Strength improvement of LPS–SiC ceramics by oxidation treatment

Strength improvement of LPS–SiC ceramics by oxidation treatment

Int. Journal of Refractory Metals & Hard Materials 28 (2010) 484–488 Contents lists available at ScienceDirect Int. Journal of Refractory Metals & H...

692KB Sizes 2 Downloads 68 Views

Int. Journal of Refractory Metals & Hard Materials 28 (2010) 484–488

Contents lists available at ScienceDirect

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

Strength improvement of LPS–SiC ceramics by oxidation treatment B.G. Simba a, C. Santos a, M.J. Bondioli b,*, K. Strecker b, E.S. Lima c, M.H. Prado da Silva c a

Universidade de São Paulo – Escola de Engenharia de Lorena/Departamento de Engenharia de Materiais, USP-EEL/DEMAR, Pólo Urbo-Industrial, Gleba AI-6, s/n, Lorena – SP CEP 12600-000, Brazil b Universidade Federal de São João del-Rei, Departamento de Mecânica, UFSJ-CENEN, Campus Sto Antônio – Praça Frei Orlando 170 – Centro, S.J. del-Rei – MG CEP 36307-352, Brazil c IME – Instituto Militar de Engenharia, Pça. General Tibúrcio, 80, Praia Vermelha, Rio de Janeiro – RJ CEP 22290-270, Brazil

a r t i c l e

i n f o

Article history: Received 16 January 2009 Accepted 11 August 2009

Keywords: Silicon carbide Oxidation Parabolic growth rate Bending strength

a b s t r a c t In this work, SiC ceramics were liquid phase sintered (LPS), using AlN–Y2O3 as additives, and oxidized at 1400 °C in air for up to 120 h. Oxidation was monitored by the weight gain of the samples as function of exposition time and temperature. A parabolic growth of the oxidation layer has been observed and the coefficient of the growth rate has been determined by relating the weight gain and the surface area. The effect of oxidation on strength has been determined by 4-point bending tests. Phase analysis by Xray diffraction and microstructural observation by scanning electron microscopy indicated the formation of a uniform and dense oxidation layer. The elimination of surface flaws and pores and the generation of compressive stresses in the surface resulted in a strength increase of the oxidized samples. Ó 2009 Published by Elsevier Ltd.

1. Introduction Silicon carbide, SiC, is a ceramic material of relatively low density, very high hardness and good thermal and chemical stability. Due to these properties, SiC is widely used as abrasive and refractory material. Commonly, SiC ceramics are solid-state sintered with small additions of boron or aluminum and its compounds such as B4C, Al4C, AlN, etc. [1–3]. The small amount of these additives of about 1–2 wt% used in the sintering of SiC renders more difficult the preparation of homogeneous powder mixtures by conventional mixing operations, which may cause non-uniform densification or inhomogeneous microstructures. Furthermore, the high sintering temperatures, in the range of 2000–2200 °C may cause exaggerated grain growth. The absence of a secondary intergranular phase leads to a strong bonding of the grains and results in a low fracture toughness, turning SiC quite brittle [4,5]. Alternatively, the investigation of the liquid phase sintering of SiC started in the 1980’s by Omori and Takei [6], using Al2O3– Y2O3 mixtures as additives. Since then, the interest in liquid phase sintered SiC, LPS–SiC, has steadily risen, because this type of material offers the possibility of an increased fracture toughness by microstructural control [7–10], similar to Si3N4-based ceramics. However, the formation of a secondary intergranular glassy phase decreases the high temperature properties, specifically the creep resistance. This effect may be compensated by the use of more

* Corresponding author. Tel.: +55 (32) 33722514. E-mail address: [email protected] (M.J. Bondioli). 0263-4368/$ - see front matter Ó 2009 Published by Elsevier Ltd. doi:10.1016/j.ijrmhm.2009.08.004

refractory glass forming additives being more refractory and/or crystallization of the amorphous intergranular phase [11]. Recently, several works on the liquid phase sintering of SiC using AlN/Y2O3 mixtures as additives have been published [12,16–19]. Rixecker et al. [12] also report that a significant increase of the strength has been observed after an oxidation treatment at 1200 °C for 10 min. The strength increase has been attributed to the formation of crystal phases at the surfaces of the specimens, generating compressive stresses. The objective of this work has been to investigate the effect of time of oxidation and the amount of additive on the bending strength of LPS–SiC, using AlN–Y2O3 as sintering aid.

