Available online at www.sciencedirect.com
ScienceDirect Scripta Materialia 92 (2014) 55–58 www.elsevier.com/locate/scriptamat
In situ TEM observation of domain switch in LaNbO4 added to NiO–Y2O3 stabilized ZrO2 ceramic composite ⇑
Ben Ma,a,b Bo Chi,a Jian Pua and Jian Lia, a
Center for Fuel Cell Innovation, State Key Lab for Material Processing and Tooling Technology, School of Materials Science and Engineering, Huazhong University of Science and Technology, 1037 Luo Yu Road, Wuhan, Hubei 430074, People’s Republic of China b College of Sciences, Wuhan University of Science and Technology, 947 He Ping Avenue, Wuhan, Hubei 430080, People’s Republic of China Received 3 July 2014; accepted 13 August 2014 Available online 27 August 2014
The morphology of domain structure in LaNbO4 added to YSZ–NiO composite was examined using a focused ion beam prepared specimen in a transmission electron microscope with a straining stage. It is confirmed for the first time that the morphology is changed under load through domain boundary movement. The driving force for this domain switch is the difference in Gibbs free energy between domains with different crystallographic orientations. This domain switch improves toughness by dissipating stress build-up at the crack tips. Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Ceramic composite; LaNbO4; Domain switch; In situ observation; Transmission electron microscopy
The conventional approach to toughening ceramic materials is to introduce a secondary phase in various shapes, such as particles, whiskers or fibers [1–3], into the ceramic matrix to make use of the toughening mechanism of phase transformation, micro-cracking, residual stress, fiber/whisker reinforcement or grain bridging. Recently, domain switch has been considered as another toughening mechanism in ceramic composites containing ferroics as the secondary phase [4–11]. The microstructure of ferroics contains domains in different orientations. The domain preferably oriented will grow at the expense of the other via domain boundary switch, while it is stressed by an external force. Thus it is expected that the domain switch can serve the purpose of toughening ceramic composite by dissipating the stress build-up at the tip of a propagating crack. BaTiO3 [8,10,11], LiTaO3 [6] and LaNbO4 [4,12] have been proved to be effective in enhancing the toughness of ceramic materials, even if the domain switch has not been experimentally observed so far. The purpose of this study is to confirm the stressinduced domain switch in LaNbO4 by in situ observation using transmission electron microscopy (TEM). The LaNbO4 was added to toughen NiO–Y2O3 stabilized ZrO2 (YSZ) composite, which is a state-of-the-art substrate material for anode-supported solid oxide fuel cells.
⇑ Corresponding
author. Tel.: +86 27 87557496; fax: +86 27 87558142; e-mail:
[email protected]
LaNbO4 powder was prepared by solid state reaction using La2O3 (99.9% purity, Sino-Pharm Chemical Reagent) and Nb2O5 (99.99% purity, Aladdin Chemistry) as the raw materials. The detailed procedure was reported previously in Ref. [13]: 30 wt.% of LaNbO4 was added to 47 wt.% NiO (Type standard, Inco)–53 wt.% YSZ (TZ-8YS, Tosoh) powder mixture. The LaNbO4–YSZ–NiO powder mixture was ball-milled, dry-pressed using hardened steel die and sintered at 1500 °C in air for 6 h to obtain a densified bar specimen (99% [13]) in a rectangular shape. The preparation of the specimen for TEM in situ observation was performed inside a dual-beam focused ion beam (FIB) system (FEI Helios Nanolab 600i), equipped with a Schottky field emission gun column, a Ga ion beam column, Pt gas injection system and micromanipulator (Omniprobe). The procedure was as follows: (1) a Pt protective strip, 10 3 1 lm, was deposited on the surface of the sintered bar specimen; (2) both sides of the strip were trenched by ion beam at 30 kV and 20 nA (Fig. 1a) to shape a thin “wall” between the two trenches; (3) the thin specimen was cut off from the bulk specimen by ion beam at a tilted (7°) position and placed on a Cu specimen holder by the micromanipulator (Fig. 1b); and (4) the thin specimen was further thinned sidewise to an appropriate thickness by ion beam for TEM observation (Fig. 1c). The bonding between the thin specimen and the specimen holder was established by deposited Pt. Figure 1d is the top view of Figure 3c, showing the final thickness of the
http://dx.doi.org/10.1016/j.scriptamat.2014.08.013 1359-6462/Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
56
B. Ma et al. / Scripta Materialia 92 (2014) 55–58
Figure 1. FIB preparation of specimen used for in situ TEM observation under tensile load.
Figure 2. Schematic illustration showing the configuration of specimen used for in situ TEM observation under tensile load.
Figure 3. TEM microstructure of LaNbO4 added NiO–YSZ ceramic composite and selected area diffraction pattern obtained from the grain of monoclinic LaNbO4.
