A novel preparation of calcia fully stabilised zirconia from molten alkali-metal nitrate Huda Al-Raihani,' Bernard Durand,b.c Fernand Chassagneuxb and Douglas Inman' 'Department of Materials, Imperial College of Science, Technology and Medicine, Prince Consort Road, London, UK SW7 2BP bLaboratoire de Chimie Minerale 3, URA CNRS no. 116, ISIDT, Universitk Claude Bernard Lyon I, 43 Boulevard du 11 Novembre 191 8, 69622, Villeurbanne Ckdex, France 'Laboratoire de Matkriaux Minkraux, URA CNRS no. 428, Ecole Nationale Supkrieure de Chimie de Mulhouse, 3 Rue Alfred Werner, 68093 Mulhouse Cedex, France A novel preparation of cubic calcia stabilised zirconia (CaO-ZrO,) which employs the coprecipitation of the constituents from zirconium oxychloride (Zr01.22C11.56)and calcium nitrate [Ca( NO,),] with eutectic lithium nitrate-potassium nitrate (LiN0,-KNO,) at around 600 "C is described.The optimisation and advantages of the preparation are discussed. Pure and stabilised zirconia (ZrO,) have been the objectives of numerous investigations, initially as refractories' and more recently as solid electrolytes in fuel cells., Pure ZrO, exhibits three polymorphs, monoclinic (m), tetragonal (t) and cubic (c).,-~The polymorphic transformation from monoclinic to tetragonal ZrO, involves a considerable contraction, and vice versa an expansion, creating a sharp volume change of about 9 Yo resulting in substantial weakening and fra~ture.~ The brittleness problem has been partially overcome by the addition of certain oxides such as magnesia (MgO), calcia (CaO), yttria (Y203)or rare-earth-metal oxides (e.g.CeO,), to stabilise the ZrO,, either partially or fully, depending on the quantity of oxide dopant added.' These materials have shown good thermal shock resistance, high degree of toughness, chemical inertness and refractory behaviour and these proper- ties render them useful in many application^.^-^' Calcia fully stabilised zirconia (Ca-FSZ) exhibits excellent ionic conductivity and low thermal conductivity and, therefore, it is being used as a solid electrolyte in galvanic cells, fuel cells, gas sensors; in the near future it may find use as a membrane in water thermolysis for the production of hydrogen." CaO-Zr0, has been produced by several well known methods, ranging from the conventional to the very advanced new routes.These include mixing powders with ball milling and coprecipitation in aqueous solutions followed by decompo- sition, whilst new methods are hydrothermal, spray-drying, freeze-drying and sol-gel methods.', So far, most of the fabrication routes are either expensive or have employed very high temperatures. On the grounds of cost, quality of the powder and time, it is desirable to avoid the necessity for very high temperatures in the fabrication process. Molten salt routes are possible alternatives, where metal oxides can be readily precipitated from oxyanion melts, (e.g. and nitrite^'^), or from simple ionic melts (e.g. chlorides''). This can be achieved by the oxide anion produced either from the dissociation of the melt itself'' or by the addition of alkali-metal bases,,' (e.g.hydroxide, peroxide or carbonate). The preparation of ceramic powders uia low-temperature molten salt routes has just begun to be explored. Lux-Flood acid-base reactions can be applied, e.g. in the precipitation of pure Zr0213,20,21 or coprecipitation of Y203-Zr02,22*23from molten nitrates. The purpose of the present study is to investigate the possibility of coprecipitating CaO-ZrO, via a molten salt route. Experimental Materials Zirconium oxychloride. Commercial hydrated zirconium oxychloride ZrOC1,.8H2O (Aldrich) was used after dehy- dration in an oven at 195 f5 "C for about 48 h. From thermo- gravimetric investigation it was initially thought that this treatment led to the anhydrous oxychloride ZrOC1,.24 In fact, further IR spectroscopy studies have shown that it is quite difficult (almost impossible) to completely dehydrate zirconium oxychloride