Available online at www.sciencedirect.com
Scripta Materialia 59 (2008) 139–142 www.elsevier.com/locate/scriptamat
Two-step sintering of titania nanoceramics assisted by anatase-to-rutile phase transformation Mehdi Mazaheri,* Z. Razavi Hesabi and S.K. Sadrnezhaad Materials and Energy Research Center (MERC), P.O. Box 14155-4777, Tehran, Iran Received 30 December 2007; revised 18 February 2008; accepted 26 February 2008 Available online 6 March 2008
A remarkable suppression of grain growth was achieved by taking advantage of the anatase-to-rutile phase transformation during the final stage of a two-step sintering process. The minimum grain size obtainable via two-step sintering of a sample with fullrutile (98%) phase at the end of the first sintering step was around 250 nm, whereas the anatase-to-rutile transformation at the end of the second step facilitated a reduction in grain size to around 100 nm. Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: TiO2; Phase transformation; Two-step sintering; Grain growth
Two-step sintering is a novel technique used to suppress the accelerated grain growth that usually occurs during the final firing stage [1]. High-temperature heating followed by structural freezing via rapid cooling to a constant temperature levels off the grain growth but does not stop the densification. Triple junctions provide drag forces that control grain-boundary migration most effectively [2]. A smaller grain size at the end of the first step thus results in a higher triple junction density that pins the grain boundaries, preventing them from further advancement. There is, however, a critical density up to which the continued densification generally occurs. No remarkable densification is usually possible below this critical value of compactness [3]. The question arises as to whether this critical value of compactness can be decreased any further, perhaps thereby obtaining a finer grain distribution. In order to suppress the grain growth course of action, the sintering temperature needs lowering to an extent that no interruption to the densification occurs. Applying pressure can stimulate such a process, assisting the formation of a fully dense structure even at a very low temperature [4]. As well as pressure, an incidental enhancement in atomic mobility can also increase the densification when a phase transformation has to occur due to the inevitable heating up of the sample when sin* Corresponding author. Tel.: +98 912 1691309; fax: +98 261 4412303; e-mail addresses: mmazaheri@gmail;
[email protected]
tering takes place [5]. As a consequence, transformationassisted sintering facilitates the lowering of the sintering temperature with nanophase formation retention and near theoretical-density compaction. Several researchers have successfully applied the twostep sintering procedure to exhaust grain growth in nanoceramic specimens [1–3,6]. A few others have used transformation-assisted sintering [5,7] with the same purpose. No one has, however, designed any system of significant grain growth suppression based on simultaneous two-step sintering and phase transformation compaction. In the present study, different regimes are envisaged to reveal the role of a combined phase transformation/two-step sintering procedure on the grain growth and microstructural evolution of the titania nanoceramic exploiting the anatase-to-rutile phase transformation. TiO2 nanopowder (P25, Degussa Co., Frankfurt, Germany) with a particle size ranging from 11 to 27 nm measured via transmission electron microscopy (TEM, CM200 FEG, Philips, The Netherlands) was used. Phase analysis of as-received powder was conducted by an X-ray diffractometer (XRD, Philips X’Pert). The maximum peak intensities in the XRD pattern using Cu Ka radiation gave about 77% anatase content and 23% rutile phase. Powder samples were uniaxially pressed under 100 MPa in a rigid die (5 mm diameter). By measuring the weight and dimensions of the pressed powders, a green compactness of 53 ± 2% of the theoretical-density (TD) was obtained.
1359-6462/$ - see front matter Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.scriptamat.2008.02.041
M. Mazaheri et al. / Scripta Materialia 59 (2008) 139–142
where IA and IR are the respecty X-ray intensity of the anatase (1 0 1) and the rutile (1 1 0) peaks. Figure 1 indicates the fractional density, the fraction of the rutile phase and the grain size of a normally sintered sample vs. the sintering temperature. Note that the phase transformation starts above 600 °C and is completed at about 800 °C while no remarkable change in the grain size occurs up to 88% rutile formation. The density reaches 83% TD at 88% rutile content. The change of the density with the rutile content at an almost constant grain size up to about 90% rutile content demonstrates that while the grain size is independent of the phase change, the fractional density depends on the latter (Fig. 2). An explosive grain growth accompanied by densification occurs at the final stage of sintering where density increases from 90% TD to 98% TD. Similar
100
Fractional density Rutile content Grain size
100 Fractional density Grain size
2000
90
1500
80
1000
70
60
500
50 0 20
40
60
80
100
Rutile, %
Figure 2. Fractional density and grain size of sintered nanocrystalline TiO2 powder compacts as a function of rutile phase percentage formed during transformation under NS.
