Critical zirconia amount to enhance the strength of alumina

(1)

Critical zirconia amount to enhance the strength of alumina

Wei-Hsing Tuan

*

, Jiang-Rung Chen, Chang-Ju Ho

Department of Materials Science and Engineering, National Taiwan University, Taipei 106, Taiwan Received 23 July 2006; received in revised form 24 April 2007; accepted 20 August 2007

Available online 29 September 2007

Abstract

A small amount of zirconia particles (5 vol.%) is added into alumina in the present study. The grain size of alumina is reduced; the strength of alumina is therefore enhanced. Though the theoretical analysis demonstrates that an addition of 1 vol.% of fine zirconia particles is sufficient to prohibit the coarsening of alumina grains; the experimental measurements indicate that a minimum amount of 2 vol.% is required to reduce the coarsening of alumina matrix and its size distribution. This discrepancy is due to the separation between the zirconia particles and alumina grain boundaries, which takes place when the alumina grain size increases above a critical value.

Keywords: A. Sintering; B. Microstructure-final; C. Strength; D. Al2O3; D. ZrO2

1. Introduction

Ceramics are brittle in nature. Such brittleness limits many applications involving using ceramics as engineering compo-nents. To improve the toughness of ceramics, there has been much work in the last several decades. Since an increase of strength also improves the possibility of brittle ceramics to survive external impacts, increasing the strength of ceramic is also an important task.

Several relationships have been derived to describe the strength, s, of brittle ceramics[1,2]. One relationship is that proposed by Griffith:

s¼ KIC

YpffiffiffiffiC (1)

In the above equation, KIC is the toughness, Y a geometrical

constant, and C is the flaw size. The above relationship demon-strates that the strength of a ceramic depends strongly on its flaw size and grain size.

The reduction of flaw size can be achieved by controlling each processing step carefully[3]. The decrease of grain size can be reached through the addition of second phase particles

[4–8]. The inhibitors can be oxides [5], non-oxides [4,6] or metallic particles [7,8]. These particles should be relatively

inert to the ceramic matrix during sintering at elevated temperature. Among the composite systems investigated, alumina–zirconia systems have attracted wide attention. The solubility between alumina and zirconia is low (2000 ppm at 1450 8C[9]), and the engineering potential of the composites is high. In the present study, the alumina-zirconia system is also used as a model system to explore the rule of microstructure design for strength enhancement.

The first study on the Al2O3–ZrO2 system employed a

minimum amount of 2 vol.% ZrO2 particles to enhance the

toughness of Al2O3 [5]. The study demonstrated that the

toughness of Al2O3 could be enhanced only when the ZrO2

content was higher than 5 vol.%. Many studies were then performed to confirm this approach[10–14]. Most studies have added more than 5 vol.% ZrO2 into Al2O3. They have

demonstrated that both the toughness and strength of alumina are improved; nevertheless, the sinterability of the ceramic matrix is reduced when a large amount of second phase is added

[15]. Furthermore, the cost of the composites is also increased. In order to cope with the drop of sintering activity and the increase of cost, a small amount, <5 vol.%, of second phase is therefore of interest. Such an approach has received little attention in previous studies. One limited study indicated that the properties were not improved by using a small amount, <5 vol.%, of particles[5]. The dilemma on designing ceramic matrix composite is thus apparent. On the one hand, as the second phase content is low, the sintering activity is little affected. But the properties improvement is suspected to be also

www.elsevier.com/locate/ceramint Available online at www.sciencedirect.com

Ceramics International 34 (2008) 2129–2135

* Corresponding author. Tel.: +886 2 23659800; fax: +886 2 23634562. E-mail address:tuan@ccms.ntu.edu.tw(W.-H. Tuan).

(2)

small (it will be demonstrated that this may not be the case). On the other hand, when the second phase content is high, the mechanical properties are improved. However, the cost increase associated with the material hinders the application potential of the composites.

In the present study, a small amount of ZrO2 particles,

5 vol.%, is added into Al2O3. The microstructure of the

composites is carefully investigated. The minimum required amount of second phase is suggested in terms of achieving strength enhancement.

2. Experimental

An alumina (TM–DAR, d50= 140 nm, Taimei Chem. Co.

