Abstract
Alumina-based composites were fabricated by reaction sintering from two different sintering powder mixtures: alumina with silica (SiO2) and alumina with silicon carbide (SiC; to allow oxidation to form SiO2). After sintering, SiO2 underwent complete reaction to form alumina/mullite composites. In terms of microstructure, the density and grain size of ceramic samples were investigated. The density of the composites prepared by alumina and SiC was lower than those of alumina and the composites prepared by alumina and SiO2. The grain size increased as the sintering temperature increased. In terms of mechanical properties, fracture surfaces, hardness, and fracture toughness were investigated. It was found that the fracture surface of alumina was rather intergranular, whereas the fracture surface of the composites was more transgranular. The hardness of the composites was higher than that of alumina at the same sintering temperature. However, the fracture toughness of the composites was not significantly different compared to that of alumina.
1 Introduction
The most common method to improve the mechanical properties of ceramics is to incorporate second-phase particle into the based ceramic, so-called ceramic matrix composites (CMCs). Alumina-based composites are widely used in many applications because alumina has high melting point, hardness, wear, and corrosion resistance. Second or reinforcement phase added to alumina results in an increase of fracture toughness, hardness, and strength [1, 2], which can be non-oxide [e.g. silicon carbide (SiC), Si3N4, and TiC] [1, 3, 4] or oxide (e.g. ZrO2 and TiO2) [5, 6]. When silica (SiO2) is added to alumina, the alumina/mullite composites can be formed due to the reaction between alumina and SiO2 at high temperatures to create mullite (3Al2O3·2SiO2) as the product.
There has been rather extensive research of alumina/mullite composites in terms of both processing and characterization. Reaction sintering is the most common method to fabricate alumina/mullite composites, and in this case, starting powders are required to produce ceramic green compact. The starting powders can be prepared by ball milling or sol-gel followed by calcination. She et al. [7] fabricated the composites using α-Al2O3, Al, and β-SiC as the starting materials and ZrO2 as the sintering aid. The powders were ball milled then uniaxially pressed and sintered in air so that SiC underwent oxidation to form SiO2, which can further react with alumina to form mullite. The composite with 15 vol% ZrO2 exhibited high density after sintering at 1580°C for 2 h with intergranular fracture mode and high strength. Sedaghat et al. [8] used the sol-gel method, where starting precursors were aluminum chloride haxahydrate (AlCl3·6H2O) and tetraethyl orthosilicate (TEOS). The composite powders were obtained by calcination at 900°C for 2 h followed by attrition milling, drying, and sieving. The resulting composites contained intergranular and intragranular mullite, and the alumina matrix grain size was decreased. Alumina/mullite composites can also be produced by infiltration technique. Marple and Green [9] prepared partially sintered alumina sample, and the sample was infiltrated with hydrolyzed ethyl silicate solution. The composites had only mullite and alumina, where the grain growth of alumina matrix was restricted by mullite. The Young’s modulus of the particular composites was lower than that of pure alumina, and fracture toughness was increased up to 60% in some samples [10].
In comparison, it can be seen that reaction sintering is the simplest route of the production. Furthermore, reaction sintering is more economical compared to other methods of ceramic processing and can also ensure the homogeneous distribution of the second phase (i.e. mullite) [11]. SiC is generally used as the starting powder to form SiO2 via oxidation. Sakka et al. [11] fabricated alumina-mullite-SiC nanocomposites from alumina and SiC powders. The SiC powder was oxidized at the surface and the SiO2 resulting from the oxidation reacted to alumina at the surface between Al2O3 and SiO2 to form mullite, leaving the unreacted alumina and remaining unoxidized SiC surrounded by mullite. The study and modeling of oxidation behavior of SiC in alumina-based CMCs was also carried out, and it was found that SiO2 and mullite formation depends mainly on the rate of oxidation (governed by sintering temperature), SiC geometry, and SiC particle size (i.e. surface area) [12, 13]. However, the fabrication of alumina/mullite composites from the mixture of alumina and SiO2 is also possible, and there is yet no study to compare the microstructure and mechanical properties of the composites from different starting powders.
