- Email: [email protected]
Preparation of carbon fiber-reinforced zirconium carbide matrix composites by reactive melt infiltration at relative low temperature
Available online at www.sciencedirect.com
Scripta Materialia 67 (2012) 822–825 www.elsevier.com/locate/scriptamat
Preparation of carbon fiber-reinforced zirconium carbide matrix composites by reactive melt infiltration at relative low temperature Yulin Zhu,⇑ Song Wang, Wei Li, Shouming Zhang and Zhaohui Chen National Key Laboratory of Science and Technology on Advanced Ceramic Fibers and Composites, College of Aerospace and Materials Engineering, National University of Defense Technology, Changsha 410073, China Received 19 April 2012; revised 11 July 2012; accepted 30 July 2012 Available online 4 August 2012
An attractive way to prepare carbon fiber-reinforced ZrC matrix composites was proposed and confirmed experimentally. The experimental results showed that the resulting composites were fabricated by immersing carbon/carbon preforms in molten Zr2Cu at 1200 °C for 3 h. ZrC was the main phase, with a content of 42.2 ± 1.3 vol.%. The composites exhibited excellent ablation resistance when undergoing ablation with an oxyacetylene torch, the mass loss rate and linear recession rate being 0.0006 ± 0.0001 g s1 and 0.0003 ± 0.0001 mm s1, respectively. Ó 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Carbides; Fiber-reinforced composites; Infiltration; Microstructure; Ablation resistance
Zirconium carbide (ZrC), which is a member of the ultrahigh-temperature ceramics family, has a number of unique properties, including an extremely high melting point, high strength and hardness, relatively low density, and high retained strength at high temperatures [1–5], which make it attractive for aerospace applications associated with hypersonic flight and rocket propulsion. However, ZrC ceramics are brittle and display little to no plasticity across a broad temperature range [5,6]. It has been proven in earlier works that introducing carbon fibers is an effective way to improve the fracture resistance of ZrC [5–7]. Known as a rapid and low-cost manufacturing process [5,8–14], the reactive melt infiltration (RMI) process has been widely used to produce structural components with complex geometries made of SiC- [14,15], ZrC[5,11,16] and HfC-matrix [13] composites. Recently, carbon fiber-reinforced zirconium carbide (Cf/ZrC) composites with porous carbon/carbon (Cf/C) preforms and pure Zr metal have been prepared by this method [5,7]. Due to its high melting point (1850 °C) [17], however, the Zr melt reacts strongly with fibers, which is detrimental to the reinforcement effect. Introducing copper to form Zr2Cu (melting point 1025 °C) can reduce the processing temperature effectively, and can wet and
⇑ Corresponding
author. Tel.: +86 731 84576441; fax: +86 731 84573165; e-mail: [email protected]
infiltrate well into a porous preform [2,10,18]. Zr2Cu instead of Zr has been used as an infiltrator to produce W/ZrC composites at temperatures as low as 1200 °C [11,16,19–21]. However, Cf/ZrC composites fabricated by the RMI process using Zr2Cu as the infiltrator have not yet been reported. In this study, an attractive way to prepare Cf/ZrC composites by RMI with Zr2Cu as a reactive infiltrator is proposed for the first time. The microstructure and composition of the composites were investigated. Furthermore, the ablation properties of the composites were also determined using an oxyacetylene torch. Felts with a fiber volume fraction of about 30 vol.% (T300, with an average diameter of about 7 lm, a tow size of 3 K, a carbon content of 93% (impurities of O + H + N + S 6 7%) and a bulk density of 1.76 g cm3; Toray, Japan) were prepared by a needle-punching technique with alternatively stacked 0° weftless piles, short-cut-fiber webs and 90° weftless piles. The preparation of Cf/C preforms included the following stages: first, phenolic resins (flaxen solid, manufactured by Taihang Co. in Xian) were smashed into powder and dissolved into ethanol with the weight ratio of 1:1; next, the carbon felts fixed by modular molds were located in a hermetic container; then the phenolic resins/ethanol solution were infused into the container and overwhelmed the felts when the container pressure was below 10 Pa. After dwelling there for 6 h, the saturated felts were cured at 80 °C for 4 h and then at 150 °C
1359-6462/$ - see front matter Ó 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.scriptamat.2012.07.044
Y. Zhu et al. / Scripta Materialia 67 (2012) 822–825
