J. MATER. CHEM., 1991, 1(6), 955-963 Growth of ‘124’ and ‘247’ Phases studied by High-resolution Transmission Electron Microscopy in HoBa,Cu,O,- Ceramics prepared under Normal Oxygen Pressure Yong Van and Marie-Genevieve Blanchin* Departement de Physique des Materiaux, (UA CNRS no 7 72),Universite Claude Bernard, 69622 Villeurbanne Cedex, France High-resolution transmission electron microscopy (HREM) has been used to study the structure of superconductor HoBa,Cu,O,-, (‘123’) ceramics prepared under normal oxygen pressure. HREM images reveal that lattice disorder as well as high densities of stacking faults occur locally in the perfect ‘123’ structure. HREM micrographs obtained from thin areas of those faulted regions show a good correspondence to projections of the HoBa,Cu,O, (‘124’) and Ho,B~,CU,O,~ (‘247’) structures proposed from neutron diffraction studies.The b/2 shift between double Cu-0 chains occurring in both structures is confirmed by comparison between experimental and simulated HREM images corresponding to different projections of these structures. This demonstrates that owing to departure from nominal composition, stacking sequences occur locally in the specimens, which can give rise to growth of ‘124’ and ‘247’ phases normally obtained under higher oxygen pressure. Keywords: Superconductor; ‘123’ phase ; ’I24’ phase ; ‘247’ phase ; High-resolution transmission electron microscopy Since superconductivity around liquid-nitrogen temperature was discovered,’ the R,Ba,Cu, (with rare-earth- metal elements R =Y, Ho etc.)family proved to be one of the most interesting groups of high- T, superconductors, from the points of view of fundamental research and applications.In the Y2Ba4Cu6 +” family, three members were disco-vered.’ The first member (n=0) is YBa,Cu,O, --x,well known as the ‘123’ phase, and is prepared under normal oxygen pressure. Two other members, i.e. for n= 1, Y2Ba4C~7015+x (the ‘247’ phase) and for n=2, YBa2Cu408 (the ‘124’ phase), are normally obtained under higher oxygen The ‘123’ phase has an oxygen-deficient perovskite structure. Two kinds of oxygen vacancies are ordered in the YBa2Cu307 structure: the first forms a two-dimensional arrangement in the Y planes whereas the other forms vacancy chains alternat- ing with Cu-0 chains parallel to the b axis in the Cu(1) planes.The YBa2Cu408 structure differs from YBa2Cu307 in that single linear Cu-0 chains parallel to the b axis are replaced by double Cu-0 chains with edge-sharing, square- planar oxygen coordination; this induces interesting properties (a strong dependence of T, on hydrostatic pressure, thermal expansion and compressibility anomalies and anisotropy in the a-b The Y,B~,CU,O~~+~ compounds have a structure that consists of alternating ‘123’ and ‘124‘ units and it was reported that T, could be varied between 14 and 70 K by changing the stoichiometry in the range+0.092 x 2 -0.72.6-9 It follows that combined investigation of ‘123’, ‘124’ and ‘247’ phases allows the study of the influence of the single and double Cu-0 chains on high-T, supercon- ductivity.Like the YB~,CU,O,-~ compounds, the high-T, HoBa2Cu307--x superconductors have an orthorhombic structure corresponding to an oxygen-deficient perovskite structure (depicted in Fig. 3, later) with lattice parameters a = 0.3819 nm, b=0.3893 nm, c= 1.1653 nm at room tempera-ture,” but they exhibit quite high critical current densities.” They can contain high densities of stacking faults correspond- ing to the occurrence of double Cu-0 chains, whose structure and mechanism of formation have been detailedI2 with respect to similar features for the Y-Ba-Cu-0 ~ystem.’