Russian Journal of Applied Chemistry, Vol. 78, No. 3, 2005, pp. 347!350. Translated from Zhurnal Prikladnoi Khimii, Vol. 78, No. 3, 2005, pp. 353!356. Original Russian Text Copyright + 2005 by Kamaev, Archugov, Mikhailov.

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ

INORGANIC SYNTHESIS AND INDUSTRIAL INORGANIC CHEMISTRY

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ

Behavior of the Al2O3!ZrO2 System at High Temperatures D. N. Kamaev, S. A. Archugov, and G. G. Mikhailov South-Ural State University, Chelyabinsk, Russia

-

Received September 17, 2004

Abstract The system Al2O33ZrO2 at temperatures of up to 2100oC was studied by differential thermal analysis.

The oxide system Al2O33ZrO2 is of great practical interest. Data on its phase diagram are significant for the production of abrasive and structural materials operating at high temperatures. These materials are produced either by fusion of the components with their subsequent crystallization or by sintering of finely dispersed powders. The phase diagram furnishes information on the material crystallization, and the use of appropriate cooling mode makes it possible to control the product quality. However, the high liquidus temperature in the Al2O33ZrO2 system significantly complicates studying this system and obtaining reliable results. Therefore, data on this system are far from being exhaustive. Most authors [137] believe that the system components form a simple eutectic and state that any intermediates are absent from the system. However, there is no agreement in the literature regarding the eutectic composition, its temperature, and shape of liquidus lines. In particular, according to the above-mentioned data, the temperature of the eutectic transition varies from 1710 to 1920oC, and the eutectic composition, from 30 to 55 wt % ZrO2.

EXPERIMENTAL The experiment was carried out on a computercontrolled upgraded installation for high-temperature three-crucible differential thermal analysis, with data processing on a PC [12]. Analytically pure grade aluminum oxide (Al2O3) and zirconium nitrate Zr(NO3)4 . 5H2O were used as starting reagents. According to thermographic analysis, zirconium nitrate loses water of crystallization at low temperatures and decomposes to the oxide at high temperatures. Therefore, it was annealed in air for 3 h at 1200oC. Before weighing, the starting oxides were additionally annealed for several hours to remove moisture and adsorbed impurities. Then they were immediately placed in weighing bottles, which were subsequently stored in a desiccator under dry air, which excluded the saturation of the oxides with water vapor and the corresponding weighing errors. The grade of accuracy of the balance in use warranted that the deviation of the obtained composition from the calculated composition did not exceed 0.01 wt %. Samples containing from 0 to 90 wt % ZrO2 in 10 wt % steps were prepared. In the range from 40 to 60 wt % ZrO2, the step was 2 wt %. After mixing, the samples were pressed in a special hard-alloy mold.

According to [5], the eutectic transition takes place at 1890oC, and the eutectic contains about 40 wt % ZrO2. The eutectic composition was confirmed in [8] by the method of three-crucible thermal analysis, which makes it possible to obtain high-precision data [8310]; however, the eutectic temperature was found to be 1866 + 7oC. In addition, according to [8], the liquidus line has a complicated shape, suggesting the tendency of the system to phase separation at temperatures slightly exceeding those of the liquidus line.

Despite the fact that ZrO2 and Al2O3 practically do not react with atmospheric moisture and CO2, the pressed samples were annealed once more at 1200oC under 1 atm of argon. To obtain data required for the subsequent calculations, the sample pellets were weighed before placing in a crucible of the thermographic installation.

Taking into account the importance of the phase diagram of the Al2O33ZrO2 system for the production of refractories and abrasives and also the ambiguous character of the previously obtained data, we have reexamined the phase diagram of this system.

The samples under study were homogenized by step-by-step annealing. In the first step, the sample was heated at a rate of 60 deg min!1 to 2100oC. In the second step, it was sharply cooled to 80031000oC. In the third step, the sample was heated at a rate of

1070-4272/05/7803-0347 + 2005 Pleiades Publishing, Inc.

348

KAMAEV et al.

Multistep heating and cooling allowed us to obtain highly reproducible results. To collect a statistically significant representative dataset, each sample after homogenization was subjected 7313 times to the subsequent treatment steps. Thermocouple readings were corrected during an experiment by the Al2O3 melting point and further by the copper melting point [13].

DT

b

c c T, oC Fig. 1. Typical DTA curves recorded at a heating rate of 60 deg min!1. (T) Temperature; the same for Fig. 2. Temperature measured: (a) solidus, (b) onset of phase separation, and (c) liquidus. ZrO2 content in the sample (wt %): (1) 46, (2) 50, (3) 60, and (4) 30.

T, oC

c d

e

b

wt %

Fig. 2. Phase diagram of the Al2O3!ZrO2 system. Experimental and calculated data: (a) model of perfect solutions, (b) model of regular solutions, (c, d) model of subregular ionic solutions, and (e) experimental points.

