z zy zyx Phase Relations in the System Ce,O,-AI,O, in Inert and Reducing Atmospheres A. Cuneyt Tas* Department of Metallurgical Engineering, Middle East Technical University, Ankara 0653 1 , Turkey Mufit Akinc* zyxwvutsrqp zyxwvut zyxwvut zyxwvutsrqp zyxwvut Department of Materials Science and Engineering, Iowa State University, Ames, Iowa The 1:l compound, CeAlO,, in the system Ce,O,-AI,O, has been synthesized from the oxides and shown to have a perovskite-like tetragonal unitacell with the lattice pararneters a = 3.763 and c = 3.792 A. A new XRD pattern is suggested for CeAIO,. This compound is shown to be stable up to 1950°C. The 1:11 compound, CeAI,,O,,, has also been synthesized and shown to possess a magnetoplumbite-like hexagonal unit cellDwiththe lattice parameters a = 5.558 and c = 22.012 A. An XRD pattern is suggested for CeAl,,O,, for the first time. The evolution of eutectic-like microstructures was observed and reported in the Ce,O,rich side of this binary system. Nd (1747°C) aluminates. Scott' predicted, by Rainan spectroscopy, the transition temperatures as 1047" and 1367°C for Pr and Nd aluminates, respectively. Mizuno et al.' reported the cerium monoaluminate to be cubic at T 2 150°C in inert atmogphere (JCPDS PDF 28-0260) with a lattice constant of 3.767 A. They also claimed that the rhombohedral polymorph could be synthesized at temperatures less than 100°C. More recently, Kaufherr et a/.'"prepared CeAIO, in a flowing hydrogen atmosphere at 1000°C. The structure of the slowly cooled product was found to be pseudo cukic, tetragonal with the lattic? parameters a = 3.760 k 0.004 A and c = 3.787 k 0.004 A. The XRD pattein of the tetragonal structure is not in the JCPDS files and to our knowledge has never been published. The second compound, cerium hexaaluminate (Ce,O,. 1lA120,), has been included in the first phase diagram for this system suggested by Leonov et ~ l .and , ~it has been reported to melt incongruently in a reducing atmosphere at about 1950°C and to be stable down to room temperature upon cooling. In the more recent phase diagram suggested by Mizuno et ul.,' the hexaaluminate field was indicated by dashed lines, due to the difficulty in preparing it as a single phase; it was shown to melt incongruently at 1890°C in a reducing atmosphere. It was reported to bave a hexagonal st;ucture with the lattice constants a = 5.543 A and c = 21.979 A and that the sample contained CeAlO, and a-A1,0, as minor phases. This compound has been labeled as Ce-P-A1,O3 in this phase diagram due mainly to the compositional similarity with Na-p'-Al,O, (Na,O. 1 1A120,). The synthesis and characterization of cerium hexaaluminate, in its pure form, have been attempted, to our knowledge, only in the above two studies in literature. In both of these, this conipound was assumed to have the p-alumina structure. A similar type of compound (LaAl,,O,,) was reported"." to be present in the La@-Al,O, system. It has been realized"-" that the addition of one divalent cation, such as Mg?', Mn2+,Co2+,or Ni' ' , to the above formula unit would produce a series of new compounds, LaMAIllOlY, with structures quite similar to that of the magnetoplumbite (PbFe,,O,,) phase shown to be present in PbO-Fe,O, system. It so appeared that the divalent M ions selected to be used as dopants should be chosen from among the M oxides that were able to form spinel-type (MAl,O,) phases with oc-alumina.L8According to this scheme,Is Al" substitutes for Fe", La3'- (or Cei+) substitutes for Pb", and the divalent cation replaces one Fe" to conserve the electrical neutrality. The magnetoplumbite and Na-P'-alumina structures are both composed of spinel blocks formed by the close packing of Al and 0 atoms. The only significant difference between the two structures is in the two mirror planes (5 = 0.25 and 0.75). Each mirror plane in the p-alumina structure contains one large cation (Na') and one oxygen.".