Transferred arc plasma processing of mullite–zirconia composite from natural bauxite and zircon sand

Vacuum 83 (2009) 353–359

Contents lists available at ScienceDirect

Vacuum journal homepage: www.elsevier.com/locate/vacuum

Transferred arc plasma processing of mullite–zirconia composite from natural bauxite and zircon sand S. Yugeswaran a, V. Selvarajan a, *, P. Dhanasekaran a, L. Lusvarghi b a b

Plasma Physics Laboratory, Department of Physics, Bharathiar University, Coimbatore 641 046, India Department of Materials and Environmental Engineering, University of Modena e Reggio Emilia, Via Vignolese, 905, 41100 Modena, Italy

a r t i c l e i n f o

a b s t r a c t

Article history: Received 24 February 2008 Received in revised form 9 May 2008 Accepted 23 May 2008

Low cost mullite–zirconia composites were prepared from the mixtures of natural bauxite and zircon sand by using transferred arc plasma processing. In this paper, a mixture of natural bauxite and zircon in the ratio of 7:3 by weight (based on composition of 3:2 mullite) was ball milled for 4 h and melted in the transferred arc plasma for 2 and 4 min. Argon was used as plasma forming gas. The torch was operated at 5 kW input power. The phase and microstructure formation of melted samples were investigated by XRD and SEM images. The results show that the processing time is a key factor to get a single phase mullite– zirconia composites with required microstructure. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Zircon Bauxite ore Mullite Composite Arc plasma X-ray techniques

1. Introduction ‘‘Bauxite residue’’, also known as ‘‘red mud’’, is the main waste generated during the production of alumina by Bayer method. At present, various attempts are made to explore the application of bauxite residues for agricultural, mining, industrial and waste treatment. However, a viable process is yet to be developed for utilizing the bauxite residue; consequently this large quantity of waste material is required to be disposed economically and safely to the environment [1]. To overcome the above problem, bauxite is directly used as a source of alumina to synthesize mullite based ceramics for high temperature applications. Mullite ceramics have been extensively studied because of their excellent properties of high melting point (1830  C), moderate thermal expansion coefficient, high resistance to thermal shock, good chemical durability, low thermal conductivity, low dielectric constant, excellent creep resistance and sufficient mechanical strength [2]. However, the low fracture toughness of mullite and the difficulties in sintering to full density are the main obstacles for mullite materials for more widespread engineering applications. ZrO2 addition is effective in improving the strength and fracture toughness of mullite at intermediate temperatures [3]. Various processing routes can be

* Corresponding author. Fax: þ91 422 2422387. E-mail address: [email protected] (V. Selvarajan). 0042-207X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.vacuum.2008.05.033

used to prepare mullite–zirconia composite [4–7]. Reaction sintering of zircon and alumina is a relatively simple and inexpensive route to obtain homogeneous mullite–zirconia ceramic with enhanced mechanical properties. However, the conventional reaction sintering methods take long time to produce the composites. Alternately, transferred arc plasma (TAP) processing produces the composite within a short time period. Also cost effectiveness is lower compared to the conventional sintering methods. There are various environmental technologies including energy saving, clean energy, waste management and material recycling, ceramics materials technology will be one of the promising technologies to establish the practical environmental technology for better global environment [8]. Transferred arc plasma is one such widely used technology for melting and vitrification of hazardous wastes due to their high temperatures and simplicity of their generation and control. The transferred arc plasma is assumed to be stable, steady, axi-symmetric, optically thin and in a local thermodynamic equilibrium (LTE) in an atmospheric–pressure environment [9,10]. The process is characterized by extremely high temperatures (up to 20,000–30,000 K), excellent arc stability and low environmental impact (low oxides emissions, low percentage of ultra-fine powder). High power density allows high production rate with evident time savings [11–14]. In this study, natural Indian bauxite mineral and zircon sand were used as raw materials (substitute for alumina and zirconia, silica, respectively) to prepare mullite–zirconia composite.