~-’~In the present paper, we focus our attention on local growth of ‘124’ Fig. 1 Medium-magnification micrograph from a thin area of a HoBa,Cu,O, -x specimen.Regions P and P’ correspond to the perfect ‘123’ structure whereas high densities of stacking faults are observed Fig. 2 Enlargement of an HREM image along the a or b axis from in regions D, D‘ and D” region D” in Fig. 1 and ‘247’ phases in HoBa2Cu307 -x ceramic superconductors prepared under normal oxygen pressure. The structure of these phases has been investigated by high-resolution trans- mission electron microscopy (HREM). The results are dis- cussed through comparison of HREM experimental images with simulated images computed from the structural models proposed in the literature. Experimental High-T, HoBa2Cu307 -x ceramic samples were produced by ALCATEL-ALSTHOM Recherche Laboratories, Marcoussis, France.The samples were made from powder synthesized by a solid-state reaction at 950 “C under atmospheric pressure for 48 h. Cylinders (4~4.5mm, 1=40 mm) were shaped by isostatic pressing and were sintered, i.e. the cylinders were heated for ca. 13 h at 920 “C and maintained at this tempera- ture for 3 h under pure oxygen at normal pressure. The specimens were then slowly cooled to 450 “C over 10 h, then maintained for 12 h at 450 “C for oxidation and slowly cooled to room temperature over 12 h under the same atmosphere. According to the data available in the literature,20 such a procedure will give a value for xin HoBa,Cu307 --x of ca. 6.95. Susceptibility measurements gave T,z86 K. The specimens t = 0.388nm projected potential t= 1.164 nm C 1 00 0 cu t = 1.940 nm 1-2.716 nm J.MATER. CHEM., 1991, VOL. I for electron microscopy were prepared by the standard tech- nique for ceramic oxides. HREM observations were performed using a JEOL 200CX-UHP2S microscope at 200 kV with a top entry stage. The pictures were obtained within times short enough to avoid deterioration of the oxygen content.I3 The objective lens spherical aberration constant was determined as C,=O.89 mm. The simulation of the HREM images was performed on a microcomputer using a program based on multislice theory.13 Results Stacking Faults in the ‘123’ Phase The specimens were processed having the nominal compo- sition HoBa,Cu,O, -x. X-Ray diffraction experiments showed that the crystals were chiefly isostructural with HoBa,Cu307 (space group Pmmm).Nevertheless local departures from that structure were revealed by HREM images at higher spatial resolution.Fig. 1 is a medium-magnification electron micro- graph from one of the samples. Regions P and P’ correspond to the perfect ‘123’ phase, whereas high densities of planar defects are observed in regions D, D’ and D”. Fig. 2 is a high-magnification HREM image from the region D” of Fig. 1, seen along the [100) or the [OlO] direction. This Af= -45 nm Af = -50 nm Af = -55 nm.. Fig. 3(a) J. MATER. CHEM., 1991, VOL. 1 (b) Af=-45 nm Af=-50 nrn Af = -55 nrn t-0.388 nrn projected potential t= 1.164 nrn 1 00 t = 1.940 nrn0C" 0 0Ho t=2.716nrn Fig.3 Series of computer simulated images of the '123' structure for increasing crystal thickness, t, depending on defocus of the objective lens A$ (a) [loo] projection, (b) [OlO] projection image was taken close to the Scherzer defocus ~ondition'~ and represents the projected potential of the structure within a good approximation in such a thin foil. The large dark dots correspond to columns of Ho and Ba atoms, and Cu columns are seen as weaker dark spots between the Ba and Ho columns. However, the HREM images along the [loo] or [OlO] direction of the ordered '123' structure cannot be discriminated without difficulty owing to the limited resolution of the 200 kV electron microscope used for the observations.As shown by the images simulated for thin crystals observed close to the Scherzer defocus condition (Fig. 3), the occupied oxygen columns ([loo] projection), or the oxygen vacancy columns ([OIO] projection) located in the Cu(1) planes (between the Ba planes) are always seen as the brightest dots for both projections. Despite the lack of resolution with respect to imaging of the oxygen sublattice, such a row of very bright dots can be considered as the fingerprint of the Cu(1) planes in the corresponding HREM images. In the high- magnification HREM micrograph of Fig. 2, the planar faults readily appear as double rows of bright dots between the Ba planes. Correspondingly, the planar defects may be thought to consist of a pair of Cu(1) planes, and thus the additional material is expected to have the