2. Experimental procedure 2.1. Processing As starting powders, b-SiC (BF-12, H.C. Starck), a-SiC (UF15, Lonza), AlN (type C, H.C. Starck) and Y2O3 (type C, H.C. Starck) were used. Three powder mixtures composed of b-SiC and 1 wt% a-SiC, acting as nucleating seeds for the b to a-SiC phase transformation, and AlN–Y2O3 as liquid phase forming additives were prepared. The additive amount was varied from 10 to 20 wt% and changing the AlN–Y2O3 proportion from 80 mol% AlN to 20 mol% Y2O3– 40 mol% AlN and 60 mol% Y2O3. The compositions prepared are summarized in Table 1. The powder mixtures were prepared by attrition milling at 500 rpm for 4 h, using isopropylic alcohol as vehicle and Si3N4 jar

B.G. Simba et al. / Int. Journal of Refractory Metals & Hard Materials 28 (2010) 484–488

with

Table 1 Composition of the powder mixtures. Samples

Composition (in wt%)

S10Y20 S20Y20 S20Y60

485

b-SiC

a-SiC

AlN

Y2 O 3

89 79

1 1

4.21 8.42 2.16

5.79 11.58 17.84

and balls as milling media. After sintering, the powder mixtures were dried at 100 °C for 24 h and sieved for deagglomeration. Green bodies of approximately 20  6  60 mm3 were obtained by uniaxial pressing in a steel die under a pressure of 30 MPa and by further cold isostatic pressing under 100 MPa. The samples were sintered in a graphite resistance heated furnace, (Thermal Technologies Inc., 1000–4560-FP20) in nitrogen atmosphere. The samples were placed in a graphite crucible, using a powder bed of identical composition as the samples, in order to minimize weight loss during sintering. The sintering cycle consisted in heating up in vacuum to 1000 °C with a rate of 20 °C/min, injecting 0.1 MPa N2 at this temperature, further heating up to 1600 °C maintaining this temperature for 30 min, increasing the N2 pressure to 0.2 MPa at the end of this stage and heating up to the final sintering temperature of 2080 °C, with a rate of 10 °C/min. An isothermal soaking time of 1 h was employed at the maximum sinter temperature. After sintering the plates were ground and cut into specimens of 3  4  4.5 mm3 bars for the oxidation experiments and the determination of the 4-point bending strength values.

KIC: actual fracture toughness of the material; KIC0: fracture toughness of material with zero porosity; B: constant (in the case of Si3N4 and SiC equal to 2.5); P: porosity of the material. The oxidation treatments were conducted in an electrical oven in air, suspending the samples by Kanthal wires. At least 10 samples were submitted to oxidation for each composition and oxidation exposure time investigated. Oxidation was conducted at 1400 °C for 4 and 120 h. The weight gain by oxidation was measured and analyzed by plotting the specific weight change (weight change divided by the exposed sample surface) versus exposure time. The strength of the sintered specimen was determined by 4point bending tests, using a 4-point bending device with outer space width of 40 mm and inner span width of 20 mm and a constant crosshead speed of 0.5 mm/s. The bending strength has been calculated by Eq. (E):

3 2

rB ¼  P 

J1  J2 bh

2

ðEÞ

where rB represents the 4-point bending strength [MPa], P the rupture load [N], b the sample width, h the sample height, J1 the outer span width and J2 the inner span width. 3. Results and discussion

2.2. Characterization

3.1. Sintering and mechanical properties

The density of the sintered samples was determined by the immersion method, using Archimedes principle, Eq. (A):

In Table 2 the results of the densities of the sintered specimens (qsint), relative density (qrel), weight change (Dm), Vickers’ hardness (HV), the measured fracture toughness (assuming zero porosity), KIC0, and the fracture toughness (corrected by Eq. (D)), KIC, of the sintered samples are summarized. Analyzing the results presented in Table 2, it is noted that specimen with 20 wt% additive content, S20Y20 and S20Y60, reached higher final densities than the specimens with 10 wt% additives. Fig. 1 presents a SEM micrograph of the fracture surface of S20Y60, which presents a uniform and homogeneous microstructure, composed of SiC grains with average grain size near 1.2 lm. Similar microstructures were observed with different compositions, S10Y20 and S20Y20. In consequence, the fracture toughness of the samples with higher additive content is also higher, about 5.1 MPa m1/2 (S20Y20 and S20Y60), in comparison to 4.5 MPa m1/2 (S10Y20). Vickers’ hardness is strongly influenced by the amount of secondary intergranular phase as can be concluded by comparison of the higher hardness of the samples S10Y20, containing 10 wt% additive and the lower hardness of samples S20Y20 containing 20 wt% of additives of the same AlN–Y2O3 proportions, despite its higher relative densities. The higher hardness of samples S20Y60 is attributed to their high relative densities, compensating the effect of an increased secondary intergranular phase amount. Furthermore, samples S20Y60, containing an Y2O3-rich mixture, performed better in all aspects than the other compositions studied; highest relative density, lowest weight loss during sintering and, consequently, highest hardness and fracture toughness.

