Creep behavior of multi-cation α-SiAlON partially stabilized, produced with an yttrium- rare earth oxide mixture (CRE2O3)

C. Santos1; K. Strecker1; M.J. R. Barboza1; S. A. Baldacim2; F. Piorino Neto2; O. M.M.Silva2; C.R.M.Silva2

1 Departamento de Engenharia de Materiais DEMAR-FAENQUIL, Polo Urbo-industrial, gleba AI-6, s/n, cep 12600-000, Lorena-SP Brazil. 2 Centro Técnico Aeroespacial – Divisão de Materiais AMR-CTA, Pça. Marechal do Ar Eduardo Gomes, 50, cep. 12228-904, S. J. Campos –SP Brazil

Key-words: α-SiAlON, Si3N4, sintering additives, CRE2O3, compressive creep.

ABSTRACT

α−SiAlON (α’) is a solid solution of α−Si3N4, where Si and N are substituted by Al and O, respectively. The principal stabilizers of the α’-phase are Mg, Ca, Y and rare earth cations. In this way, the possible use of the yttrium-rare earth oxide mixture, CRE2O3, produced at FAENQUIL, in obtaining these SiAlONs was investigated. Samples were sintered by hot- pressing at 17500C, for 30 minutes, using a sintering pressure of 20 MPa. Creep behavior of the hot-pressed CRE-α-SiAlON/β-Si3N4 ceramic was investigated, using compressive creep tests, in air, at 1280 to 13400C, under stresses of 200 to 350 MPa, for 70 hours. This type of ceramic exhibited high creep and oxidation resistance. Its improved high-temperature properties are mainly due to the absence or reduced amount of intergranular phases, because of the incorporation of the metallic cations from the liquid phase formed during sintering into the Si3N4 structure, forming a α’/β composite.

1. INTRODUCTION

Silicon nitride (Si3N4) and its solid solutions (SiAlONs) are ceramics widely used for structural applications due to its physical and mechanical properties, such as high wear, hardness and creep resistance [1,2]. The liquid phase sintering of Si3N4 is governed by a solution-precipitation (α−β−Si3N4 transformation) mechanism, that occurs in temperatures 0 between 1400 and 1800 C and is related, among other factors, to the initial α−Si3N4 phase content of the starting powder, amount and type of additives used and sintering parameters such as temperature, time and atmosphere [2-3]. The utilization of the additives based on Al2O3, AlN demonstrate great potential for obtaining of SiAlONs [4,5]. The microstructural aspects and the intergranular phase content are preponderant factors for the high creep resistance of these materials [6,7]. Due to their characteristics previously mentioned, the technological applications of silicon nitride ceramics (Si3N4) are numerous, and in this context, the study and understanding of the mechanical behavior of these materials at high temperatures is of primary interest. It has been also demonstrated that the secondary glassy phases, which appear during sintering, have a strong influence on the creep resistance [8]. In these ceramics the remaining glassy phase softens at high temperatures and reduces the high temperature strength. The creep deformation of materials containing intergranular glassy phase is generally thought to occur as a combination of viscous flow, grain boundary sliding, solution-precipitation and cavitation, mechanisms [6-8]. Previous works [9,10] have shown that the yttrium-rare earth oxide mixture, CRE2O3, consisting mainly of Y2O3 (44 wt.%), Yb2O3 (17 wt.%), Er2O3 (14 wt.%) and Dy2O3 (10 wt.%) produced at DEMAR-FAENQUIL [11], is an effective and cheap substitute for pure Y2O3 as sintering additive for Si3N4 ceramics, presenting a similar sinterability and mechanical properties at room temperature behavior. The objective of the present work has been to characterize the creep behavior of the hot- pressed multi-cation α-SiAlON/β-Si3N4 ceramics produced with AlN and CRE2O3 as sintering additives.

