Evaluation of Phase Diagrams for the Al 2 O 3 -CaO-SrO System by In-Situ Observation Using Confocal Laser Microscope

Transcription

1 Materials Transactions, Vol. 5, No. 2 (29) pp. 254 to 26 #29 The Japan Institute of Metals Evaluation of Phase Diagrams for the Al 2 O 3 -CaO-SrO System by In-Situ Observation Using Confocal Laser Microscope Tomonori Kuroki ; *, Yoshitoshi Saito 2, Taijiro Matsui 2 and Kazuki Morita Institute of Industrial Science, The University of Tokyo, Tokyo , Japan 2 Refractory Ceramics R&D Division, Environment & Process Technology Center, Nippon Steel Corporation, Futtsu , Japan Phase relations for the Al 2 O 3 -CaO-SrO ternary oxide system were clarified. This ternary system has a high liquidus temperature in most compositions and contained many types of solid solutions with a wide composition range. Hence, it was difficult to predict the precise phase diagrams by using conventional techniques such as chemical equilibration method. In fact, the phase diagram reported formerly contains a considerable amount of prediction and surmise and is far from satisfaction. In the present study, an in-situ observation at increasing temperatures was performed using a high temperature observational system that included a confocal scanning laser microscope (CSLM) in combination with an infrared image furnace as a heating device. As a result, definite images of the samples during melting were obtained, and the advantages of this method for the measurement of this type of ternary oxide system were confirmed. Further, a more appropriate ternary phase diagram and its 973 K isothermal cross section were predicted. [doi:32/matertrans.mra28352] (Received September 26, 28; Accepted November 8, 28; Published January 8, 29) Keywords: confocal laser microscope, in-situ observation, phase diagram, equilibrium measurement, strontium oxide. Introduction A confocal scanning laser microscope (CSLM) has been used in various researches, mainly in biological studies ) such as cytology, histology, and cell physiology. Its features such as high definition, a narrow focal depth, and harmlessness to cellular activation have been exploited in these researches. On the other hand, it is known that a CSLM image remains nearly unaffected by a strong emission light at a high temperature. Hence, a high temperature observational system, which consists of a CSLM in combination with an infrared image furnace, has recently been used with the central focus on the in-situ observation at a high temperature in iron and steelmaking researches, such as for the evaluation of the peritectic reaction rate in ferroalloys, 2) investigation of the inclusions on molten iron surfaces, 3,4) investigation of microholedrilling in hot tool steel via a micro-edm process, 5) etc. As for oxides, Kimura et al. 6) directly observed the melting and solidification behavior of the CaO-SiO 2 - FeO x slag system under varying oxygen partial pressure. In order to predict phase diagrams, chemical equilibration method, differential thermal analysis (DTA), differential scanning calorimetry (DSC), hot thermocouple technique, etc. have been conventionally used. However, a high temperature and the existence of solid solutions make it difficult for the chemical equilibration method to be employed, and constant variations in the state are difficult to detect using DTA and DSC. As surface variations can be directly detected using a CSLM, a method employing the CSLM is considered to have certain advantages over a conventional method. In this research, equilibrium measurement through an in-situ observation with the help of a CSLM was attempted. In this paper, we have focused on the Al 2 O 3 -CaO-SrO *Graduate Student, The University of Tokyo ternary oxide system with a high liquidus temperature and various solid solutions with a wide composition range. Masazza et al. 7,8) have investigated this ternary oxide system through an optical microscope observation. Although they rectified the ternary phase diagram (Fig. ) and its isothermal cross section at 673 K, the high temperature region in the phase diagram was not well investigated and the composition range employed by them was limited. Hence, many portions of the phase diagram must have been predicted on the basis of speculation or surmise. Further, a more appropriate phase diagram for the Al 2 O 3 -CaO-SrO system is required when this ternary system is to be used in practical applications such as heat-resistant ceramics, 9) electric industrial materials, ) etc. Therefore, the validity of equilibrium measurement using a CSLM for an in-situ observation was clarified by measuring the equilibria of the Al 2 O 3 -CaO-SrO ternary oxide system in this study. 2. Experimental 2. Preparation of experimental materials The Al 2 O 3 -CaO-SrO slag system was used in this experiment. The compositions of the samples are shown in Table and Fig. 2. Firstly, the CaO-Al 2 O 3 binary system samples, which have been well investigated in the past, were investigated to clarify the validation of equilibrium measurement using the CSLM. Thereafter, the slag system containing SrO was tested, and the effects of SrO substitution were clarified. They were mixed well, calcined at 373 K for h, and premelted in a graphite crucible in an Ar gas atmosphere at K using an electric induction furnace. In this premelting procedure, the temperature of the samples was increased as fast as possible and maintained within a range of K for a few minutes in order to prevent variations in the composition. After premelting, a chemical analysis of the sample was performed via ICP-AES.

