Calcium Cobal Oxide: the first oxide‐based misfit nanotube ‐ from nanostructure to properties

Abstract

Though misfit layered compounds (MLC) have been known for some time, 1 the interest in these materials has increased in recent years. Their unique structure allows them to behave like electron crystal, showing a phonon glass behavior and then, exhibiting superior thermoelectric properties. 2 MLC can be considered as inter‐grown materials with a general formula [(MX) 1+x ] m [TX 2 ] n , where M is a rare earth (Pb, Sb, etc); T is Ti, V, Cr, Nb, etc. and X is S, Se. 1 In 2011, nanotubes (NTs) based on MLC were synthetized for the first time in large amounts. 3 Later, the syntheses were generalized to many other chalcogenide systems which were extensively studied at the local scale by TEM and associated techniques. 4 However, until now, analogous synthesis of oxide‐based MLC NTs has not been demonstrated yet. Here, we report a chemical strategy for the synthesis of calcium cobalt oxide‐based misfit NTs. A combination of high‐resolution STEM (HRSTEM)‐HAADF imaging (including image simulations), spatially‐resolved EELS (SR‐EELS), electron diffraction, and density functional theory (DFT) calculations are used to discover the formation of new phase within these nanotubes. This new phase significantly differs from bulk starting material, inducing different electronic properties. The bulk starting material, Ca 3 Co 4 O 9 , has a misfit layered structure ( Fig. 1 ), consisting of alternate stacking of CoO 2 layers and Ca 2 CoO 3 layers with a periodicity of 1.05 nm along the c axis. After the hydrothermal synthesis of Ca 3 Co 4 O 9 , several NTs (inset of Fig. 2a ) are observed with typical lengths of several hundred nanometers. Sandwiched between bright layers (spaced by 0.86 nm), two other atomic layers, with weaker intensities, can be distinguished (purple arrows in Fig. 2a ). These facts suggest that one layer from the initial bulk structure is now missing. To achieve further insight into the chemical nature of this missing layer, SR‐EELSelemental quantification was performed by using the Ca‐L 2,3 and Co‐L 2,3 edges ( Fig. 2b ). The bright layers perfectly match the areas with the lowest Ca/Co ratio (red arrows in Fig. 2b ) and correspond to one of the two cobalt sub‐systems (CoO 2 or CoO). Following the same reasoning, the two layers are related to the Ca sub‐system (CaO) and the Co sub‐system (CoO 2 or CoO), respectively. Thus, it can be concluded that one CaO layer is missing in the NT. A novel structure CaCoO 2 ‐CoO 2 was created by removing one CaO layer from the bulk structure which was then optimized by DFT structural relaxation. To confirm this new structure, we have performed a comparison between the experimental and simulated HRSTEM‐HAADF images, using the DFT‐relaxed structure as input for the simulation. As it can been seen from Fig. 3 , there is an excellent agreement between the simulated and experimental micrographs. The excellent consistency between HRSTEM, SR‐EELS, and DFT calculations support and confirm our proposed structure for the new calcium cobaltite phase. In light of this new structural and chemical information, a growth mechanism for these NTs is proposed. In addition, we will detail the electronic properties of the new MLC phase which we predict as semiconducting in nature in contrast with the bulk phase which is metallic. 5,6