composition CuO.Detailed contrast studies of such extrinsic stacking faults in the HoBa- cuo'2 and the YBaCu0'3-'9 systems have been published together with the corresponding structural models. Clearly some faults exhibit a mirror symmetry (marked M in Fig. 2) Fig. (a) High-magnification HREM image along the a or axis across the regions P and D in Fig. 2. (b)Optical diffraction pattern from the area '124' in the negative of (a). (c) Optical diffraction pattern from the area '247' in the negative of (a) J. MATER. CHEM., 1991, VOL. 1 Y Ba2Cu408 (Ammm) Fig. 5 (a)Enlargement of the HREM image of the region ‘124‘ in Fig. 4(a). The projection of the ‘124‘ unit cell is outlined, (b) Structure of YBa,Cu,O,, showing the double CuO, chains connecting planes of CuO, pyramids (from ref.2) whereas other faults exhibit a glide symmetry (G in Fig. 2). A gradual change from mirror to glide character can occur due to interaction with permutation t~inning’~?’~.’~ (marked PT in Fig, 2). On the other hand, a change from single CuO plane to double CuO layers can also be observed (circled in Fig. 2). Local Growth of the ‘124’ Phase It should be noted that the density of the CuO double layers can be quite high in some regions of the crystals, as seen in Af= -40 nm Af = -45 nm Af = -50 nm Af = -55 nm Af I-60 nm Fig. 2. The intervals between the stacking faults are then only a few unit cells of ‘123’ phase. The new phase can be thus expected to grow from the ordered arrangement of the faults, as shown by the HREM image of Fig.4(a), obtained across the regions P and D of Fig. 1. This micrograph clearly reveals that region P corresponds to the perfect ‘123’ phase whereas region D corresponds to ‘124’ and ‘247’ phases. The ‘124‘ phase, which has been extensively studiedin the YBaCuO has the orthorhombic structure Ammm at room temperature [depicted in Fig. S(b)].It differs from the ‘123’ phase by the intercalation of a double CuO layer, i.e. double t= 1.54 nm t= 2.32 nm f = 3.09 nm Fig. qa) J. MATER. CHEM., 1991, VOL. 1 t= 1.54 nrn t= 2.32 nrn t = 3.09 nrn Af= -40 nm Af= -50 nm Af= -55 nm Af= -60 nm Fig. 6 Series of computer-simulated images of the ‘124‘ structure for increasing crystal thickness, t, depending on defocus of the objective lens AJ (a)[IOO] projection, (b)[OlO] projection Cu-0 chains, between the Ba planes [Fig.5(b)]. of this structure, close to the Scherzer defocus condition and In HoBa,Cu,08 the Y atoms are readily replaced by the Ho for increasing crystal thickness, are reproduced in Fig. 6. In atoms. The simulated images observed along the a and b axes this case the HREM images along the a and b axes can be c =5 Y2Ba4Cu& 5 (Arnmrn ) Fig. 7(a) Enlargement of the HREM image of region ‘247’ in Fig. qa);inset is the simulated image. (b)Structure of Ho,Ba,Cu,O1, consisting of alternating ‘123’ and ‘124‘ units (from ref. 2) discriminated easily from the b/2 shift within the double Cu-0 chains.21 Fig.5(a) is an enlargement of the region '124' in Fig. 4(4.Part of the micrograph of Fig. 5(a), as delineated, exhibits a good correspondence to the projection of HoBa,Cu,O, along the b axis [Fig. 6(b)]:the metal columns are seen as dark dots and the double CuO layers are imaged as rows of very bright dots exhibiting a mirror symmetry. According to Fig. 6(b),the experimental image matches best (a 1 Ba--c Hod I Ba4 . , ,. c . . ., . , Ba -.,.-....,-, , .. Af= -45 nm HO-Ba-Ba-Ho-Ba-., .. . Ba ----c Ho 4 Ba-c projected potential Af= -55 nm Af= -65nm J. MATER. CHEM., 1991, VOL. 1 with the simulated images along [010] direction for crystal thickness in the range 1.54-2.32 nm and at defocus values between -45 and -55 nm.Fig. 5(a) clearly reveals that the perfect stacking sequence of double CuO layers, leading to local growth of the '124' phase, does not extend over very large volumes of crystal, but is limited by the occurrence of structural defects, i.e. double CuO layers exhibiting a glide character [labelled G in t= 1.54 nm t = 3.08 nm t ~4.61nm Fig. 8(a) J. MATER. CHEM., 1991, VOL. 1 96 1 (b1 t= 1.55 nm t= 3.10 nm t = 4.65 nm Af=-&I In-l projected potential Af= -55 nm Af= -65nm