without losing a small amount of hydrochloric acid which involves the formation of oxychlorides correspond- ing to the formula ZrO,, +x.C12(1-x).25 Chemical analysis of zirconium and chlorine contents corroborated the conclusion that the thermal treatment at 195°C for 48 h led to the oxychloride ZrOl.22Cll.56. Calcium nitrate.Ca( NO,), (BDH general purpose reagent) was used after a thermal treatment similar to that used for zirconium oxychloride. No nitrate decomposition was noticed. Zirconium sulfate.Commercial zirconium sulfate, Zr(S0,)2.4H20 (BDH, AnalaR) was dehydrated according to Bear's method.26 The dehydration was carried out by heating Zr(SO,), in the presence of H2S04 (AnalaR grade, 98%), at temperatures between 350 and 410°C for several hours, in a similar manner to that previously indicated for the preparation of zirconia by reaction of zirconium sulfate with molten alkali- metal ni tra tes.,' Lithium nitratepotassium nitrate eutectic mixture. LiN0,-KNO, eutectic, 132 "C, was also prepared according to the method described in ref. 20. The starting materials were LiN03.3H20 (Fisons general purpose reagent) and KNO, (BDH, AnalaR). Sodium nitrate. NaNO, (BDH, AnalaR), mp 308 "C, was used as received. Procedure Most of the preliminary experiments were carried out in a small diameter furnace with a vertical steel pot, where a chromel-alumel thermocouple was located between the reac- tion tube and the furnace wall.A few reactions were performed in a larger diameter mullite furnace, but with a Pt/Pt-13% J. Muter. Chem., 1996, 6(3), 495-500 495 Rh thermocouple placed within the hot zone outside the furnace All reactions were performed in a Pyrex Quickfit tube joined to a drying tube containing glass wool and silica gel to provide a dry atmosphere Some experiments involved the use of ultra-centrifugation The precipitate-melt mixture was spun for a few minutes in order to remove the unreacted melt, without the use of water washing A special apparatus was designed and used for the filtration of the melt, consisting of an inner tube with a sintered disc and an outer tube in which the filtrate melt could be received All experiments were performed from mixtures of the start- ing metal salts [Ca(N0,)2, ZrO, 22C11 56 and (Li,K)NO,] containing an excess of alkali-metal nitrates compared to stoichiometric ratio Techniques Thermogravimetry.The reactivity of metal salts towards alkali-metal nitrates was investigated with a Stanton-Redcroft TR- 1 thermobalance which revealed mass losses corresponding to various reactions including precipitation, coprecipitation, dehydration and decomposition Ultra-centrifugation. An MSElO centrifuge was modified for high-temperature requirements The apparatus rotates four heated capsules The maximum working temperature could be up to 1000"C at a maximum speed of 2400 rpm X-Ray powder diffractometry.A Philips X-ray diffractometer was employed to identify the insoluble solids, whether unwashed or washed The resulting diffraction patterns were recorded and analysed according to the standard data listed in the JCPDS files27 Chemical analysis. Elemental chemical analysis (Ca, Zr, N, C1, Li, K or Na) was performed by the Micro-analysis Center of Solaize (CNRS, France) Results Preliminary experiments Reaction of Ca (N03)2 with eutectic LiN0,-KN03. Anhydrous Ca( N03)2 and LiN0,-KNO, were premixed thor- oughly before reaction The experiment was programmed to start heating from room temperature (25 "C) with a heating rate of 12"C min-l upto 630 "C, and to maintain this tempera- ture for 30 min No changes were observed below 195 "C, at which tempera- ture the eutectic salt started to melt This is rather higher than the literature value of the mp (132°C) Between 300 and 500"C, Ca(N03)2 was observed to be completely soluble in the melt Above 500 "C, the melt became yellowish, and colour- less O2began to evolve When the temperature reached 540 "C, the release of both brown NOz and colourless O2was intensi- fied