Fig. 1 Particle-size distribution of Al(OH)3 ((○) volume and (●) cumulative volume).

Cumulative volume (%)

Volume (%)

Cumulative volume (%)

Volume (%)

Particle size (µm)

Particle size (µm) Fig. 2 Particle-size distribution of kaolinite gangue ((○) volume and (●) cumulative volume).

Pore-Size Distribution and Strength of Porous Mullite Ceramics MgCO3 additive has only a slight influence on open porosity but significant effect on pore-size distribution, average pore size and crushing strength of mullite ceramics made from Al(OH)3 and kaolinite gangue. Wen Yan and Nan Li

Porous ceramics are finding increasing applications as catalyst carriers, hot-gas collectors, molten-metal filters, separation membranes, water cleaners, bioceramics and thermally insulated materials.1–6 Usually, a porous structure can be achieved by a conventional powder-processing route with the incorporation of some pore-forming agents, such as sawdust, starch, graphite or organic particulates,7 or by injection molding8 or gelcasting.9 Deng and co-workers10,11 made porous alumina ceramics by decomposition of Al(OH)3 to form pores insitu. This pore-forming in-situ technique is a good way to prepare porous ceramics with well-distributed pores. Moreover, these porous ceramics are environment friendly, because they do not produce carbon oxides. Mullite has a low coefficient of thermal expansion and good thermal-shock resistance as well as excellent mechanical and chemical stability. In our earlier papers,12,13 we used the in-situ decomposition pore-forming technique to prepare porous corundum–mullite ceramics. We used Al(OH)3 and kaolinite gangue or Al(OH)3 and microsilica as raw materials. However, the strength of the porous ceramics needed to be improved. MgO has been reported as a good sintering aid for fused mullite.14–16 This means that introduction of MgO into the porous ceramic prepared from Al(OH)3 and kaolinite gangue may promote sintering ability and improve mechanical properties. Therefore, we studied the results of introducing MgCO3 into a powder mixture of Al(OH)3 and kaolinite gangue to prepare a mullite porous ceramic.

9401

American Ceramic Society Bulletin, Vol. 85, No. 12

Porous Mullite Ceramics

Preparation and Analysis

Weight (%) Temperature (°C)

Fig. 3 DSC and TGA curves of Al(OH)3.

Weight (%)

Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) analysis of the Al(OH)3 and kaolinite gangue were performed (Model STA 449C, Netzsch, Bayern, Germany) (Figs. 3 and 4). These analyses were conducted in air. The heating rate of the powders from room temperature to 1500°C was 10°C/min. A total weight loss of ~33% of Al(OH)3 powder was measured, which resulted from the decomposition of Al(OH)3 at 300°C (Fig. 3). A total weight loss of ~13% of kaolinite gangue powder was measured, which resulted from the burning of carbon and organic substance in gangue and dehydration of kaolinite at temperature of 450–650°C (Fig. 4). The weight loss ratio of Al(OH)3 was more pronounced.

Heat flow (W/g)

The starting raw materials were Al(OH)3, kaolinite gangue and MgCO3 (Table 1). The particle-size distributions of Al(OH)3 and kaolinite gangue were measured using laser particle size analysis (Matersizer 2000) (Figs.1 and 2). The average particle sizes of the Al(OH)3 and kaolinite gangue were 54.3 and 42.3 µm, respectively. The average particle size of MgCO3 was 16.2 µm.