*" Figure 1 displays half of the unit cells of the two structures drawn according to the atomic coordinates reported by Gasperin eta/." for LaMnAl, ,O,, and those reported by Felschezo for NaAI, ,O,, . The non-close-packed mirror planes join the stable spinel blocks. Moreover, the sites I. Introduction 0 5001 1 of the two previously reported compounds of the Ce,O,-AI,O, system, cerium monoaluminate (CeAIO,), was first synthesized from the oxides by Zachariasen' at 1600°C in a helium atmosphere. The crystal structure of the compound was reported to be percvskite-like, tetragonal pseuGocubic with a = 3.760 k 0.004 A and c = 3.787 t 0.004 A. Keith and Roy' later confirmed the presence of this compound at 1:l molar ratio in this system. They used Ce2(C,0,),.9H,0 and AI,O, as the starting materials and observed the splitting in the (1 1 1) reflection that would indicate primitive hexagonal symmetry, contrary to Zachariasen.' Roth3 observed the formation of CeAIO, in the perovskite structure at 1600°C in abou! 1 h in helium but with rhombohedral symmetry (a = 3.766 A, OL = 90.2"). Leo90v4 reported the synthesis of bright green, cubic (u = 3.77 A) CeAIO, from 99.85% CeO, and hydrated aluminum oxide in inert or slightly reducing atmospheres at elevated temperatures. It was also noted that CeAIO, would completely decompose to CeO, and A1,0, in air at 800°C at 1 h. The binary phase diagram given by Leonov et ~ 1for. the ~ Ce,O,-A1,0, system showed two polymorphic transformations. The first occurs at 90°C (of unidentified nature) and the second was reported to be a rhombic-to-cubic transition at 980" t 20°C. Kim6indexed this compound, synthesized from Ce(C,O,),-xH,O and alumina under vacuum at 1600"C, on the basis of a hexagonal cell of rhombohedral symmetry LJCPDS PDF 21-0175) with the lattice constants of CI = 5.35 A and c = 13.02 A. He reported the X-ray density as 6.62 g/cm'. Geller and Raccah' estimated the rhombohedral-to-cubic transition for CeAIO,, if there is any, to occur at about 960°C by extrapolating the data they obtained for the transition temperatures of the La (522"C), Pr (1370°C), and NE zyx R. S . Roth4ontrihuting editor Manuscript No. 194x10. Received March 4, 1993; approved December 26, 1993. Supported in part by the McKniglit Foundation. St. Paul, MN, Iowa State University, Ames, IA: and Middle East Technical University, Ankara, Turkey. 'Member, American Ceramic Society. 296 1 2962 zy zyxwvutsrqpon zyxwvutsr zyxwvu Journal of the American Ceramic Society-Tas 3+, Vol. 77, No. 11 and Akinc zyxwvuts zyxwv Mz+ Mirror plane 4 b ( z = 0.75 ) 02- Spinel block 4 zyxwvu Mirror plane 4 b ( z = 0.25 ) Na-beta-alumina (see Ref. 20) Magnetoplumbite (see Ref. 17) Fig. 1. Half unit cells of magnetoplumbiteand sodium p-alumina structures. of the large cation and one oxygen in these planes are not fully occupied, which leads to a disorder in the mirror planes caused by a possible indeterminacy in the position of the large ion. The mirror planes in the magnetoplumbite structure contain one large cation (La3+or Ce'+) and three oxygens lying within the plane and four aluminums at the intersection of the c-cell edge with the mirror ulanes." The mirror planes in the magnetoplumbite-like structure were found," for LaMAl, ,O,-type compounds, to be disordered; the single, fully occupied, ideal site (2/3, 1/3, 1/4) for the large cation is split into two partially occupied, close sites in the plane. This gave the large cation of such structures a rather large volume in which it could move in a diffuse manner. The packings of the spinel blocks are virtually the same in both structures. The added divalent ions (M"), for LaMAl, ,O,,-type compounds, were commonly found to localize themselves among the tetrahedrally coordinated A1 sites within the spinel blocks17 of the structure rather than achieving a statistical distribution over all the available A1 sites. Hence, prior to this study, the previous literature has shown that the lanthanide-hexaaluminates, excluding that of cerium, exhibited structures similar to that of magnetoplumbite rather than that of Na-P'-alumina. Moreover, there have been no attempts to suggest an XRD pattern for the pure compound Ce,O,. 