354

S. Yugeswaran et al. / Vacuum 83 (2009) 353–359

2. Experimental set up and procedures

Table 1 Operating parameters

In this work, natural bauxite (M/s Carborundum Universal Ltd., India) and zircon sand (IREL, India) were used as raw materials to prepare a mullite–zirconia composite. The mixture of natural bauxite and zircon in the ratio of 7:3 by weight (based on composition of 3:2 mullite) was milled for 4 h with corundum ball mill media using the Planetary Mill (Insmart, India). In this experiment a 5 kW plasma torch (Ion Arc Technologies, India) was used for melting purpose. The schematic diagram of the experimental set up is shown in Fig. 1. The experimental set up consists of plasma torch, power supply, gas and water lines. In torch, a graphite bowl with 100 mm depth and 50 mm inner diameter serves as an anode and melting bed. The cathode is made of a graphite rod 250 mm long and 50 mm diameter. Its end tapers to a conical shape for better electron emission. The cathode is enclosed in a hollow brass cylinder and provisions are made for water circulation (for cooling) and gas flow. The system has multiple inlets for plasma gas at the cathode end. The melted samples were cooled by forced air. The operating parameters for the experiments are given in Table 1. The TAP processed samples were milled down to a few micrometer size powders in achate mortar prior to X-ray diffraction (XRD) analysis. For scanning electron microscope (SEM) investigation, the samples were polished beginning with a set of silicon–carbide papers ending with diamond paste. Then the samples were chemically etched using 30 wt.% HF solution for 30 s. The XRD pattern of the sample was performed using a PW Philips with Cu-Ka radiation. The microstructures of the samples were studied by the SEM (Philips XL40) and EDX spectrometry (EDX, INCA, Oxford Instruments). Differential thermal analysis (DTA) was carried out up to 1400  C using a heating rate of 10  C/min with nitrogen atmosphere. The density and porosity of the TAP processed mullite–zirconia composite were measured, respectively, by a conventional method using Archimedes’ principle in water medium (the theoretic density of mullite–zirconia composite is 3.55 g cm3) [15–17].

Input power Plasma gas and flow rate Cooling water flow rate Processing time Plasma temperature Cooling time Cooling medium

Porosityð%Þ ¼ 100 

sample density ¼

sample density  100 bulk density

D W S

(1)

(2)

5 kW Argon; 10 lpm 10 lpm 2 and 4 min w2980 K 10 min Air

where D, W and S are dry, saturated and suspended weights of the material, respectively. The excitation temperature of plasma is generally called as plasma temperature. The TAP temperature (T) just above the anode was determined by the atomic Boltzmann plot method from the following equation [18]:

    Il 625Ek ¼ C log gk Ak T

(3)

or

T ¼

625 slope value

(4)

A plot between log(Il/gkAk) and Ek yields a straight line by best fit method. The inverse of the slope of the line gives the temperature of the plasma arc along the line of sight. In the equation I is the intensity, l is the wave length of the line, C is the intercept on the ordinate axis, Ak is transition probability and Ek and gk are the excitation energy and statistical weight of atomic state k, respectively. The emission spectra of TAP were recorded by the monochromator (Thermo Oriel, ¼ M) based diagnostic arrangements along with a photomultiplier tube, power supplies, oscilloscope, X– Y recorder and a computer for running the monoutility program and optical fiber with xyz positioner for collecting radiation from the plasma arc. The values of intensities were calculated from the spectrum and gk, Ak, and Ek values of corresponding lines were taken from the literature [19]. Using a software log(Il/gkAk) vs. Ek was plotted. The slope was obtained from the straight line of the best fit and the correlation coefficient was more than 0.85 to give the best fit for random points. Different spectral lines were scanned from 400 to 452 nm in steps of 0.2 nm with 1 ms dwell time per step by using monoutility program which was created by Thermo Oriel. The radiation was collected at four different parts of the plasma column.

3. Results and discussion 3.1. Phase and chemical composition of raw materials

Fig. 1. Schematic diagram of the experimental set up.

Figs. 2 and 3 show the XRD pattern of zircon sand and bauxite ore, respectively. The pattern shows that the gibbsite (g-Al(OH)3) is the main phase of the bauxite ore combined with small amount of boehmite (g-AlO(OH)), nacrite (Al2Si2O5(OH)4), anatase TiO2 and cristobalite. The chemical compositions of the as-received natural bauxite ore and zircon sand characterized by ICP spectroscopy are shown in Table 2. The bauxite is mainly composed of alumina (57.5 wt.%) and small amount of silica (5.1 wt.%), titania (6.4 wt.%) and ferric oxide (3.7 wt.%). In addition to this, very little amounts of alkali and alkaline-earth metal oxides such as Na2O, K2O, MgO and CaO, also exist. The major content of zircon sand consists of ZrO2 (62–64 wt.%) and SiO2 (35 wt.%) along with very small amounts of transition metal oxides (1 wt.% Al2O3, 0.15 wt.%TiO2 and 0.015 wt.% Fe2O3).