Heat flow (W/g)

Raw Materials Analysis

Powder mixtures consisted of 54.3 mass% Al(OH)3 and 45.7 mass% kaolinite gangue, which was consistent with the mullite stoichiometric ratio Al2O3:SiO2. The amount of MgCO3 added to the starting powders varied from 0 to 3.51 mass%, which corresponded to 0 Temperature (°C) to 1.5 mass% MgO added (Table 2). The above starting powders were mixed for 4 h in polyurethane pots Fig. 4 DSC and TGA curves of kaolinite gangue. using alumina balls. The particle-size distribution of milled powder A was measured using laser particle-size analysis (Matersizer 2000) (Fig. 5). The average particle size of powder A was 30.8 µm. The milled powders were pressed in cylinders with height of 36 mm and diameter of 36 mm at a pressure of ~100 MPa. The green compacts were dried at 110°C and then heated at 1500°C for 180 min in an electric furnace. Phase analysis was conducted using X-ray diffractometry (XRD; Model Xpert TMP, Philips, Eindhoven, Netherlands) with a scanning speed of 2 deg/min. Relative content of mullite in the sintered specimens was obtained using X’PERT HIGHSORE (V.1.0a; Philips Analytical B.V.). The glass-phase content in the specimens was measured using HF acid that had a concentration of 40 wt% and particles that were from 150 to 180 µm in size. The density and porosity of the calcined specimens were measured using the Archimedes principle with water as medium. The pore-size distribution was measured using mercury porosimetry (Model AutoPore IV 9500, Micromeritics Instrument Corp., Norcross, Ga.). The crushing strength of sintered specimens at room temperature was measured. The morphology and pore structures of these specimens were observed using scanning electron microscopy (SEM; Model XL 30, Philips). Generally, three specimens were tested, and an average value of the results was taken.

Phase Identification XRD patterns of specimens sintered at 1500°C for 3 h were prepared to identify the formed phases (Fig. 6). In specimen A, the phases were mullite (3Al2O3·2SiO2), corundum (3Al2O3) and cristobalite (SiO2). However, for specimens B, C and D, the cristobalite phase disappeared, the corundum-phase peaks 9402

American Ceramic Society Bulletin, Vol. 85, No. 12

Porous Mullite Ceramics decreased and the mullitephase peaks increased.

Table 1 Chemical Compositions of Al(OH)3, Kaolinite Gangue and MgCO3

The relative content of mullite in the sintered speciComponent Al2O3 mens was determined (Fig. 7). The relative content of mul64.86 Al(OH)3 lite increased with increased Kaolinite gangue 36.48 amount of MgCO3 added from MgCO3 0.19 0 to 2.93 mass% but decreased slightly when the amount of MgCO3 added was further increased to 3.51 mass%.

Composition (mass%) SiO2

Fe2O3

CaO

MgO

K 2O

Na2O

LOI

0.08

0.06

0.04

0.02

0.04

0.01

34.44

44.52 0.05

0.21 0.02

0.15 0.04

0.15 42.69

0.084

0.028

17.89 57.06

Kaolinite undergoes a series of reactions during heating. At temperatures 1400°C, secondary mullite formation takes place rapidly by the dissolution of alumina into a transitory liquid-silica phase, followed by the precipitation of mullite crystals.19 The rate of secondary mullite formation is slow until eutectic liquid formation at 1587°C and is extremely fast at 1600°C.18 This means that the liquid-phase content is important in the formation of mullite. Specimen A, which was sintered at 1500°C for 180 min without MgCO3 additives, contained cristobalite (Fig. 6). However, when only 1.76 mass% MgCO3 was added (specimen B), which corresponds to ~0.75 mass% MgO, the cristobalite disappeared. When the amount of MgCO3 added was increased from 0 to 1.76 mass%, the mullite content increased rapidly, but when the amount of MgCO3 added was increased from 1.76 to 2.93 mass% (Fig.7), the mullite content increased slowly. This means that the liquid content formed was with addition of 1.76 mass% MgCO3. When the amount of MgCO3 added was increased to 3.51 mass%, the mullite content decreased slightly (Fig.7), and a small amount of MgO·Al2O3 spinel formed (Fig. 6). The glassy content of specimen C is 6.35 mass%, which is not significantly high. This level would not have a strong effect on the properties of the materials at elevated temperatures.

Pore Characterization

Particle size (µm) Fig. 5 Particle-size distribution of mixture powder A ((○) volume and (●) cumulative volume).