Fig. 8 Series of computer simulated images of the '247' structure for increasing crystal thickness, t, depending on defocus of the objective lens Af (a)[loo] projection, (b)[OlO] projection Fig.5(a)] or terminating in the crystal [circled in Fig. 5(u)]. Local Growth of the '247' Phase Moreover some local lattice disorder is superimposed with It can be seen from Fig. 4(a)that the '247' phase can be found that large density of extended defects, this being apparent besides regions corresponding to the '123' and '124' phases. showed that from the corresponding variations in the dot intensity of the Previous studies of the YBaCuO ~ystem~-~*~~ experimental HREM images. the '247' structure consists of alternating units of '123' and J. MATER. CHEM., 1991, VOL. 1 -1 t= 1.'09nm t=2.18 nm t = 3.27 nm ............... Ho ............ Ba Af= -55 nm Ba'Ba Ho'Ho Ba Ba projected potential Af= -75 nrn I Af = -95 nm Af= -1 15 nrn Fig. 9 (a) High-magnification HREM image of a '247' region observed along the [l 101 direction; inset is the simulated image.Seen along [OOI] direction, the shift in the lattice fringes (marked by dashed white lines) can be observed which shows a glide character of the CuO double layers in this projection. (b)Computer-simulated images of the '247' structure in the [l lo] projection for increasing crystal thickness, t, depending on defocus of the objective lens Af J. MATER. CHEM., 1991, VOL. 1 ‘124‘ as depicted in Fig. 7(b).The unit cell is A-centred (space group Ammm), the neighbouring ‘1 23’ blocks being translated by b/2 with respect to each other. A doubling of the stacking sequence along [OOI] is thus necessary, leading to a large c parameter [Fig.7(b)]. Fig. 4 reproduces optical diffraction patterns from areas of the negative of Fig. 4(a) corresponding to ‘124‘ [Fig. qb)] and ‘247’ [Fig. 4(c)]. The superstructure spots along the c axis are clearly viewed in Fig. qc), which corresponds to the long period arising from ‘123’ and ‘124‘ alternating units in the ‘247’ phase. This is consistent with the results given for the YBaCuO system by Beeli et aL21 Fig. 7(a) is a high-magnification micrograph of the region ‘247’ in Fig. 4(a), in which alternation of ‘123’ and ‘124’ units can be clearly seen (white arrows); the double CuO layers exhibit a glide character as in the projection of the ‘247’ structure along the a axis [Fig. 7(b)]. Simulated images of the ‘247’ phase observed close to Scherzer defocus condition along the a and b axes are reproduced in Fig.8. The double CuO layers exhibit a mirror symmetry in the simulated images along the b axis, whereas they show a glide symmetry along the a axis as observed in the experimental image of Fig. 7(a). Here again the extension of the ‘247’ phase appears to be limited by stacking faults. The existence of the ‘247’ phase in our specimens was confirmed by HREM observations along other directions. Fig. 9(a) is an HREM image from a region observed along the [l 101direction, in which the alternation of ‘123’ and ‘124‘ units is found again. The series of simulated images of the ‘247’ phase viewed along the [llO] direction at different defocus values and for increasing crystal thickness are shown in Fig.9(b). A good match with the experimental image is found for defocus values between -95 nm and -105 nm in the thickness range 1.09-3.27 nm. Accordingly variation in the contrast of the experimental image is observed owing to the change in the foil thickness. Although no atomic structure image can be achieved along this direction owing to the limited resolution of the 200 kV electron microscope used for the observation, the shift in the lattice fringe imaged in Fig. 9(a) clearly reveals that the double CuO layers exhibit a glide character. This confirms the existence of a b/2 shift within the double Cu-0 chains of the ‘247’ phase present in our specimens, in agreement with the results obtained for the YBaCuO system for neutron diffraction6-’ and electron microscopy.2o Discussion HREM observations showed that our specimens consisted of a matrix having the ‘123’ structure in which ‘124‘ and ‘247’ structures