qs ¼

md  qH2 O md  mi

ðAÞ

where qs represents the density of the sintered sample, md its dry weight, mi its weight immersed in water and qH2 O the density of water. The weight loss of the samples during sintering was determined by the weight change prior and after sintering. The Vickers hardness was determined on polished surfaces of the samples under a load of 10 kg applied for 30 s. The hardness HV10 was calculated by Eq. (B):

HV ¼

1:8544  P

ðBÞ

2

d

where HV represents the Vickers’ hardness, P the applied load and d the diagonal of the indentation mark. The fracture toughness was determined by measuring the length of the cracks induced by the Vickers’ indentation, according Eq. (C) [13]:

K IC ¼ 0:018  HV 

pffiffiffi a



0:4  0:5 E c  1 HV a

ðCÞ

where KIC represents the fracture toughness, HV the Vickers’ hardness, a the half diagonal of the Vickers’ indentation mark, c the crack length and E the elastic modulus (400 MPa were assumed for LPS– SiC). Eq. (C) is valid for Palmqvist type cracks if the c/a ratio is smaller than 3.5 [13]. In order to eliminate the effect of porosity on fracture toughness, Eq. (D) [20,21] has been used:

K IC ¼ K IC0 eBP

ðDÞ

3.2. Oxidation The results of the weight gain measurements (weight gain per surface area) after 120 h oxidation treatment at 1400 °C in air

486

B.G. Simba et al. / Int. Journal of Refractory Metals & Hard Materials 28 (2010) 484–488

Table 2 Properties of the sintered samples. Sample

qsint [g/cm3]

qrel [%]

Dm [%]

HV [MPa]

KIC0 [MPa m1/2]

KIC [MPa m1/2]

S10Y20 S20Y20 S20Y60

3.05 ± 0.4 3.22 ± 0.03 3.38 ± 0.07

92.74 ± 1.21 95.85 ± 0.82 98.41 ± 2.10

3.26 ± 0.17 3.06 ± 0.74 1.76 ± 0.15

18.99 ± 0.92 16.58 ± 1.01 22.10 ± 0.16

4.5 ± 0.2 5.1 ± 0.5 5.2 ± 0.5

3.7 ± 0.2 4.6 ± 0.5 4.9 ± 0.5

paring samples S10Y20 and S20Y20, which are samples of identical additive composition but increased additive contents, an increased weight gain of the samples with higher additive content is observed, see Table 3, probably due to the oxidation of AlN containing intergranular phases. Despite the higher amount of the additive, the S20Y60 samples are much more oxidation resistant. 3.3. Microstructural characterization of the oxidized samples

Fig. 1. SEM micrograph of the fracture surface of S20Y60 sintered samples.

Table 3 Specific weight gain and oxidation rate constants of the samples oxidized at 1400 °C in air, after 120 h. Sample

Specific weight gain [mg/cm2]

Oxidation rate constant, Kp [107 mg2 cm4 s1]

S10Y20 S20Y20 S20Y60

4.25 9.73 2.37

402.50 2177.06 130.01

and the parabolic rate constants of oxidation, kp, are listed in Table 3. Comparing the results presented in Table 3 it can be noted that the AlN-rich samples, S10Y20 and S20Y20, presented higher weight gains and also higher oxidation rate constants than the Y2O3-rich samples, S20Y60, indicating that the Y2O3-rich samples form a much more oxidation resistant intergranular phase. This is particularly evident comparing the samples with 10 wt% of the AlN-rich additive mixture, S10Y20, with the sample containing 20 wt% of the Y2O3-rich additive composition. Furthermore, com-

