derived from BaCO 3 . Small particle size has been shown to suppress a martensitic phase transition, and in this context microwave plasma synthesis appears rapid enough to prevent particle sintering and subsequent transition to the monoclinic phase. Prolonged exposure (1 h) to an O 2 or Ar plasma of either the orthorhombic or monoclinic phases did not result in a phase change as judged by powder XRD.
In conclusion, we have used a simple microwave plasma reactor for reproducible bulk synthesis of several ternary niobate and titanate phases. MIP synthesis does not require direct interaction between a solid and microwave radiation and is, therefore, amenable to any solid-state reaction. Importantly, for some reactions the plasma can also be used to heat solids into a temperature regime where additional dielectric heating can occur, resulting in a local temperature much greater than the equilibrium MIP temperature.
Experimental
Precursor oxides (Nb 2 O 5 , TiO 2 , and BaO 99.99 %) and carbonates (Li 2 CO 3 99.99 %, Na 2 CO 3 99.995 %, K 2 CO 3 99.995 %, CaCO 3 99.995+ %, BaCO 3 99.999 %, and PbCO 3 99.99+ %) were purchased from Aldrich and used as received. Plasma gases (Ar and O 2 ) were purchased from BOC. In a typical synthesis, a total mass of 2 g of precursors was ground in a drybox and pressed at 10 tonnes into a 13 mm pellet, and the pellet was transferred to the microwave apparatus via a desiccator. The pellet was placed in an alumina boat and exposed to a plasma at 13 mbar (1mbar = 10 2 Pa), 900 W, and a gas flow rate of 200 cm 3 min -1 for argon and 142 cm 3 min -1 for dioxygen. After microwave plasma irradiation, products were transferred directly into a drybox for storage. Products were characterized by powder X-ray diffraction (Philips 1800 diffractometer PH 03, with Cu Ka radiation (k = 1.54 Å) in reflection mode), scanning electron microscopy (FEI Sirion 200), and analytical transmission electron microscopy (FEI CM200 FEGTEM) that included electron energy loss spectroscopy (EELS).