American Ceramic Society Bulletin, Vol. 85, No. 12

Intensity

Volume (%)

Cumulative (%)

The apparent porosity and average pore size versus the amount of MgCO3 added to the starting powders were determined (Fig. 8). The porosity decreased slightly from 42.3 to 39.8% when the amount of MgCO3 added was increased from 0 to 1.76 mass%. It increased slightly from 39.8 to 41.4% when the amount of MgCO3 added was further increased to 3.51%. This means that the MgCO3 added had a slight influence on porosity. Average pore diameter increased from 0.326 to 3.535 µm when the amount of MgCO3 added was increased from 0 to 2.93 mass%. However, it decreased slightly from 3.535 to 3.421 µm when the amount of MgCO3 added was further increased from 2.93 to 3.51 mass%.

2θ (deg) Fig. 6. XRD pattern of specimens sintered at 1500°C for 3 h ((○) corundum, (▲) mullite, (■) cristobalite and (☼) spinel).

9403

Porous Mullite Ceramics The pore-size distributions of the sintered specimens prepared from various starting powders were determined (Figs. 9–12). The multipeak mode of pore-size distribution (MUPD) observed (Fig. 9) for specimens without MgCO3 additive consisted of three peaks. A bipeak modal of pore-size distribution (BIPD) was observed in specimen B, which contained 1.76 mass% MgCO3 added to the starting powders (Fig. 10). When the amount of MgCO3 added was 2.93 and 3.51 mass%, a monopeak mode of pore-size distribution (MOPD) was observed (Figs. 11 and 12). There was little difference between the pore-size distribution curves of specimens C and D. Table 2 Amount of MgCO3 Added in Starting Powders

MgCO3

A (mass%)

B (mass%)

C (mass%)

D (mass%)

0

1.76

2.93

3.51

Curves of cumulative porous volume (%) versus pore diameter were plotted (Fig. 13). The pore size shifted toward higher values with increased amount of MgCO3 added. However, the cumulative pore-size distribution curves of specimen C and D were almost the same. The d50 values of the pores were 0.31, 2.2, 3.5 and 3.5 µm for specimens A, B, C and D, respectively.

The proportion of micropores (pore diameter ≤0.45 µm) in the specimens was determined (Fig. 14). The ratio of micropore volume to total-pore volume in the sintered specimens decreased when the amount of MgCO3 added increased from 0 to 2.93 mass%. However, it changed slightly when the amount was increased from 2.93 to 3.51 mass%, which was in good agreement with other trends observed (Figs. 9–12).

• One type was what we called “primary pores.” They were in the pseudomorphs of Al(OH)3 and kaolinite gangue grains. Their size was 0.45 µm (peak No. 2 in Fig. 9). We did not find a third peak in our earlier work. This difference between the two works may result from the difference of Al(OH)3 and kaolinite gangue particle sizes used in the two research projects.

Relative content of mullite (mass %)

The microstructures of the specimens with various amounts of MgCO3 added were studied to determine changes of pore distribution (Figs. 15–18). The microstructures of the specimens varied. In our earlier papers,12,13 we reported that there were two types of pores in the specimens made from Al(OH)3 and kaolinite gangue:

The presence of primary pores depends on the presence and integrity of pseudomorphs. In specimen A, Amount MgCO3 added in the starting powders (mass%) pseudomorphs were more clear, had better integrity and seemed more dense (Fig.15), but the apparent porosity Fig. 7 Relative content of mullite in sintered specimens. of this specimen was the highest. This meant that there were some primary pores in the pseudomorphs. When the amount of MgCO3 added was increased to 1.76 mass%, the liquid phase formed improved the formation of mullite. The expansion resulted from mullite formation that filled in the primary pores when the expansion was small and destroyed the pseudomorphs when the expansion was bigger. On the other hand, the sintering of mullite crystallites in the pseudomorphs after mullite formation became less important and decreased the volume of primary pores in the pseudomorphs (Figs. 10 and 14) and the pseudomorph size (Fig. 16). When the amount of MgCO3 added was increased to 2.93 mass%, the liquid content was high enough to improve mullite formation and sintering and to lead to disappearance of primary pore. Then monopeak modal pore-size distribution appeared. The slight increase of apparent porosity and pore sizes of specimen D may have resulted from a small amount of spinel formation.