60 deg min!1 to a temperature higher by 100 3200oC than the expected liquidus point. Finally, after keeping for 335 min at the maximal temperature, which provided complete homogenization of the sample, it was cooled at the same rate to the temperature 30 3 100oC lower than the eutectic point and kept at this temperature for 335 min. X-ray spectral microanalysis of the quenched alloys indicates that such a treatment is sufficient for the homogenization. The resulting sample was subjected to thermal analysis. The sample was heated at a rate of 30 deg min!1 to 2100oC and then was cooled at the same rate to a temperature 30 3100oC lower than the eutectic transition point. At this temperature, it was kept for 33 5 min. In the next step, the annealed sample was sharply cooled to room temperature. In the final step, the sample was heated at a rate of 20 3 60 deg min!1 to 2100oC and then cooled to room temperature.

The DTA curves of the samples heated at various rates (20 3 60 deg min!1) demonstrate the effect of the heating rate and allow us to extrapolate the data obtained to the zero heating rate and finally to determine the equilibrium liquidus and eutectic lines in the phase diagram (Figs. 1, 2). The shape of the experimental liquidus line suggests that either a chemical compound is formed in this system or it separates into layers. To answer this question, we specially studied several samples containing 50 wt % ZrO2. Each sample was heated to approximately 1900oC and kept at this temperature for 3 min. Then the temperature was decreased to 1890oC (according to the data obtained, this temperature corresponded to a point in the liquidus line) and the sample was kept at this temperature for 3 min. Then the temperature was further decreased by 10oC, and the sample was kept for about 10 min. If the sample actually separates into layers, such keeping at 1880oC must ensure efficient formation of separable liquids. Then the sample was sharply cooled by purging the furnace with cold argon. The further study has shown that such a sharp cooling conserved to room temperature the phase composition of the samples characteristic of 1880oC. The quenched samples were studied by the X-ray diffraction analysis, which showed no formation of any chemical compounds in the sample containing 50 wt % ZrO2. The distribution of elements in the quenched sample with a total ZrO2 content of 50 wt % was analyzed with a scanning electron microscope. Phase conglomerates with high and low Zr(IV) content and, correspondingly, with low and high Al(III) content were observed (Fig. 3). No primary crystals of either ZrO2 or Al2O3 were found in this sample. Samples with a ZrO2 concentration of 60 and 30 wt % kept at the same temperature (1880oC) were quenched for comparison. Primary crystals of pure ZrO2 and a phase with the composition corresponding to the melt in equilibrium with ZrO2 were found in the quenched sample containing 60 wt % ZrO2 (Fig. 4a). The distribution of elements in these phases was analyzed. In one phase

RUSSIAN JOURNAL OF APPLIED CHEMISTRY

Vol. 78

No. 3

2005

BEHAVIOR OF THE Al2O33ZrO2 SYSTEM AT HIGH TEMPERATURES

(quenched melt), the ratio of intensities of Al(III) and Zr(IV) characteristic radiation was invariable. In the other phase (primary ZrO2 crystals), aluminum was not found at all, but the intensity of Zr(IV) radiation was much higher than in the phase of the quenched melt. The quenched sample containing 30 wt % ZrO2 was also analyzed. In the quenched melt, the ratio of the intensities of Al(III) and Zr(IV) characteristic radiation was invariable. The other phase appeared to be primary corundum crystals without Zr(IV) (Figs. 4b).

349 B

Direction of scanning beam Fig. 3. Distribution of Zr(IV) and Al(III) in quenched samples containing 50 wt % ZrO2. (A, B) Intensities of characteristic Zr(IV) and Al(III) radiation, respectively; the same for Fig. 4.

The eutectic temperature was determined by well reproducible bases of the DTA peaks for all the experiments. It was estimated at 1861 + 6oC. The eutectic composition was estimated at 40 wt % ZrO2 by the interception point of the liquidus and solidus lines.

(a) Primary ZrO2 crystals

B

The detected invariant monotectic transition has the following parameters: temperature 1861 + 6oC and melt compositions from 45 to 58 wt % ZrO2. Reliable experimental data obtained make it possible to calculate the phase diagram using various mathematical theories of oxide melts, such as theories of perfect, regular, and subregular ionic solutions.

