journal - Wiley Online Library

J. Am. Ceram. Soc., 87 [9] 1706 –1713 (2004)

journal

Neutron Thermodiffractometry Study of Calcium Zirconate/Magnesium Oxide Formation in the ZrO2–CaO–MgO System Sara Serena, Angel Caballero, and M. Antonia Sainz Instituto de Cera´mica y Vidrio, C.S.I.C Campus de Cantoblanco, 28049, Madrid, Spain

Pierre Convert and Javier Campo Institut Laue Langevin, 38042 Grenoble, France

Xavier Turrillas Instituto de Ciencias de la Construccioń Eduardo Torroja, C.S.I.C., 28032 Madrid, Spain The reaction process to obtain CaZrO3/MgO specimens in the ZrO2–CaO–MgO system from two mixtures of natural and synthetic raw materials were analyzed by neutron thermodiffractometry; differential thermal analysis–thermogravimetry (DTA-TG) also was used. The results from the different techniques provided a complete description of the decomposition and reaction process in the samples and proved that high-temperature neutron diffraction is a powerful tool for analyzing in situ reactions up to 1250°C in both mixtures. In the present work, the evolution of the crystalline phases, the mechanism and temperature of formation of CaZrO3, and the activation energy were properly established in terms of the neutron thermodiffraction study. I.

Previous studies on the kinetics of formation of CaZrO3 by solid-state reaction17–19 have established that the reaction mechanism between t-ZrO2 and CaO is diffusional.17,18 When reaction occurs between m-ZrO2 and CaO, the mechanism is diffusional for temperatures ⬍1170°C, and at ⬃1170°C, the kinetics are mixed, resulting in both diffusion and nucleation mechanisms.19 However, the techniques used in these previous works do not allow an in situ description of the process, so that possible intermediate phases involved in the reaction have not yet been clearly identified. The pyrolysis behavior of dolomite, which is an attractive alternative for the production of low-cost, high-service materials, and its reaction with m-ZrO2 at temperatures ⬍1400°C were studied earlier20 using high-temperature X-ray powder diffractometry. In that study, despite in situ observation, the reactions were followed in terms of the diffraction patterns for mixed powders. In addition, overlap between the diffraction patterns of CaZrO3 and the raw materials used made the temperature of first formation of CaZrO3 difficult to establish. Recent DTA-TG21 studies have suggested that CaZrO3 formation in samples of dolomite and m-ZrO2 occurs by the direct reaction of CaCO3 and m-ZrO2. In the present work, the mechanism of formation of CaZrO3/ MgO materials from two mixtures, m-ZrO2 and dolomite and m-ZrO2, CaCO3 and magnesium hydroxicarbonate, was established by neutron thermodiffractometry and DTA-TG studies. The first advantage of using neutron diffraction on these materials is related to its high penetration capabilities, which allow examination of the bulk of the material. Experimental data collected from reaction in a bulk sample are truly representative, because the diffraction signal comes from the whole specimen and not the surface layer alone. The intensity of the neutron source used for measurement provided a time resolution of ⬃1 min, enough to follow the solid-state reactions. On the other hand, the high scattering length of light atoms increased the sensitivity of magnesium detection under this technique. In this context, neutron thermodiffractometry was used to study the evolution of crystalline phases in the sample during thermal treatment, to establish the mechanism and temperature of formation of CaZrO3, as well as the activation energy of the process. The evolution of the samples also was followed by differential thermal analysis–thermogravimetry (DTA-TG), to obtain an initial description of the process. Comparison between the results collected from the different techniques provided a detailed description of the decomposition/reaction process.

Introduction

T

high-temperature melting point of the subsystem MgO– CaZrO31,2 and the good bonding between the MgO and CaZrO3 phases,3,4 combined with the properties of CaZrO3,5– 8 permit the application of CaZrO3/MgO materials in different areas that have since advanced to include traditional ceramic materials. The characteristic machinability and quasi-plasticity behavior of CaZrO3/MgO nanocomposites has been pointed out recently.9 The good properties of these materials, e.g., high-temperature stability, good erosion/corrosion resistance, and high specific surface area, also have prompted the study of porous composites based on CaZrO3/MgO for applications as hot-gas-filter materials with good structural stability10 or fluid flow filters (for industrial wastes, beverages, etc.), lightweight structural components, hightemperature insulators, or membrane supports.11 Because of the excellent gas sensitivity of the CaZrO3 phase doped with In2O3,12 the application of porous composites of CaZrO3/MgO as selective methane sensors also has been reported recently.13 On the other hand, the compatibility of both phases with the solid phases of cement clinker3 at high temperatures allows the development of refractory materials for cement kilns.4,14 Refractory materials of MgO bonded with CaZrO3 are a real alternative to replace the magnesia/chromium-based refractories presently forbidden by environmental regulations.15,16 HE

B. Derby—contributing editor

II.

Experimental Procedure

The starting materials were high-purity powders of m-ZrO2 (TZ-0, 99.9 wt%, Tosoh Corp., Tokyo, Japan) with an average size d50 ⬍ 0.87 ␮m; high-purity dolomite (MgCa(CO3)2; Micro15,

Manuscript No. 10320. Received July 7, 2003; approved March 16, 2004. This work received financial support from the CICYT, Spain, under Project No. MAT-2000-0941 and from CAM Project No. CAM 07N/0038/2001.

1706

September 2004

Calcium Zirconate/Magnesium Oxide Formation in the ZrO2–CaO–MgO System

99.9 wt%, Prodomasa, Ma´ laga, Spain) with d50 ⬍ 4.87 ␮m; pure CaCO3 (⬎99 wt%, Fluka Chemical Corp., Ronkonkoma, NY) with d50 ⬍ 4 ␮m; and pure magnesium hydroxycarbonate (⬎99 wt%, Merck, Darmstadt, Germany) with d50 ⬍ 1 ␮m. Two samples with equimolar ZrO2:CaO:MgO ratios of 1:1:1 were formulated from mixtures of m-ZrO2, CaCO3, and magnesium hydroxycarbonate (denoted “ZC”) and m-ZrO2 with dolomite (denoted “ZD”). The proposed processing route was as follows: The powders of both compositions were first homogenized, by attrition milling in isopropyl alcohol media for 2 h, using ZrO2 balls as milling media; dried at 60°C; and then sieved through a 60 ␮m screen and cold isostatically pressed at 200 MPa, to obtain green compacts. DTA-TG was performed (Model STA 409, Netzsch-Gera¨ tebau GmbH, Selb, Germany) at a 5°C/min constant heating rate up to 1500°C, in air, using open platinum crucibles. Neutron powder diffraction data were collected using two-axis diffractometer equipment (Model D1B; ␭ ⫽ 2.522 Å) at the Laue-Langevin Institute (ILL; Grenoble, France). To conduct the experiment, an ellipsoid mirror furnace,22 developed by ILL, was placed in the diffractometer. Specimens 5 mm in diameter and 10 mm high were introduced in platinum crucibles covered with platinum lids, to ensure uniform heat distribution in the whole sample. Heating was conducted in air, in two steps: (i) from room temperature to 700°C at 10°C/min and then (ii) from 700° to 1250°C at 2°C/min. A hole in the sidewall of the crucible allowed the introduction of another thermocouple (Pt/Pt–10% Rd) into the center of the specimen. Temperatures were recorded using the thermocouple introduced inside the sample, and the temperature stability was within 10°C over 1000°C. Simultaneously, neutron diffraction patterns were collected over the 2␪ range of 2°– 81° every 150 s during the whole experiment. The diffraction data collected were represented as a sequence of patterns in a pseudo-three-dimensional diagram, resulting in an I(2␪,T) plot, with the help of the commercial package NOESYS.23 To identify the phases involved in the reaction process and determine the temperature ranges in which each phase was present, the contour projection of I(2␪,T) on the 2␪–T plane was studied. The reaction was monitored by following the diffraction intensities of the selected peaks. To obtain the variation with temperature of the integrated intensities for selected Bragg reflections of the crystalline phases, a quantitative study was performed, using the Simpson rule and subtracting the background before and after the peak. The peaks used in the quantification were selected by their intensity and their separation from the characteristic peaks of any other crystalline phase. III.

Results

(1) DTA-TG Study Figures 1(a) and (b) show the DTA-TG curves for the ZC and ZD samples, respectively. In the ZC sample (Fig. 1(a)), peaks associated with weight loss corresponding to the decomposition of magnesium hydroxycarbonate appear in the range 240°–570°C. The first endothermic peak, at 310°C, corresponds to the loss of constitutional water; the second endothermic peak, at 440°C, is attributable to the (OH)⫺ group; and the third exothermic peak, at 490°C, is characteristic of the decomposition processes of the hydroxycarbonates.24,25 A peak at 540°C, resulting from the decarbonation reaction, also is shown. These results agree with the following decomposition reaction for magnesium hydroxycarbonate: Mg5共CO3兲4共OH兲2䡠4共H2O兲共s兲 7 5MgO共s兲 ⫹ 5H2O共 g兲 ⫹ 4CO2共 g兲

(1)

The consequent weight loss observed at 860°C is attributed to the decomposition reaction of the CaCO3,26 according to the well-known reaction CaCO3共s兲 7 CaO共s兲 ⫹ CO2共 g兲

(2)

In the ZD sample (Fig. 1(b)), only two endothermic peaks associated with weight loss occur. The first peak, detected at

1707

Fig. 1. (a) DTA-TG curves for ZC sample. (b) DTA-TG curves for ZD sample.

750°C, corresponds to decomposition of the dolomite. The second endothermic peak, at 880°C, is attributable to the decomposition reaction of the CaCO3. The weight losses are consistent with a two-step decomposition reaction of dolomite: MgCa共CO3兲2共s兲 7 MgO共s兲 ⫹ CaCO3共s兲 ⫹ CO2共 g兲

(3a)

CaCO3共s兲 7 CaO共s兲 ⫹ CO2共 g兲

(3b)

The decomposition of dolomite, as revealed by DTA-TG analysis, has been widely reported under various atmospheric conditions,27,28 and Fig. 1(b) shows the expected behavior for decomposition in air. The total weight loss for sample ZC was ⬃30.6% up to 910°C, whereas the ZD sample exhibited a total loss weight of ⬃28.8% up to 940°C. The DTA curves of the samples revealed no evidence of endothermic or exothermic effects other than those mentioned. (2) Neutron Thermodiffraction Studies The evolution of the reactions in samples ZC and ZD can be qualitatively analyzed by taking into account the contour projection of the neutron thermodiffractometry results, I(2␪,T), on the 2␪–T plane; the results for the two samples, ZC and ZD, are shown in Figs. 2 and 3, respectively. The domains of existence for the different phases (Figs. 2 and 3) show that the decomposition/reaction process in both samples occurred in sequential stages. On the other hand, it is important to remark the displacement of certain diffraction peaks of some of the phases with increasing temperature. This effect was especially pronounced in the diffraction peaks corresponding to the (104) and (006) reflections of CaCO3, the (200) reflection of MgO, and the (112) reflection of CaZrO3 in the ZC sample (Fig. 2) and the (006) reflection of dolomite, the (104) reflection of CaCO3, the (200)

1708

Journal of the American Ceramic Society—Serena et al.

Fig. 2.

Sequence of neutron diffraction patterns for ZC sample as a function of temperature I(2␪,T).

reflection of MgO, and the (112) reflection of CaZrO3 in the ZD sample (Fig. 3). (A) Effect of Temperature on the Crystalline Structure of the Phases: Before discussing the reaction process, it is important to note that neutron thermodiffraction study shows not only the phase changes resulting from the reaction process but also the effect of temperature on their crystalline structure. To analyze this effect, it is very useful to calculate initially the diffraction pattern of the crystalline phase as a function of temperature. In the present paper, use of the lattice parameters and agitation factors of the atoms of the different phases is proposed as a means of obtaining a dynamic vision of the evolution of the neutron diffraction patterns of the crystalline phases with temperature. The indicated crystallographic parameters are calculated in terms of temperature by interpolating data collected from the literature. The dynamic pattern of the crystalline phases is generated using the commercial package Crystallographica.29 The lattice parameters and agitation factors determined by Markgraf and Reeder30 for the CaCO3 phase can be used to obtain the

Fig. 3.

Vol. 87, No. 9

dynamic diffraction pattern for this phase, as shown in Fig. 4. Figure 4 reveals the displacement of the reflections and the decrease in their intensities caused by increasing temperature. The differences among the displacements observed for each reflection are explained by taking into account the expansion coefficient behavior for the various sorts of crystal bonds, which results in a mean thermal expansion coefficient of ⫺2.8 ⫻ 10⫺6/°C along the a-axis and 32.3 ⫻ 10⫺6/°C along the c-axis. The decrease in intensity is basically a consequence of the increasing thermal motion of the oxygen atoms with increasing temperature. The dynamic diffraction pattern calculated for the CaCO3 shows that the (012) reflection is the least affected by temperature and the (006) reflection the most dramatically modified and provides a description of the evolution of intensities as a function of temperature. Comparing Fig. 4 with the contour projection, I(2␪,T), of both samples (Figs. 2 and 3) shows that the experimentally observed displacements in the characteristic CaCO3 reflections are caused principally by the effect of temperature on the CaCO3 crystalline structure and are in good agreement with the calculated behavior.

Sequence of neutron diffraction patterns for ZD sample as a function of temperature I(2␪,T).

September 2004


Fig. 4. Dynamical neutron diffraction pattern for CaCO3, calculated from the reported lattice parameters and agitation factors.

The same study has been conducted for MgCa(CO3)2,31 and the experimentally observed displacements in the characteristic reflections were in agreement with those calculated in the dynamic diffraction pattern for this phase. The displacements observed in the reflections were caused mainly by temperature and are not attributable to the formation of solid solutions in the crystalline phases alone. (B) Reaction Process for ZC and ZD Samples: The evolution of the reaction process in both samples was studied from the integrated intensity of selected peaks calculated for each phase as a function of temperature. ZC Sample: Fig. 5 shows the results obtained for the ZC sample. The peaks selected for quantitative analysis of the different phases were the (221) for magnesium hydroxycarbonate, the (104) for Ca(CO)3, the (111) for m-ZrO2, the (200) for MgO, the (200) for CaO, and the (101) for CaZrO3. As shown in Fig. 5, the different reaction processes can be properly analyzed and described by three stages: (i) decomposition of the Mg5(CO3)4 (OH)2䡠4H2O and formation of MgO; (ii) decomposition of the CaCO3 and formation of CaO; and (iii) formation of CaZrO3. No other crystalline phase was detected in the samples during thermal treatment. The different stages are described as follows: (i) Decomposition of the magnesium hydroxycarbonate: The decomposition reaction began with the breakdown of the magnesium hydroxycarbonate, at 350°–500°C (Fig. 5). The formation of MgO was first noticeable at 500°C, and its intensity increased steeply up to a temperature of 1000°C. At this stage, no intermediate crystalline phase constituted by magnesium, carbon, and oxygen was present.

Fig. 5. Evolution of integrated intensities for selected reflections of the phases presents in the sample ZC along the thermal treatment. MgCOH is refereed to the phase Mg5(CO3)4(OH)2䡠4H2O.

1709

(ii) Decomposition of the CaCO3: The integrated intensity of the (104) reflection of the CaCO3 phase decreased continuously and slightly at temperatures ⬍675°C, but a significant decrease in intensity occurred between 675° and 800°C. The observed behavior of the intensity of CaCO3 at lower temperatures was expected, considering the dynamic pattern calculated earlier in the present work (Fig. 4). Figure 6(a) shows the calculated intensities for the (104) reflection, along with the experimentally observed intensities. The agreement between the calculated and experimental intensities for this reflection proves that the initial decrease resulted from the effect of temperature on the crystal structure of the CaCO3. At temperatures ⬎680°C, the behavior of the experimental intensities contradicted that of the calculated intensities, making it possible to deduce the beginning of the decomposition reaction of CaCO3 at this temperature and to establish that the breakdown of CaCO3 occurred between 680° and 800°C. CaO was produced rapidly above 680°C, and its highest intensity was observed at 775°C. (iii) Formation of CaZrO3 (reaction between the m-ZrO2 and the CaO): The sequential diffactograms in Fig. 7 clearly indicate the presence of CaO before the appearance of CaZrO3 in the sample. The beginning of the reaction between CaO and m-ZrO2 was established at temperatures ⬎680°C, taking into account the decrease in intensity of m-ZrO2 (Fig. 5). CaZrO3 was first detected at 730°C, after the decomposition of CaCO3 had begun. This means that the formation of CaZrO3 occurs after a short period of induction. These results clearly show that CaZrO3 formation occurred according to the following reaction: m-ZrO2共s兲 ⫹ CaO共s兲 f CaZrO3共s兲

(4)

The intensity of the CaZrO3 continued to increase in the sample after the decomposition of CaCO3 had ended, verifying that reaction (4) occurred. No evidence of diffraction peaks corresponding to CaO and m-ZrO2 phases was found at temperatures ⬎1010°C, and a change in the slope of the curve of the integrated

Fig. 6. Integrated intensity for the reflection (104) of the CaCO3 and reflection (104) of MgCa(CO3)2. Squares and cross represent experimental values for CaCO3 and MgCa(CO3)2, respectively, and lines the resulting of the calculates dynamical patterns.

1710


Fig. 7. Sequence of neutron diffraction pattern on heating the ZC sample: (䡺) CaCO3, (‚) m-ZrO2, (f) CaO, (E) MgO, (Œ) CaZrO3.

intensity of CaZrO3 at 1025°C was observed. Both results indicated the end of the reaction. ZD Sample: A parallel study was conducted for the ZD sample. The variations with temperature of the integrated intensities of the crystalline phases for the ZD sample are shown in Fig. 8. The peaks selected for the semiquantitative analysis of the different phases were the (104) for dolomite, the (104) for CaCO3, the (111) for m-ZrO2, the (200) for MgO, the (200) for CaO, and the (101) for CaZrO3. A detailed analysis of the process of decomposition and reaction can be obtained from the analysis of Fig. 8. The different processes can be described in two stages: (i) decomposition of the dolomite and formation of MgO and CaCO3, decomposition of the CaCO3 and formation of CaO; and (ii) formation of CaZrO3. No other crystalline phase was detected by neutron thermodiffraction. The different stages of the process are described as follows: (i) Decomposition of the dolomite: Figure 8 shows a significant decrease of the integrated intensity of the dolomite at temperatures ⬎500°C. However, the integrated intensity also shows a slight decrease at temperatures ⬎450°C. Phases resulting from the decomposition, MgO and CaCO3, were noticed at temperatures ⬎500°C. The decrease in the intensity of dolomite at lower temperatures was predicted by the dynamic pattern calculated for the dolomite, as shown in Fig. 6(b), which represents the calculated intensity for the (104) reflection and the experimentally observed intensity for

Fig. 8. Evolution of integrated intensities for selected reflections of the phases presents in the sample ZD along the thermal treatment.

Vol. 87, No. 9

that peak. At temperatures ⱖ450°C, the decrease observed experimentally was higher than that expected to be caused by thermal agitation; thus, it can be concluded that dolomite began to decompose at this temperature. No evidence of dolomite appeared in the sample at temperatures ⬎725°C; therefore, decomposition had ended. The integrated intensity of the MgO increased up to 725°C, after which a marked diminution in the slope of the curve was observed (Fig. 8), indicating that MgO formation had ended. The integrated intensity of the CaCO3 phase increased up to 680°C. CaCO3 breakdown occurred between 680° and 870°C, with CaO first noticed at 680°C and its highest intensity detected at 840°C. (ii) Formation of CaZrO3 (reaction between the m-ZrO2 and the CaO): As in the case of the ZC sample, the sequential diffractograms of the ZD sample in the temperature range of 643°–739°C clearly show the presence of CaO before the appearance of CaZrO3 in the sample. The CaZrO3 formation began after the CaO phase appeared and before the decomposition of CaCO3 ended. The beginning of the reaction between CaO and m-ZrO2 can be deduced from Fig. 8. Figure 8 shows a decrease in the intensity of m-ZrO2 at temperatures ⬎680°C, coinciding with the appearance of CaO in the sample, and the beginning of the detection of CaZrO3 at 765°C. Again, these results show that the CaZrO3 formation occurred according to reaction (4). At temperatures ⬎1010°C, no diffraction peaks corresponding to the CaO and m-ZrO2 phases were detected; therefore, it can be concluded that reaction (4) was complete.

IV.

Discussion

(1) Decomposition of the Raw Material The kinetic behavior of metal carbonate decomposition is sensitive to, among other factors, the availability and ease of removal of CO2. Because the rate of escape of volatile product by intergranular and/or intragranular diffusion is controlled by the structure and porosity of the solid product and its relationship with the reactant phase,17 even the size of the specimens is relevant to the escape of CO2 and, thus, to the kinetics of the process. In general, the mechanism and temperatures of decomposition of carbonate phases are very sensitive to atmospheric conditions, especially the CO2 pressure.27,32,33 Extensive studies on the decomposition of dolomite were conducted by Haul and Heysyek34 and Otsuka,27 using DTA-TG, X-ray diffraction, and carbon isotope techniques. Those results showed that, at a CO2 pressure of ⬍260 mmHg, decomposition proceeds in a single step to yield CaO and MgO. When the partial pressure of CO2 increases to the range of 100 –760 mmHg, this phase decomposes in two steps. Recently, De Aza et al.28 used neutron thermodiffraction methods to study the evolution, in real time, of dolomite decomposition. They reported that the decomposition of dolomite occurs in two stages, in which MgO and CaCO3 appear at 550°C, and the CaCO3 decomposes at 950°C, to yield CaO. In the present work, the neutron diffractometry equipment used enabled us to obtain data under real conditions at normal atmospheric pressure. The results showed the decomposition of dolomite in two steps, in good agreement with previous studies. The temperature values obtained are in accord with those observed by Rodriguez et al.35 during their study of the reaction between dolomite and zircon, which was monitored by neutron thermodiffraction under experimental conditions equal to those of the present study. However, the temperature observed for CaCO3 decomposition in that study was 780°C, lower than that reported by De Aza et al.28 The temperature established for CaCO3 decomposition in the ZC sample of the present work (735°C) also was lower than that usually reported for this reaction in air (⬃850°C26,32,33). These discrepancies are related to the atmospheric conditions in the crucible during neutron thermodiffraction measurement. The experimental design of the crucible, with a hole

September 2004


on the side wall to introduce the second thermocouple, enabled escape of the CO2 gas phase produced during decomposition of the raw materials and prevented the increase of the CO2 pressure, PCO2, inside the crucible. The present neutron thermodiffraction results are in good agreement with the decomposition process observed by DTA-TG, but the decomposition temperatures established by DTA-TG are higher than those determined by neutron thermodiffractometry. Again, the discrepancies are justified by taking into account the differences in PCO2 in the crucibles during the experiment. Therefore, other factors contributed to the mentioned differences, such as the higher rate of heating used during the DTA-TG registration, which caused decomposition to occur at higher temperatures.36 On the other hand, this discrepancy can also be explained by taking into account differences in the two methods of measurement: In the case of neutron diffraction, the amount of crystalline phase is directly measured, whereas DTA-TG analysis is based on the measurement of the amount of gaseous phase that leaves the sample; thus, in the course of the reaction (reactions (1), (2), and (4) above), breakdown of the carbonate crystalline phases could occur before the escape of the gaseous CO2 generated during the reaction. (2) Formation of MgO Figures 5 and 8 show that the two respective samples exhibit different behaviors in the formation of MgO. These differences can be explained in terms of the raw materials used. In the case of the MgO obtained from Mg5(CO3)4(OH)2䡠4H2O, the complete decomposition of this phase (reaction (3)) is a complex process that occurs in sequential stages, involving reactions of dehydration and dehydroxylation.25,37,38 These reactions result in the formation of intermediate phases, often amorphous or pseudoamorphous, and certain anhydrous initial products, which undergo subsequent recrystallization. All of these results cause the observed delay in the formation of the final stable oxide resulting from the decomposition of Mg5(CO3)4(OH)2䡠4H2O, i.e., MgO. However, in the case of the decomposition of the dolomite, it has been observed previously20,28 that the “crystallization” of MgO occurs quickly. (3) Formation of CaZrO3 The DTA curves for both samples lack evidence of endothermic or exothermic effects that could be attributed to the formation of CaZrO3; thus, direct observation of the DTA-TG analysis was not decisive in this respect. On the other hand, the results obtained by neutron thermodiffraction offer a detailed and conclusive description of the reaction process. The following reaction has been proposed recently21 for the formation of CaZrO3: CaCO3共s兲 ⫹ m-ZrO2共s兲 3 CaZrO3共s兲 ⫹ CO2共 g兲

1711

(2) and (5) and the exothermic effect associated with reaction (4). The values of ⌬G show that reactions (2) and (5) are thermodynamically favorable at temperatures higher than 900° and 600°C, respectively, whereas reaction (4) is thermodynamically favorable at all temperatures. If the DTA-TG curve alone is considered, the endothermic peaks at 860°C in ZC and at 880°C in ZD can be attributed to reaction (5), leading to the conclusion that CaZrO3 is formed from CaCO3 and m-ZrO2, as proposed by Suzuki et al.21 However, in the present study, neutron thermodiffraction revealed the presence of a CaO phase in both samples before the formation of CaZrO3, and no CaZrO3 was observed before the decomposition of the CaCO3. Therefore, the formation of CaZrO3 from CaO has been established according to reaction (4). Nevertheless, the neutron thermodiffraction results show a certain overlap between reactions (2) and (4) in the samples, so that the marked endothermic character of reaction (2) might be preventing the detection of exothermic reaction (4) in the DTA curves. The presence of isolated characteristic reflections of CaZrO3 in the neutron diffractograms makes it possible to establish the temperature of first formation of this phase. The high reactivity of the CaO resulting from the decomposition of the CaCO3 causes the reaction between m-ZrO2 and CaO, to form CaZrO3, to begin at temperatures as low as 730°C for ZC and 765°C for ZD. No t-ZrO2 stabilized with CaO or MgO is revealed by neutron diffraction; thus, in the reaction between ZrO2 and CaO, the ZrO2 remains in the monoclinic phase, although the temperature of formation of t-ZrO2 is ⬃930°C40 in the ZrO2–CaO system and 1130°C40 in the ZrO2–MgO system. This finding reveals that the activation energy for the formation of CaZrO3 is lower than that for the stabilization of t-ZrO2, confirming the data given by Sekkina.41 (4) Kinetics of CaZrO3 Formation The integrated intensities for CaZrO3 can be used to conduct a dynamic study of its kinetics26,42 of formation. Using the Ginstling model43 for the solid-state reaction controlled by diffusion,44 the relation between the advance of the reaction (fraction of CaZrO3 formed) and the temperature can be expressed as f共␣兲 ⫽

冋

册

ART 2 ␤RT ⫺E/RT 1⫺ e 2 ␤E E

(6)

Here, ␣ is the advance of the reaction, ␤ the heating rate, E the activation energy, A the frequency factor, R the Boltzmann constant, and T the temperature (K); the expression of f(␣) according to the Ginstling43 model is 1 ⫺ 2/3␣ ⫺ (1 ⫺ ␣)2/3.

(5)

Table I shows the changes in enthalpy and Gibbs free energy39 associated with the possible reactions involved in the formation of CaZrO3 in the ZC and ZD samples. The ⌬H value observed for the three reactions establishes the endothermic character of reactions

Table I. ⌬H (kJ/mol) and ⌬G (kJ/mol) for the Indicated Reactions at Temperatures from 500° to 1000°C† Reactions CaCO3 3 CaO ⫹ CO2 T (°C)

500 600 700 800 900 1000 †

⌬H

⌬G

173.596 53.275 171.987 37.811 170.263 22.539 168.431 7.449 166.499 ⫺7.466 164.470 ⫺22.211

CaO ⫹ m-ZrO2 3 CaZrO3

CaCO3 ⫹ m-ZrO2 3 CaZrO3 ⫹ CO2

⌬H

⌬G

⌬H

⌬G

⫺33.472 ⫺33.472 ⫺33.472 ⫺33.472 ⫺33.472 ⫺33.472

⫺36.267 ⫺36.629 ⫺36.990 ⫺37.351 ⫺37.713 ⫺38.074

140.124 138.515 136.791 134.959 133.027 130.999

17.008 1.183 ⫺14.451 ⫺29.903 ⫺45.179 ⫺60.285

Data obtained from SGTE Substances Database.39

Fig. 9. Plot to obtain the activation energy from the slope. Kinetic fitted to Gistling model. Circles and triangles represent experimental values for CaZrO3 phase in ZC and ZD, respectively.

1712


Figure 9 shows a representation of F(␣) ⫽ ln [f(␣)] - 2 ln (T) ⫹ ln (␤) versus 1/T for the reaction of formation of CaZrO3 in the ZC and ZD samples. For both samples, the curves represent two ranges of different behavior, at temperatures of 800°C for ZC and 840°C for ZD. There is a satisfactory agreement with the kinetic model at the higher temperatures, but not at the lower temperatures. The values calculated for the activation energy for the formation of CaZrO3 at higher temperatures are 28.9 ⫾ 1.5 kcal/mol in ZC and 27.6 ⫾ 1.5 kcal/mol in ZD. The change in the behavior of the curves may be related to the finalization of the decomposition reaction of CaCO3. The higher slope of the curves for lower temperatures points to higher values of activation energy for the formation of CaZrO3 in the temperature ranges in which this process overlaps the decomposition of CaCO3. In this range of temperatures, the decomposition/reaction process is very complex and does not obey a kinetic model fully controlled by diffusion. After the decomposition of CaCO3 has ended, the diffusion of Ca2⫹ through the formed CaZrO3 controls the kinetics of the process.17,18 There are no differences in the mechanism of formation of CaZrO3 for the two samples, but the formation of CaZrO3 occurs at a lower temperature in the ZC sample than in the ZD sample, because the decomposition of CaCO3 occurs earlier in the ZC than in the ZD. The results calculated in the present work can be compared with those reported by Oprea,19 because the processing of the samples in that study was similar to our method. However, the activation energies obtained in the present study were lower than those reported by Oprea. This difference can be attributed mainly to the smaller particle size of the powder obtained by attrition milling and to the higher isostatic pressure applied to obtain the green compacts during processing in the present work, in comparison with the method used by Oprea.19 Both effects contributed considerably to decreasing the activation energy of CaZrO3 formation in the samples. V.

Conclusions

Analysis by DTA-TG of the reaction process for a mixture of m-ZrO2, CaCO3, and Mg5(CO3)4(OH)2䡠4H2O (ZC) and a mixture of m-ZrO2 and MgCa(CO3)2 (ZD) provides an initial description of the decomposition process of the raw materials but allows no conclusions about the formation of CaZrO3 and MgO. However, neutron thermodiffraction has permitted, for the first time, a complete description of the mechanism of formation of CaZrO3. The reaction mechanism in m-ZrO2, CaCO3, Mg5(CO3)4(OH)2䡠4H2O and m-ZrO2, MgCa(CO3)2 mixtures up to 1250°C was established by an in situ neutron thermodiffraction technique. The results obtained demonstrated that reaction occurs by decomposition and subsequent reaction. Under the atmospheric conditions of the present neutron diffraction experiment, the decomposition of dolomite was shown to occur at 500°–725°C and the resulting CaCO3 to decompose at 725°– 870°C. The breakdown of calcite also was detected at temperatures ⬃150°C lower than those reported for its decomposition in air. The first formation of CaZrO3 occurred in the presence of CaO, during the decomposition of CaCO3 by reaction between m-ZrO2 and CaO at 730° and 765°C in the ZC and ZD samples, respectively. The activation energy for this process was determined from neutron thermodiffraction results. The present study proved that in situ high-temperature neutron diffraction analysis is a powerful tool for tracking phase assemblage in the preparation of CaZrO3/MgO composites and, in general, for establishing the reaction mechanism in these ceramic materials. References 1

Y. Du, Z. Jin, and P. Huang, “Calculation of the ZrO2–CaO–MgO System,” CALPHAD: Comput. Coupling Phase Diagrams Thermochem., 16 [3] 221–30 (1992). 2 Y. Yin and B. B. Argent, “Phase Diagrams and Thermodynamics of the ZrO2–CaO–MgO and CaO–MgO Systems,” J. Phase Equilib., 14 [5] 588 – 600 (1993).

3

Vol. 87, No. 9

S. De Aza, C. Richmond, and J. White, “Compatibility Relationships of Periclase in the System CaO–MgO–ZrO2–SiO2,” Trans. J. Br. Ceram. Soc., 73 [4] 109 –16 (1974). 4 H. Kozuka, Y. Kaita, Y. Tuchina, T. Honda, and S. Otha, “New Kind of Chrome-Free Bricks for Burning Zone of Rotary Cement Kiln”; pp. 1027–37 in Unitecr-93 Congress, Sao Paulo, Brazil. Asociacio´ n Latinoamericana de Fabricantes de Refractarios, Brazil, 1993. 5 A. De Pretis, F. Ricciardiello, and O. Sbaizero, “Mechanical Behaviour of Various Polycrystalline CaZrO3 Materials,” Mater. Sci. Eng., 71, 166 (1985). 6 T. I. Hou and W. M. Kriven, “Mechanical Properties and Microstructure of Ca2SiO4–CaZrO3 Composites,” J. Am. Ceram. Soc., 77 [1] 65–72 (1994). 7 K. Kobayashi, S. Yamaguchi, and Y. Iguchi, “Electrical Transport Properties of Calcium Zirconate at High Temperature,” Solid State Ionics, 108, 355– 62 (1998). 8 S. Yamaguchi, K. Kobayashi, T. Higuchi, S. Shin, and Y. Iguchi, “Electronic Transport Properties and Electronic Structure of InO1.5-Doped CaZrO3,” Solid State Ionics, 136 –137, 305–11 (2000). 9 Y. Suzuki, P. E. D. Morgan, and T. Ohji, “Synthesis and Deformation Behavior of Nano-Diphasic Materials from Natural Dolomite,” Mater. Sci. Eng. A, 304 –306, 780 – 84 (2001). 10 Y. Suzuki, P. E. D. Morgan, and T. Ohji, “New Uniformly Porous CaZrO3/MgO Composites with Three-Dimensional Network Structure from Natural Dolomite,” J. Am. Ceram. Soc., 83 [8] 2091–93 (2000). 11 N. Corbitt, “Inorganic Membranes: Markets, Technologies, Players,” Business Communication Co., Norwalk, CT, 1997. 12 Y. Chiang, C. C. Wang, and S. A. Akbar, “Calcium Zirconate for the Monitoring of Hydrocarbons,” Sens. Actuators B, 46, 208 –12 (1998). 13 Y. Suziki, M. Awano, N. Kondo, and T. Ohji, “CH4-Sensing and HighTemperature Mechanical Properties of Porous CaZrO3/MgO Composites with ThreeDimensional Network Structure,” J. Ceram. Soc. Jpn., 109 [1] 79 – 81 (2001). 14 H. Kozuka, Y. Kaita, K. Tokunaga, K. Sakakibara, and S. Ohta, “Further Improvements of MgO–CaO–ZrO2 Bricks for Burning Zone of Rotary Cement Kiln”; pp. 256 – 63 in Unitecr. 95, Kyoto, Japan. Technical Association of Refractories, Tokyo, Japan, 1995. 15 P. Bartha and H. J. Klischat, “Present State of the Refractory Lining for Cement Kilns,” CN-Refractories, 6 [3] 31–38 (1999.) 16 Y. Guotian and X. Yanquing, “Refractories for Cement Rotary Kilns,” China’s Refract., 10 [1] 20 –26 (2001). 17 Y. Angers, R. Tremblay, and A. C. D. Chaklader, “Formation of CaZrO3 by Solid-State Reaction between CaO and ZrO2,” J. Am. Ceram. Soc., 55 [8] 425–26 (1972). 18 Y. Angers, R. Tremblay, and A. C. D. Chaklader, “Effect of Pressure on Formation of CaZrO3 by Solid-State Reaction between CaO and ZrO2,” J. Am. Ceram. Soc., 57 [5] 231–32 (1974). 19 G. Oprea, “Solid State Reaction Kinetics for the Calcium Zirconate Synthesis,” Solid State Phenom., 25–26, 125–32 (1992). 20 Y. Suzuki, P. E. D. Morgan, T. Sekino, and K. Niihara, “Manufacturing Nano-Diphasic Materials from Natural Dolomite: In Situ Observation of Nanophase Formation Behaviour,” J. Am. Ceram. Soc., 80 [11] 2949 –53 (1997). 21 Y. Suzuki, P. E. D. Morgan, and T. Ohji, “Processing of Porous CaZrO3/MgO/Pt Composites via In Situ Reactions,” Ceram. Eng. Sci. Proc., 22, 217–23 (2002). 22 Z. Mursic, T. Vogt, H. Boysen, and F. Frey, “Diffraction Study of Metamict Zircon up to 2000 K,” J. Appl. Crystallogr., 25, 519 –23 (1992). 23 Program NOESYS Version 1.2, Fortner Software LLC, Sterling, VA. 24 R. C. MacKenzie, Differential Thermal Analysis, Vol. I, Fundamental Aspects. Academic Press, London, U.K., 1970 –1972. 25 Y. Sawada, J. Yamaguchi, O. Sakurai, K. Uematsu, N. Mizutani, and M. Kato “Isothermal Differential Scanning Calorimetry on an Exothermic Phenomenon during Thermal Decomposition of Hydromagnesite 4MgCO3Mg(OH)2䡠4H2O,” Thermochim. Acta, 34 [2] 233–37 (1979). 26 C. H. Bamford and C. F. H. Tipper, Chemical Kinetics, Vol, 22, Reactions in the Solid State. Elsevier Scientific Publishing Co., Amsterdam, Holland, 1980. 27 R. Otsuka, “Recent Studies on the Decomposition of the Dolomite Group by Thermal Analysis,” Thermochim. Acta, 100, 69 – 80 (1986). 28 A. H. de Aza, M. A. Rodriguez, J. L. Rodriguez, S. de Aza, P. Pena, P. Convert, T. Hansen, and X. Turrillas, “Decomposition of Dolomite Monitored by Neutron Thermodiffractometry,” J. Am. Ceram. Soc., 85 [4] 881– 88 (2002). 29 Crystallographica v1.51, Oxford Cryosystems 1995–99. 30 S. A. Markgraf and R. J. Reeder, “High Temperature Structure Refinements of Calcite and Magnesite,” Am. Mineral., 70, 590 – 600 (1985). 31 S. A. Markgraf and R. J. Reeder, “High-Temperature Crystal Chemistry of Dolomite,” Am. Mineral., 71, 795– 804 (1986). 32 I. Carrizosa, J. M. Criado, F. Gonza´ lez Garcıá, and M. Gonza´lez, “Influencia de las Condiciones Experimentales en la Cine´tica de la Reaccioń de Descomposicioń Te´rmica del CaCO3,” Bol. Soc. Esp. Ceram. Vidrio, 17 [1] 23–27 (1978). 33 B. V. L’vov, “Mechanism and Kinetics of Thermal Decomposition of Carbonates,” Thermochim. Acta, 386, 1–16 (2002). 34 R. A. W. Haul and H. Heysyek, “Differential Thermal Analysis of the Dolomite Decomposition,” Am. Mineral., 37, 166 –79 (1952). 35 J. L. Rodriguez, A. H. de Aza, P. Pena, J. Campo, P. Convert, and X. Turrillas “Study of Zircon–Dolomite Reactions Monitored by Neutron Thermodiffractometry,” J. Solid State Chem., 166, 426 –33 (2002). 36 J. P. Sanders and P. K. Gallagher, “Kinetic Analyses Using Simultaneous TG/DSC Measurements. Part I: Decomposition of Calcium Carbonate in Argon,” Thermochim. Acta, 388, 115–28 (2002). 37 Y. Sawada, J. Yamaguchi, O. Sakurai, K. Uematsu, N. Mizutani, and M. Kato, “Isothermal Differential Scanning Calorimetry on an Exothermic Phenomenon during Thermal Decomposition of Hydromagnesite 4MgCO3Mg(OH)2䡠4H2O,” Thermochim. Acta, 34 [2] 233–37 (1979).

September 2004 38


H. Be´ arat, M. J. McKelvy, A. V. G. Chizmeshya, R. Sharma, and R. W. Carpenter, “Magnesium Dehydroxilation/Carbonation Reaction Processes: Implications for Carbon Dioxide Mineral Sequestration,” J. Am. Ceram. Soc., 85 [4] 742– 48 (2002). 39 SGTE Substance database, Thermo-Calc software AB version P (1995–2003), Foundation of Computational Thermodynamics, Stockholm, Sweden. 40 Y. Yin and B. B. Argent, “Phase Diagrams and Thermodynamics of the System ZrO2–CaO and ZrO2–MgO,” J. Phase Equilib., 14 [4] 439 –50 (1993). 41 M. M. Abou Sekkina, “Realization about the Solid State Reaction of Zirconia

1713

with Calcia throught the Production of Best Quality Highly Refractory Oxide,” Indian Ceram., 30 [6] 113–17 (1987). 42 A. W. Coast and J. P. Redfern, “Kinetic Parameters from Thermogravimetric Data,” Nature (London), 201 [1] 68 – 69 (1964). 43 A. M. Ginstling and B. I. Brounshtein, “Concerning the Diffusion Kinetics of Reactions in Spherical Particles,” J. Appl. Chem. U.S.S.R. (Engl. Transl.), 23, 1327 (1950). 44 S. F. Hulbert, “Models for Solid-State Reactions in Powdered Compacts: A Review,” Trans. J. Br. Ceram. Soc., 6, 5–20 (1969). 䡺