for 3 h. Finally, the cured felts were pyrolyzed at 1200 °C for 60 min in N2 (purity P 99.999 vol.%, O2 + H2O 6 0.001 vol.%) under atmospheric pressure, with a heating rate of 10–15 °C min1. Four cycles of infiltration and pyrolysis were necessary to obtain the Cf/C preforms with appropriate porosity. Commercial Zr2Cu ingots (purity P 99.2 wt.%, impurities (Hf + Fe + Ti + Si + Mg + Al + Ni) 6 0.8 wt.%; Hunan Rare Earth Metal & Material Institute, Hunan, China) were used as the reactive infiltrant. The ingots were prepared with spongy Zr pieces (purity P 99.6 wt.%) and copper cylinders (purity P 99.99 wt.%) by arc melting. To allow for homogenization, each ingot was flipped and remelted three times. The preparation of Cf/ZrC composites included the following stages: the Zr2Cu alloy was rapidly heated up to 1200 °C, the reactor keeping the preforms separated from the alloy during heating. After the alloy had melted completely, the porous Cf/C preforms were mechanically driven into the melt and kept there for 3 h, before being separated from the liquid Zr2Cu bath and left to cool naturally to room temperature. Any excess solidified melt adhering to the sample was removed by a grinding machining with a diamond wheel. The porosity of the specimen was equal to one minus the quotient of the bulk density to the theoretical density. The bulk density of the specimen was obtained from measurements of its dry weight and external dimensions. The theoretical densities were calculated using rule of mixtures calculations, based on the theoretical densities of 6.63 g cm3 for ZrC, 1.55 g cm3 for the deposited C, 8.96 g cm3 for Cu, 6.49 g cm3 for Zr and 1.76 g cm3 for the T300 fiber. The volume fractions of the solid phases were measured by inductively coupled plasma (ICP) and chemolysis analysis, which were described previously [22]. The density and open porosity of the specimen was measured using the Archimedes method, with distilled water as the immersing medium, after open porosity was saturated under vacuum for 12 h. The density was measured for at least five samples. In order to analyze the contents of each phase, the resulting composites were crushed into powder by milling machining with a diamond milling cutter and granulated using a 100-mesh sieve, before being immersed in an acid solution (1 part HCl, 1 part HNO3 and 2 parts distilled water) for 30 min. When tested by reacting with pure Zr, Cu and ZrC, respectively, this solution can dissolve Zr and Cu but not ZrC. After filtering, the fluid underwent ICP analysis to determine the amount of residual Zr and Cu; the solids left on the filter were dried in an oven at 80 °C for 5 h and then kept in muffle furnace at 1200 °C for 5 h, where the solid powders were oxidized thoroughly into ZrO2. The amount of ZrC was calculated from the content of ZrO2, and the amount of the residual carbon (containing carbon matrix and fiber) was equal to the weight difference between the drying powders and ZrC. Since the fiber fraction was a known quantity, the content of the residual carbon matrix could be calculated. The phases were analyzed by X-ray diffraction (XRD) with a Bruker D8 Advance instrument. The microstructures were observed by scanning electronic microscopy (SEM; Quanta-200) and energy-dispersive
823
spectroscopy (EDS). All specimens for SEM/EDS analyses were prepared by incised, ground and polished with a series of diamond pastes to a surface finish of 0.5 lm. Image analysis software package (ImageJ, National Institutes of Health, Bethesda, MD) was used to determine the average diameters of the fibers in the composites. The average diameter was calculated by measuring 30 fibers per image of at least 10 different SEM images. The ablation properties were tested with an oxyacetylene flame according to the GJB323A-96 Ablation Standard. The ablation tests were conducted in air, and the pressures of oxygen and acetylene were 0.4 and 0.095 MPa, respectively. The surface temperature of the sample was monitored with an optical pyrometer. During the test, the ablation gun, with a gunpoint diameter of 2 mm, was first ignited. When the fluxes of acetylene and oxygen were turned to 1.116 and 1.512 m3 h1, respectively, the gun was moved perpendicular to the sample surface. The distance between the tip of the gun and the top of the sample was 10 mm and the exposure time under the torch flame was 30 s. The heat flux was calculated from the heat absorbed by water flowing through a calorimeter, and was found to be 4200 kW m2 (10% error). At least three samples with dimensions of 30 mm 30 mm 3.5 mm were examined in each test. The mass and linear ablation rates (Rm and Rl) were defined as follows: Rm ¼ ðm0 m1 Þ=t Rl ¼ ðl0 l1 Þ=t where m0 and m1 represent the weight before and after ablation, respectively; l0 and l1 represent the thickness before and after ablation, respectively; and t is the ablation time. In detail, l1 is the average thickness of seven equidistant points along the longest chord of the ablated region with an approximately circular shape. The overall ablation rates of the composites were taken from three specimens on average. The reaction between carbon and Zr2Cu melt can be expressed as Reaction (1). Thermodynamics calculations were conducted in the standard state to estimate the direction of this reaction. The thermodynamics data of the Zr2Cu phase were calculated using Pandat 7.0 software based on the database optimized by Zeng et al. [23]. The thermodynamics data of the other reactant and products are from Chase [24]. The changes in Gibbs’ free energy and enthalpy above the melting point of Zr2Cu are shown in Figure 1, which reveals that
Figure 1. Changes in Gibbs’ free energy and enthalpy as a function of temperature above 1200 °C for Reaction (1).
