Unit-cell symmetries and Raman spectra of calcium- and neodymium-doped barium cerate proton-conducting ceramic electrolytes Robert C. T. Slade,*" Sara D. Flint," Alison Holloway," Narendra Singh,"*b Lubomir Smrcok' and Daniel Tunega' aDepartmentof Chemistry] University of Exeter, Exeter, UK, EX4 4QD bDepartment of Physics, Gaya College, Gaya 823001, India 'Institute of Inorganic Chemistry, Slovak Academy of Sciences, SK-842 36 Bratislava, Slovak Republic Combined X-ray powder diffraction and Raman spectroscopic investigations of the systems BaCe, -,Ca,O, -a and BaCe, -,Nd,O, -o: at ambient temperature are presented. X-Ray studies involving in-depth investigation of diffraction profiles show that the unit cell (a crystal average) is truly orthorhombic throughout the composition ranges of both systems.The Raman spectra for BaCe, -xCax03-a are similar to that for BaCeO, itself, indicating little distortion of the structure when calcium is dopant. A feature at around 610 cm-' in the spectra for BaCe, -,Nd,O,-,, and not evident in the spectra of Ca-doped samples, is associated with modes involving the rare-earth-metal dopant. Studies of proton-conducting ceramics are comparatively recent. In 1981 Iwahara et al.' found that certain perovskite- type oxides exhibit appreciable ionic conductivity (crzlop2S cm-l), at temperatures in the range 600 <T/"C <800, in atmospheres containing hydrogen or water vapour. Protonic conduction in such materials is a result of reactions of oxygen vacancies (which are produced by partial substitution by an aliovalent cation, e.g.YIrl on Ce" sites) and/or holes with moisture or H, to generate OH- ions, the protons of which can migrate by jumping to a neighbouring (unprotonated) oxygen. The first such materials studied were of the type SrCel-xM1ll,O,-o: (MI1'= trivalent metal and a= number of oxide ion vacancies in the lattice per formula unit).' Further studies have included other systems with trivalent dopants (BaCe, -xM111x03 CaZr, -xM111x03 -a, -a, BaZr, -xM111x03-a, BaTh, -xGd,032-12 and SrZr, -xM111x03-a, also the system BaCel-,Ca,0,-a13-16 (Ca is a divalent dopant). Recent developments have included novel complex perovskites of the types A2BC06 and A,BC209 (A =Sr, Ba; B, C =metals, e.g.BC, =CaNb,), with oxide vacancies introduced by varying the B :C atomic These materials exhibit protonic conductivities greater than those of previous systems. The unit-cell symmetries appropriate to the structures of BaCeO, and the derived protonic conductors BaCe, -,M,03 -a have been the subject of much confusion in the literature, this confusion being consequent, in part, on possible slight distor- tions (rotations of CeO, octahedra) which would result in a symmetry lower than cubic. Early studies proposed a cubic (Pm3m),199" tetragonal (P4/mbm)21 or orthorhombic (Pbnm)22*23cell. The results of X-ray powder diffraction studies are determined largely by the positions of the 'heavy' (high atomic number) atoms present and are 'insensitive' to distor- tions originating largely in the repositioning of 'light' atoms (0in cerates). The details of X-ray diffraction profiles do not, however, appear to have been the subject of studies at high resolution.A single-crystal X-ray diffraction study of SrCeO, showed that compound to belong to the space group Pbr~rn.,~ In experimental studies of cell symmetries in the systems BaCe, -xMx03-a no studies combining high-resolution diffraction techniques and spectroscopic (Raman) measure- ments on the same samples have been reported. Confusion in the previous literature could have arisen from comparison of the results of X-ray and Raman studies carried out by separate teams on different samples with possibly differing thermal histories.Combined studies provide definition of unit-cell symmetry (from diffraction) prior to interpretation of the appearances of Raman spectra, and we now present such studies of the systems BaCe, -,CaXO3-a and BaCe, -xNd,O, -a at ambient temperature. Experimenta1 Materials Doped barium cerates BaCe, -,Ca,O, -a (x =0.02, 0.05, 0.10, 0.15) and BaCe, -xNd,O,-a (x =0.02,0.05,0.10) were prepared from dry powders by standard ceramic routes as described previo~sly.~*~*'~Sample compositions were verified by both X- ray fluorescence (XRF) and energy dispersive analysis by X- rays (EDAX) techniques, with BaCeO, as standard. Intended (by design of reaction mixtures) and empirical compositions agreed well within experimental error (at the relatively high mass percentages of Ba and Ce in these materials, and with the isostructural parent BaCeO, as standard, uncertainties in elemental percentages are small fractions of one percent of the empirical values; derived uncertainties in x are small, even for x =0.02).X-Ray powder diffraction studies Initial X-ray powder diffraction studies (in Exeter) wete per- formed using Ni-filtered Cu-Ka radiation (A= 1.54178 A) and a Philips PW1050 goniometer adapted in-house for step- scanning and digital acquisition of diffraction profiles at ambi- ent temperature. Initial profiles, recorded with a step size of 28=0.01" and a dwell time of 8 s, were as reported in the literature (e.g.refs. 6,15), consistent with the formation of single perovskite phases.Close examination of these profiles (by computer expansion of the axes in 28) revealed evidence for possible structure in the 'lines' at 28>40". Further profiles were recorded over the range 39 <20 <61", with a step size of 0.01" and a dwell time of 160s. Fig. 1 and 2 show sections of these profiles for single phases in the systems BaCe, -,Ca,O, -a and BaCe, -,NdXO3 -a respectively. Each 'line' in the range 39 <28 <61" is at least a triplet, and it follows that, as for the BaCeO, parent, all the phases studied are of lower symmetry than cubic or tetragonal at ambient temperature. High-resolution powder diffraction studies were performed in Bratislava using a Stoe STADI-P diffractometer in trans- mission mode and configured with a curved Ge(ll1) primary beam monochromator and linear position-sensitive detecto:.Strictly monochromated Cu-Ka, radiation (A=1.540598 A) J. Mater. Chem., 1996, 6(4), 629-633 629 I I I I 1 400 40.5 41 0 41.5 420 500 505 51 0 515 520 5815 59'0 595 600 605 26/degrees Fig. 1 Sections of the X-ray powder diffraction profiles (Exeter) for the system BaCe, ,$2aXO3 for (top to bottom) x =O 15,O 10 0 05,O 02,O 0 I I I I I 400 405 41 0 41 5 420 500 505 51 0 51 5 520 58 5 590 595 60.0 605 2Megrees Fig. 2 Sections of the X-ray powder diffraction profiles (Exeter) for the system BaCe, xNd,O, a for (top to bottom) x =O 10, 0 05, 0 02, 0 0 (see above) Full structure determinations, and was used to scan the diffraction profiles at ambient temperature and Niel~en~~ in the angular range 20 <28 <120" in steps of 0 02" All samples in particular the fine detail of 0-atom positions, would require thus analysed were strong absorbers and the orthorhombic neutron powder diffraction data The purpose of this study unit-cell symmetry led to profiles which, relative to those for was to examine symmetry and determine lattice constants a related cubic cell, had additional fine structure and a plethora of new peaks of low intensity, rendering it difficult to maintain Raman spectra an equal precision for all estimated peak positions Rietveld Spectra were collected in Bratislava at ambient temperature techniquesz5 were therefore applied (giving the advantage of using a JEOL JRS-SZ Raman spectrophotometer equipped an additional, structural, constraint in lattice parameter with a double-grating monochromator An argon ion laser refinement procedures) via a local version of the program was used as the source, providing excitation at 488 0 nm (blue DBW3226 Atomic positions, taken to be those determined line) The power of the beam at the sample was ca 350mWfrom single-crystal data for SrCeO, (space group Pbnrn) by and the scattered light from the powder samples was collected Ranlrav and Niel~en,~~ were held constant The lattice param- at the right-angle scattering geometry eters, the so-called profile parameters and the overall isotropic temperature factors were allowed to vary, and the background was modelled with a polynomial in 28 Rietveld refinements Results and Discussion converged with R, =8-1 1% Fig 3 shows a typical empirical diffraction profile obtained Results of previous studies -a) and the subsequent fit The use of neutron powder diffraction techniques might have in Bratislava (for BaCeo 98Nd0 0203 to the profile after Rietveld refinement The profile at high been expected to resolve the problem of unit-cell symmetry in angle (95 d28 d 120") is of particular interest, clearly showing these systems A neutron study by Jacobsen et showed splittings that would not arise from a cubic cell In fitting the the structure of BaCeO, itself at ambient temperature to profile, atom positions were constrained to be those of Ranlrav belong to the orthorhombic space group Pbnrn, the later studies 630 J Muter Chem , 1996,6(4), 629-633 15K 1OK 5000 7 I v) v)C s 0Y .-a v)t o 0.0 40.0 L 60.O 80.0 c-. 100.- i s c.- II In 111 11 1111 II 1111 111 I I ni I I 111 I II u ir IIIP 11 I 111 1111 111 144 II 111 III II *0° I 0 0.o 95.0 100. 105. 110. 115. 2Wdegrees Fig. 3 X-Ray powder diffraction profile (Cu-Ka, radiation) for BaCeo,,,Nd,~o,03~a at ambient temperature. The solid lines show the fit and difference plot obtained after constrained Rietveld refinement (see text) of the data. Vertical bars denote the positions (2O/degrees) corresponding to predicted d spacings. DMPLOT software was used for plotting.36 by Longo et aLZ7 were, however, interpreted in terms of a tetragonal cell at ambient temperature, and those of Predu and Dinescu2* in terms of a tetragonal cell at T< 427 "C and a cubic cell at higher temperatures. A recent neutron diffraction study by Knight29 has indicated that the correct interpretation could be the following sequence of structural changes: Pbnm (orthorhombic) at ambient-290 "C, Incn (orthorhombic) at 290-400 "C, F32/n (rhombohedral) at 400-900 "C, Pmjm (cubic) at T >900 "C.There is also confusion as to the unit-cell symmetries appro- priate to doped perovskites of the type BaCe, -xMx03-a. Early studies assumed cubic cells (e.g. refs. 6,20), but neutron diffrac- tion studies of Y-and Gd-doped samples (x=O.lO) at ambient temperature demonstrated the same Pbnm (orthorhombic) structure found for the parent BaCeO, under the same con- dition~.~~Scherban et al. 31-33 measured Raman spectra as functions of dopant, temperature and composition. They inferred a variable unit-cell symmetry (but did not report diffraction studies), the symmetry observed depending on temperature, dopant identity and level of doping (Pbnrn=>P4/rnbrn*Prn3rn with increasing temperature and/or x).It follows that the ambient-temperature defect structure may not be linked directly to the high-temperature conduction mechanism (applying in a phase at high temperature) and, further, that conductivity studies pertaining to different tem- perature or composition regimes should not be inter-related in too simplistic a manner (structural changes have generally been neglected).Scherban et inferred a simple composition dependence of unit-cell symmetry at ambient temperature for Nd as dopant (uiz.x=O.O2, Pbnrn; 0.05, P4/mbrn; 0.10, Prngrn), but Knight and B~nanos~~ found no evidence in neutron diffraction studies for Nd-concentration-dependent phase transitions. Raman spectra for SrCeO, and for the phases SrCe,-,Yb,O,-, have been interpreted on the basis of the factor group D2,, for the space group Pbnrn by Scherban et a1.31,33 and Kosacki et aL3' respectively. Scherban et aL3, found no apparent temperature dependence of unit-cell sym- metry for SrCeO,. Unit cells determined in this work For all materials studied in this work, the ambient temperature structures have orthorhombic unit cells (see above).The lattice parameters determined by Rietveld refinement techniques in the space group Pbnm (see above) for compositions in the systems BaCe, -,Ca,O, -,and BaCe, -,Nd,O, -a are given in Table 1. The parameters for the system BaCe,-,Nd,O,-, fit closely into the trends evident in the neutron diffraction work of Knight and Bonano~,~~ who worked at intervals of 0.04 in x in the range O.O<x<O.2. Raman spectra with Ca as dopant Raman spectra for phases in the system BaCe,-,Ca,O,-, are presented in Fig. 4. The general appearances of the spectra are very similar for all values of x and closely resemble those in spectra given by Scherban et al. for BaCeO, at ambient temperature,, and for other systems they assigned as ortho- rhombi~.~~,,~This is not surprising in view of the chemical J.Muter. Chem., 1996, 6(4), 629-633 631 Table 1 Lattice parameters for the orthorhombic unit cells" character- istic of the doped cerates BaCe, *CaXO3 a and BaCe, ,Nd,03 a Ca 0 02 0 05 0 10 6 2320( 2) 6 2297 (2) 6 2294(4) 6 2156(2) 6 2149( 3) 6 2171(4) 8 7774( 3) 8 7760(3) 8 7776(4) 0 15 6 2295( 3) 6 2143(4) 8 7733(4) Nd 0 02 0 05 6 2318(2) 6 2319(2) 6 2163(3) 6 2166(3) 8 7782( 3) 8 7791(3) 0 10 6 2294( 3) 6 2129(3) 8 7737(4) "All observed XRD lines are indexable on the basis of the space group Pbmn, numbers in parentheses denote the standard deviation in the last figure 150 250 350 450 550 650 750 wavenumber /an' Fig. 4 Raman spectra at ambient temperature for the system BaCe, ,CaXO3 a for (top to bottom) x =0 15, 0 10,O05, 0 02 similarity to those materials and our demonstration (above) of orthorhombic cells in this study The multiplet around 350cm-1 arises from the Ce-0 vibrational modes associated with the CeO, octahedra of the perovskite structure 31 35 The broad band previously reported3' at ca 200 cm-' is also evident, but peaks at lower frequencies are not observable The latter is a consequence of the available instrumentation, which employed a double monochromator (Scherban et a1 had access to a triple mono~hromator~~) Raman spectra with Nd as dopant Samples containing Nd are brown and are much stronger absorbers in the visible region than are the off-white samples containing Ca as dopant In the case of Nd-doped samples, a new feature at ca 610cm-' is seen in the spectra (Fig 5) A further difference is the higher background and signal-to-noise levels in the Nd case, possibly arising simply from higher dielectric screening in the Nd case (this would be consistent both with the brown colouration of the samples and with the disappearance of the 300 cm-' feature into the background at high Nd levels) The diffraction studies in this work demonstrated ortho- rhombic unit cells for all compositions in the system BaCe, -,Nd,03 at ambient temperature, in agreement with -GI the work of Knight and Bonanos34 The Raman spectra in Fig 4, nonetheless, have the same general appearance as those reported by Scherban et a2 32 33 Their inference, from interpret- ation of the Raman spectra, of a composition dependence of unit-cell symmetry at ambient temperature is, therefore, not supported I I 250 350 450 550 650 750 wavenumber Icm-' Fig.5 Raman spectra at ambient temperature for the system BaCe, ,Nd,O, a for (top to bottom) x=O 10, 005, 0 02 The feature at around 610cm-', not evident in the spectra of Ca-doped samples (above), is, as stated by Scherban et a1 ,32 33 associated with modes involving the rare-earth-metal dopant, a similar band and assignment have been reported by Kosacki et al 35 for SrCe1-,Yb,O3-, Consistent with that assignment, the relative intensity of that band increases with increasing dopant level in this work Conclusions Combined X-ray diffraction and Raman spectrcscopic studies have been carried out on samples within the systems BaCe, -,Ca,03 -a and BaCe, -,Nd,03 at ambient tempera- --a ture Confusion in the previous literature could have arisen from comparison of the results of separate X-ray and Raman studies of different samples with possibly differing thermal histories The following conclusions can be drawn As shown by high-resolution powder diffraction, the unit cell (a crystal average) for all compositions in the systems (with either Ca or Nd as dopant) in this study is truly orthorhombic at ambient temperature The Raman spectra for BaCe, -,Ca,03 -oL are similar to that for BaCeO, itself, indicating little distortion of the structure when Ca is dopant The postulate of Scherban et a1 ,33 that unit-cell symmetry in the system BaCe, -,Nd,03 at ambient temperature depends on x,is not supported by this study Raman spectroscopic investigations of single crystals within these systems could remove some of the ambiguities inherent in interpretation of the Raman spectra of powders We thank referees for constructive comments concerning the Raman spectra We thank the EPSRC and British Gas plc for a CASE studentship for SDF We thank the University of Exeter for a travel grant (to Exeter) for NS We thank the Earth Resources centre (University of Exeter) for analyses using XRF and EDAX techniques We thank the Royal Society for a travel grant (to Bratislava) for R C T S The programme in Bratislava is funded by the Slovak Grant Agency for Science References 1 H Iwahara, T Esaka, H Uchida and N Maeda, Solid State Ionzcs, 1981,314,359 2 A Mitsui, M Miyayama and H Yanagida, Solid State iunics, 1987, 22,213 3 H Iwahara, H Uchida, K Ono and K Ogaki, J Electrochem SOC,1988,135,529 632 J Muter Chern, 1996, 6(4), 629-633 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 H Iwahara, H Uchida and K Morimoto, J Electrochem Soc, 1990,137,462 T Yajima, H Kazeoka, T Yogo and H Iwahara, Solid State Zonics, 1991,47,271 R C T Slade and N Singh, Solid State Zonics, 1991,46, 11 1 R C T Slade and N Singh, J Muter Chem, 1991,1,441 T Hibino, K Mizutani, T Yajima and H Iwahara, Solid State Zonics, 1992,57, 303 H Iwahara, T Yajima, T Hibino, K Ozaki and H Suzuki, Solid State Zonics, 1993,61, 65 H H Huang and M Ishigame, Solid State Ionics, 1991,47,251 A F Sammells, R L Cook, J H White, J J Osborne and R C Macduff, Solid State Ionics, 1992, 52, 11 1 R L Cook, J J Osborne, J H White, R C Macduff and A F Sammells, J Electrochem Soc, 1992,139, L19 T Yajima, H Iwahara and H Uchida, Solid State Zonics, 1991, 47,177 J F Liu and A S Nowick, in Solid State Zonics IZ,eds G A Nazri, D F Shriver, R A Huggins and M Balkanski, MRS Symp Proc , 1991 210,675 R C T Slade, S D Flint and N Singh, J Muter Chem, 1994, 4,509 S D Flint and R C T Slade, Solid State Zonics, 1995,77,215 K C Liang and A S Nowick, Solid State Zonics, 1993,61,77 K C Liang, Y Du and A S Nowick, Solid State Ionics, 1994, 69,117 A J Smith and A J E Welch, Acta Crystallogr, 1960, 13,653 N Bonanos, B Ellis, K S Knight and M N Mahmood, Solid State Zonrcs, 1989,35, 179 S Shin, H H Huang, M Ishigame and H Iwahara, Solid State lonics, 1990, 40/41,910 22 A J Jacobsen, B C Tofield and B E F Fender, Acta Crystallogr Sect B, 1972,28,956 23 K S Knight, M Soars and N Bonanos, J Muter Chem, 1992, 2,709 24 J Ranlarv and K Nielsen, J Muter Chem , 1994,4, 867 25 H M Rietveld, J Appl Crystallogr ,1969,2,65 26 The Rietveld Method, ed R A Young, Oxford University Press, Oxford, 1993 27 V Longo, F Ricciarcello and D Mmichelh, J Muter Sci, 1981, 16,3503 28 M Predu and R Dinescu, Rev Roum Chrm ,1976,21,1023 29 K S Knight, Solid State Zonics, 1994,94, 109 30 K S Knight, M Soar and N Bonanos, J Muter Chem, 1992, 2,709 31 T Scherban, R Villeneuve, L Abello and G Lucazeau, Solid State Commun ,1992,84,341 32 T Scherban, R Villeneuve, L Abello and G Lucazeau, Solid State Ionics, 1993,61, 93 33 T Scherban, R Villeneuve, L Abello and G Lucazeau, J Raman Spectrosc , 1993,24,805 34 K S Knight and N Bonanos, Solid State Zontcs, 1995,77, 189 35 I Kosacki, J Schoonman and M Balkansi, Solid State Zonics, 1992,57,345 36 H Marciniak, S Gierlotka and B Palosz, Advanced graphics tools for the analysis of powder difraction data XVIEW and DMPLOT, in Collected Abstracts of the 4th European Powder Difraction Meeting, Chester, 1995, p 103 Paper 5/04231C, Received 30th June, 1995 J Muter Chem, 1996, 6(4), 629-633 633