Direction of scanning beam (b) Primary Al2O3 crystals

The models of perfect and regular solutions [14, 15] failed to describe satisfactorily the liquidus line of the system (Fig. 2, a); therefore, we have attempted to calculate coordinates of the liquidus line within the framework of the theory of subregular ionic solutions [16]. According to this model, the chemical potential of a liquid solution in a binary system is m1 = m01 + n1[RT ln x1 + 3x21 x22 Q1112 + (2 3 3x1) x1 x2 Q1122

+ (1 3 3x1) x32 Q1222], m2 = m02 + n1[RT ln x2 + (1 3 3x2) x31 Q1112

+ (2 3

3x2) x21 x2 Q1122

+

3x21 x22 Q1222],

where R is the universal gas constant; ni, number of cations in an oxide molecule; m 0i , standard chemical potential; Q1112, Q1122, and Q1222, energy parameters; and xi, mole fraction, with index 1 referring to ZrO2 and index 2, to Al2O3. To construct liquidus lines, we choose the experimental reference points without taking into account separation into layers: 20, 40, and 60 wt % ZrO2; then we wrote down the expressions for the equilibria in these points and solved the resulting system of equations. Using the parameters obtained, we calculated the liquidus line coordinates. The resulting liquidus line points to the separation of the system into layers RUSSIAN JOURNAL OF APPLIED CHEMISTRY

Vol. 78

B

Direction of scanning beam Fig. 4. Distribution of Zr(IV) and Al(III) in quenched samples containing (a) 60 wt % ZrO2 and (b) 30 wt % ZrO2: equilibrium of primary ZrO2 crystals with a melt.

(Fig. 2, c), but it agrees with the experimental data insufficiently. Then we carried out calculations taking into account the separation into layers found experimentally. The following points were chosen: 40 wt % ZrO2, 1861oC; 43 wt % ZrO2, 1872oC; 56 wt % ZrO2, 1872oC; and 60 wt % ZrO2. For these points, we composed equilibrium expressions and solved the system of the resulting equations. The following energy parameters were found: Q1112 = 9897, Q1122 = 71 280, and Q1222 = 20 490 J mol!1. The liquidus line calculated with the parameters found is shown in Fig. 2, d. It differs to some extent from the experimental liquidus line, which may be due to the averaging of the calculated interaction parameters for a relatively wide temperature range. Thus, the experimental study of the system Al2O33 No. 3

2005

350

KAMAEV et al.

ZrO2 and the application of the theory of subregular ionic solutions to this system suggest that the possibility of phase separation in the melt.

8. 9.

REFERENCES 10. 1. Suzuki, H., Kimura, H., Yamada, S., and Yamauchi, T., J. Ceram. Assoc. Jpn., 1961, vol. 69, no. 782 (2), p. 52. 2. Suzuki, H., Kimura, H., Yamada S., and Yamauchi, T., J. Ceram. Assoc. Jpn., 1961, vol. 69, no. 790 (10), pp. 3453350. 3. Cevales, G., Ber. Deutsch. Keram. Ges., 1968, vol. 45, no. 5, pp. 216 3219. 4. Fisher, G.R., Manfredo, L.J., McNally R., and Doman, R.C., J. Mater. Sci., 1981, vol. 16, no. 12, pp. 344733451. 5. Alper, A.M., Science of Ceramics: Proc. Int. Conf., July 538, 1965, vol. 3, pp. 3353369. 6. Volkova, I.Yu., Semenov, S.S., Kravchik, A.E., et al., Izv. Akad. Nauk SSSR, Neorg. Mater., 1987, vol. 23, no. 3, pp. 4483 451. 7. Wartenberg, H., Linde, H., and Jung, R., Z. Anorg.

11. 12.

13. 14.

15.

16.

Chem., 1928, vol. 176, p. 349. Jerebtsov, D.A., Mikhailov, G.G., and Sverdina, S.V., Ceram. Int., 2000, vol. 26, pp. 8213823. Zherebtsov, D.A. and Mikhailov, G.G., Prib. Tekh. Eksp., 2000, no. 5, pp. 140 3148. Jerebtsov, D.A. and Mikhailov, G.G., Ceram. Int., 2001, vol. 27, pp. 25328. Jerebtsov, D.A. and Mikhailov, G.G., Ceram. Int., 2001, vol. 27, pp. 2473250. Kamaev, D.N., Archugov, S.A., and Mikhailov, G.G., Izv. Vyssh. Uchebn. Zaved., Chern. Metall., 2003, no. 6, pp. 46 3 49. Bragin, B.K., Tr. Vses. Nauchno-Issled. Inst. Metrol. im. D.I. Mendeleeva, 1963, issue 71 (131), pp. 973100. Kozheurov, V.A., Statisticheskaya termodinamika (Statistical Thermodynamics), Moscow: Metallurgiya, 1975. Gerasimov, Ya.I. and Geiderikh, V.A., Termodinamika rastvorov (Thermodynamics of Solutions), Moscow: Mosk. Gos. Univ., 1980. Mikhailov, G.G. and Povolotskii, D.Ya., Termodinamika raskisleniya stali (Steel Deoxidation Thermodynamics), Moscow: Metallurgiya, 1993.

RUSSIAN JOURNAL OF APPLIED CHEMISTRY

Vol. 78

No. 3

2005