In this study, alumina-based composites were fabricated by reaction sintering using alumina, SiO2, and SiC (to be oxidized to form SiO2) as the starting powders. The objective is to investigate the effect of sintering temperature on the microstructure and mechanical properties of alumina-based composites prepared by different starting powders with the same route of processing.
2 Materials and methods
The starting powders consist of Al2O3 (200 nm), SiC (3 μm), and SiO2 (0.007 μm). To fabricate alumina-based composites, two powder mixtures (Al2O3+4.1 wt% SiC and Al2O3+6 wt% SiO2) were used. The amount of SiC added (4.1 wt%) would be oxidized to 6 wt% SiO2, provided that SiC undergoes complete oxidation. The ceramic powders were mechanically mixed by ball milling using ethanol as dispersing media and zirconia balls as milling media. To prevent abnormal grain growth, 0.25 wt% MgO was also added in the powders. Ball milling was performed for 5 h and then the slurry was dried in an oven for 24 h. After drying, the powder was obtained by crushing in mortar and pestle and then sieving through a 150 μm sieve. Ceramic pellets were produced by uniaxial pressing at 150 MPa, and the ceramic samples were sintered in air at the temperatures between 1400°C and 1550°C for 30 min, where the heating and cooling rate was 5°C/min. The pure alumina samples are denoted as A, the alumina-based composites prepared from alumina and SiO2 powders are denoted as ASO, and the alumina-based composites prepared from alumina and SiC powders are denoted as ASC. The numbers of the samples are sintering temperatures, where 40, 45, 50, and 55 are 1400°C, 1450°C, 1500°C, and 1550°C, respectively.
The phase identification of the sintered samples was carried out using X-ray diffraction (XRD) technique. The scanning step of the diffractometer (Miniflex 600, Rigaku, Japan) was 0.05°, and 2θ was between 20° and 100°. The diffraction patterns were matched with patterns in the JCPDS file. The investigation of the microstructure and mechanical properties include density, grain size, fracture surface, hardness, and fracture toughness. Sample density was measured by the Archimedes method. The morphology of ceramic grain and fracture surface was examined by a scanning electron microscope (SEM-LV5910, JEOL, Japan). To measure the grain size and hardness of the ceramic samples, metallographic preparation for polishing was carried out as follows: grinding using 25 μm diamond suspension until flat and polishing using 6, 3, and 1 μm diamond suspension to the final surface finish. The samples were prepared for grain size determination by polishing and thermal etching at 50°C lower than the sintering temperature for 30 min. The linear interception method was used to calculate the grain size of alumina. The fracture surfaces were prepared by breaking the samples with no further surface treatment. The samples for SEM observation were sputtered with gold to prevent surface charging. The polished surfaces were subjected to hardness measurement using Vickers microhardness tester (Wolpert W Group, UK) with 1 kg load (9.8 N) and a dwell time of 10 s. Fracture toughness was calculated from the formula:
where KIc is fracture toughness, E is Young’s modulus, HV is Vickers hardness, P is indentor load, and c is radial crack length. The Young’s modulus values are obtained from Asmani et al. [14], where the porosity of the alumina samples is taken into account.
3 Results and discussion
The XRD patterns of alumina and alumina/mullite composites made from the two mixtures sintered at 1400°C are shown in Figure 1. It can be seen that the composites consisted of only alumina (JCPDS file no. 42-1468) and mullite (JCPDS file no. 79-1456) with no peaks of SiO2 or SiC. It can be concluded that mullite was completely formed by SiO2 (either from deliberate addition or SiO2 as a product of oxidation from SiC). In the case of SiC addition, a temperature of 1400°C or higher could result in complete oxidation for the SiC with a particle size of 3 μm because the oxidation of SiC started at approximately 820°C for micron-sized SiC [15].

XRD patterns of alumina and alumina/mullite composites at 1400°C (dots show the presence of mullite).
