HREM Study of Epitaxially Stabilized Hexagonal Rare Earth Manganites I. E. Graboy, ² A. A. Bosak,* ,²,‡ O. Yu. Gorbenko, ² A. R. Kaul, ² C. Dubourdieu, J.-P. Se´nateur, V. L. Svetchnikov, § and H. W. Zandbergen § Department of Chemistry, Moscow State University, 119899 Moscow, Russia, LMGP, UMR 5628 CNRS, ENSPG, BP 46, 38402 St. Martin d’He` res, France, and National Centre for HREM, Laboratory of Materials Science, Delft University of Technology, Rotterdamseweg 137, 2628 AL Delft, The Netherlands Received September 23, 2002. Revised Manuscript Received February 25, 2003 The formation of the high-temperature hexagonal modification of DyMnO 3 and nonexisting as bulk hexagonal EuMnO 3 , GdMnO 3 , and SmMnO 3 was observed on ZrO 2 (Y 2 O 3 ) (111) substrates at 900 °C due to epitaxial stabilization. HREM study reveals epitaxial growth of the hexagonal film of limited thickness depending on the nature of the rare earth cation. For thickness exceeding critical, the oriented stable perovskite form grows semicoherently on the hexagonal phase. The interface of two polymorphs is not abrupt and involves the formation of the transition zone with the characteristic pyramid-like shapes on the top of the hexagonal layer. The typical structural defects in the hexagonal RMnO 3 films are described. 1. Introduction RMnO 3 compounds, where R is a trivalent rare earth cation, possess perovskite structure for rare earth cations with larger ionic radius. Stable hexagonal LuMnO 3 -type structure (space group P6 3 cm) has been found for RMnO 3 compounds in the case of R with small ionic radius (Ho-Lu, Y, Sc). This structure can be described as dense oxygen-ion packing (ABCACB) with Mn 3+ ions having coordination number CN ) 5 (5-fold trigonal bipyramidal coordination) and R 3+ with CN ) 7 (7-fold monocapped octahedral coordination) 1 (see Figure 1a). YMnO 3 , which belongs to this structural type, is considered to be potential ferroelectric material for electronic applications. 2 A perovskite phase with space group Pnma (see Figure 1b) can be obtained for R ) Y, Ho-Lu instead of hexagonal RMnO 3 by high-pressure synthesis, 3 by “soft chemistry” synthesis, 4 or by the epitaxy on per- ovskite substrates. 5 The requirement of high pressure can be easily understood, taking into account the significant decrease of the unit cell volume (8-9%) from hexagonal to perovskite structure, 1 as shown in Figure 2, where the normalized unit cell volume is traced vs tolerance factor t. When considering the RBO 3 series of various 3d elements, the stable hexagonal structure is only observed for RMnO 3 compounds, but metastable hexagonal phases are known for some gallates and aluminates. It is interesting to note that both the perovskite and the hexagonal series correspond to nearly linear plots in Figure 2. As both series are observed in the same tolerance factor t range, it seems that there is no geometrical limitation on hexagonal phase formation. So the criteria of a polymorph forma- tion should be of an energetic nature correlating with an energy difference between an octahedron in perovs- kite and a less symmetrical polyhedron of 3d ionss trigonal bipyramidsin hexagonal phases. 6 For DyMnO 3 the free energies of two polymorphs are very close: stable modification is perovskite, and the hexagonal phase was obtained by quenching from high temperature (g1600 °C). 4 But for larger rare earth ions such as Gd 3+ direct extrapolation of transition temper- ature gives too high phase transition temperature (about 2800 °C, which is well above the melting point of manganites). In our previous work 7 we have demonstrated that a suitable method for synthesis of metastable hexagonal manganites is epitaxial stabilization. The calculations made using available data for stable hexagonal phases show that ZrO 2 (Y 2 O 3 ) (111) cubic substrate (the atomic ratio Y/(Y + Zr) ) 0.15) has in-plane lattice parameters closest to those of hypothetical hexagonal phases (R > R Dy ) (Figure 3) and excellent coincidence of oxygen crystallographic positions at the interface. Thin epi- taxial films of hexagonal RMnO 3 , which are unstable in bulk under usual conditions, were deposited by the MOCVD technique on appropriate (111) ZrO 2 (Y 2 O 3 ) * Corresponding author: Tel.: +007 (095) 9391492. Fax: +007 (095) 9391492. E-mail: bossak@inorg.chem.msu.ru. ² Moscow State University. CNRS. § Delft University of Technology. (1) Yakel, H. L.; Koehler, W. C.; Bertaud, E. F.; Forrat, E. F. Acta Crystallogr. 1963, 16, 957. (2) Fujimura, N.; Ishida, T.; Yoshimura, T.; Ito, T. Appl. Phys. Lett. 1996, 69, 1011. (3) Waintal, A.; Chenavas, J. Mater. Res. Bull. 1967, 2, 819. (4) Szabo, G. The`se, University of Lyon, Lyon, France, 1969. (5) Bosak, A. A.; Kamenev, A. A.; Graboy, I. E.; Antonov, S. V.; Gorbenko, O. Yu.; Kaul, A. R.; Dubourdieu, C.; Senateur, J.-P.; Svechnikov, V. L.; Zandbergen, H. W.; Holla¨ nder, B. Thin Solid Films 2001, 400, 149. (6) Reznitskii, L. A. Neorg. Mater. (in Russian) 1984, 20, 743. (7) Bosak, A. A.; Dubourdieu, C.; Se´ nateur, J.-P.; Gorbenko, O. Yu.; Kaul, A. R. J. Mater. Chem. 2002, 12, 800. 2632 Chem. Mater. 2003, 15, 2632-2637 10.1021/cm021315b CCC: $25.00 © 2003 American Chemical Society Published on Web 05/31/2003