2 Evaluation of Phase Diagrams for the Al 2 O 3 -CaO-SrO System by In-Situ Observation Using Confocal Laser Microscope 255 Fig. Phase diagram for the Al 2 O 3 -CaO-SrO system reported by Masazza et al. 8) Table Compositions of samples. Sample Al 2 O 3 CaO SrO x Sample Al 2 O 3 CaO SrO x A D A D A D A D A D A D A D A D A E A E A E A E A E A E A E A E B F B F B F B F B G C G C G C G C C C C C Fig. 2.4 A Observation and temperature measurement A high temperature observational system (VL-2/ Lasertec Co. Ltd.) was used in this research. An infrared image furnace was equipped as a heating device wherein the samples were set on a platinum plate (Fig. 3). In order to eliminate the effect of temperature distribution due to the unequal incidence of emission from a halogen lamp in the furnace, small size samples were used, i.e., a few crushed samples (approximately mm each) were used in this experiment. While the temperature of the samples was measured using a B-type thermocouple welded beneath a sample holder, it was impossible to detect the accurate temperature of the samples due to the large temperature gradient around the samples. Accordingly, the temperature was clarified using the congruent melting temperature of intermediate compounds of C F G D B E CaAl 2 O 4 (CaO Al 2 O 3 ).4 Ca 2 Al 4 O 33 (2CaO 7Al 2 O 3 ) Composition of the grouped samples in the Al 2 O 3 -CaO-SrO system.

3 256 T. Kuroki, Y. Saito, T. Matsui and K. Morita Quartz Window B-Type Thermocouple Sample Holder Ar Gas Outlet Laser Microscope (Objective Lens) Ar Gas Inlet Furnace Cooling Water Outlet Lamp Cooling Air Control Valve Furnace Cooling Water Inlet Lamp Cooling Air Inlet Halogen Lamp (Max..5 kw) Platinum Plate Alumina Bar Platinum Wire Alumina Plate Fig. 4 Observable Area Relationship between sample shape and observable region. B-Type Thermocouple Sample Fig. 3 Experimental setup of image furnace and enlarged view of sample holder. Table 2 Intermediate compounds used in temperature calibration. Melting Point/K Reference Measured Value 2CaO7Al 2 O CaOSiO MgOTiO binary oxide systems listed in Table 2. Although the temperature difference between the thermocouple and standard samples was significant due to the steep temperature distribution, reproducibility of their values was as good as within 5 K. The focal depth of the CSLM was so narrow that it was difficult to observe the wide surface area of the crushed samples. Moreover, the sample shape varied with the change in the liquid fraction, making it difficult to continue with the focus adjustment on the sample surface. Furthermore, it was not advisable to process the samples for a long time at room temperature due to their hygroscopic and CO 2 absorbent properties. In order to solve these issues, the samples were quickly crushed in air and then remelted on the platinum plate in the image furnace to create droplets (Fig. 4), followed by quenching, and finally, equilibrium observation was performed. An example of the CSLM images obtained for a sample during the observation is shown in Fig. 5. When the temperature was increased to T, the liquid phase appeared, and thereafter, the liquid fraction increased with the temperature. When the temperature reached T 2, the solid phase disappeared. Here, T and T 2 were determined as the liquid formation temperature and liquidus temperature, respectively. The measurement of temperatures during cooling was considered difficult because supercooling and/or vitrification might have caused failures, following equilibrium solidification. In order to accurately measure the temperatures, the samples were required to exist in their equilibrium crystalline phases before liquid formation, although almost all the samples existed in a glassy phase after remelting. Hence, the samples were heated up to 5 K below the liquid formation temperature for 5 min. This temperature was maintained until the variation in transparency and morphology became static, and then it was again increased at a rate of 3 K/min and both the temperatures were determined. Subsequently, the samples were quenched to room temperature by switching off the image furnace. This procedure was repeated several times. Except during the first measurement, the temperature increase rate near the liquid formation temperature and liquidus temperature was maintained at approximately K/min to obtain more accurate measurement values. 2.3 Identification of equilibrium crystalline phases in molten oxide Since the transition temperatures were obtained using the CSLM in this research, the equilibrium crystalline phases and the compositions of the solid solutions were investigated through XRD. When a solid solution is formed, its composition can be roughly estimated by studying the XRD peak shifts caused by the variations in a lattice parameter. Using this method, an SrO substitution ratio x ¼ SrO=ðCaO þ SrOÞ was obtained in this experiment. Generally, samples obtained after quenching do not always reflect their states before quenching. However, it was assumed that there was little variation in the state because the crystal morphology did not vary during quenching and most of the liquid phase remained as glassy phase in this experiment. For XRD measurement, the CaAl 2 O 4 -SrAl 2 O 4 pseudobinary oxide system prepared from reagent grade chemicals such as Al 2 O 3, CaCO 3, and SrCO 3 was used to prepare the standard samples. The samples were mixed well, pressed at 3.8 MPa, calcined at 373 K for 8 h, and then sintered for 8 h. The compositions of the standard samples and the sintering temperatures are listed in Table 3. The sintered lumps were grounded to powder and subjected to XRD measurement.