The density, grain size, hardness, and fracture toughness of all ceramic samples are summarized in Table 1. The graph of density of samples as a function of sintering temperature is shown in Figure 2. It can be seen that alumina (A) and alumina-based composites prepared by alumina and SiO2 mixture (ASO) had higher density as the sintering temperature was increased, whereas the final density of alumina-based composites prepared by ASC tended to be insensitive to sintering temperature. The density of alumina-based composites from alumina and SiO2 was slightly higher than that of alumina at the same sintering temperature (0.4–1.2%). The increase of final density of alumina due to an addition of SiO2 to form alumina/mullite composites is also reported by Svancarek et al. [16], where the addition of up to 10 wt% SiO2 could slightly improve the density of alumina samples. The density of the composites prepared by alumina and SiC was significantly lower than that of alumina at the same sintering temperature (2.6–6.2%). Luo et al. [17] reported the decrease in final density when alumina-based composites were prepared by the same method as this work. However, the decrease was quite small (~0.5%) and this was a lot smaller than this work. The possible explanation for the decrease of final density is that SiC before oxidation to SiO2 hinders the densification of alumina ceramics during sintering [18]. Also, oxidation of SiC results in CO2 as a product, which could create open pores [12] and thus reduce the final density of the samples fabricated by alumina and SiC as the starting powders.
Physical and mechanical properties of all samples.
| Temperature (°C) | Sample | Density (%) | G (μm) | HV (GPa) | KIc (MPam1/2) |
|---|---|---|---|---|---|
| 1400 | A40 | 95.0±0.1 | 0.7±0.1 | 9.4±1.0 | 2.7±0.1 |
| ASO40 | 96.2±0.1 | 0.6±0.1 | 12.0±0.8 | 2.7±0.1 | |
| ASC40 | 92.4±0.1 | 0.9±0.2 | 13.9±1.1 | 2.3±0.1 | |
| 1450 | A45 | 96.8±0.1 | 1.0±0.1 | 9.8±0.5 | 2.4±0.1 |
| ASO45 | 97.2±0.1 | 0.6±0.1 | 12.6±1.7 | 2.6±0.2 | |
| ASC45 | 91.0±0.1 | 1.3±0.2 | 17.5±1.3 | 2.2±0.1 | |
| 1500 | A50 | 97.4±0.1 | 1.1±0.2 | 15.3±2.2 | 2.4±0.2 |
| ASO50 | 98.1±0.1 | 3.9±0.1 | 17.4±1.7 | 2.4±0.2 | |
| ASC50 | 92.3±0.1 | 5.0±0.1 | 17.7±1.6 | 2.4±0.2 | |
| 1550 | A55 | 98.7±0.1 | 2.9±0.3 | 11.9±1.7 | 3.1±0.2 |
| ASO55 | 99.4±0.1 | 6.0±0.2 | 18.5±1.8 | 2.5±0.2 | |
| ASC55 | 92.5±0.1 | 3.6±0.4 | 16.6±1.7 | 2.5±0.2 |

Graph of density as a function of sintering temperature.
Figure 3 shows the graph of alumina grain size versus sintering temperature. It can be seen that the grain size increases as the sintering temperature is increased. All samples had a grain size of ~1 μm when the sintering temperature was 1400°C and 1450°C. At 1500°C, alumina (A50) had a grain size of 1.1 μm, but the alumina-based composites (ASO50 and ASC50) had larger grain size. The grain size of all samples was increased at the sintering temperature of 1550°C, where alumina (A55) had a smaller grain size than the alumina-based samples (ASO55 and ASC55). It should be noted that, at higher temperatures (i.e. 1500°C and 1550°C), alumina/mullite from alumina and SiC powders had larger grain size than pure alumina because micron-sized SiC particles before oxidation were not effective to inhibit grain growth from grain boundary pinning like nanosized SiC particles [19, 20]. Moreover, SiO2 in the powder mixture from both deliberate addition and oxidation could also act as a sintering aid and promote grain growth during sintering due to the formation of glassy thin films [21], resulting in the larger grain size and higher final density compared to pure alumina.

Graph of alumina grain size as a function of sintering temperature.