Ltd., Nagno-ken, Japan) powder was ball milled together with 0–5 vol.% zirconia powder (TZ-3Y, ZrO2+ 3 mol.% Y2O3,

d50= 230 nm, Tosoh Co., Japan) in ethyl alcohol for 24 h, using

10 mm diameter zirconia balls as grinding media. The slurry of the powder mixtures was dried with a rotary evaporator, and then the dried lumps were crushed and passed through a plastic sieve. Powder compacts with dimensions of 7 mm 6 mm 50 mm were formed by uniaxially pressing at 44 MPa. Sintering was carried out in a box furnace at 1600 8C for 1 h in air with heating rate of 5 8C/min. A number of disc samples with a diameter of 6 mm in diameter were prepared for a kinetic study with a dilatometer (SETSYS 1600, TMA, Setaram Co., Caluire, France). The heating rate for the dilatometer was also 5 8C/min.

The sintered rectangular bars were ground longitudinally with a 325 grit resin-bonded diamond wheel at a depth of 5 mm per pass. The final dimensions of the specimens were 3 mm 4 mm 36 mm. The strength of the specimens was determined using a four-point bend test carried out at ambient conditions. The upper and lower spans were 10 mm and 30 mm, respectively. The rate of loading was 0.5 mm/min. The fracture toughness was determined by the single-edge-notched-beam (SENB) technique. The notch was generated by cutting with a diamond saw. The width of the notch was approximately 0.3 mm.

Phase identification was performed on sintered and fractured surfaces by X-ray diffractometry (XRD) with Cu Ka radiation. The relative phase content of zirconia was estimated by using the method proposed by Evans et al.[16]. The final density of the specimens was determined by the Archimedes method. The solubility between the materials used in the present study was low; the relative density of the sintered composites was estimated by using the theoretical density of 3.98 g/cm3 for Al2O3 and 6.05 g/cm3 for ZrO2. Polished

surfaces for microstructure observation were prepared by grinding and polishing with diamond paste to 6 mm and with silica suspension to 0.05 mm. The polished specimens were thermally etched at 1450 8C for 0.5 h to reveal the grain boundaries of matrix grains. Microstructure characterization was conducted using scanning electron microscopy (SEM). Image analysis was conducted on SEM micrographs to determine the average size of Al2O3 grains and their size

distribution.

3. Results

The density of the green powder compacts increases slightly from 59.5% to 60.5% with the ZrO2content varies from zero to

five vol.% (seeFig. 1). The densities of the sintered Al2O3/ZrO2

composites are all higher than 99%. Fig. 2 shows the XRD patterns of the Al2O3 and Al2O3/5% ZrO2 composite. XRD

analysis shows only Al2O3 and ZrO2 phases in the sintered

Al2O3/ZrO2composites.Fig. 3(a) shows the linear shrinkage of

the Al2O3/ZrO2composites during sintering. The presence of

the ZrO2 particles delays the shrinkage of Al2O3 in the

temperature range from 1000 to 1400 8C. However, the temperature at the peak densification rate (the second peak in Fig. 3b) remains the same. The densification for all the composites is almost complete as the temperature reaches 1450 8C.

Fig. 4shows typical SEM images of the specimens. These micrographs demonstrate that the addition of ZrO2 particles

Fig. 1. Green and sintered density of the Al2O3/ZrO2composites.

(3)

reduces the size of Al2O3grains (seeTable 1). In the Al2O3/1%

ZrO2 composite (see Fig. 4b), most ZrO2 particles are

embedded within the Al2O3 grains. For the composites

containing more than 2 vol.% ZrO2, the particles are mainly

at the boundaries of the Al2O3matrix grains (seeFig. 4c–f). The

size of ZrO2particles is about 0.3 mm which is more or less the

same in all the Al2O3/ZrO2 composites. Nevertheless, some

ZrO2agglomerates are found occasionally (see Fig. 4e). The

ZrO2grains in the agglomerate are larger than the isolated ZrO2

particles.Fig. 5shows the size distribution of Al2O3grains for

the Al2O3and Al2O3/ZrO2specimens. When the ZrO2content

is higher than 2 vol.%, the addition of a small amount of ZrO2

reduces not only the average size of Al2O3grains but also their

distribution. In the Al2O3/ZrO2 composite with 1% ZrO2,

several large grains are still present (Figs. 4b and 5).

Fig. 6 shows the flexural strength of the Al2O3/ZrO2

composites as a function of ZrO2 content. The flexural

strength of the composites increases with the increase of ZrO2

content. Apart from the dependence on the ZrO2content, the

strength as expected also shows strong dependence on the size of Al2O3 grains (see Fig. 7). Fig. 8 shows the fracture

toughness of the Al2O3/ZrO2composites as a function of ZrO2

content. The XRD analysis shows that around 30% of the ZrO2

particles on the fracture surface of the Al2O3/5% ZrO2

composites is transformed from tetragonal to monoclinic phase, indicating that transformation toughening is active in the composites.