trends has previously been reported for normal sintering of 8YSZ nanopowder [6] as well as TiO2 nanoparticles [9]. To find how effectively two-step sintering would hinder grain growth regardless of any phase transformation, T1 was taken as 800 °C. The final grain size obtained through TSS1 was 7.5 times less than that made by the conventional sintering procedure. Figure 3 reveals that an active sintering mechanism causes an enhancement in density without significant grain growth having to occur. These results are comparable with those obtained by spark plasma sintering reported by other investigators [10]. In other words, in the present study, without applying any pressure, complete densification occurs at a lower temperature without any significant grain growth. Referring to the open literature, the residual pores after the first sintering step of TSS ought to be subcritical and unstable against shrinkage, but dependent on starting density. Although Kingery [11] calculated the critical pore-size theoretically by considering a specific powder particle size, there are many reports in the literature which indicate the formation of some agglomerates that retard the densification. Chen and co-workers [2] have reported, for instance, that the relative density of 75% TD after the first step is sufficient for complete densification of nanosized yttria during soaking at lower temperatures, while Bodisova et al. [12] have shown that in the case of alumina, samples with starting density of lower than 92%
400
2000 100
80
40 500
Rutile content Fractional density Grain size
96 %
1000
98 300
94
250
Grain size, nm
60
350 Grain size, nm
1500 %
Fractional density, %
Normal sintering (NS) was conducted at temperatures between 500 and 1000 °C for 1 h in air with a heating ramp of 5 °C min1. Two-step sintering consisted of: (a) heating at 5 °C min1 to T1, (b) holding for 1 h at T1, (c) cooling at 50 °C min1 to T2 = 700 °C, and (d) soaking up to 30 h at T2. To reveal the role of phase transformation on structural evolution of the sample, the amount of rutile at the end of the first step was changed. To do this, T1 was decreased from 800 °C (TSS1) to 750 °C (TSS2), yielding, respectively a full-rutile and a mixed anatase–rutile structure. The density of the sintered sample was measured by the water displacement (Archimedes) method. At least three samples were used for each density measurement and the average value was reported for each test. To track the mechanism of grain growth, sintered pellets were fractured and then characterized by SEM (Philips XL30, The Netherlands). An image analyzer was used to calculate the mean grain size of each sample. To detect the phase transformation during every sintering stage, XRD analysis was performed in the 2h range from 24° to 29° to distinguish anatase (1 0 1) occurring at 2h = 27.5° from rutile (1 1 0) occurring at 2h = 25.4°. Based on the respective peak intensities, the rutile weight fractions (x) of the samples were evaluated from the following equation [8]: 1 IA ð1Þ x ¼ 1 þ 0:8 IR
Grain size, nm
140
92
20
0 500
600
700 800 Temperature, ºC
900
200
1000
Figure 1. Fractional density, rutile content and grain size of sintered nanocrystalline TiO2 powder compacts as a function of sintering temperature under NS condition with a holding time of 1 h.
90 0
5
10 15 Holding time, h
20
25
Figure 3. Structural evolution of nanocrystalline TiO2 powder compacts sintered under TSS1 condition vs. the holding time.
M. Mazaheri et al. / Scripta Materialia 59 (2008) 139–142
TD would never be densified even after prolonged heat treatment. Interestingly, Li et al. [13] observed a significant increase in density in the second step for samples with starting density of 82% TD. The morphology, size distribution and agglomeration degree of the nanopowder used, as well as the shaping methods, strongly affect pore-size distribution within a green body specimen. Accordingly, it is not guaranteed that complete densification can be achieved even above the theoretical critical density. In order to obtain a fully dense structure, densification mechanisms should therefore be motivated. For this purpose one can use the benefits of increased atomic mobility assisted by a phase transformation occurring in a system. For instance, Kumar et al. [5] have shown that for titania nanopowder, an enhanced-sintering procedure can occur near the phase transformation temperature. It is worth noting that while the nucleation rate is comparable to growth rate, one can take advantage of phase transformation for densification and grain refinement. By using rutalization, a nearly fully dense structure with grain size of 100 nm has been produced. Figure 150
100 98
Rutile content Fractional density
96
140
%
92
120
90 110
88 86
Grain size, nm
130
94
100
84 90
82 0
5
10 15 20 Holding time, h
25
30
Figure 4. Structural evolution of nanocrystalline TiO2 powder compacts sintered under TSS2 condition vs. the holding time.
