Analysis of the rhombohedral–tetragonal symmetries coexistence in lead-free 0.94(Bi 0.5 Na 0.5 )TiO 3 –0.06BaTiO 3 ceramics from nanopowders

Abstract

Ceramics with perovskite-type structure and 0.94(Bi_0.5Na_0.5)TiO₃–0.06BaTiO₃ (BNBT) composition have been studied by conventional powder X-ray diffraction in Bragg–Brentano geometry. Ceramics were obtained from sol–gel autocombustion nanopowders and processed either by hot pressing and subsequent recrystallisation or pressureless sintering in two steps. These methods provided single-phase, sub-micron grain size (<700 nm), dense ceramics with good piezoelectric performance (96–94% of theoretical density and d₃₃ = 143–124 pC N^–1, respectively). For the considered ceramics, the splitting of the peaks of the cubic perovskite-type structure with 111 and 200 Miller indices has been repeatedly used as a symmetry identification criterion. In this work a simple, yet powerful, procedure to validate the consistency of the mentioned splitting interpretation is presented. Based on peaks fitting and well-known crystallographic expressions, the rhombohedral and tetragonal symmetries’ coexistence is verified. The suggested procedure can be applied to the study of peak splitting in ceramics at Morphotropic Phase Boundaries in a general way. In a given series of BNBT ceramics, inconsistencies for interplanar distances, intensities’ ratios and the evolution of these from not-poled to poled samples have been found. In poled ceramics, special care has been taken when carrying out this analysis, due to the anisotropic strains arising from ferroelectric (FE) domain orientation. Poling gives rise to a displacement of the peaks angular positions and modification of the intensity ratios. However, the interplanar distance changes associated with the angular deviations here observed are one order of magnitude higher than those expected from anisotropic strains. These results set up a doubt on the sufficiency of the [rhombohedral + tetragonal] model to characterise the considered ceramics. A model of a mesoscopic FE phase with rhombohedral symmetry, a mesoscopic and globally weakly polar phase, with cubic symmetry, and a nanosised phase, also cubic, is presented as a plausible alternative.