Compressive Strength The relationship between the amount of MgCO3 added to the starting powders and crushing strength of porous mullite ceramic was determined (Fig. 19). The strength increased with increased amount of MgCO3

9404

American Ceramic Society Bulletin, Vol. 85, No. 12

Average pore size (µm)

added from 0 to 2.93 mass%. However, it decreased slightly when the amount increased from 2.93 to 3.51 mass%. When the amount of MgCO3 added was 2.93 mass%, the specimens achieved the highest strength of 52 MPa at a porosity of ~40%. Fortunately, the porosities of specimens B and C were similar, but the crushing strength of the latter was higher than that of the former. This means an appropriate amount of MgO addition may improve strength without decreasing porosity.

Apparent porosity (vol %)

Porous Mullite Ceramics

There are three reasons for the increase of strength: • The formation of a well-developed neck. The liquid phase improves formation and development of neck during sintering.

Amount of MgCO3 added in the starting powders (mass %) Fig. 8 Apparent porosity and average pore size vs MgCO3 added (%) in the starting powders ((□) apparent porosity and (●) average pore size).Fig. 3 DSC and TGA curves of Al(OH)3.

• The well-distributed pore size. The pore-size distribution is a single and sharp peak distribution (Fig. 11), which corresponds to a homogeneous and integrated solid net in the microstructure of pore ceramics. This is favorable to the strength of porous ceramics. • The strength of the mullite grains. The liquid phase improves formation and sintering of mullite crystallites in the pseudomorphs and decreases primary pore porosity to improve the strength of grains. The slight decrease of strength of specimen D may result from the formation of spinel, which weakens the neck among grains and increases of porosity in the porous ceramics. I

About the Authors Dr. Wen Yan ([email protected]) and Prof. Nan Li ([email protected]) are faculty members of the Hubei Province Key Laboratory of Refractories and Ceramics, Wuhan University of Science & Technology, Wuhan, Peoples’ Republic of China. Correspondence should be directed to Dr. Yan.

American Ceramic Society Bulletin, Vol. 85, No. 12

9405

Porous Mullite Ceramics References 1S.M.

Yang, N. Coombs and G.A. Ozin, “Micromolding in Inverted Polymer Opals (MIPO): Synthesis of Hexagonal Mesoporous Silica Opals,” Adv. Mater., 12 [24] 1940–44 (2000).

2J.B.

Davis, A. Kristoffersson, et al., “Fabrication and Crack Deflection in Ceramic Laminates with Porous Interlayers,” J. Am. Ceram. Soc., 83, 2369–74 (2000).

3P.A.

Vityaz, V.M. Shelekhina, et al., “Development of the Porous Substrates of Ceramic Membrane Filters,” J. Eng. Phys. Thermophys., 77 [4] 797–801 (2004).

4N.O.

Engin and A.C. Tas, “Manufacture of Macroporous Calcium Hydroxyapatite Bioceramics,” J. Eur. Ceram. Soc., 19 [13–14] 2569–72 (1999). 5Y.M. Jo, R.B. Hutchison and J.A. Raper, “Characterization of Ceramic Composite Membrane Filters for Hot Gas Cleaning,” Powder Technol., 91 [1] 55–62 (1997). 6J.

Coronas, M. Menendez and J. Santamaria, “Methane Oxidative Coupling Using Porous Ceramic Membrane Reactors—II. Reaction Studies,” Chem. Eng. Sci., 49 [12] 2015–25 (1994).

7J.H.

She, et al., “Fabrication and Characterization of Highly Porous Mullite Ceramics,” Mater. Chem. Phys., 80, 610–14 (2003).

8Z.