and this continued up to around 600"C, when a white precipitate could be seen The precipitate-melt mixture was then cooled in a desiccator, washed with distilled water and finally subjected to XRD analysis The X-ray diffraction pat- terns for the washed product showed several lines correspond- ing to CaO, Ca(OH), and CaCO, (Table 1) It was thought that calcium hydroxide was formed by hydrolysis of calcium oxide during water washing, and calcium carbonate by car- bonation of calcium hydroxide The analysis, by X-ray diffractometry, of the unwashed cooled reaction mixture cor- roborated this assumption Lines for calcium hydroxide and calcium carbonate disappeared from the pattern, which revealed only lines of calcium oxide and of the eutectic LiNO,-KNO, (Table 2) Note that experimental lines of the 496 J Mater Chem ,1996, 6(3), 495-500 Table 1 XRD results of the white precipitate obtained from the reaction of Ca(NO,), with eutectic LiN0,-KNO, CaO Ca(OH), CaCO, washed (JCPDS (JCPDS (JCPDS precipitate 27-775) 4-733) 5-586) d/A I/I, d/A I/I, d/A I/I, djA I/I, 4 90 61 490 74 303 100 300 100 304 100 2 78 36 2 62 15 263 100 2 49 13 2 49 14 2 28 13 2 29 18 2 09 12 2 10 18 191 15 191 60 193 42 191 17 187 15 188 60 187 17 Table 2 XRD results of the unwashed solid obtained from the reaction of Ca(NO,), with eutectic LiN0,-KNO, CaO LlNO3-KNO3 KNO, unwashed (JCPDS eutectic (JCPDS solid 28-775) (experimental) 5-377) d/A Ill, d/A I/I, d/A I/I, d/A I/Io 4 63 9 4 63 13 466 23 4 54 9 4 54 13 458 11 3 75 40 3 75 79 378 100 3 70 26 3 70 56 373 56 3 21 6 321 5 307 15 3 02 100 300 100 302 100 303 55 2 75 17 2 75 35 276 28 2 70 6 2 69 9 271 17 266 41 2 64 29 2 64 47 265 55 2 63 15 2 63 35 263 20 2 40 5 241 7 2 36 6 237 4 2 32 6 233 9 2 18 20 2 18 47 219 41 2 15 13 2 15 12 216 20 2 06 4 2 06 7 207 13 2 04 10 2 04 16 205 18 194 13 194 26 195 24 193 4 193 7 194 6 191 20 191 60 188 20 188 60 eutectic are very close in position and intensity to those of orthorhombic KNO, Reaction of ZrOl 22C11 56 with eutectic LiN03-KN03.56Thermogravimetry of ZrO, 22c11 with the eutectic LiN03-KN03 showed a total mass loss of 488% of the starting ZrO, 22C1, 56 This occurred in two stages (Fig 1) I $1? 40 t Ls) -' Q) Ig 0 -----' : . '' U 100 200 300 400 500 600 700 800 TPC Fig. 1 TG data for the reaction of ZrO, &ll 56 with a LiN0,-KNO, eutectic mixture (1)from 120 to 380 "C, the experimental mass loss was 23.9% (maximum rate at 365 "C); (2) from 380 to 500 "C, the exper- imental mass loss was 24.9% (maximum rate at 450°C).Between 500 and 600"C, there was no change in mass, but this was followed by a sharp mass loss which corresponds to the decomposition of the melt, probably starting just above 600°C to give a yellow melt. X-Ray diffraction identified the reaction product as mainly tetragonal zirconia, with sometimes a small proportion of monoclinic zirconia. Zirconium oxychloride has been reported previously to react with the NaN0,-KNO, mixture in a similar manner.21,24,28 Co-precipitation of CaO-ZrO, from Ca (NO3),, Zr0,.22C1,,, and the eutectic LiNO,-KNO, CaO-ZrO, samples with different proportions of CaO, (4.7, 12.2 and 20.8 mol%) were coprecipitated from the reactions between Ca(NO,),, ZrOl.22Cll.56 and LiNO,-KNO,.For the three reactions, the programme was set to heat the reaction mixture from room temperature to around 630°C with a heating rate of 12 "C rnin-'. The temperature was maintained at 630 "C for 30 min. Four stages could be identified during the furnace runs: (1) from room temperature to 200"C, NO, (light brown fumes) started to evolve. The mass loss revealed by thermogravimetry is negligible. (2) From 200 to 500"C, reaction was established. A continuous release of NOz and 0, was observed simultaneously with the formation of a white precipitate. The thermogravimetric curve (Fig.2), quite similar to the one for the reaction of ZrOl.22Cll.56 alone (Fig. l), indicates a two-step transformation with maximum rates at 350 and 450 "C. (3) From 500 to 600 "C, reaction