4. Discussion

Two peaks, a small peak at 1160 8C and a larger one at 1400 8C, are found in the sintering kinetic curves (seeFig. 3b). The first peak can be related to the presence of Al2O3

agglomerates due to the fact that intra-agglomerate pores are usually small [17]. The addition of more than 2 vol.% ZrO2

particles reduces the height of the first peak, suggesting that the Al2O3agglomerates are dispersed due to the presence of ZrO2

particles. The reduction of agglomerate can contribute to the reduction of flaw size [3]; the strength of the alumina is therefore enhanced.

Through microstructure observation shown in Fig. 4

indicates that the ZrO2particles act as effective grain growth

inhibitors to Al2O3. The size of alumina grains reduces from

5.5 mm (for pure alumina) to 2.2 mm (for Al2O3/5% ZrO2

composite). Except the Al2O3/1% ZrO2 composite, the size

scatter of the Al2O3grains in the composites is also reduced.

Most ZrO2 particles are well separated within the Al2O3

matrix. Since the coarsening of ZrO2particles can then take

place only through the diffusion within the Al2O3matrix, this is

a relatively slow process. Therefore, the size of zirconia particles is more or less the same in all the Al2O3/ZrO2

composites, indicating that the coarsening of well-dispersed ZrO2 inclusions is limited. Such limited coarsening can be

related to the low ZrO2content used in the present study. The

coarsening of ZrO2 particles is found only in the ZrO2

agglomerate (see Fig. 4e).

For a composite containing mono-sized inclusions, the grain size decreases with the increase of particle content (Zener effect). The nearest neighbor distance (or mean free path), l, between randomly distributed particles can be estimated with the following relationship proposed by Westmacott et al.[18]

and Kock[19]; l¼ _p 6 1=2 d F1=2 (2)

where d is the size of particle and F is the volume fraction. The calculated values for the distance between nearest neighboring ZrO2particles in Al2O3matrix are shown inTable 1. The table

indicates that the mean free path between the ZrO2particles in

all the Al2O3/ZrO2composites is shorter than the corresponding

size of the Al2O3grains.

Fig. 3. (a) Percent shrinkage and (b) densification rate as function of tempera-ture for various Al2O3/ZrO2composites.

(4)

It is thus of interest to estimate the minimum amount of second phase particles needed to prohibit the growth of matrix grains. By assuming that one particle is enough to prohibit the movement of one grain boundary, then l G (G = size of matrix grains). The amount of particle needed to prohibit the growth of matrix grains can be estimated by re-arranging Eq.(2)as,

F1=2¼ p 6 1=2 d G (3)

For the microstructure shown in Fig. 9(a), each grain boundary of matrix grains (white grains) is pinned by one

particle (dark particle). As the size of particle is the same as that of matrix grains, the particle content is around 50 vol.%. Though the microstructure should be stable throughout the sintering process, such high particle content can, not only prohibit the grain growth but also limit the sintering activity of the ceramic matrix grains. Therefore, the amount of the second phase should be lower than 50 vol.%.

As the particle size is reduced to one tenth that of matrix grains,Fig. 9(b), a volume fraction of 0.5 vol.% is obtained by using Eq.(3). If the d/G ratio is further decreased to 0.01, the amount of second phase can then be reduced to 0.005 vol.%. The analysis implies that the amount of 1 vol.% particle may be

Fig. 4. SEM images of the (a) Al2O3, (b) Al2O3/1% ZrO2, (c) Al2O3/2% ZrO2, (d) Al2O3/3% ZrO2, (e) Al2O3/4% ZrO2, and (f) Al2O3/5% ZrO2composites. An

agglomerate of ZrO2in (e) is indicated with arrow.

Table 1

Microstructure characteristics of the Al2O3/ZrO2composites

Average size of Al2O3grains (mm)

Standard deviation/ coefficient of variationa

Mean free path between ZrO2particlesb(mm) Al2O3 5.5 2.5/45% – Al2O3/1% ZrO2 3.8 1.9/50% 2.2 Al2O3/2% ZrO2 3.0 1.2/40% 1.5 Al2O3/3% ZrO2 3.0 1.2/40% 1.3 Al2O3/4% ZrO2 2.6 1.0/38% 1.1 Al2O3/5% ZrO2 2.2 0.76/35% 1.0 a

Coefficient of variation = standard deviation/average value.

b

(5)

more than enough to prohibit the grain growth of Al2O3matrix.