S. Yugeswaran et al. / Vacuum 83 (2009) 353–359 Table 2 Chemical composition (by ICP method) of raw materials

z - zircon

Intensity (a.u.)

z

z z z z z

10

20

z z

30

40

z

z

z z z

50

z

60

355

z

70

2θ Fig. 2. XRD pattern of zircon sand.

3.2. Phase formation Fig. 4 shows the XRD patterns of the TAP processed mullite– zirconia composite at two different processing times (2 and 4 min). The first evidence for the decomposition of zircon and major bauxite hydrated phases, such as gibbsite, boehmite and goethite appears in the XRD pattern of the sample that was processed at 2 min. It shows the predominant presence of monoclinic phase zirconia and mullite along with small amounts of quartz, corundum and zircon. The presence of monoclinic phase is taken as an indication for the beginning of the decomposition of zircon according to the following reaction: ZrSiO4 / SiO2 D ZrO2 Corundum crystallizes from the decomposition products of major Al-bearing hydrates, mostly gibbsite and boehmite. a-Quartz was formed from cristobalite (b-SiO2) in the bauxite and dissociated zircon. The initial step in the thermal decomposition of gibbsite is the diffusion of protons and the reaction with hydroxyl ions

Materials (Oxides)

Zircon sand (wt.%)

Bauxite ore (wt.%)

ZrO2 SiO2 Al2O3 CaO MgO Na2O K2O Fe2O3 TiO2 Other traces

62–64 35 1 – – – – 0.015 0.15 Balance

– 5.1 57.5 0.30 0.64 0.32 0.17 3.7 6.4 Balance

to form water. This process removes the binding forces between subsequent units in the gibbsite structure and causes changes in the chemical composition and density. At w280  C, the boehmite was formed via an hydrothermal reaction due to the retarded diffusion of water out of the larger gibbsite particles. This reaction was not possible in finely grained gibbsite. There is general agreement that boehmite and a disordered transition alumina (g-alumina) are formed upon the thermal treatment of coarse gibbsite up to 400  C [20]. The Boehmite converts into ‘transition alumina’ up to 600  C, and at higher temperatures (1100  C) converts to alumina (a-Al2O3) with the corundum structure. This transformation temperature varied with boehmite’s crystallite size [21]. The result shows that the processing time (2 min) was not enough to get a complete dissociation and transformation of the raw materials. The intensities of the peaks of mullite and zirconia increase when the processing time is increased from 2 to 4 min. Simultaneously, the intensities of the peaks of zircon, corundum and quartz phases have totally disappeared in 4 min processed samples. The absence of peaks of silica in this XRD pattern indicates that this phase is either amorphous or that it is completely consumed in the reaction with alumina to yield mullite. This indicates that corundum reacted with quartz for secondary mullitization by solid state reaction. In this case, corundum disappeared due to its dissolution into transitory liquid glassy phase. At high temperature, more secondary mullite crystals precipitated from the liquid glassy phase [22,23]. Hence increasing the processing time increases the intensities of mullite lines due to the high concentration of mullite phase in the processed samples.

b

4 minutes g - Gibbsite b - Boehmite c - Cristobalite a - Anatace TiO2 n - Nacrite

m - mullite b - monoclinic zirconia c - corundum q - quartz z - zircon

b

m

m

m

Intensity (a.u.)

Intensity (a.u.)

g

2 minutes

n

c

gg bc

g n

g

20

30

40

50

60

m m

m

m

b c

c m m m b

m m

cm

z

m m m m c

b g 10

10

q z

g gg g g g n bg c b

m mb m

m

b

g

b

b

m g

c

m m

b

g b

m

mm

m

a

m

mb

b

70

20

30

40

50

60

70

2θ

2θ Fig. 3. XRD pattern of bauxite ore.