could be found locally.The present specimens were prepared by standard ceramic processing under normal oxy- gen pressure to have the nominal composition HoBa2Cu307-x. Obviously excess of CuO could occur locally, and could be accommodated by intercalation of double CuO layers. The ordering of such double layers, which can alternate with simple layers, gives rise to the stacking sequence of the ‘124‘ and ‘247’ phases. The ‘124‘ phase was first observed2 in partially decomposed powders and then as a component of multiphase thin films.Later it was shown that polycrystalline ‘124‘ can be synthesized under normal pressure using a mineralizer and very fine powders from the decomposition of nitrates.2 On the other hand the existence of the ‘247’ phase was only reported in multicomponent samples with nominal composition YBa2Cu30, -solidified under high pressures of oxygen (2900 bar) or in small single crystals grown with a high-pressure travelling solvent method at oxygen pressure of 130 bars2 The present study reveals that small amounts of ‘247’ phase can be formed locally in polycrystalline ceramic samples with nominal composition HoBa2Cu307 -synthe-sized under normal oxygen pressure. Conclusion The present study confirms that carefully controlled reactions and thermal treatments are crucial to obtain chemically and structurally homogeneous polycrystalline samples.The key role played by the ordering of double Cu04 chains in trans- formation from ‘123’ phase into ‘124‘ and ‘247’ phases is confirmed. Many other stacking sequences of ‘123’ or ‘124‘ units could be obtained in the family of homologous RBa2CU6+ compounds. Control of complex chemical reactions under moderate if not normal oxygen pressure should be investigated to generate such a series of new superconductor compounds. We thank Dr. P. Dubots and Mr. A. Wicker (Alcatel-Alsthom Recherche, France), who provided the samples, for stimulating discussions on the subject. Dr. B. Poumellec (Laboratoire des Composes Non-Stoechiometriques, Universite Paris-Sud, ORSAY, France) is gratefully acknowledged for his interest in the present study.References 1 M. K. Wu, J. R. Ashburn, C. J. Torng, P. H. Hor, R. L. Meng, L. Gao, Z. J. Hung, Y. Q. Wang and C. W. Chu, Phys. Rev. Lett., 1987, 58, 908. 2 E. Kaldis and J. Karpinski, Eur. J. Solid State lnorg. Chem., 1990, 27, 143. 3 J. Karpinski, E. Kaldis, E. Jilek, S. Rusiecki and B. Bucher, Nature (London), 1988, 336, 660. 4 P. Marsh, R. M. Fleming, M. L. Mandich, A.M. De Santolo, J. Kwo, M. Hong and L. J. Martinez-miranda, Nature (London), 1988, 334, 141. 5 P. Fischer, J. Karpinski, E. Kaldis, E. Jilek and S. Rusiecki, Solid State Commun., 1989, 69, 531. 6 J. Karpinski, C. Beeli, E. Kaldis, A. Wisard and E.Jilek, Physica C, 1988, 153-155, 830. 7 J. Karpinski and E. Kaldis, Nature (London), 1988,331, 242. 8 P. Bordet, C. Chaillout, J. Chenavas, J. L. Hodeau, M. Marezio, J. Karpinski and E. Kaldis, Nature (London), 1989, 334, 596. 9 J. Karpinski, S. Rusicki, B. Bucher, E. Kaldis and E. Jilek, Physica C, 1989, 161, 618. 10 P. Fischer, K. Kakurai, M. Steiner, K. N. Clausen, B. Lebech, F. Hulliger, H. R. Ott, P. Bruesch and P. Unternahrer, Physica C, 1988, 152, 145. 11 M. Wacennovsky, H. W. Weber, 0.B. Hyun, D. K. Finnemore and K. Ereiter, Physica C, 1989, 160, 55. 12 Y. Yan, M.G. Blanchin and A. Wicker, Physica C, 1991, 175, 651. 13 Y. Yan and M. G. Blanchin, Philos. Mag. A, 1990, 61, 513. 14 Y.Matsui, E. Takayama-Muromachi and K. Kato, Jpn. J. Appl. Phys., 1988, 27, L 350. 15 Y. Matsui, E. Takayama-Muromachi and A. Ono, Jpn. J. Appl. Phys., 1988, 26, L 777. 16 B. Domenges, M. Hervieu, C. Michel and B. Raveau, Europhys. Lett., 1987, 4, 21 1. 17 H. W. Zandbergen, R. Gronsky, K. Wang and G. Thomas, Nature (London), 1988,331, 596. 18 G. Van Tendeloo and S. Amelinckx, Phys. Status Solidi (a), 1987, K1,103. 19 G. Van Tendeloo, D. Broddin, H. W. Zandbergen and S. Amelinckx, Physica C, 1990, 167, 627. 20 B. Poumellec, Physica C, 1990, 166, 289. 21 C. Beeli, H-U. Nissen, Y. Kawamata and P. Stadelmann, Z. Phys. B, 1988, 73, 313. Paper 1/01966J; Received 26th April, 1991