11A1,0,, although the JCPDS files did contain an entry for the divalent cation-doped version, CeMgAl,,O,,, of the magnetoplumbite type (JCPDS PDF 26-0872)." The present study focuses on the structural characterization of CeAIO, and CeAI,,O,,, using powder X-ray diffraction, and the high-temperature phase equilibria in Ce,O,-AI,O, system in reducing (Ar 10% H,), inert (Ar), and vacuum atmospheres. + 11. Experimental Procedure Compositions in the Ce,O,-AI,O, system were prepared by mixing and stirring the appropriate amounts of the starting oxides, CeO, (99.87%, 0.7 pm, Cerac, Milwaukee, WI) and Al,O, (99.96%, 0.5 pm, Reynolds, Inc.) in ethanol in glass jars for about an hour followed by 15 min of ultrasonification. The contents of the jars were dried overnight at 70°C in air. The recovered cakes were then calcined in air at 950°C for 12 h to remove any remaining carbonaceous residues. The calcined mixtures were then lightly ground in an agate mortar for about 30 min. The details of the green pellet preparation, equilibration heatings (in reducing, inert and vacuum atmospheres), quench practices, and DTA analyses employed were reported elsewhere.22 After equilibration and quenching, each sample was ground to a mean particle size of 7 km. Phase analysis was made by zyxwvutsrqpo Table I. Results of Equilibration Runs in the System Ce,O,-AI,O, Composition CeA10, CeAIO, CeAIO, CeAIO, CeA10, CeA10, CeAIO, CeAIO, CeA10, CeAIO, CeAIO, Ce,O,. 1 I A120, Ce20,.l 1A1,0, Ce,O,. 1 lAl,O, Ce,O,. I 1Al,O, Ce,O,. 1 1AI,O, Ce,O,. 1 1A1,0, Ce,O,.l 1A120, 12 moi% Ce,O,-88 12 mol% Ce,0,-88 74 mol% Ce,O,-26 79 mol% Ce2O,-2l 83 mol% Ce,03-17 Temp ("C) 1450 1550 1450 1550 1650 1750 1850 1950 1550 I550 1650 mol% A120, mol% AI,O, mol% AI,O, mol% A1,0, mol% A1,0, 1550 1550 1775 1880 1880 191s 1915 1700 1820 I875 1940 1940 Time (h) 50 40 50 40 20 10 3 1 50 50 25 30 30 12 2 2 0.1 0.1 12 8 4 3 3 Atmosphere Cooling Argon Argon Argon + 10%H, Argon 1 O%H, Vacuum Vacuum Vacuum Vacuum Argon + 10%H2 Argon Vacuum Argon Argon + 1O%H, Vacuum Vacuum Vacuum Vacuum Vacuum Vacuum Vacuum Vacuum Vacuum Vacuum Quench Quench Quench Quench Furnace quench Furnace quench Furnace quench Furnace quench 15"Clh to RT 15"C/h to RT 15"CIh to RT Quench Quench Furnace quench Furnace quench 1O"C/m to RT Furnace quench 1OWm to RT Furnace quench Furnace quench Furnace quench Furnace quench Furnace quench + Phases observed Tetr. CeAIO, Tetr. CeA10, Tetr. CeAIO, Tetr. CeAIO, Tetr. CeAIO, Tetr. CeAIO, Tetr. CeAlO, Tetr. CeAiO, Tetr. CeAIO, Tetr. CeAIO, Tetr. CeAIO, Ce,O,.l I AI,O, + A&O, + CeA10, Ce,O,~llAI,O, + AI,O, CeAIO, Ce,O,.llAI,O, + A1,0, + CeAIO, Ce,O,.l 1AI20, +Al,O, Ce,O,~llAI,O, + AI,O, Liquid A1,0, Ce,0,~11Al,03+ trace AI,O, Ce,O,.l IAI,O, + CeAlO, Liquid Ce,0,~11A1,03 CeA10, + Ce20, CeO,-, CeAIO, Ce,O, CeO,_, CeAIO, Ce,O, CeO,-, + + + + + + + + November 1994 z zyxwvut zyxwvutsrqpo zyxwvut 2963 P h a x Relations in the System Ce@-AI,O, in Inert and Reducing Atmospheres XRD in the 5"-160" 2 range m k g a Cu-tube powder diffractometer (XDS 2000, Scintag, Santa Clara, CA) operated at 45 kV and 30 mA with a scanning rate range of 0.125%.6" 2 /%n. Least-squares cell refinement, indexing, and lattice constant determinations were performed using the Appleman and Evans" and T R E O R * ~routines. Elemental silicon (NBS 640a) was used as an external standard. Microstructural analyses were carried out by scanning electron microscopy (SEM, JSM-6100, JEOL, Peabody, MA) on polished or as-is surfaces (in order not to damage the samples) of the fired pellets. The chemical compositions of the microstructural features observed in SEM images of the polished samples were examined with energy-dispersive X-ray spectroscopy (EDXS, Model Delta-5, Kevex, Foster City, CA) using single-phase CeAIO, as a standard, and the information thus obtained was believed to be accurate to within t3 at.%. of samples of this stoichiometry consistently produced a singlephase (without any free Ce,O, and ct-Al,O, remaining) substance with a typical XRD pattern given in Table 11.' The pattern shown in this table was obtained from a sample heated at 1550°C for 50 h followed by slow cooling (15"C/h) to room temperature (21"-23"C) in the furnace under a flow of argon + 10% H, gas. The structure has be$n determined to be tetragonal, with lattice constants a = 3.763 A and c = 3.792 A. This phase therefore is not the one covered by either of the previously published JCPDS files (21-0175, hexagonal and 23-0260, cubic) for CeA10,. However, our lattice constants agree well with those published by Zachariasen' and by Kaufherr rt al." The splitting of the peaks and the