Fig. 2 shows the X-ray diffraction patterns of the sample surfaces, prior and after oxidation at 1400 °C in air for 120 h. Prior to oxidation, Fig. 2a, a and b-SiC have been identified as majority phases, while Y2O3 has also been detected as minority phase. The AlN phase has not been detected in the sintered samples, probably due to the incorporation into secondary glassy phases. Furthermore, the peak intensity of Y2O3 in sample S20Y60 is higher due to the increased Y2O3 content compared to the samples S10Y20 and S20Y20. After oxidation, see Fig. 2b, the crystal phases encountered in the surface are YAlO3 and Y2Si2O7, besides SiC. Micrographs of the oxidized samples, surface and cross-section are shown in Figs. 3 and 4. In the samples with 20 wt% additives, S20Y20 and S20Y60, pores in the oxidized layer can be observed, see Fig. 3, probably due to the formation of gases or volatilization of compounds during the oxidation treatment. Guinel and Norton [14] suggest as possible cause the reaction between SiC and SiO2, resulting in the formation of CO and/or CO2. In general, the cross-sections of the oxidized layer, Fig. 4, show a smooth oxidation layer of relatively constant thickness for all compositions studied. Furthermore, it can be seen that the thickness of the oxidized surface layer in the sample with 20 wt% AlN-rich composition of the additive, S20Y20 is much larger than in the case of the other two investigated sample compositions, in agreement with the results of the oxidation weight gain listed in Table 3. Analysis of the composition by EDS of the oxide layer revealed that the grey phase is predominantly composed of Si and O, indicating the presence of SiO2, while the white phase is composed of Y, Si and O, indicating the presence of Y2Si2O7, as identified by X-ray diffraction.

Fig. 2. XRD-patterns of the sintered samples: (a) before oxidation tests, and (b) after oxidation tests at 1400 °C.

487

B.G. Simba et al. / Int. Journal of Refractory Metals & Hard Materials 28 (2010) 484–488

Fig. 3. Surface of the oxidized samples after 120 h at 1400 °C. (a) S10Y20, (b) S20Y20, and (c) S20Y60. Fig. 4. Cross sections of the oxidized samples oxidized for 120 h at 1400 °C. (a) S10Y20, (b) S20Y20, and (c) S20Y60.

3.4. Bending strength 500 450

Bending strength (MPa) (4-point bending testing)

The bending strength of the samples prior and after oxidation treatment is shown in Fig. 5. The bending strength of the samples studied increased with oxidation, consistent with previous work [15]. In the case of the AlNrich samples, S10Y20 and S20Y20, the strength increased up to 4 h, oxidation treatment, probably as a result of the elimination of damages introduced during machining and of the closing of residual pores, Fig. 6, while after the treatment for 120 h a slight strength decrease has been observed, probably because of the formation of flaws in the oxidized surface, see Fig. 7. In the case of the Y2O3-rich sample, S20Y60, a continuous strength increase up to 120 h has been observed. The strength increase is due to crack bending of surface flaws or the oxidation layer generates a very smooth and continuous surface. Furthermore, thermal compressive stresses in the surface layer are generated by the thermal mismatch between the SiC matrix (aSiC = 3.0  106/°C) and the oxidation layer (aSiO2 = 10.0  106/°C and aY2 Si2 O7 = 6.5  106/°C).

400 350 300 250

S10Y20 200

S20Y20 S20Y60

150 100 0

10

20

100

110

120

Oxidation exposure time (hours) Fig. 5. Effect of the oxidation exposure time on bending strength of the sintered SiC ceramics.

488

B.G. Simba et al. / Int. Journal of Refractory Metals & Hard Materials 28 (2010) 484–488

The strength increase by oxidation has been attributed to the healing of surface flaws and the generation of compressive stresses in the oxidation layer. For longer oxidation treatments, the increasing thickness of the oxidation layer reduces the stress concentration thus a slight decrease in the bending strength has been observed. Acknowledgement The authors would like to thank CAPES, CNPq and FAPESP for financial support granted. References

Fig. 6. Details of the S10Y20 oxidized surface after 4 h at 1400 °C.

Fig. 7. Details of the S10Y20 oxidized surface after 120 h at 1400 °C.

The slight reduction of the bending strength observed in the samples S10Y20 and S20Y20 between 4 and 120 h oxidation is attributed to the increasing thickness of the oxidation layer, resulting in a better distribution of the stress in the surface layer and thus, reducing the beneficial effect of the compressive stress concentration. The Y2O3-rich samples, S20Y60, demonstrated a continuous increase of the strength during oxidation. In this case, the thickness of the oxidation layer is much smaller due to the higher oxidation resistance, and therefore, a significant stress distribution has not yet occurred, even after 120 h. 4. Conclusions This work demonstrated the possibility of increasing the bending strength of LPS–SiC ceramics by oxidation treatments. The amount of the additive and the additive system strongly influence the oxidation resistance of the SiC ceramics studied in this work. The Y2O3-rich additive system proved to be much more oxidation resistant than the AlN-rich system.