However, as the d/G value is too low, it may encourage the separation of the particle and grain boundary, as demonstrated in the Al2O3/1 vol.% ZrO2 composite (see Fig. 4b). In the

composite, most ZrO2 particles are separated from the grain

boundaries of Al2O3grains. The ZrO2particles with such low

amount can no longer prohibit the growth of matrix grains; several coarse grains are thus survived after sintering (see

Fig. 5). In order to prohibit the movement of grain boundary, the particle is preferably located at the grain boundary as the case shown inFig. 9(a) and (b). The case as demonstrated inFig. 9(c) should be avoided. Therefore, the criteria for the separation of secondary particles and grain boundary should be investigated.

During sintering, with the help of a number of transport mechanisms the matrix grains can grow with the help from the movement of grain boundary. The velocity of the grain boundary movement is virtually the same as the grain growth rate. As the velocity of grain boundary, vb, is much higher than

that of particle, vp, the separation can take place (seeFig. 10).

The following relationship defines the criteria for the separation to take place:

vb> vmaxp (4)

In order for the particle and grain boundary to move, the velocity is contributed by mobility, M, and force, F, as

v¼ MF (5)

Fig. 6. Flexural strength of the composites as function of ZrO2content.

Fig. 7. Flexural strength of the Al2O3/ZrO2composites as function of Al2O3

grain size.

Fig. 8. Fracture toughness of the composites as function of ZrO2content.

(6)

For the grain boundary, its mobility is controlled by the diffusion across the grain boundary, D*, the Boltzmann’s constant, k, and the absolute temperature, T. The force on the grain boundary is controlled by the grain boundary energy, ggb,

and grain size, G. The velocity of boundary, vb, can then be

expressed as[20]; vb¼ MbFb¼ D kT g_gb G (6)

For the particle, mass can transport along the interface or the interior of the particle or both. Therefore, an effective diffusion coefficient, Deff, to illustrate the mobility of the particle is used

instead as[20];

Mp¼

DeffV

kTdn (7)

where V is the volume of controlling ions and n is a constant depending on the mechanism. The force on the particle is also provided by the grain boundary energy. By using the relation-ship to estimate the force on an isolated pore[20], the force on a particle is estimated as

Fp¼ 2pdggb (8)

The velocity of particle, vp, is

vp ¼ MpFp¼

DeffV

kTdn 2pdggb (9) By comparing Eqs.(6) and(9), the following equation is obtained:

D 2pDeffV

> G

dn1 (10)

The analysis above demonstrates that there is a critical value for the ratio of grain size over particle size. Since the grains grow rapidly as the density is higher than around 80%

[20]. Along with the grain growth, the ratio of grain size over particle size also increases. Since the coarsening of particles is limited as the second phase amount is low, the separation between the grain boundary and particle is virtually controlled by the growth of matrix grains. From the microstructure observation, the critical size for the Al2O3grain to separate

from the ZrO2 particle is around 4 mm. Since the ZrO2

particles in all the Al2O3/ZrO2composite are similar in size,

the increase of ZrO2 content increases the number of ZrO2

particles. More ZrO2particles at each grain boundary exert

higher dragging force to the movement of grain boundary. The size of alumina grains is thus reduced with the increase of ZrO2content.

Fig. 9. Interactions between grain boundary and second phase particles; showing (a) both matrix and the particles are the same size and (b) much smaller interboundary particles. The particles in (c) fail to pin the grain boundaries due to the matrix grains are larger than a critical value.

(7)

The analysis above demonstrates that there is a critical amount for the second phase to prohibit the coarsening of matrix grains. The critical amount of the second phase is a function of the ratio of grain size over particle size. In the present study, it demonstrates that the critical amount of second phase depends on the growth of matrix grains, provided that the coarsening of particles is limited. For the present system under the processing conditions used, the critical amount of ZrO2

particles is 2 vol.%. The analysis also suggests that the decrease of particle size can further encourage the separation of grain boundary and particles. If the particles are preferable to be swallowed by the matrix grains in order to leave behind wavy grain boundaries, as is the case for the Al2O3/SiC

nanocompo-site [21], such target possibly can be achieved either by reducing the particle size or by increasing the grain size through the rise of sintering temperature.

5. Conclusions

There are several advantages of using a small amount (<5 vol.%) of ZrO2 particles as the strengthening agent to

Al2O3ceramics.