Fig. 4. XRD pattern for the transferred arc plasma processed bauxite–zircon mixture at two different processing times (2 and 4 min).

356

S. Yugeswaran et al. / Vacuum 83 (2009) 353–359

According to Al2O3–SiO2 binary phase diagram, liquid phase can be formed at the above eutectic temperature (1590  10  C) [24]. In our case, excessive transition metallic oxides such as TiO2 and Fe2O3 in the starting materials enhance liquid glassy phase at a relatively lower than actual temperature. In addition, the co-existing alkali and alkaline-earth metal oxides (Na2O, K2O, MgO, CaO, etc.), acted as glass modifiers, not only lowered the viscosity of liquid glassy phase by forming weak Si–O bonds but also changed its wetting characteristic [25]. In both cases, zirconia appears only as a monoclinic phase due to the normal quenching. It has been reported that conversion of tzirconia to m-zirconia is noticeably inhibited by the presence of SiO2 in the ZrO2–SiO2 system [26]. At low temperatures, ZrO2 and amorphous SiO2 were formed due to ZrSiO4 dissociation, and ZrO2 particles were immersed in the amorphous SiO2 matrix; the SiO2 stabilized ZrO2 as t-ZrO2 phase. At high temperatures, however, the amorphous SiO2 was finally consumed to form mullite, losing the ability to stabilize ZrO2; thus, the ZrO2 in the samples obtained at high temperatures mainly existed as m-ZrO2. Another factor of importance is to understand the phase change of ZrO2. It is well known that ZrO2 phase transformation is subjected to the particle size effect. When the particles are larger than a critical size, upon cooling the transformation from t-ZrO2 to m-ZrO2 occurs. The critical particle size is in the order of 0.6–1 mm for mullite–zirconia composites [27]. It is reasonable to assume that the zirconia particles that precipitated from zircon through transferred arc plasma processing are larger than the critical size. In the present reaction process, ZrO2 grains would grow relatively larger and easily get converted to m-ZrO2 phase when cooled down to room temperature. In the mullite–zirconia formation, melted bauxite is a source of alumina and dissociated zircon is a source of silica and zirconia. Mullite formation is more complex in the reaction sintering of

alumina–zircon mixtures because it involves a solid state dissociation of zircon. At least four processes are involved in mullite grain growth. (1) Initial formation of glassy phase from the present impurities with partially thermal dissociated zircon, (2) dissolution of zircon into this glassy phase, (3) dissociation of zircon and the formation of zirconia and higher SiO2-rich glassy phases, and (4) dissolution of alumina in this glassy phase. The concentration of Al2O3 in the melt increases until the stoichiometric composition of mullite is obtained. All these facts may imply that Al2O3 dissolution proceeds more easily than ZrSiO4 dissociation in the present system, and that zircon dissociation is a rate-controlling step in mullite grain growth. During the processes, the zircon dissociation began at melting and formed amorphous SiO2. The amorphous SiO2 formed softened with increasing temperature, then penetrated into the Al2O3 agglomerates and dissolved the Al2O3 to form an amorphous aluminosilicate phase. Mullite nucleating will occur when the concentration of alumina in the amorphous phase exceeds the critical nucleation concentration (CNC) or when the liquid structure becomes saturated with mullite. There are several facts supporting the supposed mechanism. As noted earlier, the temperature for the onset of mullitization is higher than that obtained by thermodynamic calculation, suggesting the existence of an incubation period prior to mullite nucleating. The incubation period is indicative of the process of Al2O3 dissolution into amorphous SiO2 phase, which results in the lag time required for alumina concentration in amorphous phase to reach the CNC. However, ZrSiO4 will dissociate rapidly at high temperatures. There also exists some point, above which Al3þ or Si4þ diffusion process through mullite layer controls the mullite growth. At the final stage of mullitization, the Al2O3-rich mullite or fine Al2O3 particles within mullite grain and SiO2-rich amorphous phase would react to form stoichiometric mullite. This process

Fig. 5. Different surface morphologies of transferred arc plasma processed bauxite–zircon mixture: (a and b) 2 min melted; (c and d) 4 min melted.