[1] Prochazka S. Sintering of silicon carbide. In: Burke JJ, Gorum AE, Katz RM, editors. Proceedings of the conference on ceramics for high performance applications. Hyanuis, MA: Brook Hill Publishing Co.; 1975. p. 07–13. [2] Prochazka S. The role of boron and carbon in the sintering of silicon carbide. In: Special ceramics 6. British Ceramic Res Assoc., Stoke-on-Trent; 1975. p. 171– 81. [3] Johnson CA, Prochazka S. Microstructures of sintered SiC. In: Fulrath Park S, editor. Ceramic microstructures, vol. 76; 1977. p. 366–78. [4] Kingery WD, Bowen HK, Uhlmann DR. Introduction to ceramics. 7th ed. New York: John Wiley; 1974. [5] Murata Y, Smoak RH. Densification of silicon carbide by the addition of BN, BP and B4C and correlation to their solid solubilities. In: Proceedings of the international symposium of factors in densification and sintering of oxide and non-oxide ceramics, Hakone-Japan; 1978. p. 382–99. [6] Omori M, Takei H. Pressureless sintering of SiC. J Am Ceram Soc 1982;65(6):92. [7] Padture NP. In situ-toughened silicon carbide. J Am Ceram Soc 1994;77(2):519–23. [8] Kim JY, Kim YW, Mitomo M, Zhang D, Lee JG. Microstructure and mechanical properties of a – silicon carbide sintered with yttrium–aluminium garnet and silica. J Am Ceram Soc 1999;82(2):441–4. [9] Nader M, Aldinger F, Hoffmann MJ. Influence of the ab-SiC phase transformation on microstructural development and mechanical properties of liquid phase sintered silicon carbide. J Mater Sci 1999;34:1197–204. [10] Zhan GD, Xie RJ, Mitomo M, Kim YW. Effect of b-to-a phase transformation on microstructural development and mechanical properties of fine-grained silicon carbide ceramics. J Am Ceram Soc 2001;84(5):945–50. [11] Guo S, Hirosaki N, Tanaka H, Yamamoto Y, Nishimura T. Oxidation behaviour of liquid-phase sintered SiC with AlN and Er2O3 additives between 1200 and 1400 °C. J Eur Ceram Soc 2003;23:2023–9. [12] Rixecker G, Wiedmann I, Rosinus A, Aldinger F. High-temperature effects in the fracture mechanical behaviour of silicon carbide liquid-phase sintered with AlN–Y2O3 additives. J Eur Ceram Soc 2001;21(8):1013–9. [13] Niihara K, Morena R, Hasselmann DPH. Evaluation of KIC of brittle solids by the indentation method with low crack-to-indent ratios. J Mater Sci Lett 1982;01:13–6. [14] Guinel MJF, Norton MG. Blowing of silica microforms on silicon carbide. J NonCryst Solids 2005;351:251–7. [15] Biswas K, Rixecker G, Aldinger F. Effect of rare-earth cation additions on the high temperature oxidation behaviour of LPS–SiC. Mater Sci Eng A 2004;374:56–63. [16] Izhevskyi VA, Genova LA, Bressiani AHA, Bressiani JC. Microstructure and properties tailoring of liquid-phase sintered SiC. Int J Refract Met Hard Mater 2001;19:409–17. [17] Balestra RM, Ribeiro S, Taguchi SP, Motta FV, Nunes CB. Wetting behaviour of Y2O3/AlN additive on SiC ceramics. J Eur Ceram Soc 2006;26:3881–6. [18] Taguchi SP, Ribeiro S, Balestra RM, Rodrigues Jr D. Infiltration of Al2O3/Y2O3 and AlN/Y2O3 mixes into SiC performs. Mater Sci Eng A 2007;454–455:24–9. [19] Izhevskyi VA, Bressiani AHA, Bressiani JC. Effect of liquid phase sintering on microstructure and mechanical properties of Yb2O3–AlN containing SiC-based ceramics. J Am Ceram Soc 2005;88(5):1115–21. [20] Knudsen FP. Dependence of mechanical strength of brittle polycrystalline specimens on porosity and grain size. J Am Ceram Soc 1959;42:376–87. [21] Spriggs RM. Expression for effect of porosity on elastic modulus of polycrystalline refractory materials particularly aluminum oxide. J Am Ceram Soc 1961;44:628–9.