Formation and densification behavior of reaction sintered alumina–20 wt.% aluminium titanate nano-composites

Int. Journal of Refractory Metals and Hard Materials 47 (2014) 49–53

Contents lists available at ScienceDirect

Int. Journal of Refractory Metals and Hard Materials journal homepage: www.elsevier.com/locate/IJRMHM

Formation and densification behavior of reaction sintered alumina–20 wt.% aluminium titanate nano-composites M. Sobhani ⁎, T. Ebadzadeh, M.R. Rahimipour Ceramic Department of Materials and Energy Research Center, Alborz, Iran

a r t i c l e

i n f o

Article history: Received 27 April 2014 Accepted 24 June 2014 Available online 1 July 2014 Keywords: Alumina/aluminium titanate Thermal expansion Reaction sintering Composites

a b s t r a c t Densification and reaction sintering behavior of the alumina–20 wt.% aluminium titanate nano-composites were studied. Nano TiO 2 powders were used in achieving the aluminium titanate phase during reaction sintering of alumina and titania. High resolution X-ray diffraction (HR-XRD) and scanning electron microscopy (SEM) results revealed that the aluminium titanate formation was affected by heating rate and soaking time at the elevated temperatures. The content of aluminium titanate phase decreased during increment of heating rate and increased during increment of soaking time at sintering temperature. Final densities of the sintered composites at 1550 °C were lower than those sintered at 1450 °C which can be explained by the grain cracking phenomenon induced by thermal expansion mismatch. The residual strain of the aluminium titanate grains released after grain cracking. The Al2TiO5 grains cracked along the (010) crystallographic planes because of the most positive expansion of the b-axis, and the lattice parameter b reverted to its original value after grain cracking. © 2014 Elsevier Ltd. All rights reserved.

Introduction Alumina as an industrial ceramics is widely used due to its attractive properties. Strength, melting point, hardness and Young's modulus of alumina are higher than most well known ceramics. Aluminium titanate (Al2TiO5) as a second phase in alumina ceramics, with lower Young's modulus and thermal expansion coefficient than the alumina, can therefore enhance the composite's toughness and resistance to thermal shock [1–4]. Additionally, the limited solution of titania can improve the densification of alumina [5,6]. The alumina/ aluminium titanate (A/AT) composites can be fabricated by direct mixing of alumina and aluminium titanate (tialite) powders or by in-situ reaction sintering of mixed alumina and TiO 2 powders. In spite of its low thermal expansion coefficient (~ 1 × 10− 6 °C− 1 for polycrystalline bodies), tialite similar to other pseudobrookite materials exhibits high thermal expansion anisotropy (TEA) (αa ≈ 11, αb ≈ 21, αc ≈ − 3 × 10−6 °C−1) [7–9]. Accordingly, thermal stresses induced during cooling from the sintering temperature in bulk can lead to spontaneous microcracking phenomenon. At the first time, Kuszyk and Bradt [10] proposed that an energy criterion controls the formation of microcracks in ceramics with high thermal expansion anisotropy. They introduce a basic relation between the critical grain size needed for spontaneous microcracking and energy criterion (strain and fracture

⁎ Corresponding author. Tel.: +98 9155022379; fax: +98 2636204139. E-mail address: [email protected] (M. Sobhani).

http://dx.doi.org/10.1016/j.ijrmhm.2014.06.018 0263-4368/© 2014 Elsevier Ltd. All rights reserved.

surface energy). Afterwards, Cleveland and Bradt [11] offer the following developed equation: Dgr ¼