TEM specimen. As shown schematically in Figure 2, the Cu specimen holder was 2.4 mm wide, 8 mm long and 0.1 mm thick with two 1.5 mm holes at both ends to fit the straining stage; and the thin specimen was seated right on a preformed V-shape slot that facilitates the tensile experiment. In situ observation of the tensile load was performed by TEM, using a Jeol 2010 instrument at 200 kV with a
Gatan single-tilt straining holder under a slowly increased tensile load. Figure 3 demonstrates the microstructure of 30 wt.% LaNbO4–NiO–YSZ composite, in which the monoclinic LaNbO4 is featured by two domains with widths varying from 20 to 50 nm. Two sets of diffraction spots shown by the inset in Figure 3 indicates that the two domains are oriented in different crystallographic orientations, with a domain boundary located at a lattice plane in the {2 0 4}I/{4 0 2}II family [14,15]. Figure 4 compares the domain structure in a LaNbO4 grain before and after tensile load. It is noticed that the overall morphology of the domain structure was obviously changed (Fig. 4a and b) as a result of the tensile load, and more specifically the width of domain in the circled area was narrowed (Fig. 4c and d). These phenomena confirm for the first time that the domain boundary was indeed moved by the applied tensile load. Tsunekawa [16] postulated a model for this domain switch. It suggests a sharp domain boundary and that all the atoms are moved from the positions in one domain to the positions in the other by a shearing on the domain boundary plane, followed by atomic shuffles to reach the final positions. This model may be crystallographically correct and may not be thermodynamically realistic, since high energy is needed to move all the atoms simultaneously through a large displacement. It is observed that there is a fine diffraction line connecting two corresponding diffraction spots from each domain, which implies that a diffuse domain boundary separating the two domains [15] and the atoms located in the boundary zone are progressively changing their positions from one domain to the other. In other words, a very short distance atom movement at the edge of the boundary zone can move the boundary as a whole, which makes the domain boundary easily moveable under a small load. With an external uniaxial stress rij, the
B. Ma et al. / Scripta Materialia 92 (2014) 55–58
57
Figure 4. Morphology change in domain structure caused external tensile load in monoclinic LaNbO4 in situ observed using a TEM with a straining stage.
Gibbs free energy density G for a domain can be written in differential form as [15] dG ¼ SdT eij drij
ð1Þ
where S, T and eij denote the entropy, temperature and strain, respectively. And the strain can be further expressed as eij ¼ esij þ sijkl drkl
ð2Þ
where esij is the component of the spontaneous strain tensor, and sijkl is the component of the elastic compliance tensor. The difference in integrated free energy density (DG) for the domains in two different orientations can be written as DG ¼ G2 G1 ¼ esij rij
ð3Þ
which determines the preferred domain orientation and acts as the driving force for domain switch under load between these two orientations. Toughness improvement was obtained by adding LaNbO4 into NiO–YSZ composite material, which was attributed to the domain switch in monoclinic LaNbO4, which changed the stress state at the tip of propagating cracks [13]. In fact, the stress state at the tip of a crack is characterized by the stress intensity factor Ktip, which can be described as K tip ¼ K applied K shield
ð4Þ
where Kapplied is the stress intensity factor caused by an externally applied stress, and Kshield is the reduction in stress intensity factor at crack tip as a result of secondary phase addition. Once Ktip is larger than the intrinsic critical stress intensity factor (KC) of the material, i.e. K c > K tip
ð5Þ
or K applied > K c þ K shield
ð6Þ
crack propagation is initiated. Inequality (6) suggests that, if cracks in a composite material are shielded by the added secondary phase that provides Kshield, a larger applied stress is needed to generate higher Kapplied for initiating crack propagation. In other words, the material is toughened by adding the secondary phase. In the case of the LaNbO4– NiO–YSZ composite, the crack tip is shielded by added LaNbO4 through its domain switch, as observed above, resulting in enhanced toughness of the composite. In summary, the morphology change in domain structure caused by the domain switch in LaNbO4 added to NiO– YSZ composite was observed in situ using a TEM with a straining stage. The driving force for the domain switch under load is the difference in Gibbs free energy between domains with different crystallographic orientations. This domain switch shields the propagating cracks by dissipating the stress build-up at the crack tip, resulting in an improvement in toughness of the composite.
58
B. Ma et al. / Scripta Materialia 92 (2014) 55–58
The authors gratefully acknowledge the financial support to this study provided by the National Natural Science Foundation of China under the project 51172082. The specimen preparation and TEM observation were performed with the assistance of the Analytical and Testing Center of Shanghai University.
[1] S.A. Baldacim, C.A.A. Cairo, C.R.M. Silva, J. Mater. Process. Technol. 119 (2001) 273. [2] I.E. Reimanis, Mater. Sci. Eng., A 237 (1997) 159. [3] V. Kostopoulos, Y.P. Markopoulos, Mater. Sci. Eng., A 250 (1998) 303. [4] Z.L. Zhang, L. Zhou, Y.G. Hu, L. Jiang, Scripta Mater. 47 (2002) 637. [5] S. Yuh, Y. Lai, C.C. Chou, J. Mater. Sci. 36 (2001) 2303. [6] Y.G. Liu, Y. Zhou, D.C. Jia, Q.C. Meng, Y.H. Chen, Scripta Mater. 47 (2002) 63.
[7] X.M. Chen, X.Q. Liu, F. Liu, X.B. Zhang, J. Eur. Ceram. Soc. 21 (2001) 477. [8] B. Yang, X.M. Chen, X.Q. Liu, J. Eur. Ceram. Soc. 20 (2000) 1153. [9] B. Yang, X.M. Chen, J. Eur. Ceram. Soc. 20 (2000) 1687. [10] G.D. Zhan, J. Kuntz, J. Wan, J. Garay, A.K. Mukherjee, Mater. Sci. Eng., A 356 (2003) 443. [11] J.Y. Li, H. Dai, X.H. Zhong, Y.F. Zhang, X.F. Ma, J. Meng, X.Q. Cao, J. Alloys Compd. 452 (2008) 406. [12] S. Maschio, G. Pezzotti, O. Sbaizero, J. Eur. Ceram. Soc. 18 (1998) 1779. [13] B. Ma, B. Chi, J. Pu, J. Li, Int. J. Hydrogen Energy 38 (2013) 4776. [14] J. Li, C.M. Huang, G.B. Xu, C.M. Wayman, Mater. Lett. 21 (1994) 105. [15] J. Li, C.M. Wayman, J. Am. Ceram. Soc. 79 (1996) 1642. [16] S. Tsunekawa, Sci. Rep. Res. Inst. Tohoku Univ., Ser., A 29 (1980) 1.