was complete and a negligible amount of NO, was evolved. During this time, the melt became yellowish in colour. (4) Above 600"C, a further loss was noticed due to the decomposition of the excess of nitrates, producing a NO, release. The precipitate-melt mixture was cooled in a desiccator and both unwashed and washed samples were analysed by XRD. Whatever the ratio Ca(NO,), :Zr01.22C11.56, the unwashed powders were identified as a poorly crystallised cubic phase similar to cubic zirconia (JCPDS 27-997) and also to cubic 85Ca, 15Zro 8501 (JCPDS 26-341) and a well crystallised LiN0,-KNO, mixture.Nitrates were eliminated by washing and the three washed powders contained only cubic calcia stabilised zirconia as shown in Fig. 3 for 12.2 mol% CaO- ZrO,. Investigation of the influence of ultra-centrifugation on the coprecipitation of CaO-ZrO, from Ca( NO3), and Zr01.22C11.56 in molten LiNO3-KNO3 In order to test the efficiency of ultra-centrifugation at high temperature to separate the reaction product from the unreacted nitrate melts, an experiment was performed using '0° i 1 8ok600 70 .-o, 500 i t 0\ I03 400 1 ,-I I 2 300 200 300 400 500 600 700 5 TI'C Fig.2 TG data for the simultaneous reaction of Ca(NO,), and ZrO, ,,C1, 56 with eutectic LiN0,-KNO,. -, Am; ---,AmlAT 10 20 30 40 50 60 70 28ldegrees Fig.3 XRD pattern of washed 12.2 mol% CaO-ZrO, Ca(NO,),, Zr01.22C11.56 and LiN0,-KNO, in the same pro- portions as in the previous section to prepare 20.8 mol% CaO- ZrO,. The heating profile was slightly changed to ensure complete dryness of the starting materials. The mixture was first maintained at 50°C for about 15 min, then heated to 630°C at a rate of 200°C h-' and finally maintained at this temperature for 13h. The precipitate-melt mixture produced from the reaction was divided into three portions and treated as follows: (a) spun/unwashed: this portion was spun in the ultra-centrifuge for 5 min at 800 rpm at a temperature of 300 "C to filter off the melt; (b)spun/washed: this portion was spun in a similar manner, but was then washed with a very small quantity of distilled water and then dried in a desiccator; (c) spun/unwashed and sintered: this portion was also spun in a similar manner, but for a longer time (ca.15 min) without washing. It was then pressed into a pellet and sintered at 1100 "C in a muffle furnace. The XRD patterns of the three samples, as in the previous section are comparable to either cubic zirconia (JCPDS 27-997) or 15 mol% Ca0-85 mol% ZrO,, (JSPDS 26-341), as shown in Fig. 4 for samples (a) and (c). The strongest lines of the LiN0,-KNO, mixture emerge slightly in sample (a). They are detected neither for the washed sample (b),nor for the sintered sample (c). Discussion Ca(NO,), was reported to decompose after heating for 2 h in air at 580"C, giving NO,, 0, and CaO which was confirmed by X-ray powder diffra~tion.,~ In eutectic LiN0,-KNO,, Ca(NO,), was found to be soluble and stable up to 500"C,29 which is in agreement with our work.The fact that Ca(NO,), did not react with LiN03-KN0, at these low temperatures can be explained in terms of the relative polarising powers /I 10 20 30 40 50 60 70 80 2Bldegrees Fig. 4 XRD pattern of CaO-ZrO, samples (a) and (c) prepared from reactions of Ca(NO,), and ZrO, 22Cll 56 with eutectic LiN0,-KNO, using ultra-centrifugation J. Muter. Chem., 1996, 6(3),495-500 497 (acidities) of the Ca2+ and Li+ cations At higher temperatures (above 500 "C), CaO is precipitated according to eqn (1) Ca2++ 2N03- +CaO + 2N0, +to, (1) In this study, CaO has been precipitated between 540 and 600 "C ZrOC1, was reported to react with eutectic NaN03-KN03 mixture (mp 225 "C) and with pure NaNO, (mp 308 "C) giving ZrO, either in the tetragonal or tetragonal + monoclinic forms 23 29 Thermogravimetry of the reaction of ZrO, ,,ell 56 with LiNO3-KNO, showed a total mass loss of 48 8%, which is closer to the value of 507% calculated for eqn (2) which leads to oxygen than to the value of 38 3% calculated