S. Yugeswaran et al. / Vacuum 83 (2009) 353–359

necessitates Al3þ or Si4þ diffusion through crystalline mullite layer. It should be pointed out that, due to the complexity of the present system, the kinetic mechanism of mullite growth proposed above is only a reasonable hypothesis to some degree, and needs to be confirmed by further research work. A number of SiO2–ZrO2 phase diagrams have been reported in the literature [28,29]. The authors indicate that zircon is converted, usually by thermal means at about 1676  C to zirconia and silica. Above 1173  C, zirconia takes the

357

tetragonal form, and below this temperature it exists in the monoclinic phase. 3.3. Microstructure analysis The microstructure of the TAP processed mullite–zirconia composite was investigated by SEM and EDS. In transferred arc melting, initially the plasma arc strikes the top wall of graphite

Fig. 6. EDS spectra of transferred arc plasma processed bauxite–zircon mixture: (a) 2 min melted; (b) 4 min melted.

358

S. Yugeswaran et al. / Vacuum 83 (2009) 353–359

anode (melting bed) and the tip of cathode, later the arc randomly spreads to all parts of the melting bed and melt the powders. The melted powders form a semi-sphere shape slag due to the design of the melting bed and the direction of plasma arc formation (see Fig. 1). During solidification (after cut of the plasma arc) the edges solidify sooner than the middle of the molten semi-sphere due to the temperature gradient of anode surfaces. At the same time the sluggishly solidifying slag produces pores and cracks due to the release of strain and stresses. This solidification also plays a major role in the microstructure of mullite growth. Fig. 5a and b shows the surface morphology of 2 min TAP processed mullite–zirconia composite. The surface morphology shows that zirconia appears in white color with uniformly arranged needle shape along with black color alumina rich mullite in the form of slag. Fig. 5c and d shows the surface morphology of 4 min TAP processed mullite–zirconia composite. In this case, both the zirconia and mullite appear in needle shape with white and grey colors, respectively. The structure and color of the zirconia and mullite were confirmed by EDS spectrum (Fig. 6 a and b) with corresponding SEM images. The EDS results show that 2 min processing time is not enough to get complete melting of the bauxite and zircon mixture. Hence the alumina, mullite and zirconia randomly appear along with small amounts of Ca, Ti and Fe oxides. In the case of 4 min processing, the mullite and zirconia are well distinguished and appear with uniform microstructure due to the sufficient heating of plasma arc. Further it was found that increasing processing time from 2 to 4 min will increase the density of the bauxite–zircon mixture from 2.95 to 3.12 g/cm3 and decreased the porosity from 16.9 to 12.11%. 3.4. Differential thermal analysis The DTA curve of the TAP processed mullite–zirconia composite at two different processing times (2 and 4 min) are shown in Fig. 7. The DTA curve of 2 min processed composite appears with three exothermic peaks at 340, 945 and 1300  C. The processing time (2 min) was not enough to get complete dissociation and transformation of the raw materials. The first exothermic peak at 340  C is due to the burning of organic materials [30] or transformation of b-quartzite (cristobalite) to a-quartzite [31]. The second exothermic peak at 945  C is due to the formation of primary mullite and the third exothermic peak at 1300  C is due to the secondary mullite formation [32]. The DTA curve of the 4 min processed is almost

0.6

2 minutes 4 minutes

0.5

DTA (mw/mg)

0.4 340 °C

0.3 0.2 0.1

945 °C 1300 °C

0.0 -0.1 -0.2 0

200

400

600

800

1000

1200

1400

Temperature ( °C ) Fig. 7. DTA curve of transferred arc plasma processed bauxite–zircon mixture at two different processing times (2 and 4 min).