Figure 4 shows the micrographs of fracture surface of the alumina and alumina-based samples with similar grain size. Intergranular fracture is observed in alumina where facetted fracture surface is the characteristic. There is also a small area of transgranular fracture in alumina. In alumina-based composite, the fracture surface is rather transgranular with lesser extent of intergranular fracture compared to pure alumina. The morphology of alumina/mullite fracture surface is in good agreement to a previous report [22]. The change of fracture mode was caused by the presence of mullite from reaction sintering, which resulted in grain boundary strengthening [17]. From the micrographs, it can be seen that the fracture surface of ASO (Figure 4B) is more transgranular than ASC (Figure 4C). This could be due to the fact that mullite from SiO2 obtained by oxidation of SiC (particle size of 3 μm) had a larger size than that obtained by direct addition (0.007 μm), as previously reported by Burgos-Montes et al. [22], or the mullite in ASO was more intergranular than that in ASC [23].

Fracture surfaces of alumina and alumina/mullite composites at 1550°C: (A) alumina (A55), (B) alumina/mullite from alumina and SiO2 (ASO55), and (C) alumina/mullite from alumina and SiC (ASC55).
The arrow indicates intergranular fracture and the area in dotted circle indicates transgranular fracture.
The graph of hardness as a function of sintering temperature is shown in Figure 5. The hardness of A and ASO samples increased as the sintering temperature increased (except for the A50 sample, where the hardness was higher than that of the A55 sample). It can be seen that the increase of density plays an important role to improve the hardness of the ceramics [24]. Comparing at the same sintering temperature, the hardness of alumina was lower than those of alumina-based composites, which shows the same trend as in a previous report [16, 17]. However, alumina-based composites prepared from SiC starting powder (ASC) had higher hardness than that prepared from SiO2 starting powder (ASO). It could be due to the fact that the higher hardness resulted from the greater amount of mullite formed or the more homogeneous distribution of mullite in the ASC ceramic sample [16, 17]. Figure 6 shows the graph of fracture toughness versus sintering temperature. It can be seen that there is no significant difference in fracture toughness between alumina and alumina-based composites. The value of fracture toughness of alumina sintered at 1550°C is higher than the other samples because of the high value of (E/HV)1/2. In this work, the calculated volume fraction of mullite in the composites is ~14% (from the reaction sintering with 6 wt% SiO2 from direct addition or oxidation) and the result of hardness and fracture toughness can be compared with the report from Luo et al. [17] and Svancarek et al. [16]. The value of hardness is agreeable and the value of fracture toughness is slightly lower because the calculation in this work takes the effect of porosity into account [14].

Graph of hardness versus sintering temperature.

Graph of fracture toughness versus sintering temperature.
To summarize the results, the relationship between mechanical properties and microstructure is discussed in this paragraph. It can be seen that alumina-based composites could be obtained by ceramic processing from two different starting powder mixtures. The larger grain size of alumina-based composites compared to alumina, in most cases, at the same sintering temperature was caused by SiO2 from the direct addition or oxidation of SiC [21]. The ASC ceramic sample had lower density than alumina due to the delay of densification resulting from SiC before oxidation or formation of CO2 gas during oxidation, whereas the ASO ceramic sample had slightly higher density than alumina due to SiO2. The subsequent formation of mullite resulted in an increase of hardness and the change of fracture mode to transgranular. However, the mullite formation process in ASO and ASC was different, resulting in the difference of hardness and degree of transgranular fracture. When large SiC particles are used as starting powder for reaction sintering, mullite tends to form inside alumina matrix [11], whereas small SiO2 particles react with alumina to form mullite at grain boundaries [22]. As a result, mullite in ASC was distributed more evenly whereas mullite in ASO tended to be situated at the grain boundary. Therefore, the hardness of ASC was higher with less transgranular fracture compared to ASO. In this work, it can be seen that reaction sintering process is beneficial to the processing of alumina-based composites, where the particle size of starting powders can be used to tailor the position and distribution of mullite, which affect the mechanical properties (i.e. fracture mode and hardness).
4 Conclusions
The effect of sintering temperature on the microstructure and mechanical properties of alumina-based composites was studied. It was found that:
The density of pure alumina and alumina-based composites prepared by alumina with SiO2 increased with sintering temperature, but the density of the composites was not changed when the powder mixture was alumina with SiC.
The grain size increased with sintering temperature in both alumina and alumina-based composites.
The hardness of alumina-based composites was higher than that of alumina. Higher sintering temperature resulted in higher hardness because of the increase in the density of the ceramic sample.