1. The addition of the fine ZrO2particles reduces the amount of

Al2O3agglomerates.

2. The addition of a small amount of ZrO2particles can also

reduce the size of Al2O3matrix grains.

3. The strength of Al2O3is enhanced due to the reduction of

grain size.

4. The coarsening of particles within ceramic matrix is limited as the particle content is lower.

5. The critical amount of particles to prohibit the coarsening of matrix grains is a function of grain size over particle size.

Acknowledgments

The present study was supported by the National Science Council through the contract numbers of NSC94-2216-E-002-014 and NSC94-2216-E-002-016. Valuable comments given by Prof. Jay Shieh, Dept. of Mater. Sci. & Eng., National Taiwan University, are helpful.

References

[1] R.W. Rice, Microstructure dependence of mechanical behavior of cera-mics, Treatise. Mater. Sci. Tech. 11 (1977) 199–381.

[2] W.H. Tuan, M.J. Lai, M.C. Lin, C.C. Chan, S.C. Chiu, The mechanical performance of alumina as a function of grain size, Mater. Chem. Phys. 36 (1994) 246–251.

[3] W.H. Rhodes, Agglomerate and particle size effect on sintering yttria-stabilized zirconia, J. Am. Ceram. Soc. 64 (1981) 19–26.

[4] G.C. Wei, P.F. Becher, Development of SiC-whisker-reinforced ceramics, Am. Ceram. Soc. Bull. 64 (1985) 298–304.

[5] N. Claussen, J. Steeb, R.F. Fabst, Effect of induced microcracking on the fracture toughness of ceramics, Am. Ceram. Soc. Bull. 56 (1977) 559– 562.

[6] K. Niihara, New design concept of structural ceramic–ceramic nanocom-posites, J. Ceram. Soc. Jpn. 99 (1991) 974–982.

[7] W.H. Tuan, R.J. Brook, The toughening of alumina with nickel inclusions, J. Eur. Ceram. Soc. 6 (1990) 31–37.

[8] T. Sekino, T. Nakajima, S. Ueda, K. Niihara, Reduction and sintering of a nickel-dispersed-alumina composite and its properties, J. Am. Ceram. Soc. 80 (1977) 1139–1148.

[9] R. Majumdar, E. Gilbart, R.J. Brook, Kinetics of densification of alumina– zirconia ceramics, Br. Ceram. Trans. J. 85 (1993) 156–160.

[10] F.F. Lange, Transformation toughening, IV, fabrication, fracture toughness and strength of Al2O3–ZrO2composites, J. Mater. Sci. 17 (1982) 247–254.

[11] P.F. Becher, Slow crack growth behavior in transformation-toughened Al2O3–ZrO2(Y2O3) ceramics, J. Am. Ceram. Soc. 66 (1983) 485–488.

[12] M. Ruhle, A.G. Evans, R.M. McMeeking, P.G. Charalambides, J.W. Hutchinson, Microcrack toughening in alumina/zirconia, Acta Metall. 25 (1987) 2701–2710.

[13] M. Ruhle, A.G. Evans, High toughness ceramics and ceramic–matrix composites, Prog. Mater. Sci. 33 (1989) 85–167.

[14] R.H.J. Hannink, P.M. Kelly, B.C. Muddle, Transformation toughening in zirconia-containing ceramics, J. Am. Ceram. Soc. 83 (2000) 461–487. [15] O. Sudre, F.F. Lange, The effect of inclusions on densification: I,

micro-structural developments in an Al2O3matrix containing a large volume

fraction of ZrO2inclusion, J. Am. Ceram. Soc. 75 (1992) 519–524.

[16] P.A. Evans, R. Stevens, J.P. Binner, Quantitative X-ray diffraction analysis of polymorphic mixes of pure zirconia, Br. Ceram. Trans. J. 84 (1984) 39–43. [17] A. Roosen, H. Hausner, Sintering kinetics of ZrO2powders, Adv. Ceram.

12 (1984) 714–726.

[18] K.H. Westmacott, C.W. Fountain, R.J. Stirton, On the spacing of dispersed obstacles, Acta Metall. 14 (1966) 1628–1629.

[19] U.F. Kocks, On the spacing of dispersed obstacles, Acta Metall. 14 (1966) 1629–1631.

[20] R.J. Brook, Controlled grain growth, Treatise. Mater. Sci. Tech. 9 (1976) 331–364.

[21] M. Sternitzke, Review: structural ceramic nanocomposites, J. Eur. Ceram. Soc. 17 (1997) 1061–1082.