4 Evaluation of Phase Diagrams for the Al 2 O 3 -CaO-SrO System by In-Situ Observation Using Confocal Laser Microscope K 73 K T 9 K 93 K T 2 Fig. 5 CSLM images of a sample during heating. Table 3 Compositions and sintering temperatures of standard samples for XRD measurement. Sample Al 2 O 3 CaO SrO x Sintering Temp./K B B B B B B Liquidus Temp. Liquid Formation Temp. Standard Ca 2 Al 4 O 33 Existing Phase Diagram 3. Result and Discussion 3. Temperature measurement with CSLM 3.. The CaO-Al 2 O 3 binary system The fundamental oxide of the CaO-Al 2 O 3 binary system was investigated. The measured temperatures agreed rather well with the literature value (Fig. 6). The error in three times measurements for different sample was approximately 5 K. Therefore, the validity of the application of a CSLM in an insitu observation for determining the phase diagrams was confirmed The Ca 2 Al 4 O 33 -Sr 2 Al 4 O 33 and CaAl 2 O 4 - SrAl 2 O 4 cross sections Next, the measurement of the samples containing SrO, where the solid solutions existed, was performed. Figure 7 illustrates the result of the temperature measurement using the Ca 2 Al 4 O 33 -Sr 2 Al 4 O 33 cross section. Both the liquidus temperature and liquid formation temperature exhibited everincreasing curves with increasing SrO contents, while a ternary eutectic point was considered to exist in the previous Fig. 6 Result of temperature measurement with CSLM for the CaO-Al 2 O 3 binary system. study. 7,8) This was probably because SrO strongly absorbs water and CO 2 from the ambient, and this made it difficult to obtain an accurate phase diagram, as the liquidus temperature decreased due to the existence of hydroxide or carbonate. A liquidus line consists of three different curves corresponding to different crystalline phases. A liquid formation temperature curve consists of two different curves and one line. A solid solution with a variable composition would relate to the variable region, and two solid solutions and a liquid phase would cause the flat region.

5 258 T. Kuroki, Y. Saito, T. Matsui and K. Morita.4.4 Fig. 7 Result of temperature measurement with CSLM for the Ca 2 Al 4 O 33 -Sr 2 Al 4 O 33 pseudo-binary system. Fig. 8 Result of temperature measurement with CSLM for the CaAl 2 O 4 - SrAl 2 O 4 pseudo-binary system. (a) (b) Fig. 9 (a) XRD patterns and (b) main peak shift of CaAl 2 O 4 solid solution. In the CaAl 2 O 4 -SrAl 2 O 4 pseudo-binary system, both the temperatures for the samples with a high SrO content could not be measured due to the high stability of solid phases. Lens-like curves were clarified only for samples with a low SrO content (Fig. 8) Other compositions Two patterns of temperature curves were observed. For samples that contained 5% or more CaO, the temperature curves demonstrated a lens-like shape. In the other samples, the temperature curves consisted of several different curves, similar to those of the Ca 2 Al 4 O 33 -Sr 2 Al 4 O 33 cross section. 3.2 XRD measurement Figure 9 presents the XRD patterns of the standard samples. Two types of solid solutions existed in this system one was the SrO substitute of CaAl 2 O 4 (-phase) and the other was the CaO substitute of SrAl 2 O 4 (-phase). Considering the measured temperature of the CaAl 2 O 4 - SrAl 2 O 4 pseudo-binary system and the result of the high temperature XRD measurement at 573 K in the previous study, ) this system was not evaluated as a complete solid solution, but a simple peritectic system, as shown in Fig.. Temperature /K Fig. Predicted phase diagram for the CaAl 2 O 4 -SrAl 2 O 4 pseudo-binary system. The XRD patterns of the other samples were also measured. The samples were maintained at 973 K and quenched and subjected to XRD (Fig. ). Only the -phase was detected in the Ca 2 Al 4 O 33 -based samples and only the -phase in the samples near the CaAl 2 O 4 -SrAl 2 O 4 line. The compositions of the standard samples were calculated from their peak shifts, as shown in Table 4.

6 Evaluation of Phase Diagrams for the Al 2 O 3 -CaO-SrO System by In-Situ Observation Using Confocal Laser Microscope 259 Fig. 2 Liquidus and tie lines drawn in the Al 2 O 3 -CaO-SrO ternary system at 973 K. Fig. XRD patterns of some samples after CSLM measurement. Table 4 Compositions of crystalline phases in samples maintained at 973 K, as calculated from XRD peak shifts..4 Sample Average Composition A6 A7 A8 E35! Crystalline Phase -phase x ¼ :75:9 -phase x: Prediction of phase diagram for the Al 2 O 3 -CaO- SrO ternary system 3.3. The isothermal cross section at 973 K The liquidus at 973 K was predicted as described below with Fig. 2. () Liquidus at 973 K was drawn considering the liquidus temperature measured by CSLM. (2) From the cross section of the CaAl 2 O 4 -SrAl 2 O 4 pseudo-binary system (Fig. ), regions of the solid solutions, and, were determined. (3) From XRD results of the quenehed samples in two phase region in Fig. 2, solid phase was identified as and its composition was determined using Fig.. (4) Similarly, a certain composition of phase was identified for the sample E35 by XRD measurement and Fig.. Accordingly, the isothermal cross section at 973 K around Ca 2 Al 4 O 33 was predicted as shown in Fig. 3. In the existing phase diagram, the liquidus line at 973 K consists of Fig. 3 Isothermal cross section at 973 K for the Al 2 O 3 -CaO-SrO ternary system. two curves that correspond to the (CaO, SrO)Al 2 O 3 solid solution and (CaO, SrO)2Al 2 O 3 solid solution. However, in this study, the liquidus line consists of three curves that correspond to the -phase, -phase, and (CaO, SrO)2Al 2 O 3 solid solution The phase diagram for the Al 2 O 3 -CaO-SrO ternary system Using the entire measurement data set of the CSLM, the ternary phase diagram was predicted. Firstly, the liquidus temperature near Ca 2 Al 4 O 33 was considerably higher than that observed in the previous study. In a liquidus surface (Fig. 4), at least three primary crystal surfaces were considered to exist. Regions I and II were considered to correspond to the -phase and -phase. On the other hand, in the Fig. 5, there existed three regions. Just above the regions I and III, the liquid phase was considered to be relatively equilibrated with the -phase and -phase with a variable

7 26 T. Kuroki, Y. Saito, T. Matsui and K. Morita Temperature /K I II III Temperature /K II III I Fig. 4 Liquidus surface revealed by CSLM measurement. SrO content, and region II was considered to correspond to the lines that cross the Gibbs triangle. In this manner, the transition temperatures corresponding to a liquid phase were detected accurately enough to discern a phase diagram even for the existence of unclear solid solutions via an in-situ observation using a CSLM. 4. Conclusions Equilibrium measurement was performed in this study via an in-situ observation using a CSLM. The results are summarized as follows: () The images during the melting process of the Al 2 O 3 - CaO-SrO ternary oxide system were obtained up to 973 K via the CSLM in-situ observation. (2) The accuracy and validity of the evaluation of liquidus temperature and liquid formation temperature were confirmed via the CSLM in-situ observation. (3) The temperature measurement with the CSLM required less time to achieve the equilibrium due to the small size of samples, as compared with a conventional method. Fig. 5 (4) The Al 2 O 3 -CaO-SrO ternary phase diagram and its isothermal cross section at 973 K were evaluated. REFERENCES Liquid formation surface revealed by CSLM measurement. ) H. Kurata, A. Takahashi, K. Yokohama, M. Oyamada and T. Takamatsu: Acta histochem. et cytochemica 22 (997) ) N. J. McDonald and S. Sridhar: Proc. Austenite Formation and Decomposition Conf. (23) pp ) W. Yan, M. Valdez and S. Sridhar: Z. Matallkunde 93 (22) ) Y. Kang: Doctoral Thesis of Royal Institute of Technology (Stockholm, Sweden, 26). 5) T. Y. Tai, T. Masusawa and H. T. Lee: Mater. Trans. 48 (27) ) H. Kimura, S. Endo, K. Yajima and F. Tsukihashi: ISIJ Int (24) ) F. Masazza: La Chimica e L Industria XLI (959) ) F. Masazza and E. Sirchia: Ann. Chim. (Rome) 49 (959) ) F. Guirado, S. Gal and S. Chinchon: Cem. Concr. Res. 3 (2) 23. ) H. Yamamoto and T. Matsuzawa: J. Lumin. 72/74 (997) 287. ) S. Ito, S. Sakano, K. Suzuki and M. Inagaki: Yogyo-Kyokai-Shi 87 (979)