Xie, et al., “Study on Binder Removal Process of Ceramics Injection Molding,” Bull. Chin. Ceram. Soc., 2, 18–21 (1998). 9Y.F.

Liu, et al., “Porous Mullite Ceramics from National Clay Produced by Gelcasting,” Ceram. Int., 27, 1–7 (2001).

10Z.Y.

Deng, T. Fukasawa, et al., “Microstructure and Mechanical Properties of Porous Alumina Ceramics Fabricated by the Decomposition of Aluminum Hydroxide,” J. Am. Ceram. Soc., 84 [11] 2638–44 (2001). 11Z.Y.

Deng, T. Fukasawa, et al., “High-Surface-Area Alumina Ceramics Fabricated by the Decomposition of Al(OH)3,” J. Am. Ceram. Soc., 84 [3] 485–91 (2001). 12S. Li and N. Li, “Effects of Composition and Temperature on Porosity and Pore Size Distribution of Porous Ceramics Prepared from Al(OH)3 and Kaolinite Gangue,” Ceram. Int., in press; corrected proof available online 3 Mar 2006. 13S.

Li and N. Li, “Influences of Composition of Starting Powders and Sintering Temperature on the Pore Size Distribution of Porous Corundum–Mullite Ceramics,” Sci. Sinter., 37, 173–80 (2005).

14L.

Montanaro, C. Perrot, et al., “Sintering of Industrial Mullites in the Presence of Magnesia as a Sintering Aid,” J. Am. Ceram. Soc., 83, 189–96 (2000).

15L.

Montanaro, J.M. Tulliani, et al., “Sintering of Industrial Mullite,” J. Eur. Ceram. Soc., 17, 1715–23 (1997).

16M.G.M.U.

Ismail, H. Tsunatori and Z. Nakai, “Preparation of MgO-Doped Mullite by Sol–Gel Method, Powder Characteristics and Sintering,” J. Mater. Sci., 25, 2619–25 (1990).

17A.K. Chakraborty and D.K. Ghosh, “Reexamination of the Kaolinite-to-Mullite Reaction Series,” J. Am. Ceram. Soc., 61, 170–73 (1978). 18K.C.

Liu and G. Thomas, “Time–Temperature–Transformation Curves for Kaolinite–α-Alumina,” J. Am. Ceram. Soc., 77, 545–52 (1994). 19H.R. Rezaie, W.M. Rainforth and W.E. Lee, “Mullite Evolution in Ceramics Derived from Kaolinite, Kaolinite with Added α-Alumina and Sol–Gel Precursors,” Trans. Br. Ceram. Soc., 96, 181–87 (1997).

9406

American Ceramic Society Bulletin, Vol. 85, No. 12

log differential intrusion (mL/g)

log differential intrusion (mL/g)

Porous Mullite Ceramics

Pore diameter (µm)

Fig. 10 Pore-size distribution of specimen B.

log differential intrusion (mL/g)

log differential intrusion (mL/g)

Fig. 9 Pore-size distribution of specimen A.

Pore diameter (µm)

Pore diameter (µm)

Pore diameter (µm)

Fig. 11 Pore-size distribution of specimen C.

Fig. 12 Pore-size distribution of specimen D.

Cumulative volume (%)

Porous Mullite Ceramics

Pore diameter (µm)

Micropore volume (%)

Fig. 13 Cumulative porous volume (%) as a function of pore size (specimen (●) A, (▲) B, (▼) C and (♦) D).

Amount of MgCO3 added in the starting powders (mass%) Fig. 14 Volume amount of micropores (%) vs amount of MgCO3 added (mass%) in the starting powders.

Porous Mullite Ceramics

Fig. 15 SEM photograph of specimen A.

Fig. 18 SEM photograph of specimen D.

Compressive strength (MPa)

Fig. 17 SEM photograph of specimen C.

Fig. 16 SEM photograph of specimen B.

Amount of MgCO3 added in the starting powders (mass%) Fig. 19 Compressive strength vs MgCO3 added (mass%) in the starting powders.