141
4 shows the structural evolution during densification at the second step of the TSS2 sintering procedure. In comparison with TSS1, the starting structure at the beginning of the second step in the TSS2 regime is composed of 88% rutile and 12% anatase, while the former is nearly a fully rutile phase. The effect of phase transformation in the second step of TSS2 on the grain size and density is shown in Figure 4. While the density increases up to 18.7%, the grain growth percentage is 16.6%. Interestingly, although the increase of density in the second step of TSS1 is half that of TSS2, the increase in grain size exceeds 32% for the former. Returning to the starting state, the higher triple junction density in finer structure in the TSS2 regime would pin grain boundaries more efficiently rather than that in TSS1 condition. Although a smaller grain size provides a higher triple junction density, there would be a minimum junction density that hinders the growth of the grains. In other words, for smaller ranges of the grain size, this effect would not be as obvious as for the larger. Phase transformation coinciding with the densification can be considered as another effective contributor in obtaining the nearly fully dense structures with finer grain sizes. To support this conclusion, one can cite investigations showing the effect of phase transformation on densification of TiO2 nanopowders to produce a fine structure at lower temperatures. Liao et al. [7] have controlled, for example, the microstructure of the nanocrystalline TiO2 by premature transformation under a high pressure (>1 GPa). They have succeeded in producing titania nanostructure by using accelerated nucleation and growth of rutile phase in the parent phase to improve densification, similar to results obtained by Kim and Kim [14]. Seemingly, by taking the advantage of anatase-to-rutile transformation during TSS of titania nanopowder, one can produce a finer structure without deterioration of the densification. Figure 5 summarizes the results of the present study through a simplified comparison of the grain size/rutile
100
Open symbols: Rutile content Solid symbols: Grain size
3000
NS 2500
TSS1
95
2000
90
1500
Rutile content, %
Grain size, nm
TSS2
1000 85 500
0
80 82
84
86
88
90
92
94
96
98
100
Fractional density, %
Figure 5. Effect of phase transformation on the ‘‘sintering path” of TiO2 nanoceramics sintered under NS, TSS1 and transformation-assisted TSS2.
142
M. Mazaheri et al. / Scripta Materialia 59 (2008) 139–142
the results obtained, one can decrease the grain size by designing a phase transformation/two-step sintering synergistic system at temperatures low enough to complete the densification by phase transformation. Figure 6 shows SEM micrographs of the samples sintered under: (a) NS at 1000 °C, (b) TSS1, and (c) transformation-assisted TSS2 condition. In summary, the present study shows the formation of nanograins during two-step sintering of titania nanoceramics assisted by the well-known anatase-to-rutile phase transformation. The grain sizes of the normally sintered samples were 1–2 lm. While the application of two-step sintering led to a remarkable decrease in the grain sizes down to 250 nm, by taking the benefits of simultaneous phase transformation two-step sintering effects, a structure with finer grain sizes of around 100 nm was obtainable.
Figure 6. SEM micrographs of the samples sintered under: (a) NS at 1000 °C for 1 h, (b) TSS1 at T1 = 800 °C for 1 h, and T2 = 700 °C for 25 h, and (c) transformation-assisted TSS2 condition at T1 = 750 °C for 1 h and T2 = 700 °C for 30 h.
formation during various heating regimes. Most obviously, the phase transformation plays a vital role in structural evolution of the titania nanoceramics during normal and two-step sintering operations. Based on
[1] I.-W. Chen, X.-H. Wang, Nature 404 (2000) 168. [2] X.-H. Wang, P.-L. Chen, I.-W. Chen, J. Am. Ceram. Soc. 89 (2006) 431. [3] M. Mazaheri, A.M. Zahedi, S.K. Sadrnezhaad, J. Am. Ceram. Soc. 91 (2008) 56. [4] A. Weibel, R. Bouchet, R. Denoyel, P. Knauth, J. Eur. Ceram. Soc. 27 (2007) 2641. [5] K.-N.P. Kumar, K. Keizer, A.J. Burggraaf, T. Okubo, H. Nagamoto, S. Morooka, Nature 358 (1992) 48. [6] M. Mazaheri, M. Valefi, Z. Razavi Hesabi, S.K. Sadrnezhaad, Ceram. Int., in press. [7] S.-C. Liao, Y.-J. Chen, W.E. Mayo, B.H. Kear, Nanostruct. Mater. 4 (1999) 553. [8] X.-Z. Ding, X.-H. Liu, Y.-Z. He, J. Mater. Sci. Lett. 15 (1996) 1789. [9] J. Li, Y. Ye, L. Shen, J. Chen, H. Zhou, Mater. Sci. Eng., A 390 (2005) 265. [10] Y.I. Lee, J.-H. Lee, S.-H. Hong, D.-Y. Kim, Mater. Res. Bull. 38 (2003) 925. [11] W.D. Kingery, B. Francois, in: S. Somiya, Y. Moriyoshi (Eds.), Sintering Key Papers, Elsevier, London, 1990, pp. 449–466. [12] K. Bodisova, P. Sajgalik, D. Galusek, P. Svancarek, J. Am. Ceram. Soc. 90 (2007) 330. [13] J. Li, Y. Ye, J. Am. Ceram. Soc. 89 (2006) 139. [14] H.G. Kim, K.T. Kim, Acta Mater. 47 (1999) 3561.