The fracture toughness of alumina and alumina-based composites was not so different and it did not depend on sintering temperature.
Acknowledgments
A.L. would like to thank the Electroceramics Research Laboratory, Chiang Mai University (CMU), for high-temperature furnace usage. The authors are grateful to Prof. Dr. Gobwute Rujijanagul for his help. This work is supported by the CMU New Researcher Grant 2013 and Thailand Research Fund in conjunction with CMU and the Office of the Higher Education Commission grant no. MRG5680060.
References
[1] Niihara K. J. Ceram. Soc. Jpn. 1991, 99, 974–982.10.2109/jcersj.99.974Search in Google Scholar
[2] Zhao J, Stearns LC, Harmer MP, Chan HM, Miller GA. J. Am. Ceram. Soc. 1993, 76, 503–510.10.1111/j.1151-2916.1993.tb03814.xSearch in Google Scholar
[3] Walker CN, Borsa CE, Todd RI, Davidge RW, Brook RJ. Br. Ceram. Proc. 1994, 53, 249–264.Search in Google Scholar
[4] Gong J, Miao H, Zhao Z. J. Eur. Ceram. Soc. 2001, 21, 2377–2381.10.1016/S0955-2219(01)00206-0Search in Google Scholar
[5] Rittidech A, Tunkasiri T. Ceram. Int. 2012, 38, S125–S129.10.1016/j.ceramint.2011.04.065Search in Google Scholar
[6] Wang CJ, Huang CY. Mater. Sci. Eng. A 2008, 492, 306–310.10.1016/j.msea.2008.04.048Search in Google Scholar
[7] She JH, Schneider H, Inoue T, Suzuki M, Sodeoka S, Ueno K. Mater. Chem. Phys. 2001, 68, 105–109.10.1016/S0254-0584(00)00286-8Search in Google Scholar
[8] Sedaghat A, Taheri-Nassaj E, Soraru GD, Ebadzadeh T. Ceram. Int. 2014, 40, 2605–2611.10.1016/j.ceramint.2013.10.071Search in Google Scholar
[9] Marple BR, Green DJ. J. Am. Ceram. Soc. 1988, 71, C471–C473.10.1111/j.1151-2916.1988.tb07553.xSearch in Google Scholar
[10] Marple BR, Green DJ. J. Am. Ceram. Soc. 1991, 74, 2453–2459.10.1111/j.1151-2916.1991.tb06784.xSearch in Google Scholar
[11] Sakka Y, Bidinger DD, Askay IA. J. Am. Ceram. Soc. 1995, 78, 479–486.10.1111/j.1151-2916.1995.tb08827.xSearch in Google Scholar
[12] Sciti D, Bellosi A. J. Mater. Sci. 1998, 33, 3823–3830.10.1023/A:1004687016043Search in Google Scholar
[13] Mogilevsky P, Zangvil A. Mater. Sci. Eng. A 1999, 262, 16–24.10.1016/S0921-5093(98)01029-6Search in Google Scholar
[14] Asmani M, Kermel C, Leriche A, Ourak M. J. Eur. Ceram. Soc. 2001, 21, 1081–1086.10.1016/S0955-2219(00)00314-9Search in Google Scholar
[15] Wu S, Claussen N. J. Am. Ceram. Soc. 1994, 77, 2898–2904.10.1111/j.1151-2916.1994.tb04521.xSearch in Google Scholar
[16] Svancarek P, Lendvayova S, Galusek D, Hnatko M, Vavra I, Wang X. Wear 2011, 271, 760–769.10.1016/j.wear.2011.03.016Search in Google Scholar
[17] Luo HH, Zhang FC, Roberts SG. Mater. Sci. Eng. A 2008, 478, 270–275.10.1016/j.msea.2007.05.104Search in Google Scholar
[18] Stearns LC, Zhao J, Harmer MP. J. Eur. Ceram. Soc. 1992, 10, 473–477.10.1016/0955-2219(92)90022-6Search in Google Scholar
[19] Ortiz-Merino JL, Todd RI. Acta Mater. 2005, 53, 3345–3357.10.1016/j.actamat.2005.03.031Search in Google Scholar
[20] Limpichaipanit A, Todd RI. J. Eur. Ceram. Soc. 2009, 29, 2841–2848.10.1016/j.jeurceramsoc.2009.03.023Search in Google Scholar
[21] Bae SI, Baik S. J. Am. Ceram. Soc. 1994, 77, 2499–2504.10.1111/j.1151-2916.1994.tb04634.xSearch in Google Scholar
[22] Burgos-Montes O, Moreno R, Baudin C. J. Eur. Ceram. Soc. 2010, 30, 857–863.10.1016/j.jeurceramsoc.2009.10.013Search in Google Scholar
[23] Jiao S, Jenkins ML, Davidge RW. J. Microsc. 1997, 185, 259–264.10.1046/j.1365-2818.1997.d01-610.xSearch in Google Scholar
[24] Rice RW. Porosity of Ceramics. Marcel Dekker: New York, 1998.Search in Google Scholar
©2017 Walter de Gruyter GmbH, Berlin/Boston