2. EXPERIMENTAL PROCEDURE

2.1. Processing and creep tests

The powder batches prepared consisted of a mixture of 95 vol.% of commercial α−Si3N4 (HCST-Germany containing 0.1wt% Si and 5wt.% β−Si3N4), and 5 vol.% of AlN (HCST- Germany) and CRE2O3 (FAENQUIL-Brazil) as sintering additive, at a molar ratio of 9:1, respectively, in order to form a certain content of α-SiAlON with β-Si3N4 as matrix of this ceramic. The powders were mixed by planetary ball milling during 2 h using ethanol as milling media. After mixing, the powder batches were dried first in a rotary evaporator and subsequently in an oven at 120oC during 12 h. The powder mixtures were then compacted and o sintered by uniaxial hot-pressing under 20 MPa pressure in 0.1 MPa N2 atmosphere at 1750 C for 30 minutes, with a heating rate of 15 0C/minute. From the as received materials, samples of approximate 6 x 3 x 3 mm3 were cut and grounded for tests. Compressive creep tests were carried in air, at temperatures ranging between 1280 and 1340oC and under nominal stresses of 200, 300 and 350 MPa. The details of the experimental device can be found elsewhere [12]. Strain x time curves were obtained and the results were analyzed in terms of Norton‘s equation [13] as shown in Eq. (1):

• n ε ss = Aσ exp(-Q/RT) (1)

• Where, ε ss is the true steady-state strain rate, A is a constant, σ is the applied stress, T the absolute temperature and R the gas constant, n and Q are, respectively, the stress exponent and the apparent activation energy for creep in the steady-state region. The results were performed: at constant temperature and different stress, in order to evaluate the stress exponent n; at constant load and different temperatures, to evaluate the apparent activation energy Q. All tests were finished after 70 hours, without evidence of macroscopic failure.

2.2. Sample Characterization

The samples were characterized by relative density using the immersion method. Phase analysis was done by X-ray diffraction (XRD), in both samples before and after creep tests comparing the diffraction patterns with the JCPDS files. Quantitative α’/β-phase contents were evaluated using the procedure proposed by Gazzara et al [14]. XRD analysis was conducted on a plane parallel to the hot pressing direction in a bulk specimen. The ratio of β- phase (101)/(210) peak area in the XRD patterns was used as an indication of the orientation degree of elongated grains after creep tests [15,16]. For microstructural analysis, using scanning electron microscopy (SEM), the samples were chemically etched by a 1:1 mixture of NaOH and KOH at 500 oC, for 4 minutes. By SEM analysis on a plane parallel to the hot pressing direction in the crept sample, the reorientation of the grains was observed.

3. RESULTS AND DISCUSSION

3.1. Samples Characterization

The sintered samples presented average relative density of 96%. This result is related, mainly, to the reduction of the liquid phase amount in the sintering, during the solution- precipitation stage, where atoms of Al and O, besides cations Y+3 (and other rare earth cations) incorporate in the Si3N4 structure forming solid solution, substitutional and intersticial, respectively [4,5]. This reduction of the additive content hinders the obtaining of full density, due to the difficulty of performance of the densification mechanisms, in systems with little or any amount of liquid phase. XRD analysis identify the presence of α-SiAlON and β-Si3N4 as crystalline phases, in relative amounts of 10% and 90%, respectively.

3.2. Creep behavior

Table 1 presents steady-state creep rates in the appraised conditions in this work. Typical creep curve obtained in this work is presented in the Fig. 1.

Table 1– Creep parameters of samples tested between 1280 and 1340oC, under stress of 200, 300 and 350 MPa. Tests conditions • Creep Rates (ε ss ) o Temperature ( C) Stress (MPa) • -1 • -1 ε ss (h ) ε ss (s ) 1280 300 4.2 x 10-4 1.2 x 10-7 1300 200 2.1 x 10-4 5.9 x 10-8 1300 300 6.1 x 10-4 1.7 x 10-7 1300 350 7.5 x 10-4 2.1 x 10-7 1340 300 1.7 x 10-3 4.6 x 10-7

0,12 1300oC / 300 MPa 0,10

. -4 0 0,08 εSS= 6.1 x 10 (1300 C - 300 MPa) (mm/mm) 0,06

Strain 0,04

0,02

0,00 0 10203040506070 Time (h)

Fig. 1 - Typical creep curve obtained by compressive test, at 1300oC and 300 MPa.

Silicon nitride based ceramics (Si3N4), usually suffer creep deformation due to the viscous flow, with sliding of grain boundary, solution-precipitation and cavitation [6-8,16]. The creep results showed in Table 1, indicate that the α-SiAlON/β-Si3N4 (α‘/β) composites, produced in this work, present lower steady-state creep rates that β-Si3N4 ceramics produced with the same additive, CRE2O3, with the same additive content, and Al2O3 substituting AlN. The α‘/β composites present steady-state creep rates of 2.1 x 10-4 h-1 and 6.1 x 10-4 h-1 0 for stresses of 200 and 300 MPa, respectively, while β-Si3N4 ceramics, at 1300 C, the creep rates were 7.94 x 10-4 h-1 (200 MPa) and 1.19 x 10-3 h-1(300 MPa) [12]. Comparatively, these composites, in the same stresses and temperatures, are much more creep resistant for compression. The results indicated reductions of the creep rates among 50% (1300oC / 300 o MPa) and 75% (1300 C / 200 MPa) when AlN was used as additive, in substitution to Al2O3. The justification for such behavior can be in the microstructural characteristics and in the crystalline phases presents in this composite, that possesses a lower amount of phase intergranular. The values of steady-state creep rates of the α-SiAlON/β-Si3N4 ceramics are plotted logarithmically as a function applied stress in Fig. 2a, at 1300oC. By the standard regression technique, the value of the stress exponent n is determined to be 2.3 under a stress range of 200 to 350 MPa. Si3N4 ceramics, including SiAlONs (α‘and β‘) possess stress exponent values varying between 0.8 and 3, and activation energy between 350 and 1000 kJ/mol. These variations are due to the processing conditions, type of creep test (tension or compressive), temperature, chemical composition and amount of intergranular phase, and stress applied [6]. Lange et al [17] reported that Si3N4 material with small amount of the glassy phase showed a stress exponent n to be near to unity, corresponding to the diffusional creep. The apparent activation energy for creep determined from the steady-state creep rate to 300 MPa is 472 kJmol-1, as shown in Fig. 2b.

-6,2 -7,0 0 -6,4 a) σ = 300 MPa -7,2 b) 1300 C

-6,6 ) -7,4 ) -1 -1 -6,8 -7,6 (h (h

-7,0 ss -7,8

ss . ε

. ε n = 2.33 -7,2 -8,0 Q = 472 kJ/mol ln ln -7,4 -8,2 -7,6 -8,4 -7,8 -8,6 -4 -4 -4 -4 -4 5,2 5,3 5,4 5,5 5,6 5,7 5,8 5,9 6,1x10 6,2x10 6,3x10 6,4x10 6,5x10 1/T (K) ln σ (MPa)

Fig. 2 – a) Dependency of the steady-state creep rate on temperature under 300 MPa; b) Stress dependence of the creep rate for hot-pressed α‘-partially stabilized under 1300oC.

3.3. Characterizations

Fig. 3 shows the change in the ratio of β-Si3N4 (101)/(210) peak area of a bulk specimen before and after creep testing, respectively. It can be observed, through the XRD patterns, that the parallel plane (P plane) of the samples, before the creep tests, shows a ratio of the β-Si3N4 (101)/(210) peak areas of 1.4. These values indicate an alignment of the grains, due to hot-pressing. After the creep tests at 0 1300 C and under stress of 300 MPa, can be seen from the XRD patterns, that the β-Si3N4 (101)/(210) area ratio in the P plane increased from 1.4 to 1.7, indicating that there was a reorientation of the grains in the samples, by the rotation of the grains during creep. This suggests that physical grain rotation is a preferential mechanism for texture development and is due to viscous flow, where the viscous flow can redistribute the grain boundary glassy phase and thus enabling the adjacent grains of Si3N4 to be rearranged.

after creep tests (101)

β β(101)/(210) = 1.4 (210) β β (a.u.) α' α' α'α'

α' β β β α'

Intensity Intensity before creep tests (210) β (101) β(101)/(210) = 1.7 ' α' α' α α' βα' β β α'

25 30 35 40 45 50 2θ (degree)

Fig. 3 - XRD patterns of samples before and after creep tests at 1300 0C/300 MPa.

Microstructures of the samples before and after the creep tests, at 13000C and 300 MPa, are shown in the Fig. 4.

a) b)

2µm 2µm

Fig. 4 - Microstructure of the samples in the plane parallel to the hot-pressing direction: a) before and b) after creep tests at 1340 0C, 300 MPa for 70 hours.

No considerable grain growth and aspect ratio alterations occurred in the crept specimens, suggesting that the solution-precipitation phenomena is not the main creep mechanism under the conditions of the creep tests. Fig. 4 shows a larger alignment of the elongated β-Si3N4 grains, after creep test, in the parallel plane to the compressive loading. This behavior confirms that the grain-boundary sliding should be the main creep mechanism in the compressive tests, in the temperature and stress range used in the tests.

4. CONCLUSION

Ceramics composites based on α-SiAlON and β-Si3N4, also called partially stabilized α- SiAlON, were obtained by using of CRE2O3 with stabilizer of α’ phase. Based on the results it is verified that this type of ceramic presents better creep resistance that β-Si3N4–based ceramics, due to reduction of the intergranular phase amount during the sintering for formation of α-SiAlON. As the creep mechanism was determined to be for diffusion, inducing to the grain-boundary sliding, the reduction of intergranular phase reduced the creep deformation, improving like this the creep behavior of these ceramics.

ACKNOWLEDGMENT

The authors thank FAPESP for financial support, granted under process no. 01/08682-6.

REFERENCES

[1] A. Bellosi, in: Materials Science of carbides, nitrides and borides Ed by: Y.G. Gogotsi, R.A. Andrieuski, Kluwer Academic Publishers, Netherlands, Vol 68, (1999), p. 285-304. [2] F.L. Riley, J. Am. Ceram. Soc; 83 (2) (2000), p. 245-65. [3] K.R. Lai, T.Y. Tien, J. Am. Ceram. Soc. 76 (1) (1993), p. 91-96. [4] K.H. Jack, J. Mat. Sci., 11 (1976), p.1135-1158. [5] Z.-K.Huang, P. Greil, G. Petzow, Comunications of Am. Ceram. Soc., (1983), C96-97. [6] W.R. Cannon, T.G. Langdon, J. Mat. Sci. 18 (1), (1983), p. 01-50. [7] S.M. Wiederhorn, B.J. Hockley, J.D. French, J. Eur. Cer. Soc, (19) (1999), p. 2273-2284. [8] S.-Y. Yoon, T. Akatsu, E. Yasuda, J. Mat. Research, 11 (1) (1996), p. 120-126. [9] C. Santos, S. Ribeiro, K. Strecker, C.R.M. Silva, J. Mat. Proc. Techn. (2003), in press. [10] K. Strecker, R. Gonzaga, S. Ribeiro, M.J. Hoffmann, Mat. Lett., 45, (2000) 39-42. [11] S. Ribeiro, FAENQUIL-DEMAR, Lorena-SP, Brazil, Doctoral Thesis, (1997), 194p. [12] C. Santos, K. Strecker, M.R.J. Barboza, F. Piorino Neto, O. M. M. Silva, C. R. M. da Silva, Mat. Res. Bull. (2003), in revision. [13] R.W. Evans, B. Wilshire, “Introduction to Creep”, The Institute of Metals, London, (1993), 115p. [14] C. P. Gazzara and D. R. Messier, Ceram. Bulletin, 56 (9), p. 777-80. [15] F. Lee, K.J. Bowman, J. Am. Ceram. Soc.,75 (7) (1992), p. 1748-1755. [16] S.Y. Yoon, T. Akatsu, E.Yasuda, J. Mat. Sci., 32, (1997), p. 3813-3819. [17] F.F. Lange, B.I. Davis, D.R. Clarke, 1980, J. Mat. Sci, 15, (1980) p.601.