Nano–nano composite powders of lanthanum–gadolinium zirconate and gadolinia-stabilized zirconia prepared by spray pyrolysis

Nano–nano composite powders of lanthanum–gadolinium zirconate and gadolinia-stabilized zirconia prepared by spray pyrolysis

Surface & Coatings Technology 232 (2013) 419–424 Contents lists available at ScienceDirect Surface & Coatings Technology journal homepage: www.elsev...

1MB Sizes 1 Downloads 66 Views

Surface & Coatings Technology 232 (2013) 419–424

Contents lists available at ScienceDirect

Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat

Nano–nano composite powders of lanthanum–gadolinium zirconate and gadolinia-stabilized zirconia prepared by spray pyrolysis Songbai Liu a,b, Kuo Jiang b,⁎, Hongbo Zhang c, Yi Liu b, Li Zhang b, Bingye Su b, Yibao Liu a a b c

Department of Nuclear Engineering and Geophysics, East China Institute of Technology, Nanchang 330013, China Department of Defence Science and Technology, Southwest University of Science and Technology, Mianyang 621010, China Ceramics Science Institute, China Building Materials Academy, Beijing 100024, China

a r t i c l e

i n f o

Article history: Received 30 March 2013 Accepted in revised form 28 May 2013 Available online 5 June 2013 Keywords: Nanocomposites Lanthanum–gadolinium zirconate Gadolinia-stabilized zirconia Grain size Spray pyrolysis Doping behavior

a b s t r a c t Ceramics with structure on nanometer-length scales have a thermal conductivity much lower than their coarser-grained ceramics. But the relatively low activation energy for grain growth of nanoparticles limits the stability of nano-material at high temperatures. The objective of the present study was to investigate the preparation of the composite powders of nano (La,Gd)2Zr2O7 and nano (Zr,Gd)O2 − δ by a sol-spray pyrolysis method. The XRD results indicated that the reaction products of 0.2LaO1.5–xGdO1.5–(0.8 − x)ZrO2 (0.15 ≤ × ≤ 0.25) ternary system were composed of pyrochlore (La,Gd)2Zr2O7 and cubic (Zr,Gd)O2 phases. The composition with 0.2LaO1.5–0.2GdO1.5–0.6ZrO2 had a grain size much smaller than the lower GdO1.5 content counterpart for both the (La,Gd)2Zr2O7 and (Zr,Gd)O2 − δ phases. The relationship between grain size and GdO1.5 content was explained in terms of lattice matching. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Ceramics with structure on nanometer-length scales have received steadily growing attention because of their peculiar and fascinating magnetic, electrical, optical, thermal and mechanical properties superior to their coarser-grained ceramics [1–4]. For example, the room temperature thermal conductivity of a 30 nm grain-sized yttria-stabilized zirconia (YSZ) was reported to be more than 50% lower than bulk YSZ [3]. However, the relatively low activation energy for grain growth of nanoparticles limits the stability of nano-material at the service temperatures higher than 1000 °C typically required of thermal barrier coatings (TBCs) [1]. As is well known, the addition of second phase particles causes grain boundary motion to drag [5,6]. This effect was considered as a powerful tool to control the microstructure and stability of grains [6]. Recently, the nanopowders of 8 mol% La2O3–8YSZ composite oxides were reported to have a smaller grain size than 8YSZ in the temperature range of 200–1300 °C [7]. This phenomenon can be attributed to grain boundary motion to drag from the La2Zr2O7 particles of the composite powders. La2Zr2O7 with pyrochlore (P) structure is a promising candidate for the TBC of the future because of its phase stability, high melting point, and low thermal conductivity [8]. Lehmann et al. [9] systematically examined the influence of partial or complete substitution of La3+ by Nd3+, Eu3+, Gd3+, and Dy3+ on the thermo-physical properties of ⁎ Corresponding author. Tel.: +868162419010 E-mail address: [email protected] (K. Jiang). 0257-8972/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.surfcoat.2013.05.044

La2Zr2O7. Among these trivalent cations, Gd3 + ion was found to be one of the most effective rare-earth ions for lowering the thermal conductivity and improving the thermal expansion coefficient of La2Zr2O7. In addition, both the thermal conductivity and sinter shrinkage of 4 mol% Gd2O3-stabilized ZrO2 were reported to be lower than YSZ of the same composition [10]. Based on above research, it is expected that the composite powders composing of nanocrystalline (La,Gd)2Zr2O7 and nanocrystalline (Zr, Gd)O2 − δ could be of an extreme low thermal conductivity. The objective of the present study was to investigate the phase composition and crystal structure of the reaction products of LaO1.5–GdO1.5–ZrO2 ternary system from the in-situ reaction. The mechanism of controlling rare-earth oxide (REO1.5) doping behavior in the La2Zr2O7–ZrO2 system was also studied via substitution of GdO1.5 by YO1.5 and NdO1.5 in the ternary system. 2. Experimental 2.1. Powder synthesis In the present study, 99.99% purity Zr(NO3)4 · 3H2O, La(NO3)3 · 6H2O and Gd(NO3)3 · 6H2O (Aladdin Chemistry Co. Ltd., Shanghai, China) were used as the starting materials and were dissolved in the deionized water. The total concentration of Zr4+, La3+ and Gd3+ was 0.1 M, to which 80 g/L citric acid (C6H8O7 · H2O) and 50 g/L polyethylene glycol (PEG, molecular weight = 20,000) were added to the starting solution. The solution was stirred with a magnetic stirrer for 30 min to achieve complete dissolution.

420

S. Liu et al. / Surface & Coatings Technology 232 (2013) 419–424

The resulting solution was then poured into a polytetrafluoroethylene (PTFE) tube and was atomized by using an atomizer. The atomizing device consisted of a sol supply, an atomizing medium gas supply, a stainless steel nozzle and a programmable high-temperature furnace, where 99.9% purity N2 gas was used as the atomizing medium, and the pyrolysis temperature was 500 °C. The products were annealed at 900–1400 °C for 10–24 h in high-purity Al2O3 crucibles and were cooled down to room temperature in air.

2.2. Characterization X-ray diffraction (XRD) analysis for the samples was carried out on an X' Pert PRO diffractometer with Cu Kα radiation (λ = 0.15406 nm; PANalytical, Almelo, Netherlands). The XRD patterns were collected in a 2θ range of 10°–90° at room temperature at a scanning rate of 0.05°/s and a step size of 0.033°. The peak positions and FWHM of the XRD peaks were determined using an X' Pert HighScore Plus software (version 2.0, PANalytical B. V., Almelo, Netherlands). The crystallite size was calculated using the Scherrer's formula for (440) and (220) peaks of pyrochlore and cubic zirconia phases, respectively: D = 0.89λ / βcosθ, where D is the crystallite size (in nm), λ is the wavelength (in nm), β is the FWHM, and θ is the diffraction angle. Raman spectra of the samples were measured by a Renishaw InVia Raman spectrometer (Renishaw plc, New Mills, UK), with a spectrum resolution of 1 cm−1. The excitation source of the emission line used was an Ar+ ion laser of 514.5 nm. The powders annealed at 900 °C were compacted in an 8 mm diameter steel die (applied pressure ≈ 500 MPa) to give pellets with a height of ~3 mm. The compacts were sintered at 1300 °C for 48 h in air (heating and cooling rates = 2 °C/min; SSX-8-16, Y-feng Electrical Furnace Co. Ltd., Shanghai, China). The samples of the bulk were observed using a field emission scanning electron microscopy (FE-SEM, Apollo 300 FE, Obducat CamScan Ltd., Cambridge, UK) operated at 30 kV. The powders were ultrasonically dispersed in acetone and the resulting suspension was placed on a carbon-coated copper mesh grid. The TEM observation of the samples was obtained using a transmission electron microscope (TEM; JEM-2010, JEOL, Tokyo, Japan).

3. Results and discussion 3.1. Characteristics of prepared powders The XRD patterns of (La1 − xGdx)2Zr2O7 (x = 0–1.0) powders after annealing at 1200 °C for 24 h are shown in Fig. 1(a). The peaks at 2θ = 14.2°, 36.2° and 43.5° respectively correspond to the (111), (331) and (511) reflections of La2Zr2O7, characteristic of pyrochlore structure. With increasing Gd content in (La1 − xGdx)2Zr2O7 solid solution, the diffraction angle of the peaks of above patterns was altered, and the characteristic of pyrochlore structure vanished when x values were higher than 0.8. Fig. 1 (b) shows the XRD patterns and Raman spectra of Gd2Zr2O7 sample after annealing at 1200 °C for 24 h and 1400 °C for 10 h. After a heat treatment of 1400 °C, the XRD peaks correspond to the (111), (311), (331) and (511) reflections of Gd2Zr2O7 displayed, attributable to the pyrochlore structure. In the case of the Raman spectra, see the inset of Fig. 1 (b), the bands at 298, 392, 493, and 511 cm−1 were assigned to the Raman active vibration modes for the pyrochlore La2Zr2O7 [11]. The corresponding bands for the Gd2Zr2O7 annealed at 1400 °C were displayed at 317, 398, 538, and 602 cm−1, and the bands mentioned above for the Gd2Zr2O7 annealed at 1200 °C were obscure. The weakening of the XRD peaks, the broadening and merging of the Raman spectrum for the Gd2Zr2O7 annealed at 1200 °C could be attributed to the defect-, furface-, size-related features of nanoparticles [12].

Fig. 1. (a) XRD patterns of (La1 − xGdx)2Zr2O7 (x = 0–1.0) powders after annealing at 1200 °C for 24 h in air; (b) XRD patterns of Gd2Zr2O7 sample after annealing at 1200 °C for 24 h and 1400 °C for 10 h in air (Inset: Raman spectra of La2Zr2O7 and Gd2Zr2O7 annealed at 1200 °C–1400 °C; and (c) Calculated lattice parameters from Bragg's law.

The lattice parameter (a) of the P-type phase can be calculated by using the following equation on the basis of Vegard's rule: [13–15]      pffiffiffih i 3þ 4þ þ 0:875 rðOÞ þ 0:125 rðVO Þ α ¼ 8= 3 0:5 r Lnav: þ 0:5 r Zr ð1Þ

S. Liu et al. / Surface & Coatings Technology 232 (2013) 419–424

where r(O), r(VO) and r(Zr4+) are the radius of oxygen (0.1380 nm), oxygen vacancy (0.0993 nm) and Zr4+ ion (0.0840 nm), respectively [16]. The r(Ln3av.+) is the average ionic radius of the A-site in the (La1 − xLnx)2Zr2O7 system, and which can be estimated from the following equation: [13]       3þ 3þ 3þ r Lnav: ¼ ð1−xÞ r La þ x r Ln

ð2Þ

In this study, the values for r(La3+) and r(Ln3+) are 0.118 nm and 0.106 nm (ionic radius of Gd3+), respectively. Fig. 1 (c) shows the lattice parameters of (La1 − xGdx)2Zr2O7 system calculated from Bragg's law. The linear relationship between lattice parameter and x was in good agreement with the data calculated from Eq. (1–2) and the experimental results reported by Pan et al. [17]. The results of Fig. 1 indicated that GdO1.5 had been completely dissolved into La2Zr2O7 crystal by substituting Gd3+ for La3+. Fig. 2(a) and (b) show the XRD patterns and Raman spectra of (Zr1 − xGdx)O2 − x / 2 (x = 0–0.3) powders after annealing at 1200 °C for 24 h, respectively. In Fig. 2 (a), the monoclinic (m) to metastable tetragonal (t′) and the subsequent transformation of this phase to cubic (c) of (Zr1 − xGdx)O2 − x / 2 powders were confirmed by XRD measurement. In Fig. 2 (b), the bands at 263, 315, 460, and 637 cm−1 for the x = 0.10 specimen were assigned to the Raman active vibration modes for the

421

tetragonal ZrO2. With higher GdO1.5 content (x = 0.15–0.30), the lines showed broad bands, attributable to the cubic phase [18]. The data indicate that the upper boundary of the t′ phase region in ZrO2–GdO1.5 is approximately 15 mol% GdO1.5, in good agreement with that reported by Leung et al. [19]. Fig. 3(a) shows XRD patterns of 0.2LaO1.5–xGdO1.5–(0.8 − x)ZrO2 (x = 0–0.3) powders after annealing at 1200 °C for 24 h. The pyrochlore and c-ZrO2 phases were determined in the samples of x = 0.15, 0.20 and 0.25. The composition with higher GdO1.5 concentration (x = 0.30) showed a pure pyrochlore phase. The peak shifts in the patterns suggested that the structures of pyrochlore and c-ZrO2 phases were altered. Fig. 3 (b) shows the calculated lattice parameters of the pyrochlore and c-ZrO2 phases of composite powders. The data show that the lattice matching between P-(La,Gd)2Zr2O7 and c-ZrO2 phases can be improved by increasing the content of GdO1.5 in LaO1.5–GdO1.5–ZrO2 ternary system. The results also show that the La3+ in La2Zr2O7 and the Zr4 + in ZrO2 are partially substituted by Gd3+, the reaction can be expressed as: 0:2LaO1:5 þ xGdO1:5 þ ð0:8−xÞZrO2 → ðLa; GdÞ2 Zr 2 O7 þ ðZr; GdÞO2−δ ðx ≤ 0:25Þ:

ð3Þ

3.2. Grain growth and microstructure Fig. 4 shows the XRD patterns of 0.2LaO1.5–xGdO1.5–(0.8 − x) ZrO2 (x = 0.10, 0.15 and 0.20) powders after annealing at 900– 1300 °C for 24 h. Compared with the lower GdO1.5 concentrations

Fig. 2. (a) XRD patterns of (Zr1 − xGdx)O2 − x / 2 (x = 0–0.3) powders after annealing at 1200 °C for 24 h in air; (b) Raman spectra of (Zr1 − xGdx)O2 − x / 2 (x = 0–0.3) powders after annealing at 1200 °C for 24 h in air.

Fig. 3. (a) XRD patterns of 0.2LaO1.5–xGdO1.5–(0.8 − x)ZrO2 (x = 0–0.3) powders after annealing at 1200 °C for 24 h in air; (b) Calculated lattice parameters from Bragg's law.

422

S. Liu et al. / Surface & Coatings Technology 232 (2013) 419–424

Fig. 4. XRD patterns of 0.2LaO1.5–xGdO1.5–(0.8 − x)ZrO2 (x = 0.10, 0.15 and 0.20) powders after annealing at 900–1300 °C for 24 h in air (Inset: XRD pattern of 2θ = 46°–52° region of x = 0.20 sample after annealing 900 °C).

(x = 0.10 and 0.15), the broadening and merging of peaks for x = 0.20 samples were displayed. Fig. 5 shows the high-resolution TEM (HRTEM) image of 0.2LaO1.5–0.2GdO1.5–0.6ZrO2 powders after annealing at 900 °C for 24 h. The clear lattice fringes in the HRTEM image confirm the high crystallinity of the sample. The lattice fringes with 0.607 nm displayed the characteristic of pyrochlore structure, assigned to the (111) lattice plane of (La,Gd)2Zr2O7 phase [20]. And the lattice fringes with 0.261 nm assigned to the (200) lattice plane of c-ZrO2 phase. Fig. 6 shows the grain sizes of La2Zr2O7 and 0.2LaO1.5–xGdO1.5– (0.8 − x)ZrO2 (x = 0.15 and 0.20) powders after annealing at 900– 1300 °C. The results showed that the grain size of x = 0.20 sample was lower than that of x = 0.15 counterpart for both the P-(La, Gd)2Zr2O7 and c-ZrO2 phases. Fig. 7(a) and (b) show the surface SEM images of La2Zr2O7 and 0.2LaO1.5–0.2GdO1.5–0.6ZrO2 bulks after sintering at 1300 °C for 48 h, respectively. SEM images reveal that both of the samples have a uniform particle size and homogenous microstructure. SEM images also reveal that the particle size

Fig. 6. Grain sizes of La2Zr2O7 and 0.2LaO1.5–xGdO1.5–(0.8 − x)ZrO2 (x = 0.15 and 0.20) powders after annealing at 900–1300 °C for 24 h in air (Inset: XRD patterns of {440} region after annealing at 1100 °C).

and the sintered density of composite bulk are distinctly lower than those of La2Zr2O7 prepared by a similar technique. Fig. 7 (c) shows the particle size distribution of the sintered bulks, the data were obtained from the SEM images by measuring 500 different particles for both the samples. The data indicated that the sintered composite bulk consisted of porous distribution of small spherical particles approximately 400–800 nm in diameter, and the particle size distribution of La2Zr2O7 bulk was in the range of 0.5–1.4 μm. In the present study, the grain size of the composite materials was much lower than 1 μm. The chemical composition of the reaction products of 0.2LaO1.5–xGdO1.5–(0.8 − x)ZrO2 ternary system could not be determined by using SEM–energy dispersive X-ray spectroscopy (EDS), because the resolution of the SEM–EDS was in micron magnitude. On the other hand, as shown in Fig. 1 (c), the lattice parameters of the (La1−xGdx)2Zr2O7 system decreased linearly with increasing the x values. These results had proved a route for estimating the chemical composition of the composite powders. The concentration of GdO1.5 in the (La,Gd)2Zr2O7 phase of the composite powders can be calculated by using Eq. (1–2). The concentration of GdO1.5 in the (Zr,Gd)O2 − δ phase and the relative phase content (CR) of the two phases can be obtained easily based on the stoichiometry ratio of the ternary system. For this calculation, the lattice parameters of the (La,Gd)2Zr2O7 phase of the composite powders annealed at 1200 °C had been selected (see Fig. 2(a)), to avoid the impact of grain size. Table 1 lists the calculated chemical compositions of 0.2LaO1.5–xGdO1.5–(0.8 − x)ZrO2 (x = 0.10–0.25) composite powders. The data in Table 1 showed that the concentration of GdO1.5 in both (La,Gd)2Zr2O7 and (Zr,Gd)O2 − δ phases increased with increasing the x value. The data also showed that the (Zr,Gd) O2 − δ phases of the x = 0.15–0.25 samples were cubic form. 3.3. Doping behavior of REO1.5 in the La2Zr2O7–ZrO2 system

Fig. 5. HRTEM image of 0.2LaO1.5–0.2GdO1.5–0.6ZrO2 powders after annealing at 900 °C for 24 h in air.

Fig. 8 shows the XRD patterns of 0.2LaO1.5–xNdO1.5–(0.8 − x) ZrO2 (x = 0.05–0.25) powders after annealing at 1200 °C for 24 h. With increasing the Nd content in the ternary system, the relative peak intensities assigned to the (101) lattice plane of t-ZrO2 phase became weak. In contrast, the relative peak intensities assigned to the pyrochlore phase became strong. The inset of Fig. 8 shows the lattice parameters of pyrochlore phase as a function of NdO1.5 content (x). It provided proof that NdO1.5 had been completely dissolved into the pyrochlore phase by substituting Nd3+ for La3+, in good accordance

S. Liu et al. / Surface & Coatings Technology 232 (2013) 419–424

423

Fig. 8. XRD patterns of 0.2LaO1.5–xNdO1.5–(0.8 − x)ZrO2 (x = 0.05–0.25) powders after annealing at 1200 °C for 24 h in air (Inset: lattice parameters of pyrochlore phase calculated from Bragg's law).

with the results of (La1 − xNdx)2Zr2O7 solid solution [21,22]. The reaction of LaO1.5–NdO1.5–ZrO2 system can be expressed as: 0:2LaO1:5 þ xNdO1:5 þ ð0:8−xÞZrO2 → þ ð0:6−2xÞZrO2 ðx ≤ 0:25Þ:

 0:2 þ x  La 0:2 Nd x Zr 2 O7 0:2þx 0:2þx 2 2

ð4Þ

Fig. 9 shows the XRD patterns of 0.2LaO1.5–xYO1.5–(0.8 − x)ZrO2 (x = 0.05–0.15) powders after annealing at 1200 °C for 24 h. Other than the LaO1.5–NdO1.5–ZrO2 system, no clear evidence was found for the presence of peak shift for pyrochlore phase in the LaO1.5–YO1.5–ZrO2 system. The transformation of t′- to c-ZrO2 phases indicated that YO1.5 was completely doped into ZrO2 crystal lattice to play a role of stabilizer. The relationship between the lattice parameters and the YO1.5 content of c-ZrO2 phase was in line with the results of YO1.5–ZrO2 system [23]. The reaction of LaO1.5–YO1.5–ZrO2 system can be expressed as: 0:2LaO1:5  þ xYO1:5þ ð0:8−xÞZrO2 → 0:1La2 Zr 2 O7 þ 0:6 Zr 0:6−x Y 0:6x O2:4−x ðx ≤ 0:15Þ: 0:6

1:2

ð5Þ

A possible explanation for the doping behavior of REO1.5 (RE = Y, Nd and Gd) in the La2Zr2O7–ZrO2 system can be provided by

Fig. 7. SEM images of (a) La2Zr2O7 and (b) 0.2LaO1.5–0.2GdO1.5–0.6ZrO2 bulks after sintering at 1300 °C for 48 h in air; and (c) particle size distribution of La2Zr2O7 and 0.2LaO1.5–0.2GdO1.5–0.6ZrO2 bulks obtained from the SEM images.

Table 1 Chemical composition of 0.2LaO1.5–xGdO1.5–(0.8 − x)ZrO2 composite powders. x

0.10 0.15 0.20 0.25 a

(La,Gd)2Zr2O7

(Zr,Gd)O2 − δ

C Ra

Chemical composition

CRa

Chemical composition

0.25 0.29 0.37 0.60

(La0.78Gd0.22)2Zr2O7 (La0.73Gd0.27)2Zr2O7 (La0.66Gd0.34)2Zr2O7 (La0.53Gd0.47)2Zr2O7

0.75 0.71 0.63 0.40

(Zr0.91Gd0.09)O1.96 (Zr0.83Gd0.17)O1.92 (Zr0.76Gd0.24)O1.88 (Zr0.61Gd0.29)O1.86

Relative phase content in mole fraction.

Fig. 9. XRD patterns of 0.2LaO1.5–xYO1.5–(0.8 − x)ZrO2 (x = 0–0.15) powders after annealing at 1200 °C for 24 h in air (Inset: lattice parameters of pyrochlore and cubic zirconia phases calculated from Bragg's law).

424

S. Liu et al. / Surface & Coatings Technology 232 (2013) 419–424

considering the relationship between the enthalpies of formation of RE2Zr2O7 and the RE3+ ionic radius. The experimental and calculated enthalpies of formation of pyrochlore and fluorite RE2Zr2O7 (RE = Y, La, Nd, Sm, Gd, Dy and Yb) had been summarized by Zinkevich et al. [24–28]. The data show that the pyrochlore RE2Zr2O7 is more stable with decreasing the ionic radius of RE3+. However, the calculated enthalpies of formation of fluorite RE2Zr2O7 show complete different trends from pyrochlore counterparts. It is reasonable to expect that, the substitution of La3+ in La2Zr2O7 by Nd3+ ions will result in a lower increase of enthalpies than Y3+ does, because of the larger Nd3 + ionic radius compared to Y3+ [16]. And, the situation is reverse on the fluorite phases. Thus, the doping behavior of REO1.5 (RE = Y, Nd and Gd) in the La2Zr2O7–ZrO2 system may be controlled by the RE3+ ionic radius. 4. Conclusions The composite powders composing of nano P-(La,Gd)2Zr2O7 and nano c-(Zr,Gd)O2 − δ had been prepared by a sol-spray pyrolysis method. The composition with higher GdO1.5 content had a smaller grain size than the lower GdO1.5 content counterpart for both P-(La,Gd)2Zr2O7 and c-(Zr,Gd)O2 − δ phases. The sintered composite bulk was porous, and was consisted of small spherical particles approximately 500 nm in diameter after annealing at 1300 °C for 48 h. Since the La3+ in La2Zr2O7 and the Zr4+ in ZrO2 were partially substituted by Gd3+, the role of Gd3+ in lowering grain size may be explained in terms of enhancing lattice matching. The mechanism of controlling REO1.5 doping behavior in the La2Zr2O7–ZrO2 system can be attributed to the difference of RE3+ ionic radius. The chemical element of the composite powders is relatively simple. The P-(La,Gd)2Zr2O7 and c-(Zr,Gd)O2 − δ phases can be obtained in a wide GdO1.5 content range. These are beneficial to the preparation of the composite coatings for both electron beam physical vapor deposition (EB-PVD) and plasma spraying (PS) techniques. Further work is required to obtain the composite coatings, and examine the thermal conductivity of the coatings. Acknowledgments We thank Dr. Huibin Xu and Dr. Shengkai Gong for their helpful discussions during our investigation. Thanks also to Dr. Ruishi Xie,

Dr. Zhaohui Zhou and Dr. Huiping Duan for their assistance with XRD, SEM and TEM. This work was financially supported by the Educational Committee (No.09ZZ032) and the Science & Technology Department (No.2011JY0137) of Sichuan province of China.

References [1] R.M. Costescu, D.G. Cahill, F.H. Fabreguette, Z.A. Sechrist, S.M. George, Science 303 (2004) 989–990. [2] Y. Xia, P. Yang, Y. Sun, Y. Wu, B. Mayers, B. Gates, Y. Yin, F. Kim, H. Yan, Adv. Mater. 15 (2003) 353–389. [3] G. Soyez, J.A. Eastman, L.J. Thompson, G.-R. Bai, P.M. Baldo, A.W. McCormick, R.J. DiMelfi, A.A. Elmustafa, M.F. Tambwe, D.S. Stone, Appl. Phys. Lett. 77 (2000) 1155–1157. [4] N.Y.S. Somiya, N. Yanagida (Eds.), American Ceramic Society, Columbus, OH, (1988). [5] C.S. Smith, Trans. AIME 175 (1948) 15–51. [6] G. Gottstein, L.S. Shvindlerman, Scr. Mater. 63 (2010) 1089–1091. [7] C. Wang, Y. Wang, Y. Cheng, L. Zhu, B. Zou, Y. Zhao, W. Huang, X. Fan, Z.S. Khan, X. Cao, J. Cryst. Growth 335 (2011) 165–171. [8] R. Vassen, X.Q. Cao, F. Tietz, D. Basu, D. Stöver, J. Am. Ceram. Soc. 83 (2000) 2023–2028. [9] H. Lehmann, D. Pitzer, G. Pracht, R. Vassen, D. Stöver, J. Am. Ceram. Soc. 86 (2003) 1338–1344. [10] M.N. Rahaman, J.R. Gross, R.E. Dutton, H. Wang, Acta Mater. 54 (2006) 1615–1621. [11] D. Michael, M. Perez, Y. Jorba, R. Collongues, J. Raman Spectrosc. 5 (1976) 163–180. [12] G.G. Siu, M.J. Stokes, Yulong Liu, Physical Review B 59 (1999) 3173–3179. [13] H. Yamamura, H. Nishino, K. Kakinuma, K. Nomura, Solid State Ion. 158 (2003) 359–365. [14] Seong Jae Hong, Anil V. Virkar, J. Am. Ceram. Soc. 78 (1995) 433–439. [15] L. Vegard, Z. Phys. 5 (1921) 17. [16] R.D. Shannon, C.T. Prewitt, Acta Crystallogr. B25 (1969) 925–946. [17] C.L. Wan, W. Pan, Q. Xu, Y.X. Qin, J.D. Wang, Z.X. Qu, M.H. Fang, Phys. Rev. B 74 (2006) 114109. [18] M. Yashima, K. Ohtake, M. Kakihana, H. Arashi, M. Yoshimura, J. Phys. Chem. Solids 57 (1996) 17–24. [19] D.K. Leung, C.J. Chan, M. Rühle, F. Lange, J. Am. Ceram. Soc. 74 (1991) 2786–2792. [20] S.X. Wang, B.D. Begg, L.M. Wang, R.C. Ewing, W.J. Weber, K.V. Govidan Kutty, J. Mater. Res. 14 (1999) 4470–4473. [21] H. Yokoi, Y. Arita, T. Matsui, H. Ohno, K. Kobayashi, J. Nucl. Mater. 238 (1996) 163–168. [22] E.J. Harvey, K.R. Whittle, G.R. Lumpkin, R.I. Smith, S.A.T. Redfern, J. Solid State Chem. 178 (2005) 800–810. [23] H.G. Scott, J. Mater. Sci. 10 (1975) 1527–1535. [24] C. Wang, M. Zinkevich, F. Aldinger, Pure Appl. Chem. 79 (2007) 1731–1753. [25] T.A. Lee, A. Navrotsky, I. Molodetsky, J. Mater. Res. 18 (2003) 908–918. [26] K. Helean, B.D. Begg, A. Navrotsky, B. Ebbinghaus, W.J. Weber, R.C. Ewing, Mater. Res. Soc. Symp. Proc. 663 (2001) 691. [27] A. Navrotsky, J. Mater. Chem. 15 (2005) 1883–1890. [28] V.R. Korneev, V.B. Glushkova, E.K. Keler, Inorg. Mater. 7 (1971) 781–782.