1090 K ( x = 0.5 ) to −1629 K ( x = 3 ). For x > 3 a very large, ∼2×10 −3 emu/mol, temperature independent (TIP) contribution dominates. Conductivity measurements on sintered, polycrystalline samples show semiconducting behavior for all compositions. Activation energies for Mott hopping derived from high temperature data range from ∼0.1 eV for x = 0 – 1 and fall to a plateau of 0.06 eV for x = 3 – 7 . Low temperature data for x = 3 , 5 and 7 show evidence for Mott variable range hoping (VRH) with a T 1 / 4 law and in one case between 5 and 17 K, a Efros-Shklovskii correlated hopping, T 1 / 2 law, was seen, in sharp contrast to BaV 10O 15 where only the E-S law was observed up to 75 K. Seebeck coefficients are small (<35 μV/K), positive, roughly TIP and increase with increasing x up to x = 5 . This may point to a Heikes hopping of holes but a simple single carrier model is impossible. The compositions for x > 3 are remarkable in that local moment behavior is lost, yet a metallic state is not reached. The failure of this system to be driven metallic even at such high doping levels is not fully understood but it seems clear that disorder induced carrier localization plays a major role.