for eqn (3) which leads to chlorine ZrOi ,,c1, 56+ 1 56N0,- -+ ZrO, + 1 56N0, + 0 390, + 1 56C1- (2) ZrO, 56 -l-0 78NO3--+ ZrO, + 0 78N0, + 0 39C1, +0 78C1- (3) In our experiments the release of oxygen is more likely to occur than that of chlonne, in agreement with Jebrouni's earlier results 24 This may be related to the concentration of zirconium oxychloride in the molten bath, a low concentration (z e the present experiments) favouring the evolution of oxygen, whereas a high one (I e Jebrouni's experiments) promoting the evolution of chlorine Calcia stabilised zirconia can be obtained by firing, at elevated temperature (> 1000 "C), mixtures of zirconia and calcium oxide or carbonate According to Duwez et ul, cubic CaO-ZrO, is produced with powders containing between 16 and 30molYo CaO, whereas for lower contents cubic and monoclinic phases are produced 30 In 1963, Tien and Subbarao reported that cubic calcia fully stabilised zirconia (Ca-FSZ) is obtained from mixtures at compositions between 12 and 21 mol% CaO Mondal et a1 corroborated this result as in their work the addition of 20mol% CaO to ZrO, favoured the formation of cubic calcia stabilised zirconia 32 For the same ratio, Garvie suggested that the cubic solid solution 20 mol% CaO-ZrO, could be the mixed oxide CaZr,O, 33 The coprecipitation of Ca-FSZ (cubic solid solution after heat treatment of the amorphous phase obtained after drying) by freeze drying has been reported by Roosen and Hausner,,, using Ca(NO,), and ZrOCl, as starting materials The same chemicals were used for the present study but the preparation was in non-aqueous solution The main advantage of freeze- drying is the avoidance of the formation of hard agglomerates The method is, however, rather expensive, compared to the method used in the present work employing molten salts In transformations involving calcium nitrate and zirconium oxychlonde together, Ca(NO,), reacts with molten eutectic LiN0,-KNO, at lower temperatures than when it is alone, 1 e under 500 "C The simultaneous reaction leads to calcia Table 3 Influence of reaction parameters [ratio Ca(N03), ZrO, ,,Cll lines in diffraction patterns of calcia stabilised zirconia unwashed CaO-ZrO, intensity, III, stabilised zirconia, in which calcium oxide is incorporated inside the zirconia lattice On the basis of the literature, the formation of either cubic or monoclinic solid solutions is expected in the range 0-30 mol% CaO-ZrO, and, for the composition 20 mol% CaO-ZrO,, the mixed oxide CaZr,O, should be obtained The existence of domains containing a tetragonal CaO-ZrO, solid solution in Stubican's phase diagram,' 36 leads to the suggestion that this phase could be produced in a metastable state by rapid quenching, from temperatures above 1170°C to room temperature, of melts of CaO and ZrO, In our experi- ments in molten nitrates, it is shown that neither the formation of a monoclinic solid solution even for the lowest content in CaO, nor the precipitation of CaZr,O, for the highest content, take place At first glance, the XRD patterns of all the samples are comparable either to the one of cubic zirconia (JCPDS 27- 997) or to the one of cubic Ca, 15Zro s50185 (JCPDS 26-341) The main difference between cubic and tetragonal zirconia is the splitting of some cubic lines For well crystallised powders, the splitting appears clearly in XRD patterns, allowing an easy identification Thus, it can be concluded undoubtedly that the powder 20 8 mol% CaO-ZrO, separated from the molten salt by ultra-centrifugation, then pressed and sintered [sample (c), previous section] is cubic calcia stabilised zirconia, even if the intensity of lines 220 and 3 11 is a little lower than indicated in JCPDS for cubic ZrOz or cubic Ca, 15Zro 8sOi 85 (Table 3) For finely crystallised powders, the convolution of XRD peaks makes the detection of the splitting more difficult or even impossible and the question arises whether the non-sintered CaO-ZrO, samples, prepared from