featureless except dilute exothermic peak at 340  C due to the burning of organic materials and a broad exothermic peak at 1300  C due to the rapid release of energy during crystallization and this is in agreement with the results reported by other authors [32,33]. In this case, maximum mullite formation was achieved followed by the complete dissociation and transformation of raw materials during processing. Hence there is no mullite formation after the plasma processing. From the processing viewpoint, the time duration is a key factor to get a single phase mullite–zirconia composite. 3.5. Comparison between TAP and other processes The results described above show that the mixtures of zircon and bauxite processed by TAP yield low cost mullite–zirconia composite. Now, in order to further highlight the advantages of this process, in this section, the results of the TAP processed mullite– zirconia composite will be compared with the results obtained by other conventional and spark plasma sintering (SPS) processes. Reaction sintering of alumina and zircon is an easy and inexpensive route to obtain mullite–zirconia composites containing dispersed zirconia with enhanced mechanical properties and has been extensively studied by many researchers [3,34–37]. However, when conventional reaction sintering methods are employed, problems are encountered in the form of low bulk and grain boundary diffusion coefficient of mullite, and very long processing time duration (2 h). Khor et al. [38] has attempted to produce mullite–zirconia composite by spark plasma sintering (SPS). They used a plasma spheroidized zircon/alumina powder which was composed of alumina and partially dissociated zircon. The mixture was completely transformed into mullite and zirconia at an SPS temperature of 1300  C for 10 min. However, rapid coarsening of the microstructure took place and the zirconia transformed to its monoclinic variant. Their composite ended up with numerous microcracks. Most of the conventional and SPS results showed complete mullitization along with monoclinic and tetragonal phase zirconia. Cem Ozturk and Yahya Kemal Tur [39] have prepared textured mullite–zirconia composites from a reactive mixture of alumina and zircon powders sintered at 1450–1550  C for 4 h with a constant heating rate of 10  C/min in air. In this case, complete mullitization was achieved along with single phase zirconia (monoclinic) at 1550  C. Similar mullitization trend is seen for TAP processed for 4 min. Evidently, it is not easy to compare the experimental results for the TAP and the conventional processes, because the TAP temperatures were measured along the line of sight of the stabilized plasma column by a monochromator through optical fiber focused at the top surface of the plasma column, which might not be representative of the actual uniform temperature of the medium. In this case, the average temperature of the plasma medium is 2980 K. A temperature evolution during TAP processes has shown that there might be a large temperature difference between the surface of the plasma column and the powder. It can depend on number of factors, such as the dimension of graphite electrodes, input power and thermal properties and quantity of powders. Therefore, it is not possible to compare TAP temperature and other conventional processing temperature unambiguously. It is clear, however, that the TAP process accelerates the reaction. The relative density of the TAP processed mullite–zirconia composites was increased from 83.1 to 87.9% on increasing the processing time from 2 to 4 min. From TAP processing results, two phenomena are assumed to have taken place at the higher processing time, i.e., the removal of the last remnants of porosity and completion of mullitization. In comparison, the relative density of TAP processed composites is lower compared to that of others. For

S. Yugeswaran et al. / Vacuum 83 (2009) 353–359

example, the relative density of the mullite–zirconia composite prepared by conventional reaction sintering at 1550  C for 2 h was 92.1% and it was 95% in SPS processes at temperature 1420  C. The fast sintering and high densities that can be obtained by SPS at low temperatures are usually described as the main features of the SPS process. However, in application point of view the porous composites can also be used as lightweight structural components, catalysts’ carriers for chemical plants and automobiles, heat exchangers, acoustic absorbers, dust collectors, filters and electrodes [40].

4. Conclusion In the present work, a low cost synthesis of mullite–zirconia composite from natural bauxite and zircon sand by using low power (5 kW) transferred arc plasma torch has been reported. This technique allowed to completely transform the raw materials into single phase mullite–zirconia composite due to its high temperature (w2980 K) in just 4 min of processing time. The XRD and DTA results confirm that the processing time is a major parameter to get a single phase mullite–zirconia composite. The SEM images showed that a higher processing time enhances the homogeneity of the microstructure of mullite–zirconia composite. Moreover, inexpensive raw natural bauxite and zircon sand were testified as an effective substitute for industrial alumina and zirconia, respectively, for the production of mullite–zirconia composite. The reported results suggest that TAP processing method can be very competitive compared with conventional sintering methods as the processing time period is 30 times lesser than conventional method.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31]

Acknowledgements The authors acknowledge the project funding under the category of Technology Systems from Department of Science and Technology Government of India. Authors would like to thank Mr. Janarthanan Nair (Ion Arc Technologies Pvt. Ltd., India) for their help in developing torch facilities. One of the authors (S.Y.) acknowledges the Junior Research Fellowship awarded by the Department of Science and Technology, Government of India.