This article is distributed under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Articles in the same Issue
- Frontmatter
- Original articles
- Surface modification of polyimide fibers by oxygen plasma treatment and interfacial adhesion behavior of a polyimide fiber/epoxy composite
- Influence of micro- and nanofiller contents on friction and wear behavior of epoxy composites
- Sintering temperature-microstructure-property relationships of alumina matrix composites with silicon carbide and silica additives
- A study of the effect of chemical treatments on areca fiber reinforced polypropylene composite properties
- Studying the nonlinear properties and strain-rate sensitivity of SiC short fiber-reinforced Al matrix composites
- Effects of precure cycle on tensile and dynamical mechanical properties of carbon/benzoxazine laminates
- Microstructure and wear resistance of composite coating by laser cladding Ni60A/B4C pre-placed powders on Ti-6Al-4V substrate
- Elevated electrochemical corrosion behavior of a B4C/Al neutron absorber by shot peening modification
- The influence of microstructure geometry on the scale effect in mechanical behaviour of heterogeneous materials
- Design and analysis of a complementary split ring resonator (CSRR) metamaterial based antenna for wideband application
- Numerical study on the interfacial behavior of Mg/Al plate in explosive/impact welding
- The effect of calcium salts on air-void structure in air-entrained concrete – a statistical and simulated study
- Optimisation of the ceramic-like body for ceramifiable EVA-based composites
- Elasto-plastic analysis and finite element simulation of thick-walled functionally graded cylinder subjected to combined pressure and thermal loading
- Performance of a transfer beam with hybrid reinforcement of CFRP bars and steel bars under reversed cyclic loading
Articles in the same Issue
- Frontmatter
- Original articles
- Surface modification of polyimide fibers by oxygen plasma treatment and interfacial adhesion behavior of a polyimide fiber/epoxy composite
- Influence of micro- and nanofiller contents on friction and wear behavior of epoxy composites
- Sintering temperature-microstructure-property relationships of alumina matrix composites with silicon carbide and silica additives
- A study of the effect of chemical treatments on areca fiber reinforced polypropylene composite properties
- Studying the nonlinear properties and strain-rate sensitivity of SiC short fiber-reinforced Al matrix composites
- Effects of precure cycle on tensile and dynamical mechanical properties of carbon/benzoxazine laminates
- Microstructure and wear resistance of composite coating by laser cladding Ni60A/B4C pre-placed powders on Ti-6Al-4V substrate
- Elevated electrochemical corrosion behavior of a B4C/Al neutron absorber by shot peening modification
- The influence of microstructure geometry on the scale effect in mechanical behaviour of heterogeneous materials
- Design and analysis of a complementary split ring resonator (CSRR) metamaterial based antenna for wideband application
- Numerical study on the interfacial behavior of Mg/Al plate in explosive/impact welding
- The effect of calcium salts on air-void structure in air-entrained concrete – a statistical and simulated study
- Optimisation of the ceramic-like body for ceramifiable EVA-based composites
- Elasto-plastic analysis and finite element simulation of thick-walled functionally graded cylinder subjected to combined pressure and thermal loading
- Performance of a transfer beam with hybrid reinforcement of CFRP bars and steel bars under reversed cyclic loading