proper observation of those were thought to be the most important factors in the correct resolution of such a structure. Figure 2 shows a portion of the peak analysis performed for the data of the above XRD pattern. For this structure to be regarded as cubic, the splitting of the peaks displayed in Figs. 2(A), (B), (D-F) should certainly be absent. Similarly, the peak splitting observed in Figs. 2(A) and (D) and the 004:400 splitting shown in the pattern given in Table I1 should be absent to index this structure on the basis of a hexagonal cell with rhombohedral symmetry. The rhombohedral symmetry also zyxwvut 111. Results and Discussion Compositions of Ce,O,.AI,O, stoichiometry were heated either in flowing purified argon (0 H,O 5 50 ppm), argon 10% H2, or under vacuum (Po>5 6 X atm) at maximum temperatures ranging from 1450" to 1950°C for periods in the range of 1 to 50 h. The equilibration heat treatments carried out in this study are tabulated in Table I. All the thermal treatments + 23.2 23.4 + 23.6 2 theta 10 'For Tables I1 and 111, order ACSD-216 from Data Depository Scivicc, The American Ceramic Society, 735 Ceramic Place, Westetville, OH 4308 1.8720. 23.8 2 theto zyxwvutsrq . . zyxwvutsrqpon zyxwvutsrqpon zyxwvutsrqpo zyxwvutsrqponm 0 - , ' . 41 5 41.3 41.1 2 theta 41 7 9u 35 - 12.5 F 202 - 30' B 25- : E 20- 15- 10- 5- ". r 59 6 . . 590 60 2 theto 602 604 70 1 703 70 5 2 theta 70 7 Fig. 2. Splitting of the XRD peaks in tetragonal CeAIO, (a = 3.763 8,c = 3.792 8). 70 9 2964 zyxwvutsrqpon zyxwvuts zyxwvutsrqpo Journal of the American Ceramic Society-Tas requires the tetragonal 111 reflection, as we observed in Fig. 2(C), to split into 202 and 006 reflections of the hexagonal system. Our repeated attempts to observe the polymorphic transformation that was previously to occur at 960" or 100°C with differential thermal analysis have all failed. The DTA traces (4"C/min) all registered the undisturbed background from room temperature to 1950°C. Nevertheless, the enthalpy of such a transformation, if there really was one, might be too low to detect by DTA. We found that slow cooling (15"C/h) a CeAIO, sample following the equilibration at 1550°C (Ar 10% H2) or 1650°C (vacuum) to 500°C (holding there for 24 h) and then to room temperature followed by 1- to 2-h holding in the furnace (21"-23°C) also did not produce any changes at all in the suggested XRD pattem that would be indicative of a polymorphic transformation. One question that still remains is the possibility of the presence of an extremely sluggish transition that might occur below 100°C. On the other hand, it needs to be noted that the purity of the starting materials might play a much more important role, in the realm of high-temperature phase equilibria and structural chemistry, than it was once thought. For instance, it may well not be a coincidence for the two different groups of researchers (Keith and Roy' and Kim') to use the same starting material, i.e., cerous oxalate; Ce(C,O,),~nH,O, in the preparation of CeAlO, and to produce the rhombohedral "polymorph." Is the presence of carbon in the structure causing this rhombohedral symmetry? It might still be a possibility, which needs to be investigated, that the carbon would not "bum out" until the temperature reaches 960"980°C. As another possibility, to our knowledge, LaA10, does not have a reported "cubic" polymorph. In our laboratory we + and Akinc Vol. 77, No. 11 also were not able to produce a tetragonal or cubic form of LaA10,. Our quenched samples (froom above 155Q"C) of LaAIO, always showed the (202:2.190 A)-(006:2.184 A) splitting that was a prerequisite for the rhombohedral symmetry to set in and be identified. Therefore, considering the fact that the significant presence of La,O, as an impurity in cerium oxides was not so uncommon one or two decades ago, the rhombohedral LaAIO, which might form at very small quantities might have acted as a seed which would then impose the CeAIO, structure to acquire the rhombohedral symmetry of LaAlO, . Research is presently under way in our labs to test this hypothesis by investigating the chemical synthesis conditions of LaA10, and CeAlO, via aqueous precipitation. Small portions of equilibrated, single-phase CeAlO, samples were heated in Pt envelopes open on one surface in a flowing air atmosphere at 250°C ( 5 days) and 1000°C (12 h) to test the stability of the compound. The XRD analysis showed that CeA10, was still perfectly intact at 250°C; however, in the 1000°C sample, only traces of CeAlO, could be detected among the major phases of CeO, and a-Al,O, . The same tetragonal XRD pattem as that given in Table 11, observed in the sample heated in air at 250°C for 5 days, has excluded, again, for us the possibility of identifying the previously reported,6 low-temperature, rhombohedral "polymorph." If one considers the perovskite-like lattice of CeAIO, as composed of A13+ions being at the octahedral interstices (cell comers, CN = 6), 0'- ions at the edge centers, and finally the large Ce'+ being at the body center (CN = 12). then it would be noticed that the (1 11) planes of this structure will be the highest atomic density regions because of the close packing of ceriums and oxygens. The close-packed (1 1 1) faces of such a cell would zyxwvuts Fig. 3. SEM micrographs of CeA10, samples: (A) 1S50°C, argon, 35 h, quenched (bar = I pm); (B) 1650"C, vacuum, 15 h, furnace-quenched (bar = 1 pm); (C) 19SO"C, vacuum, 1 h, fumace-quenched (bar = 10 pm). zyxwvut zyxwvutsrqpo zyxwvutsrq zyxwvutsrqponmlkji zyxwvutsr zy November 1994 Phase Relations in the Svstem Ce,O,-Al,O, in Inert and Reducing Atmospheres then be expected to display the hexagonal symmetry characterized by the presence of trigonal axes while being the most thermodynamically stable especially during anomalous grain growth. As was very recently discussed by DrofenikZ5for the BaTiO, perovskite, predominant (1 11) faces displaying the trigonal axes might be the surface habit during the grain-growth stage of the densification process. We have observed such microstructures in studying CeAIO,. Figure 3(A) shows the SEM micrograph of the fracture surface of a CeA10, pellet equilibrated in argon at 1550°C for 3.5 h, followed by water quenching. The extensive grain growth and well-developed (11 1 ) faces are clearly visible, especially in the upper righthand comer of this micrograph. This sample produced the typical "tetragonal" XRD pattern of Table 11. The micrograph of Fig. 3(B) was taken from the fracture surface of a CeA10, sample equilibrated at 1650°C for 15 h under vacuum, followed by furnace quenching (by shutting off the main power, 600"C/min cooling from 1650" to IOOO'C, more sluggish exponential decay-type cooling from 1000°C to room temperature) to retain the high-temperature phases and features. A (1 11) face is visible at the center that is surrounded by microfacets and steps. The sample still produces the characteristic tetragonal XRD pattern indicating one more time that the real symmetry of a crystal does not necessarily depend upon the symmetrical shape and size of its apparent faces. Figure 3(C) shows the as-is surface of a CeA10, pellet heated at 1950°C for 1 h under vacuum followed by furnace quenching to room temperature. In this sample, it was almost impossible to find a region of the fracture surface that shows the (1 11) faces. Apparently, recrystallization 2965 has occurred and the whole surface has been covered with cuboids and tetragonal rhombs. The XRD pattern of this sample still conforms to that given in Table 11. We synthesized the binary compound of Ce,O,. 1 1A1,0, from the starting oxides CeO, and A1,0, by heating the samples either in argon, argon 10% H,, or under vacuum at temperatures above 1550°C. The resultant product was never pure; it always contained some a-Al,O, as a minor phase whose quantity decreased significantly with increasing temperatures up to 1880°C. At lower temperatures, such as 1550"C, the samples also contained small amounts of tetragonal CeAIO,. Cerium hexaaluminate was found to be an incongruently melting compound as was predicted by Mizuno et al.' To determine the first liquid formation temperature, a small piece of a preequilibrated cerium hexaaluminate sample, with several sharp edges, was hung by a molybdenum wire in the vacuum furnace that was visible through a quartz viewing port. An IR pyrometer, calibrated just before this run against the melting point of PI, was focused on the sample. The sample was heated at the rate of 3"C/min. The melting behavior of the chunk was continuously monitored through the pyrometer, and at 1915" t 25°C the sharp edges of the sample became rounded. The chunk, at this point, began to flow and then fell as a drop onto a molybdenum foil placed just below the wire. Following a 5-min hold at this state, the sample was furnace-quenched to room temperature. We assumed this temperature (i.e., the peritectic temperature) as the incongruent melting temperature of this compound. This reading agrees well with the temperature reported for the same event by Mizuno eta1.9 The final sample of this experiment + zy Fig. 4. SEM micrographs of samples of CeAI,,O,, composition: (A) and (B) backscattered electron images. 3"C/min to 1915"C,5 min, fumacequenched to RT. Darker trigonal a-alumina crystals surrounded by lighter, cerium-containing, once-liquid phase. (Bars = 100 km.) (C) 1550"C, 30 h, quenched, ground, repelletized, 1S5O0C,30 h, quenched (bar = 10 Fm). (D) Same sample as in (C) (bar = 1 Fm). 2966 zyxwvutsr zyxwvutsrqpon zyxwvu zyxwvutsrqp Journal of the American Ceramic Society-Tas showed, in its XRD analysis, a-AI,O, as the only crystalline phase present. Figures 4(A) and (B) depict the surface morphology of this sample, showing darker trigonal alumina crystals surrounded by a lighter, cerium-containing, once-liquid phase. Repeating the same experiment, but this t.ime using a slow cooling from 1915°C to room temperature, yielded the most pure hexaaluminate sample we obtained in this study. Traces of ol-Al,O, still found in that sample would be indicative of incomplete resorption. We suggest an XRD pattern for cerium hexaaluminate, CeAI, ,O,,, in Table IILt This pattern is obtained from a sample first heated at 1550°C for 30 h, water-quenched, ground, repelletized, and then heated at 1550°C for another 30 h in a flowing argon -t- 10%H, atmosphere, followed by quenching. The sample still contained some a-Al,O, that could easily be detected from the pattern and very small amounts of CeA10, which could be detected in the pattern only after a Rietveld analysis of the diffraction data. The peaks of these minor phases are excluded from the pattern given in the above table. Figures 4(C) and (D) show the details of the fracture surfaces of this sample. Platelike, large grains of CeAl,,O,, had always been the characteristic microstructural features of our hexaaluminate samples in this study. The EDXS analyses carried out on the flat surfaces of the grains of this sample (by using pure, single-phase CeAIO, as a standard) gave us the mean weight percentages of the three elements present as follows: Ce 19.0%, A1 40.8%, and 0 40.2%. When compared with the theoretical percentages (Ce 19.33%, A1 40.94%, 0 39.73%) that were calculated from CeA1,,0,,, within the limits of error of EDXS analysis, these may be considered as a good match. and Akinc Vol. 77, No. 11 The determination of the crystal structure among the two likely candidates, magnetoplumbite and Na-p'-alumina, while working with a powder sample rather than a single crystal of cerium hexaaluminate, had been a difficult question to answer in the scope of this study. The assessment of the degree of disorder and the extent of site splitting in the mirror planes of the CeAI, ,O,, structure which might cause the possible, slight variations in the phase stoichiometry, as well as the exact determination of the structure type, need to await further powder neutron or single-crystal X-ray diffraction experiments. We have confirmed that the eutectic between CeAI, ,O,, and CeAlO, occurred at 1785 k 15°C as was predicted by Mizuno et al.' The equilibrium heat treatments for a sample of composition 12-mol% Ce20,-88-mol% A1,0, at 1700" and 1820°C produced a solid-state phase mixture of CeAI,,O,, and CeAIO,, and a mixture of CeA1,,0,, and liquid, respectively, upon furnace quenching, as predicted by the available phase diagram.' DTA runs for the same composition indicated a single endothermic event at 1785°C corresponding to the first liquid formation. The heating of one of these samples to 1820°C, soaking for 6 h at that temperature. and slowly cooling (5"C/min) to 1100°C caused the posteutectic liquid matrix surrounding the elongated CeAI,,O,, grains to crystallize. A sample micrograph of this structure is given in Fig. 5(A). The XRD pattern of this sample showed the presence of two crystalline phases, CeAIO, and CeAI,,O,,. We could not synthesize the R-compound that was reported to form at 79 mol% Ce,O, and decompose at temperatures below 1850°C by Mizuno et aLYOne probable reason for this might be the inadequate quenching rates we had in the vacuum zyxw z zy Fig.5. SEM micrographs of samples in the Ce,O,-Al,O, system: (A) 12 mol% Ce,O,. (Backscattered electron image) 1820"C, vacuum, 6 h, S"C/min to 1 10O"C, furnace-quenched to RT. Eutectic matrix surrounding the elongated, dark a-alumina crystals. CeA10, is also present in the sample. (B) 74 mol% Ce,O,. (Backscattered electron image) 1875"C, vacuum, 4 h, furnace-quenched. Dark CeAIO, grains, gray CeO, matrix. (C) 79 mol% Ce,O,. 1940"C, vacuum, 3 h, furnace-quenched. Eutectic-like matrix of CeA10, and CeO,_, with large gray grains of CeO,-,. (Bars = 10 Fm.) I zyxwvutsrqpo zyxwvuts zyxwvutsrqponm zyxwvutsrqp zyxwvu zyxwvutsrqp zyxw zyxwvuts zyxwvutsrqp November 1994 Phuse Relation s in the System Ce,O,-A1,0, in Inert und Reducing Atmnsphcres furnace that would not be rapid enough to prevent the decomposition of this compound. We were only able to furnace-quench (i.e., the natural cooling of the furnace itself upon turning off the power) our samples, although Mizuno et ~ 1reported . ~ that they were able to water-quench their samples from temperatures in excess of 1900°C. Heating a 74-mol% Ce,0,-26-mol% A1,0, sample (preequilibrated at 1550°C for 24 h, followed by water-quenching, grinding, and repelletizing) to 1875°C and furnace-quenching produced a solid-state phase mixture of CeO,_,] as shown in Fig. 5(B). The CeAIO, and [Ce,O, darker phase is CeA10, in this picture. The other two compositions we studied in the Ce,O,-rich region of this system produced unique microstructures similar to that displayed in Fig. 5(C). Upon heating the 79-mol% and 83-mol% Ce,O, samples to 1940°C, followed by furnace-quenching, the liquid was formed, and during cooling the submicrometer-spaced, eutectic-like matrix evolved. At this point we choose to just report the microstructures observed, even though they may not represent the equilibrium conditions for this portion of the binary. The lighter phase in Fig. 5(C)was the crystalline mixCe0,- ,], whereas the darker phase was CeAIO,. ture [Ce,O, Although our attempts to synthesize the orthrohombic R-compound9 have failed, we still detected the traces of an orthorhombic phase embedded in the backgrounds of the XRD patterns of the above three samples. The extremely low intensities (1 to 3 out of 109) of that phasg, with the tentative lattice constants a = 19.28 A, b = 12.35 A, and c = 6.37 A, did not at this point warrant giving an accurate listing of the individual reflections. It needs to be noted that the R-compound in the Ce,O,-AI20, system reported by Mizyno et al.’ ha{ quite differen! lattice constants: a = 9.538 A, b = 5.850 A, c = 15.205 A. + + IV. Conclusions The two binary compounds, CeAIO, and CeAI,,O,,, of the system Ce,O,-Al,O, have been synthesized from the pure oxides CeOz and A1,0,. It has been found that the conditions of formation of these two compounds were not affected by the atmospheres used in this study, namely, argon, argon + 10% H,, and vacuum. CeAIO, was shown to be stable in a perovskite-lige tetragonal structure, with the lattice constants a = 3.763 A and c = 3.792 A, in the atmospheres used in this study from room temperature to 1950°C. A new XRD pattern was suggested for CeAIO,. CeAI, ,O,, was synthesized from the oxides and was shown to be an incongruently melting compound, confirming the findings of the previous researchers. It decomposed into a-Al,O, and a liquid phase at 1915” It_ 25°C upon heating. An XRD pattern is suggested, for the first time, for the hexagonalocompound CeA1,,Ol,, with the lattice constants a = 5.558 A and c = 22.012 A. The eutectic reaction between CeAI, ,O,, and CeAlO, was confirmed to occur at 1785” -C 15°C as reported by Mizuno et a1.9 Although it was almost impossible for us to form the singlephase CeAI,,O,, in the solid state, we were able to form the crystals of this compound, in the presence of a liquid phase, at a composition of 12-mol% Ce,O, (without any unreacted a-Al,O,), within a eutectic matrix of CeAl,,O,, and CeAIO,. 2967 Eutectic-like microstructures, with submicrometer-sized lamellae, were observed, for the first time in any of the Ln,O,LnA10, (Ln: Lanthanide) systems studied, in the Ce,O,-rich (74-83 mol%) side of the Ce,O,-AI,O, binary. References ‘W. H. Zachariasen, “Crystal Chemical Studies of the 5f-Series of Elements. XII. New Compounds Representing Known Structure Types,” Ai.tu Cry,~/u//o,~r., 2,388-390 (1949). ’M. L. Keith and R. Roy, “Structural Relations among Double Oxide\ of Trivalent Elements,”Am. M i m w / . , 39, 1-23 (1954). ’R. S. Roth, “Classification of Perovskite and Other ARO,-Type Corrrpounds,” J Rrs. Nutl. Bur. Srurrd. (U.S.j,58, 75 (1957). ‘A. I. Leonov, “The Valence of Cerium in Synthetic and Natural Cerium Aluminates and Silicates, Part I , Compounds of the Perovskite Group,’’ /:I.. Akud. NuukSSSR, Otd. Khini. Nuuk, 1, 8-13 (1963). ‘A. I. Leonov, A. V. Andreeva, V. E. Shvaiko-Shvaikovskii, and E. K. Keler, “High-Temperature Chemistry of Cerium in the Systems Ce,O,-AI,O,, Cr20,. Ga,O,,” liv. Akud. Nauk SSSR. Neorg. Muter., 2 , 5 17-23 (1966). ‘Y. S. Kim, “Crystallographic Study of Cerium Aluminate (CeAIO,),” Actu Crystullogr., Sect. B , Strucr. Sci., 24, 295-96 (1968). ’S. Geller and P. M. Raccah, “Phase Transitions in Perovskitelike Compounds of the Rare Earths,” Phys. Re>,.Bc Corzdens. Muter., 2, 1 167-72 ( I 970). ‘J. F. Scott, “Raman Study of Trigonal-Cubic Phase Transitions in Rare-Earth Aluminates,” Ptiys. Rev., 183. 823-25 ( I 969). ’M. Mizuno, T. Yamada, and T. Noguchi, “Phase Diagram of the System AI,O,-Ce,O, at High Temperature,” Yogyo Kyokuishi, 83,90-95 ( 1975). “’N. Kaufherr, L. Mendelovici, and M. Steinberg, “The Preparation of Cerium(ll1) Aluminate at Lower Temperatures: IR, X-ray and Electron Spin Kwonance Study,” J . Less-Common Met., 107,281-89 (1985). “M. Mizuno, T. Yamada, and T. Noguchi, “Phase Diagram of the System AIzO,-La,O, at High Temperatures,” Yogp Kyokuishr, 82,630-36 ( 1974). ‘’R. C. Ropp and G. G. Libowitz, “The Nature of the Alumina-Rich Phase in the System La,O,-AI20,,” J . Am. Cer-uni. Soc., 61,473-75 (1978). ”J. M. P. J. Verstegen, J. L. Sommerdijk, and J. G. Verriet, “Cerium and Terbium Luminescence in LaMgAl,,O,,,,”J . Lumin., 6,425-31 (1973). “J. M. P. J. Verstegen, “A Survey of a Group of Phmphors, Based on Hexagonal Aluminate and Gallate Host Lattice\,” .I. E k r r o c hcni. SOC., 121, 1623-27 ( I 974). ”A. Kahn, A. M. Lejus, M. Madsac, J. Thery, D. Vivien. and J. C. Bernier, “Preparation, Structure, Optical, and Magnetic Propertie.. of Lanthanide Aluminate Single Crystals (LnMAI, ,O,,),” J . Appl. Phvs., 52, 686469 (198 I ). IhF. Laville and A. M. Lejuc, “Crystal Growth and Characterisation of LaMAl,,O,, Lanthanum Aluminates,” J . Cryst. Growth, 63. 426-28 (1981). “M. Gasperin, M. C. Saine, A. Kahn, F. Laville, and A. M. Lejus, “Influence of M2’ Ions Substitution on the Structure of Lanthanum Hexaaluminates with Magnetoplumbite Structure,” J . Solid Srutr Chem., 54,6l-69 (1984). “X. H. Wang, A. M. Lejus, D. Vivien, and R. Collongues, “Synthesis and Characterisation of Lanthanum Aluminum Oxynitrides with Magnetoplumbitelike Structure,”Murer. Res. Bu//.,23, 4 3 4 9 (I988). ”C. A. Beevers and M. A. S. Ross, “The Crystal Structure of ‘Beta Alumina’ Na,O. I 1AI,O,,” Z. Kristuilogr., 97, 59-(~6( I 937). ’“J. Felsche, “The Alkali Problem in the Crystal Structure of Bcta Alumina,” Z. K?-ista//ogr.,127,94100 (1968). ”Powder Diffraction File, Card No. 26-0872. International Centre for Diffraction Data, Newtowne Square. PA. 22A.C. Tas and M. Akinc. “Phase Relations in the System Ce,O the Temperature Range 1150” to 1970°C in Reducing and Inert A .I. Am. Ceruin. SOC.,in press. ”D. E. Appleman and H. T. Evans, “U.S. Geological Survey, Computer Coiltribution No. 20,” NTIS Report No. PB-216188. U.S. National Technical Information Service, Springfield, VA, 1973. E. Werner, “Trial and Error Program for Indexing of llnknown Powder Patterns, TREOR.” University of Stockholm, Stockholm, Sweden. 1984. ”M. Drofenik, “Origin of the Grain Growth Anomaly in Donor-Doped Bar0 ium Titanate,”/. Am. Cernni. Soc., 76, 123-28 (1993). zyx