824
Y. Zhu et al. / Scripta Materialia 67 (2012) 822–825
Reaction (1) is strongly favored thermodynamically as low as 1200 °C. XRD results (Fig. 2a) revealed that the alloy had a stoichiometric phase of Zr2Cu. 2CðsÞ þ Zr2 CuðlÞ ! 2ZrCðsÞ þ CuðlÞ at T 1200 C
ð1Þ
The Cf/C preforms, with a fiber fraction of 31.9 ± 2.7 vol.%, had a density of 1.01 ± 0.03 g cm3 and a porosity of 39.1 ± 2.4%, with the pore size mainly in the range of 10–40 lm; however, the infiltrated specimens had a density of 4.30 ± 0.25 g cm3 and a porosity of 7.9 ± 1.1%. The weight change was about 326%, which was mainly attributed to the densification of the preforms and the transformation from C to ZrC. Since the experimental density was extremely close to the theoretical value of 4.33 ± 0.24 g cm3, the weight change was quantitatively consistent with the phase content and porosity of the reacted specimen. The XRD pattern in Figure 2b reveals that there were three phases in the resulting composite: carbon fibers, ZrC and Cu. ICP and chemolysis revealed that the Cf/ZrC composites contained 42.2 ± 1.3 vol.% ZrC, 8.8 ± 2.3 vol.% residual deposited C, 8.0 ± 1.2 vol.% Cu and 1.4 ± 0.6 vol.% Zr, which means that ZrC was the main phase and that a small quantity of metals remained in the composites. However, according to Wang’s research [7], no detectable residual zirconium was found in the C/C–ZrC composites prepared with pure Zr liquid. The residual metals in our composites could be ascribed to the gentle infiltrating temperature and the dilution of Cu. Microstructural characterization was conducted to analyze the distribution of phases in the Cf/ZrC composites. Figure 3a shows a representative microstructure of the cross-section. It reveals a dense specimen with a few fine (<1 lm) isolated pores (indicated by arrows in Fig. 3e), and carbon fibers surrounded by gray matrix. Image analyses revealed an average fiber diameter of 7.0 ± 0.3 lm, indicating that the fibers were not eroded by the Zr2Cu liquid. The gray part was magnified to observe the details and is presented in Figure 3b, which shows many dark island-like particles distributed throughout the light matrix. As demonstrated by the EDS analysis results (Fig. 3c and d), these island-like particles (indicated in region A) contained about 51 at.% C and 49 at.% Zr, while the surrounding substance (indicated in region B) contained about 14 at.% C, 6 at.% Zr and 80 at.% Cu. Taking account of the XRD results shown in Figure 2b, these island-like
Figure 2. XRD patterns of (a) the alloy and the Cf/ZrC composite (b) before ablation and (c) after ablated by oxyacetylene torch.
Figure 3. Micrographs of (a) a representative cross-section of the Cf/ZrC composites and (b) the matrix of the Cf/ZrC composites; EDS patterns of (c) elements in region A and (d) elements in region B.
