High resolution STEM analysis of temperature stable relaxors

Abstract

There is a need for dielectric materials with high relative permittivity (ε r ) values that can operate at high temperatures. Such materials have uses in new automotive, aerospace and energy technologies where there are demands for capacitors to operate in increasingly demanding high temperature environments. One particular ceramic system that shows promise is (1‐x)Ba 0.8 Ca 0.2 TiO 3 ‐(x)BiMg 0.5 Ti 0.5 O 3 (BCT‐BMT) [1]. Typical relaxor materials display a broad ε r peak when plotted against temperature [2]. Their behaviour may be explained by the presence of nano‐domains [3, 4], or polar nano‐regions (PNRs), which act to broaden the dielectric peak through a distribution of relaxation times. However, a special class of relaxors display ε r plots with flat temperature responses up to 500 ºC. These are termed temperature‐stable relaxors and BCT‐BMT is such a material. This behaviour cannot be explained by current understanding of relaxor PNRs, and so further investigation of structural and chemical differences in the nanostructure is required. One technique ideally suited to probing such features is atomic resolution scanning transmission electron microscopy (STEM). Initial studies of the material with composition (Ba 0.4 Ca 0.1 Bi 0.5 ) (Mg 0.25 Ti 0.75 )O 3 , which displays typical temperature stable relaxor behaviour, have been carried out using aberration corrected STEM using a Nion UltraSTEM 100. STEM high angle annular dark‐field (HAADF) imaging at the atomic level suggests uniform crystallinity; however measurements of displacements of B‐site atomic columns within the projected lattice on the nanoscale are detectable. These localised, coherent B‐site displacements (in the form of aligned nanodomains) may respond to changes in temperature in a different way to normal relaxor materials and their detection could go some way to explain the temperature stable behaviour. Atomically resolved electron energy loss spectroscopy (EELS) spectrum imaging (SI) was used to identify potential chemical inhomogeneity which might also contribute to a distinctive polar nanostructure (Figure 1) and help explain the dielectric behaviour. However there are challenges associated with this type of EELS analysis: the Ca L 2,3 ‐edge was obscured by the tail of the C K‐ edge, the Mg K ‐edge is weak and the Bi M 4,5 ‐edge sits at a high energy loss. Nevertheless, preliminary studies suggest significant column‐to‐column variations in HAADF intensity for both A and B sites in BCT‐BMT for the very thinnest regions of the STEM specimen, suggesting significant short‐range variations in composition. Present studies are concentrating on detecting any chemical inhomogeneities in the sample using STEM energy dispersive X‐ray (EDX) mapping, which will complement EELS by being able to map the Ca, Mg and Bi elemental signals without overlap. The new 300 kV FEI Titan 3 Themis G2 STEM installed at Leeds University with its FEI SuperX EDX system is capable of producing elemental maps with sufficient resolution and sensitivity, as demonstrated in EDX maps shown in Figure 2. Here, EDX mapping has been carried out on a different sample, in an area containing an interface between SrTiO 3 and (0.7)BiFeO 3 – (0.3)PbTiO 3 . The positions of the columns of Bi are clearly visible, and it is anticipated that such capability will mean that more information can be gained about any possible chemical segregation in BCT‐BMT.