molten salts, are cubic or tetragonal calcia stabilised zirconia Taking into account the fact that the splitting is never detected, and the results pre- viously published in the literature, it can reasonably be assumed that the 20 8 mol% CaO-ZrO, samples, either unwashed and extracted by washing (see earlier), or extracted by ultra-centrifugation without washing and with further water washing [sample (a) and (b)]are also cubic calcia stabilised zirconia, all the more so because the intensity of the XRD peaks is in fairly good agreement with those for both cubic ZrO, and cubic Ca, i5Zro 850i (Table 3) For samples with a lower85 CaO content, 4 7 and 12 2 mol% CaO-21-0, (see earlier), the situation is more complex Yet, both for unwashed and for washed powders, the increase of calcia content from 47 to 20 8 mol% does not involve any variation in the intensities of the zirconia peaks (Table 3) This can be interpreted on the basis that, whatever the calcium content in the range 4 7-20 8 mol%, the structure of calcia stabilised zirconia co- precipitated from molten alkali-metal nitrates is cubic Moreover, it is noticeable that water washing changes neither the intensities of the XRD peaks, as shown in Table3 for 122mol% CaO-ZrO,, nor the ratio CaO (CaO+ZrO,) determined from chemical analysis, as revealed for the same 56, washing, use of ultra-centrifugation] on the intensity of zirconia XRD ultra-centrifugated unwashed washed unwashed & sintered CaO-ZrO, (12 2 mol%) intensity, I/Z, d/A hkl 4 7 mol% 12 2 mol% 20 8 mol% intensity, I/Z, (4 (4 2 95 111 100 100 100 100 100 100 2 55 200 21 22 20 22 19 21 181 220 51 54 53 54 45 35 155 311 33 35 36 31 27 17 148 222 6 8 7 6 7 4 128 400 5 6 6 4 5 3 498 J Muter Chem, 1996, 6(3), 495-500 samples in Table4.This means that calcium oxide is partly hydrolysed by water when it alone is precipitated, whereas when CaO is incorporated inside the zirconia lattice, forming a solid solution, water washing has no effect. The comparison between both 12.2 mol% CaO-ZrO, samples indicates also that water washing eliminates quite completely the excess of alkali-metal nitrates and the formed alkali-metal chlorides, producing a calcia stabilised zirconia with purity ratios which are satisfactory for further uses.In the same manner, the comparison of 20.8 mol% CaO-ZrO, samples (a) and (c) extracted by ultra-centrifugation (Table 4) reveals that pressing and sintering at 1100 "C eliminates most of the alkali-metal nitrates remaining in the powders after ultra-centrifugation. From chemical analysis (Table4), it appears also that cubic calcia zirconia solid solutions coprecipitated from molten nitrates are characterised by a CaO mol% quite close to that initially introduced into the molten medium ratio Ca(N0,)2 : [Ca(NO,), + ZrOl.22Cll,56}. Sintering does not significantly change this ratio. The last point to discuss is the mechanism of the reaction in the molten nitrate medium.Transformations occurring below 500 "C, the formation of solid solutions CaO-ZrO,, when calcium nitrate and zirconium oxychloride react together, would seem difficult if the starting salts remained in the solid state. They are much more likely to be partially dissolved in the molten medium giving calcium and zirconyl ions, according to eqn. (4) and (5), which react with 0,-ions coming from the dissociation of nitrates, to precipitate the final mixed oxide according to eqn. (6): Ca(N03)2$ Ca2++ 2N03-(4) ZrO(, +x)Clz(l-x) sZrO(, +,),('-,)+ +2(1-x)c1-(5) aCa2++ bZrO(l+,,2(1-X)++ [a+b( 1--)lo2-+ca,Zr,O(,+,,) (6) where Ca,ZrbO(, + 26) is equivalent to ( 100a/a+ b) mol% CaO- ZrO,.Concerning the dissociation of nitrates, Jebrouni et ~1.~' proposed an oxidation-reduction transformation evolving chlorine via eqn. (7): 2NO,-+ 2C1- -+2N02 + 20,-+ C1, (7) But, as explained previously in connection with the reaction of zirconium sulfate with nitrates,,, a dissociation evolving oxygen via eqn. (8) can also be considered: 2N03- +2NO, +02-+0.50, (8) As stated earlier under the present experimental conditions, particularly the low concentrations of calcium nitrate and zirconium oxychloride in molten alkali-metal nitrates, the evolution of oxygen take place. The observed mass losses are in fair agreement with eqn. (9) and corroborate this assumption: aCa(NO,), + bZrOl.22Cll~56+ 1.56bN03- -+Ca,ZrbO(, + 2,) +2(a+0.78b)NO2+0.5(a+0.78b)O2+1.56bC1- (9) As an example, the results of the chemical analysis of sample (a) (Table 4) agree with the formula 20.8 mol% CaO-ZrO, which is equivalent to Cao~21Zro,7gOl 79.Replacing a and b by their values (0.21 and 0.79, respectively) in eqn. (9) gives a calculated mass loss of 56.3% of the starting quantity in calcium nitrate + zirconium oxychloride, which is relatively close to the calculated value, 59.0%. Some Morphological Characterisations of 20.8 mol%CaO-ZrO, Powders The morphology of two 20.8 mol% calcia stabilised zirconia samples was investigated by transmission electron microscopy. Both samples were obtained from the same experiment carried out as described earlier. The first sample, very close to sample 20.8 mol% formed earlier, was extracted from the excess of molten salt by water washing.The second sample, identical to sample (a),was separated by centrifugation at 300 "C. For both powders, only cubic calcia stabilised zirconia is detected from the electron diffraction patterns, as shown on Fig. 5 for the sample extracted by washing. Considering chemi- cal analysis data (Table 4), such a result could be expected for the washed sample but is more surprising for the centrifuged sample. It can be understood by assuming that the interaction of the electron beam with the powder involves an increase of the temperature sufficient to melt the alkali-metal nitrates (mp 132°C). Examination of both powders at low magnification reveals agglomerates with sizes in the range 0.2-2 pm.High magnifi- cation investigations (Fig. 6) show that crystallites are larger in the powder separated by centrifugation (40-50 nm) than in the powder extracted by water washing (10-20nm). The difference is corroborated by the XRD patterns, which exhibit lines broader for the washed powder than for the centrifuged one [full width at half maximum: washed wz3 mm, Fig. 3; centrifuged w x2 mm, Fig. 4(a)]. Conclusions The reaction of zirconium oxychloride with molten eutectic LiN0,-KN03 produces finely divided zirconia at temperatures below 500°C. Calcium oxide can also be precipitated from reaction of calcium nitrate with LiN0,-KNO,, but at tempera- tures above 500°C. The CaO obtained in this way is very unstable to water washing.Fig. 5 ED pattern of 20.8 mol% CaO-ZrO, extracted by washing Table 4 Chemical analysis of CaO-ZrO, samples obtained from reactions of Ca(N03), and ZrO, ,,C1, 56 with eutectic LiN0,-KNO, mass% Zrsample Ca K Li NO3 c1 mol% Ca : (Ca + Zr) 39.1unwashed 12.2 mol% 2.4 11.7 1.5 27.1 3.1 12.3 67.0ed 12.2 mol%wash 4.1 <0.1 <0.1 <0.1 <0.1 12.2 46.0ultra­centrifuged 20.8 mol% 5.3 7.9 1.1 18.6 2.3 20.8 65.6sintered 20.8 mol% 7.5 0.3 <0.1 <0.1 <0.1 20.7 J. Muter. Chem., 1996, 6(3),495-500 499 References 1 A. B. Searle, Refractory Materials: Their Manufacture and Uses, Griffen, London, 1940, p. 210. 2 N. Q. Minh, J. Am. Ceram. SOC., 1993,76,563. 3 J.D. McCullough and K. N. Trueblood, Acta Crystallogr., 1959, 12, 507. 4 D. K. Smith and H. W. Newkirk, Acta Crystallogr., 1965, 18,983. 5 G. Teuffer, Acta Crystallogr., 1962, 15, 1187. 6 D. K. Smith and C. F. Cline, J. Am. Ceram. SOC., 1962,45249. 7 W. D. Kingery, H. K. Bowen and D. R. Uhlmann, Introduction to Ceramics,2nd edn., Wiley, New York, 1976, p. 81. 8 R. C. Garvie, R. H. Hannink and R. T. Pasoe, Nature Phys. Sci., 1975,258,703. 9 D. W. Richerson, Modern Ceramic Engineering, Marcel Dekker, New York and Basel, 1982, p. 376. 10 S. Shinroka, Fine Ceramics, Elsevier, Oxford, 1985, p. xvi. 11 D. Gauthier, G. Olalde and A. Vialaron, Advances in Ceramics, vol. 24B, Science and Technology of Zirconia 111, Am.Ceram. SOC., Ohio, 1986, p. 879. 12 P. Cousin and R. A. Ross, Muter. Sci. Eng. A, 1990,130,119. Fig. 6 TEM images washing, and (b) by of 20.2 mol% CaO-ZrO, extracted (a) by water centrifugation at high temperature 13 14 D. H. Kerridge and J. Cancela Rey, J. Inorg. Nucl. Chem., 1977, 39,405. B. Durand, M. Jebrouni and M. Roubin, Euchem Conference on The reactions at temperatures below 500 "C, of mixtures of calcium nitrate and zirconium oxychloride, with mole ratios Ca: (Ca+Zr) in the range 4-20%, leads to calcia stabilised zirconia. Owing to the incorporation of calcium oxide inside the zirconia lattice, the obtained solid solutions are stable under water washing which allows, after cooling, an easy separation of the insoluble stabilised zirconia from the excess of alkali-metal nitrates and the salts formed during the reaction.Powders with a satisfying degree of purity are thus recovered. At the end of the transformation, the unreacted molten salt can also be separated from the formed stabilised zirconia by ultra-centrifugation at elevated temperature. The major part of the remaining alkali-metal nitrate can then be eliminated 15 16 17 18 19 20 21 22 Molten Salts, Greece, 1990. D. H. Kerridge and A. Y. Khudhari, J. Inorg. Nucl. Chem., 1975, 37, 1893. H. Frounzanfar and D. H. Kerridge, J. Inorg. Nucl. Chem., 1979, 41, 181. S. S. Alomer and D. H. Kerridge, J. Chem. SOC., Dalton Trans., 1978,1589. G. Picard, Euchem Conference on Molten Salts, Greece, 1990. D. H. Kerridge, Molten Salts as Nonaqueous Solvents, ed.J. J. Lagowski, Academic Press, NY, 1978, ch. 5, p. 229. H. Al-Raihani, B. Durand, D. H. Kerridge and D. Inman, J. Muter. Chem., 1994,4,1331. M. Jebrouni, B. Durand and M. Roubm, Ann. Chim. Fr., 1991, 16, 569. M. Jebrouni, B. Durand and M. Roubin, Ann. Chim. Fr., 1992, 17, 143. by natural sintering at 1100°C. Besides avoiding the use of water which can favour the agglomeration of the powders, the greatest interest in ultra-centrifugation is that, in the case of larger scale development, it could permit the recycling of the molten medium. 23 24 25 H. Al-Raihani, B. Durand and D. Inman, J. Muter. Chem., to be submitted. M. Jebrouni, Thesis, Universite Claude Bernard Lyon 1, France, 1990. J. L. Tosan, Thesis, Universite Claude Bernard Lyon 1, France, 1991.The reaction of calcium nitrate and zirconium oxychloride 26 I. J. Bear, Aust. J. Chem., 1966, 19, 357. with molten eutectic LiN03-KN03 proceeds according to a dissolution-precipitation process involving the dissolution of starting salts with the probable formation of complex soluble species whose decomposition induces the precipitation of the obtained solid solution. The solid solutions contain a mole ratio Ca: (Ca+Zr) which is practically identical to the one introduced in the molten medium. Whatever the CaO content, 27 28 29 30 31 32 Joint Committee on Powder Diffraction Standards, Pennsylvania, 1973. B. Durand and M. Roubin, Muter. Sci. For., 1991,73-75, 663. H. Abood, PhD Thesis, University of Southampton, 1984. P. Duwez, F. Ode11 and F. H. Bowen, Jr, J. Am. Cerum. Soc., 1952, 35, 107. T. Y. Tien and E. C. Subbarao, J. Chem. Phys., 1963,39,1041. B. Mondal, A. N. Virkar, A. B. Chattopadhyay and A. Paul, in the investigated range, the reaction in molten nitrates below 500 "Cproduces cubic calcia stabilised zirconia. 33 34 J. Mater. Sci. Lett., 1987,6,7-12, 1395. R. C. Garvie, J. Am. Ceram. SOC., 1968,51,553. A. Roosen and H. Hausner, Ceramic Powders, ed. P. Vincenzini, This work was supported by the Commission of the European Communities in the frame of a twinning contract between the Imperial College of London and the University of Lyon. The 35 36 Elsevier, Amsterdam, 1983, p. 773. V. S. Stubican and J. R. Hellmann, Adv. Ceram., 1981,3,25. V. S. Stubican and S. P. Ray, J. Am. Ceram. SOC., 1977,60,534. authors are indebted to the Commission of European Paper 5/06183K; Received 19th September, 1995 Communities for its financial support. 500 J. Muter. Chem., 1996, 6(3), 495-500