particles were ZrC grains and the surrounding material was residual Cu-enriched phase. Since the composites are used to produce high-temperature components, in which ablation is the main cause of material failure, an oxyacetylene flame ablation test was conducted to estimate the resistance. After ablation for 60 s, the Cf/ZrC composites showed a mass loss rate of 0.0006 ± 0.0001 g s1 and a linear recession rate of 0.0003 ± 0.0001 mm s1, nearly one magnitude lower than that of the C/C–ZrC composite [7], which was prepared with pure Zr at >1900 °C and had a mass loss rate of 0.004 ± 0.001 g s1 and a linear recession rate of 0.002 ± 0.001 mm s1. The excellent ablation resistance of the Cf/ZrC composites might be related to the presence of residual Cu. Figure 4b shows the optical images of the ablated Cf/ZrC composites. Compared with the raw sample shown in Figure 4a, the ablated sample remained flat and showed no apparent ablated pitting, other than a white oxide shell on the surface. In the center of the sample, a glass-like layer can be observed, and a loose layer formed around the center (Fig. 4b). These two parts are magnified in Figure 4c and d, which reveals that the glass-like layer is composed of dense and continuous molten ZrO2, while the loose layer consists of a large quantity of ZrO2 particles. XRD analysis indicated that the main component of the white ablation product was monoclinic ZrO2 (m-ZrO2) (Fig. 2c). Additionally, a copper layer was observed on the surface of a graphite anvil, which may have been caused by evaporation due to its high saturated vapor pressure [18]. During the ablation process, the evaporation of copper could absorb a large amount of heat and reduce the erosion of the sample. In this article, Cf/ZrC composites were fabricated by RMI with porous Cf/C preforms and Zr2Cu alloy at 1200 °C for 3 h in vacuum. ZrC as the main phase had a content of 42.2 ± 1.3 vol.%. The Cf/ZrC composites
Y. Zhu et al. / Scripta Materialia 67 (2012) 822–825
Figure 4. Optical images of Cf/ZrC composites (a) before ablation and (b) after ablation with an oxyacetylene torch, and micrographs of (c) the glass-like ZrO2 layer in the ablated center and (d) a loose ZrO2 particle near to the ablated center.
exhibited excellent ablation resistance after ablation with an oxyacetylene torch, with a mass loss rate of 0.0006 ± 0.0001 g s1 and a linear recession rate of 0.0003 ± 0.0001 mm s1. In the future, high-temperature properties such as flexural strength and modulus should be measured. The influence of the residual alloy and how to remove it will also be subjects for our future work. This work was supported by National Natural Science Foundation of China (No. 51002186 and 90916002). [1] S.N. Karlsdottir, J.W. Halloran, J. Am. Ceram. Soc. 90 (2007) 3233.
825
[2] S.M. Zhang, S. Wang, Y.L. Zhu, Z.H. Chen, Scripta Mater. 65 (2011) 139. [3] G.M. Song, Y.J. Wang, Y. Zhou, J. Mater. Sci. 36 (2001) 4625. [4] W. Sun, X. Xiong, B.Y. Huang, G.D. Li, H.B. Zhang, Z.K. Chen, X.L. Zheng, Carbon 47 (2009) 3365. [5] L. Zou, N. Wali, J.M. Yang, N.P. Bansal, J. Eur. Ceram. Soc. 30 (2010) 1527. [6] D. Zhao, C.R. Zhang, H.F. Hu, Y.D. Zhang, Compos. Sci. Technol. 71 (2011) 1392. [7] Y.G. Wang, X.J. Zhu, L.T. Zhang, L.F. Cheng, Ceram. Int. 37 (2011) 1277. [8] W.B. Hillig, J. Am. Ceram. Soc. 71 (1988) C-96–C-99. [9] H. Zhou, R.N. Singh, J. Am. Ceram. Soc. 78 (1995) 2456. [10] S.M. Zhang, S. Wang, W. Li, Y.L. Zhu, Z.H. Chen, J. Alloys Compd. 509 (2011) 8327. [11] D.W. Lipke, Y. Zhang, Y. Liu, B.C. Church, K.H. Sandhage, J. Eur. Ceram. Soc. 30 (2010) 2265. [12] W. Krenkel, J. Int, Appl. Ceram. Technol. 1 (2004) 188. [13] Y. Liu, D.W. Lipke, Y. Zhang, K.H. Sandhage, Acta Mater. 57 (2009) 3924. [14] J.N. Ness, T.F. Page, J. Mater. Sci. 21 (1986) 1377. [15] N.R. Calderon, M. Martı´nez-Escandell, J. Narciso, F. Rodrı´guez-Reinoso, Carbon 47 (2009) 220. [16] M.B. Dickerson, P.J. Wurm, J.R. Schorr, W.P. Hoffman, P.G. Wapner, K.H. Sandhage, J. Mater. Sci. 39 (2004) 6005. [17] W.M. Guo, J. Vleugels, G.J. Zhang, P.L. Wang, O. