J. MATER. CHEM., 1994, 4(3), 389-392 Pyrochlore-like Compounds derived from Antimonic Acid Aldo Jose Gorgatti Zarbin and Oswaldo Luiz Alves* Laboratorio de Quimica do Estado Solido, lnstituto de Quimica-UNICAMP, Caixa Postal 6754, CEP 73087-970, Campinas, Sa"o Paulo, Brazil Highly crystalline antimonic acid, H2Sb20,.1 .5H,O (CAA) with a pyrochlore-like structure has been ion-exchanged with Fe3+ and Cd2+ to form H,.,Feo.22Sb206*l .5H20 (CAA/Fe) and Ho~73Cd,,3,Sb206.1 .5H20 (CAA/Cd) phases. Thermal decomposition of these phases was studied by XRD, FTIR, TG, DSC and SEM. The final products, obtained by the thermal treatment of CAA/Fe and CAA/Cd at 11 00 "C, for 2 h, were FeSbO, and Cd2Sb207-,, respectively. The chemical homogeneity and phase purity for these compounds, and the temperature and time for reactions indicate that the use of adequate precursors has advantages, in comparison to conventional solid-solid methods involving mixtures of oxides.The search for compounds for which thermal decomposition leads to the formation of reactive or ceramic powders with technological applications, such as catalysts, semiconductors and magnetic materials, has received great attention.',2 Generally, the use of precursors has advantages over the solid-solid reaction method, in terms of temperature, time and chemical and phase purities. In this paper we investigate the possibility for FeSb04 and Cd2Sb207-compounds, which are, respectively, a catalyst for selective alkene 0xidation~3~ and a semiconductor ceramic with potential application as a gas ~ensor.~ In both cases, the powder preparation for oxides by conventional solid-state reaction requires high temperatures (ca.1100 "C) and long processing times (1-2 days).46 A precursor method can be considered if at least two cations can be atomistically distributed through a polymer7 or a particular crystal structure. In this way, the ion-exchange ( H+/Mn+) properties of crystalline hydrated oxides can lead to an interesting route for mixed-cation oxides. This paper reports the use of a pyrochlore-like phase of antimonic acid, H2Sb206.1.5H20, for the formation of ceramic oxide powders containing Fe3+ and Cd2+ as the second cation. The thermal decomposition of the ion-exchanged phases, in the range 1OO-110OoC, was monitored by X-ray diffraction (XRD), Fourier transform infrared (FTIR) spec- troscopy and scanning electron microscopy (SEM) techniques. Experimental Materials Antimony(Ir1) oxide, hydrogen peroxide solution, hydrated cadmium and iron(Ii1) nitrates, all from Merck, were used without further purification.Preparation of Crystalline Antimonic Acid (CAA) The method of synthesis of CAA was similar to that reported by Ozawa et a/.*. H,Oz (60 cm3; 31% m/m) was added to 0.04 mol of solid Sb203. The suspension was stirred vigorously for 30 h at 65 "C. After cooling to ambient temperature, the solid was separated by centrifugation and washed successively with deionized water until pH 7 was obtained. CAA was dried at 40°C under dynamic vacuum for 8 h and stored in a desiccator.Ion Exchange 100 cm3 of a 5 x lop2mol dmP3 solution of Cd(N03),-4H20 or Fe(N03),.9H,0 in nitric acid (1 x mol drnp3) was added to 1 g (2.7 mmol) of CAA. The suspension wa., stirred for various temperatures and times to achieve maximum loading. This was achieved at a 60°C temperature and times of 24 and 168 h for Cd2+ and Fe3+, respectively. The solids were separated, dried and stored in a similar way as dtxribed for CAA. Chemical Characterization The antimony, iron and cadmium determination in the ion- exchanged solids and supernatant solution were performed by flame atomic absorption using a Zeiss model FMD3 spectrometer. The H amount released after the ion-exchange + was also measured. Heat Treatment The heat treatment of the samples (0.3 g) was carried out in a furnace at temperatures varying from 100 to 1100 'C over a period of 2 h in ambient atmosphere.Then, the ,samples were cooled and stored in a desiccator. Physical Measurements Thermogravimetry (TG) and differential scanning calorimetry (DSC) curves were obtained on a Du Pont model 1090 DSC/TGA system at a heating rate of 5 'C min-' under nitrogen flux. XRD patterns were obtained using a Shimadzu. model XD-3A diffractometer utilizing Ni filters and Cu-Kx radiation with 30 kV and 20 mA, at a 2 "C min-l scan rate. Thc room- temperature measurements were carried out with the sample spread on a conventional glass sample holder. Powder silicon reflections were used for 20 calibration.Infrared spectra were recorded in the 4000-400 cm-region using Fluorolube or Nujol dispersions between alkali-metal halide windows and KBr pellets, on a Perkin-Elmer 1600 Fourier Transform Spectrometer. SEM photomicrographs were obtained on a JEOl, model JSM T-300 microscope, using the technique of coaling the dispersed samples with gold. Results and Discussion Composition, Structure and Thermal Behaviour of the Ion-exchanged Phases The antimonic acid X-ray pattern (Fig. 1)is typical of a cubic pyrochlore-like structure' and shows that the solid is well crystallized. The chemical analysis is consistent with the J. MATER. CHEM., 1994, VOL. 4 10 20 30 40 50 60 28/degrees Fig.1 XRD patterns for (a) CAA, (b)CAA/Cd and (c) CAA/Fe H,Sb,06.1.5H,0 formula. In Table 1 the composition and some characteristics of the ion-exchanged solids are presented. As can be seen in Table 1, The H+(H30f)was partially exchanged by Cd2+ and Fe3+, in contrast to Ag+ and Pb2+, whose exchanges are stoichiometric." These marked differ- ences can be associated with the selectivity of CAA towards these ions. Generally, the selectivity shows a dependence on crystallinity degree for this type of ion exchanger." The XRD patterns of CAA exchanged with Cd2+ (CAA/Cd) and Fe3+ (CAA/Fe) shown in Fig. l(b) and l(c), respectively, do not show marked differences in peak positions. This result could be taken as evidence that no structural changes occur in the pyrochlore network by the presence of the metallic cations.On the other hand, the peaks associated with the Miller indices for which the sum h +k +I is an odd number show an intensity decrease upon exchange. The latter feature has been explained in terms of a destructive interference of certain reflections caused by the presence of metallic ions in defined sites of the cubic pyrochlore-like structure.12 Similar behaviour was observed for H, -,Ag,Sb,O,nH,O phases where there exists an unequivocal dependence of the intensity decrease with the x va1~es.l~ Fig.2 illustrates the TG curves for CAA, CAA/Fe and CAA/Cd, which consist of a practically continuous 10-15% weight loss, for temperatures below 850 "C. Thermal decompo- sition of non-exchanged hydrated antimonic acid with differ- ent degrees of crystallization has been reported previously.14-16 The following steps can be proposed for the CAA prepared in this work: (i) loss of adsorbed and structural water in the 60-280 "C range yielding anhydrous H,Sb,06; (ii) loss of water and molecular oxygen (280-600 "C) as a result of the partial reduction of SbV to Sb"' with the formation of Sb6013 (Sb:"SbT0,3); and, finally, (iii) release of molecular oxygen caused by another reduction of the SbV, producing Sb,O, (Sb"'SbvO,) that is subsequently converted to Sb203 (> 1000 "C).It is important to remark that these stages are Table 1 Composition and main characteristics of ion-exchanged CAA phases ion-exchange ion loading (mequiv.Mfl+/ (M"+) composition colour g CAA) Fe3+ H,,,,Fe,,,,Sb2O6~1.5H2O brown 1.79 Cd2+ H0.73Cd0.635Sb206-1 .5H20 white 3.44 0 200 400 600 800 1000 T/"C Fig. 2 TG curves for (a) CAA, (b)CAA/Fe and (c) CAA/Cd not separate, they occur as overlapping stages. The TG curve profiles for CAA/Fe and CAA/Cd are similar to those of non- exchanged CAA; however, differences can be observed in the weight loss, mainly after step (ii). The abrupt event at tempera- tures up to 900°C was ascribed to the sublimation of Sb,03. The DSC traces (Fig. 3) show at least a similar shape for temperatures below 350 "C. For higher temperatures, the exothermic and endothermic events are related to the nature of the new phases formed.The CAA IR spectrum shows the typical broad bands in the 3400-3000 cm-I region, associated with the OH stretching of water and H30+ perturbed by hydrogen bonds; a weak band at 1660cm-l (water OH, deformation); a shoulder at ca. 1750 cm-' (H30+ deformation) and two others bands in the 780 and 450cm-' regions, attributed to the Sb-0 stretching and a combination of Sb-0 stretching and SbO, deformation, respectively. The last two bands are characteristic of pyrochlore-like structure^.'^-^^ In relation to CAA, the IR spectra of CAA/Fe and CAA/Cd show significant profile changes in the OH stretching region as a consequence of hydrogen-bond modifications. The other bands presented decreasing wavenumber shifts (<15 cm-') and widths. The last feature is particularly observable for the Sb-0 stretching band that exhibits a noticeable sensitivity to the ion-exchange loading.0 a 200 400 600 77°C Fig. 3 DSC curves for (a) CAA, (b)CAA/Fe and (c) CAA/Cd J. MATER. CHEM., 1994, VOL. 4 Thermal Decomposition CA A/Fe Phase The CAAiFe XRD patterns for the heat-treated samples at different temperatures are shown in Fig. 4. It can be seen from the patterns that the pyrochlore structure remains stable until 700°C. For higher temperatures new peaks are observed indicating the existence of a mixture containing Sb204 and FeSbO, phase^.^ At 1100 OC all the Sb204 phase is converted to Sb,O,, which sublimes, to leave only FeSbO, remaining. The FeSb0, can be indexed as a tetragonal rutile-type accord- ing to ref.19. SEM investigation of this evolution revealed, for the heat- treated sample at 1000°C, the presence of two morphologies (Fig. 5): plate-like crystallites with particle size of ca. 80 pm and aggregates of small particles (<2 pm). In fact, considering that at 1100°C only the first is observed, we can assign it to the FeSb0, phase. Fig. 6 shows the infrared spectra (1300-400 cm-l) as a function of thermal treatment. The bands near 750 and 450 cm-l undergo changes only for temperatures higher than 800°C in agreement with the XRD results. The new bands and/or splittings observed at 1000 "C are due to vibrations involving antimony and oxygen at different sites. At this temperature the powder is constituted of Sb204 and FeSbO,.The spectrum of the heat-treated sample at 1100°C is in agreement with published data for the FeSb0,20 and corrobor- ates the XRD and SEM results. Finally, it is important to remark that FeSb0, is an important catalyst for selective alkene oxidation and its performance is significantly increased if the powder is mixed with Sb204.334 This is exactly the case when CAA/Fe is heat- treated at 900 "C for 2 h. 10 20 30 40 50 60 28ldegrees Fig. 4 XRD patterns for CAA/Fe heat-treated for 2 h at (a) 100, (b)700, (c) 800, (d) 900, (e) 1000 and (f)1100°C. 0 =FeSbO, 391 Fig. 5 SEM micrograph for the sample CAA/Fe heat-trcated at 1000"C wavenum berkm-' Fig.6 IR spectra for the sample CAA/Fe heat-treated for 2 h at (a) 100, (b) 800, (c) 1000 and (d) 1100°C.Spectra (c) and id) were obtained by the KBr pellet method CAA/Cd Phase The XRD patterns for CAA/Cd samples at various stages of the heat-treatment are presented in Fig. 7. Similar to the CAA/Fe case, the results suggest that until nearly 80Ci"C the pyrochlore-like structure is maintained. However, for the heat- treated sample at 900°C new peaks were observed th& have their intensity maxima at 1000°C and which disappear at 1100"C. According to Castro et these new reflections were attributed to the presence of a CdSb,O6 phase. Therefore, at 900 and 1000°C we have a phase mixing where one of the mixture components is CdSb206. The elemental analysis of the powder originated from the 1100 "C heat-treatments is consistent with the Cd2Sb20, -x composition.Additnonally, the XRD data are in agreement with a pyrochlore-likc struc- ture, as previously reported.22 The SEM micrographs also revealed marked morphology changes for the differenr stages of heat treatment. The CdSb206 phase appears as indented 10 20 30 40 50 60 2fYdegrees Fig. 7 XRD patterns for CAA/Cd heat-treated for 2 h at (a) 100, (h)400, (c) 600, (d) 800, (e)900 (f) 1000 and (g) 1100 "C. 0 =CdSb,O, Fig. 8 SEM micrograph of the sample CAA/Cd heat-treated at 1000'C rod shape crystallites with sizes that vary by ca. 5-10 pm (Fig. 8), whereas a porous mass, composed of small crystals (< 1 pm), was observed for Cd2Sb20,-,. At 1100°C only the latter was observed.Infrared spectra for the samples treated at temperatures higher than 900 "C present very weak transmission (< 20%) in the range 4000-800 cm-'. This behaviour could be related to the semiconducting nature of these powder^.^,^ For tem- peratures lower than 900 "C the infrared spectra are similar to that for the CAA/Cd non-heat-treated sample. In this case, the use of a CAA/Cd precursor leads to the J. MATER. CHEM., 1994, VOL. 4 obtention of semiconducting ceramic powders at lower tem- peratures than the one used in the solid-solid reaction involv- ing CdO, Sb203and Sb20,.5,6321 Conclusions The present results indicate that the thermal decomposition of ion-exchanged CAA with Fe3+ and Cd2+ permits the use of new routes for the preparation of certain materials that are classically obtained by heating a mixture of oxides.We can remark on some important advantages arising from this method: (i) normally, the temperature and time reaction are less than for conventional methods (solid-solid reaction); (ii) the phases are formed from a single compound (CAA ion- exchanged), and not from a mixture of different reagents (this point eliminates problems with the reactivity of the solids, grinding, crystal habit, etc); (iii) obtention of the final products with great chemical homogeneity and phase purity. An extension of this method to other ion-exchangeable hydrated metal oxides, and studies on the correlations between loading, precursor structure type and properties of final products are under investigation in this laboratory.The authors would like to thank Conselho Nacional de Desenvolvimento Cientifico e Tecnologico (CNPq) and Fundo de Apoio ao Ensino e Pesquisa da UNICAMP (FAEP) for the fellowship to A.J.G.Z. References 1 P. A. Lessing, Ceram. Bull., 1989,68, 1002. 2 N. G. Error and H. U. Anderson, Muter. Res. Soc. Symp. Proc., 1986,73, 571. 3 M. Carbucicchio, G. Centi and F. Trifiro, J. Catal., 1985,91, 85. 4 G. I. Straguzzi, K. B. Bischoff, T. A. Koch and G. C. A. Schuit, J. Catal., 1987, 104,47. L. Biao-Rong, J. Am. Ceram. Soc., 1988,71, C-78-C-81. B. R. Li and J. L. Zhang, J. Muter. Sci.Lett., 1990,9, 109. M. Pechini, US.Pat., 3 330 697, July 11, 1967. Y. Ozawa, N. Miura, N. Yamazoe and T.Seyama, Chem. Lett., 1982, 1741. 9 W. A. England, M. G. Cross, A. Hamnet, P. J. Eisman and J. B. Goodenough, Solid State Ionics, 1980, 1,231. 10 V. A. Burmistrov, Y. M. Ryabishev, A. I. Sheykman and N. I. Hmon'kina, Izu. Akad. Nauk SSSR, Neorg. Muter., 1991, 27, 50. 11 F. A. Belinskaya and E. A. Militsina, Usp. Khim.. 1980,49, 1904. 12 V. A. Burmistrov, D. G. Kleschev, V. N. Konev and R. N. Pletnev, Zh. Neorg. Khim., 1985,30, 1959. 13 A. J. G. Zarbin, J. M. Sasaki, L. Cardoso and 0. L. Alves, in preparation. 14 D. J. Stewart, 0.Knop and C. Ayasse, Can. J. Chum., 1972,50,690. 15 M. Abe and K. Sudoh, Bull. Chem. SOC.Jpn., 1982,55,615. 16 C. Forano and J. P. Besse, Eur. J. Solid State Inorg. Chern., 1988, 25, 141. 17 M. T. Vandenborre and E. Husson, J. Solid State Chem., 1984. 53, 239. 18 M. T. Vandenborre, E. Husson and J. L. Fourquet, Spectrochirn. Acta, Part A, 1982,38,997. 19 F. J. Berry, J. G. Holden and M. H. Loretto, J. Chem. Soc. Furaduy Trans. 1, 1987,83,615. 20 C. Rocchiccioli-Deltcheff, T. Duprius, R. Frank, M. Harmelin and C. Wadier, C. R. Acad. Sci.Ser. B., 1970, 571. 21 H. Castro, 1. Rashes, M. C. Sanches-Martos and P. Garcia- Casado, Powder DifSr., 1988,3(4), 219. 22 A. J. G. Zarbin and 0.L. Alves, J. Muter. Sci. Lk'tt., in the press. Paper 3/04345B; Received 21st July, 1993