359

[32] [33] [34] [35] [36] [37] [38] [39] [40]

Clifton M, Nguyen T, Frost Ray. J Colloid Interface Sci 2007;307:572–7. Ananthakumar S, Jayasankar M, Warrier KGK. Acta Mater 2006;54:2965–73. Claussen N, Jahn J. J Am Ceram Soc 1980;63(3–4):228–9. Maitra S, Rahaman A, Sarkar A, Tarafdar A. Ceram Int 2006;32:201–6. Garrido LB, Aglietti EF, Martorello L, Camerucci MA, Cavalieri AL. Mater Sci Eng A 2006;419:290–6. Park HC, Yang TY, Yoon SY. Mater Sci Eng A 2005;405:233–8. Temoche F, Garrido1 LB, Aglietti EF. Ceram Int 2005;31:917–22. Wasa K. Bull Mater Sci 1995;18(8):937–53. Hsu KC, Etemadi K, Pfender EJ. Appl Phys 1983;54:1293–7. Coudert JF, Delalondre C, Roumilhac P, Simonin O, Fauchais P. Plasma Chem Plasma Process 1993;13:399–404. Deuis RL, Bee JV, Subramanian C. Scr Mater 1997;37(6):721–7. d’Oliveira ASCM, Vilar R, Feder CG. Appl Surf Sci 2002;201(1–4):154–60. Yang YL, Loh NL. Surf Coatings Technol 1995;71(2):196–200. Bourithis E, Tazedakis A, Papadimitriou G. J Mater Proc Tech 2002;128(1–3):169–77. Antou G, Montavon G, Hlawka F, Cornet A, Coddet C. Mater Charact 2004;53: 361–72. Kumar S, Selvarajan V, Padmanabhan PVA, Sreekumar KP. Surf Coat Technol 2006;201(3–4):1267–75. Nagano T, Kato H, Wakai F. J Mater Sci 1991;26:4985–90. Yugeswaran S, Selvarajan V. Vacuum 2006;81:347–52. Wise WL, Brault JW, Danzmann K, Helbig V, Kock M. Phys Rev A Gen Phys 1998;39:2461–71. Kloprogge JT, Ruran HD, Frost RLJ. Mater Sci 2002;37:1121–9. Laskou M, Margomenou-Leonidopoulou G, Balek V. J Therm Anal Calorim 2006;84(1):141–5. Chen CY, Lan GS, Tuan WHJ. Eur Ceram Soc 2000;20:2519–25. Sainz MA, Serrano FJ, Amigo JM, Bastida J, Caballero AJ. Eur Ceram Soc 2000; 20:403–12. Pask JAJ. Eur Ceram Soc 1996;16:101–8. Baudin C, Villar MP. J. Am Ceram Soc 1998;81:2741–5. Nagarajan VS, Rao KJ. J Mater Sci 1989;24(6):2140–6. Green DJ, Hannink RHJ, Swain MV. Transformation toughening of ceramics. Boca Raton, FL: CRC Press; 1989. Butterman W, Foster WR. Am Mineral 1967;52:880–5. Doerner P, Gauckler LJ, Krieg H, Lukas HL, Petzow G, Weiss J. Hermochem 1979;3:241–57. Kim GD, Lee DA, Lee HI, Yoon SA. J Mater Sci Eng A 1993;167:171–6. Behmanesh N, Heshmati-Manesh S, Ataie A. J Alloys Compd 2008;450(1–2): 421–5. Khor KA, Li Y. Mater Lett 2001;48:57–63. Sen S, Thiagrajan S. Ceramic Int 1988;14:77–86. Descamps P, Sakaguchi S, Poorteman M, Cambier F. J Am Ceram Soc 1991; 74(10):2476–81. Cambier F, Baudin delalastra C, Pilate P, Leriche A. Br Ceram Trans 1984;83: 196–200. Koyama T, Hayashi S, Yasumori A, Okada K, Schmucker M, Schneider H. J Eur Ceram Soc 1996;16(2–3):231–7. Wu JM, Lin CM. J Mater Sci 1991;26(17):4631–6. Khor KA, Yu LG, Li Y, Dong ZL, Munir ZA. Mater Sci Eng A 2003;339:286–9. Ozturk Cem, Tur Yahya Kemal. J Eur Ceram Soc 2007;27:1463–7. Ebadzadeh T. Ceram Int 2005;31:1091–5.