Influence of vacancy ordering on the percolative behavior of (Li(1)(-x)Na(x))(3y)La(2/3-y)TiO(3) perovskites

Influence of the vacancy concentration on the Li conductivity of the (Li(1-x)Na(x))(0.2)La(0.6)TiO(3) and (Li(1-x)Na(x)(0.5)La(0.5)TiO(3) perovskite series, with 0 < or = x < 1, has been investigated by neutron diffraction (ND), impedance spectroscopy (IS), nuclear magnetic resonance (NMR), and Monte Carlo simulations. In both series, Li(+) ions occupy unit cell faces, but Na(+) ions are located at A sites of the perovskite. From this fact, the amount of vacant A sites that participate in Li conductivity is given by the expression n(v) = [Li] + square, where square is the nominal vacancy concentration. Substitution of Li by Na decreases the amount of vacancies, reducing drastically the Li conductivity when n(v) approaches the percolation threshold of the perovskite conduction network. In disordered (Li(1-x)Na(x))(0.5)La(0.5)TiO(3) perovskites, the percolation threshold is 0.31; however, in ordered (Li(1-x)Na(x))(0.2)La(0.6)TiO(3) perovskites, this parameter changes to 0.26. Near the percolation threshold, the amount of mobile Li species deduced by (7)Li NMR spectroscopy is lower than that derived from structural formulas but higher than deduced from dc conductivity measurements. Conductivity values have been explained by Monte Carlo simulations, which assume a random walk for Li ions in the conduction network of the perovskite. In these simulations, distribution of vacancies conforms to structural models deduced from ND experiments.