14:4γ f EðΔTΔα max Þ2

ð1Þ

where γf is the fracture surface energy, E is the Young's modulus and Δαmax is the maximum difference of the thermal expansion coefficients along the crystallographic axes and ΔT is the difference between sintering and room temperatures. The estimated grain size of tialite for microcraking is about 2 μm. Therefore, the size controlling of aluminium titanate phase is the main problem, because the widespread occurrence of microcracking in coarser microstructure, resulting from thermal expansion anisotropy, decreases the mechanical properties. Lawn et al. and Padture et al. [2,3] studies showed that the R-curve behavior in fracture toughness of A/AT composites was a result of aluminium titanate grains bridging during crack propagation. Also Bueno et al. [12] reported the change of toughening mechanism from crack bridging to microcracking in A/20AT composite at the presence of sub-micrometric aluminium titanate particles. Additionally, the physical properties such as the final bulk density of monolithic aluminium titanate at room temperature can be affected by the enlarged aluminium titanate grains. The density degradation due to microcracking (obtained during cooling from the sintering temperature) is accompanied by increased grain size due to rising sintering temperature. The crack volume induced by such microcracking is proportional to D0.5 gr [13]. In view of the above, the investigation of microcracking effect on physical properties of

50

M. Sobhani et al. / Int. Journal of Refractory Metals and Hard Materials 47 (2014) 49–53

A/AT composites has drawn little attention. Therefore, in the present work, the effect of different heat treatment schedules on formation and densification of A/20AT composite was investigated. Furthermore, high resolution X-ray diffraction was used to investigate the lattice parameter changes of Al2TiO5 during grain cracking caused be thermal expansion anisotropy. Experimental procedures The starting materials were α-Al2O3 (MR70 average particle size of 0.7 μm), nanosized TiO2 powder (P25, Degussa-Evonik, Germany) with an average primary particle size of 40 nm. Fig. 1 shows the morphology of the raw materials. A mixture of alumina (≈91.2 wt.%) and titania (≈ 8.8 wt.%) was used to achieve the sintered A–20 wt.% AT composites. The mixture of desired powders milled in ethanol, using a planetary ball mill with zirconia cup and balls (during 1 h at 250 rpm). Drying was carried out using a hot-plate, while stirring continuously. Afterwards, the powders were grounded in an alumina mortar and sieved. Green specimens, 25 × 5 × 4 mm, were fabricated by uniaxially pressing at 50 MPa. Removal of any defects associated with die pressing was achieved by subsequent wet bag isostatic pressing at 175 MPa. A green density of about 60% of the theoretical limit was obtained using this method. The green bars were sintered using two heat treatment schedules. The first contains preheating at 1100 °C before sintering and the second only contains sintering schedules (listed in Table 1). Densities were determined by the Archimedes's method in deionized water (European Standard EN1389:2003) and theoretical densities were calculated by taking values of 3.99 g cm− 3 for alumina (ASTM 42-1468), 4.25 g cm−3 for rutile (ASTM 21-1276) and 3.70 g cm− 3 for aluminium titanate (ASTM 26-0040). The strength of sintered bars were measured using three point bending test supported with a span length of 15 mm and cross-head speed of 0.5 mm/min−1. At least five specimens were tested to obtain the average value along with its standard deviation for density and strength measurements. Simultaneous thermal analysis (STA) with differential scanning calorimetry (DSC) and thermo-gravimetric (TG) curves was carried out by a NETZSCH STA 409 PC/PG instrument in air atmosphere with a temperature rising rate from 5 °C/min up to 1400 °C. High resolution X-ray diffraction (HR-XRD) of the composites was determined by X-ray diffraction (Model: Siemens D-500, Germany) using CuKα (λ = 1.54 Å) radiation. The HR-XRD data were obtained using dθ = 0.005° and t = 2 s. The parameters for the zeropoint shift were refined using standard sample (silicon) diffraction lines. The results were compared with the ASTM Files (for βaluminium titanate, corundum and rutile [14–16]). Microstructure

Table 1 Different sintering conditions and determined phases for A/20AT composites. Sample

Heat treatment schedule

Phases

a b c d e f

4 4 4 2 2 4

A A A A A A

h at 1100 °C–2 h at 1450 °C (3 °C/min) h at 1100 °C–2 h at 1500 °C (3 °C/min) h at 1100 °C–2 h at 1550 °C (3 °C/min) h at 1450 °C (5 °C/min) h at 1500 °C (5 °C/min) h at 1450 °C (5 °C/min)

+ + + + + +

AT + R AT AT AT + R AT + R AT

characterization on polished (1 μm) and thermally etched (100 °C below the sintering temperature during 20 min) surfaces was performed by a TESCAN field emission scanning electron microscopy (FESEM, model MIRA ΙΙ). Results and discussions Phase development and microstructure studies The results of DSC and TG analysis of A–20 wt.% AT composite powders are presented in Fig. 2. The endothermic peak at about 90 °C in DSC curve indicates the removal of moisture. Another endothermic peak at 1309 °C is a characteristic peak of aluminium titanate formation during chemical reaction of Al2O3 and TiO2. Considering the Eq. (2) the formation temperature of Al2TiO5 in this study is much close to the thermodynamically value due to the high reactivity of the nano TiO2 powders, in comparison with the previous reports [17–19]. The total mass reduction value in TG curve (− 2.66%) is attributed to moisture evaporation. Fig. 3 demonstrates X-ray diffraction patterns of the alumina–20 wt.% aluminium titanate composites sintered at different temperatures listed in Table 1 (samples a, b and c). Considering the molar free energy equation of Al2TiO5 formation, the reaction of alumina and titania leads to the formation of aluminium titanate at above 1280 °C [20]: Al2 O3 ðαÞ þ TiO2 ðRutileÞ→Al2 TiO5 ðβÞ; ΔG ¼ 17; 000–10:95 Tð j=molÞ:



ð2Þ

Also it can be noted that the formation of β-Al2TiO5 and its solid solution compounds with MgO (MgxAl2(1 − x)Ti(1 + x)O5) using the chemical process can occur at the lower temperatures [21,22]. However, for sample (a) the reaction did not completely happen and the characteristic peak of rutile at 2θ = 27.44° can be observed. Therefore, according to XRD results unreacted titania still remains in sample (a) and formation of aluminium titanate completed after sintering at 1500 °C with a

Fig. 1. SEM micrographs of raw materials: alumina (right), nano TiO2 (left).

M. Sobhani et al. / Int. Journal of Refractory Metals and Hard Materials 47 (2014) 49–53

51

Fig. 2. DSC and TG analysis of A–20 wt.% AT composite.

heating rate of 3 °C/min. No TiO2 peaks were detected in both b and c samples after sintering at 1500 and 1550 °C, respectively. The effect of heating rate and soaking time at sintering temperature was investigated using the sintering schedule of samples (d), (e) and (f) listed in Table 1. Increasing the heating rate from 3 °C/min for sample (a) to 5 °C/min for sample (d) caused the decrease in aluminium titanate peak intensities (Fig. 4) and formation of rutile which can be observed by its strong characteristic peak. With increasing the sintering temperature from 1450 to 1500 °C for sample (e), no improvement in Al2TiO5 formation was observed, while considering Eq. (2), the formation condition of Al2TiO5 in sample (e) is better than that of samples (d) and (f). This peculiar observation in A/AT composites can be attributed to nucleation and growth phenomenon. The full formation of Al2TiO5 in samples (b), (c) and (f) confides that the decomposition of Al2TiO5 has not occurred in the temperature range from 900 to 1250 °C during cooling from sintering temperature. Assuming the zero matrix-nucleus interface energy, the necessary condition for nucleation is [23]: vol vol ΔGAl2 TiO5 ≥ΔGstrain

ð3Þ

where the ΔGvol strain (strain energy) assumed to be temperature independent and the ΔGvol Al2 TiO5 , chemical driving force per volume or ΔG° at Eq. (2), is proportional to superheating ΔT. At lower temperatures, slightly above the A12TiO5 decomposition temperature, the small

  chemical driving force ΔGvol Al2 TiO5 cannot overcome the strain energy (ΔGvol strain). Hence, at higher temperatures aluminium titanate nucleation becomes feasible and the nucleus concentration increases with rising temperature [23]. The energetic approach showed that the favorable nucleation mechanism at lower ΔGvol Al2 TiO5 appears to be in bulk TiO2 to overcome the ΔGvol strain. It can turn to nucleation in the interface of A12O3/TiO2 by increasing temperature due to the increase in ΔGvol Al2 TiO5 (proportional to superheating (ΔT)). On the other hand, just after the nucleation stage, the surface of Al2O3 and TiO2 particle is covered by Al2TiO5 and the interface is sealed by produced phase. Transportation of reactants (Al3 + and Ti4 +) across the aluminium titanate layer is slow due to the low mobilities of the cations in the product phase [23]. Therefore, the content of aluminium titanate is lower at high temperatures rather than at low temperatures due to much slower growth velocities in the former. Approving the above discussion, the listed below are suggestions for possible improvement of the aluminium titanate content in sample (e): (a) Lowering temperature: decrease in sintering temperature causes the decline in nucleus concentration. This phenomenon eases the reaction between alumina and titania, which can be approved by XRD results of sample (d). (b) Increasing soaking time: in addition to lowering temperature, an increase in sintering time brings enough opportunity for cation α-Alum ina Rutile Alum inium titanate

f

Intensity/ a. u.

Intensity/ a. u.

α-Alumina Rutile Aluminium titanate

c

b

e

a 5

d 15

25

35

45

2θ/ °

55

65

75

Fig. 3. X-ray diffraction patterns of samples: a, b and c preheated at 1100 °C and sintered at 1450, 1500 and 1550 °C, respectively.

5

15

25

35

45

2θ/ °

55

65

75

Fig. 4. X-ray diffraction patterns of samples: d, e and f sintered at 1450 (2 h), 1500 (2 h) and 1450 °C (3 h), respectively.

52

M. Sobhani et al. / Int. Journal of Refractory Metals and Hard Materials 47 (2014) 49–53

diffusion to accelerate the aluminium titanate formation. The result of (f) sample approves this suggestion. (c) Decreasing heating rate: obviously, this schedule contains both of the above suggestions simultaneously. Decrease in heating rate was applied in sample b which resulted in complete reaction of TiO2 with alumina matrix. Densification and thermal expansion anisotropy effect Density and strength variations of the sintered sample at different temperatures are presented in Fig. 5. Titania is an effective additive to densify the alumina bodies and it can promote the sintering of alumina below the normal densification temperature due to the formation ofTiAl and VAl000 [5,24]. It is known that, for the Al2O3–TiO2 system titania has the solution limit of lower than 0.5 wt.% in the temperature range of 25–1700 °C [6,25]. With increasing temperature from 1450 to 1500 °C, the density increased and with further increase in sintering temperature to 1550 °C, the density of sample (c) decreased to a minimum value in comparison with samples (a) and (b). As it is shown in Fig. 6 the spontaneous microcracking phenomenon during thermal expansion anisotropy of aluminium titanate can be observed in sample (c) sintered at 1550 °C (the lighter phase is A12TiO5 and the gray phase is Al2O3), while the other samples ((a) and (b)) are microcrack-free composites. Clearly, the grain size increases with temperature enhancement from 1450 to 1550 °C. As indicated by the arrow in Fig. 6b, the surfaces of alumina grains were covered with aluminium titanate. With regard to Eq. (1), the critical grain size (Dcr) for spontaneous microcracking of Al2TiO5 was reported about 2 μm for monolithic aluminium titanate body [11,26]. While, for alumina– aluminium titanate composite, the maximum value of Δα (determined from A1 2O3/Al2TiO5 grains) is lower than that of measured for monolithic Al2TiO5 grains. Additionally, the γf of A12O3/Al2TiO5 grain boundary is greater than that of Al2TiO5 and so Dcr is predicted to be higher than 2 μm in A/AT composites. The decrease in strength from samples (a) to (b) can be attributed to grain growth occurring due to temperature rise. Further decrease in the strength of sample (c) is related to spontaneous microcraking damage due to thermal expansion anisotropy. The HR-XRD analysis results of the (020) crystallographic planes and their related residual strains and lattice parameter, b, of Al2TiO5 for composites (a), (b) and (c) sintered at different temperatures are illustrated in Figs. 7 and 8. The residual strains induced by the thermal expansion mismatch between the aluminium titanate and alumina phases in the composite can contribute to line-shifts in XRD analysis. So the residual strength could be calculated from the following equation [27]:

Residual Strain % ¼

dref −dsample  100 dref

ð4Þ

Relative Density (%)

99.0

300

98.5 200 98.0 100

97.5

1450

1500

1550

Sintering Temperature (°C) Fig. 5. Density and strength variations with sintering temperature.

0

Strength, σ (MPa)

400

99.5

97.0 1

Fig. 6. SEM micrographs of A–20AT composites sintered at 1450 (a), 1500 (b) and 1550 °C (c).

where dref was extracted from ASTM File for aluminium titanate [14] and dsample, refining the zero-point shift, was measured from HR-XRD analysis. The measured lattice parameter, b, for samples sintered at lower temperature (Fig. 8) is bigger than the original value that is reported in previous studies [14,28]. This discrepancy may be attributed to the thermal expansion mismatch between alumina and aluminium titanate grains. The density degradation in the A–20AT composites sintered at high temperature is due to the grain cracking and the residual strain of AT grains released during cracking. Considering the most positive expansion of the b-axis, in comparison with a and c axes, the aluminium titanate grains cracked along the (010) crystallographic planes [8]. Therefore, the lattice parameter of sample (c) sintered at higher temperature reverts to initial value as a result of grain or grain boundary cracking.

M. Sobhani et al. / Int. Journal of Refractory Metals and Hard Materials 47 (2014) 49–53

53

measurements showed that the density of samples sintered at 1550 °C was lower than those sintered at 1450 °C. This phenomenon may relate to grain cracking. The lattice parameter of samples sintered at lower temperatures was bigger than those sintered at higher temperatures. References

Fig. 7. HR-XRD patterns of composites sintered at different temperatures.

Conclusions The results of this work show that the heating rate, soaking time and sintering temperature affect the formation of aluminium titanate in A/ AT composites. Thoroughly, the aluminium titanate phase content decreased during sintering temperature and heating rate increment. In contrast, it increased during sintering temperature and heating rate decrement. It was found that thermal expansion anisotropy interaction between alumina and aluminium titanate grains influences the densification behavior of these composites. Furthermore, the strain of lattice parameters resulted from thermal expansion mismatch can remain after cooling from sintering temperature to room temperature. Densification

Fig. 8. The residual strain and lattice parameter, b, of composites (a), (b) and (c) sintered at different temperatures.

[1] Runyan JL, Bennison SJ. Fabrication of flaw-tolerant aluminum-titanate reinforced alumina. J Eur Ceram Soc 1991;7:93–9. [2] Lawn BR, Padture NP, Braun LM, Bennison SJ. Model for toughness curves in twophase ceramics: I, Basic fracture mechanics. J Am Ceram Soc 1993;76(9):2235–40. [3] Padture NP, Runyan JL, Bennison SJ, Braun LM, Lawn BR. Model for toughness curves in two-phase ceramics: II. Microstructural variables. J Am Ceram Soc 1993;76(9):2241–7. [4] Hu Y, Li M, Shen Q, Xu W. Thermal shock-resistance performance of Al2TiO5/Al2O3 composites. Adv Mater Res 2009;66:108–11. [5] Baglay RD, Cutler IB, Johnson DL. Effect of TiO2 on initial sintering of Al2O3. J Am Ceram Soc 1970;53(3):136–41. [6] Sathiyakumar M, Gnanam FD. Influence of MnO and TiO2 additives on density, microstructure and mechanical properties of Al2O3. Ceram Int 2002;28:195–200. [7] Buscaglia V, Nanni P. Decomposition of Al2TiO5 and Al2(1 − x)MgxTi(1 + x)O5 ceramics. J Am Ceram Soc 1998;81(10):2645–53. [8] Bruno G, Efremov AM, Wheaton BR, Webb JE. Microcrack orientation in porous aluminum titanate. Acta Mater 2010;58(20):6649–55. [9] Skala RD, Li D, Low IM. Diffraction, structure and phase stability studies on aluminium titanate. J Eur Ceram Soc 2009;29:67–75. [10] Kuszyk JA, Bradt RC. Influence of grain size on effects of thermal expansion anisotropy in MgTi2O5. J Am Ceram Soc 1973;56(8):420–3. [11] Cleveland JJ, Bradt RC. Grain size/microcracking relations for pseudobrookite oxides. J Am Ceram Soc 1978;61(11–12):478–81. [12] Bueno S, Berger MH, Moreno R, Baudin C. Fracture behaviour of microcrackfreealumina–aluminium titanate ceramics with second phase nanoparticles at alumina grain boundaries. J Eur Ceram Soc 2008;28:1961–71. [13] Ohya Y, Nakagawa Z. Measurement of crack volume due to thermal expansion anisotropy in aluminium titanate ceramics. J Mater Sci 1996;31:1555–9. [14] ASTM File 26-0040. Database PDF (Powder Diffraction File). [15] ASTM File 21-1276. Database PDF (Powder Diffraction File). [16] ASTM File 42-1468. Database PDF (Powder Diffraction File). [17] Mani J, Ananthakumar S, Mukundan P, Warrier KG. Low-temperature fine-grained alumina–aluminum titanate composites produced from a titania-coated alumina precursor. J Am Ceram Soc 2007;90(10):3091–4. [18] Low IM, Oo Z. Reformation of phase composition in decomposed aluminium titanate. Mater Chem Phys 2008;111(1):9–12. [19] Jayasankar M, Ananthakumar S, Mukundan P, Warrier KGK. Low temperature synthesis of aluminium titanate by an aqueous sol-gel route. Mater Lett 2007;61:790–3. [20] Freudenberg B, Mocellin A. Aluminum titanate formation by solid-state reaction of fine A12O3 and TiO2 powders. J Am Ceram Soc 1987;70(1):33–8. [21] Sobhani M, Sedaghat A, Ebadzadeh T, Ebrahimi M. Preparation of nano-sized Mg0.6Al0.8Ti1.6O5 powders using the inorganic salts route. Ceram Int 2013;39:6899–905. [22] Sobhani M, Rezaie HR, Naghizadeh R. Sol-gel synthesis of aluminum titanate (Al2TiO5) nano-particles. J Mater Process Technol 2008;206:282–5. [23] Freudenberg B, Mocellin A. Aluminum titanate formation by solid-state reaction of coarse A12O3 and TiO2 powders. J Am Ceram Soc 1988;71(1):22–8. [24] Winkler ER, Sarver JP, Cutler IB. Solid solution of titanium dioxide in aluminum oxide. J Am Ceram Soc 1966;49(12):634–7. [25] Grimes RW. Solution of MgO, CaO, and TiO2 in α-Al2O3. J Am Ceram Soc 1994;77(2):378–84. [26] Ohya Y, Nakagawa Z, Hamano K. Grain-boundary microcracking due to thermal expansion anisotropy in aluminum titanate ceramics. J Am Ceram Soc 1987;70(8): C184–6. [27] Singh M, Manurung P, Low IM. Depth profiling of near-surface information in a functionally graded alumina/aluminium titanate composite using grazingincidence synchrotron radiation diffraction. Mater Lett 2002;55:344–9. [28] Morosin B, Lynch RW. Structure studies on Al2TiO5 at room temperature and at 600 °C. Acta Crystallogr B 1972;28:1040–6.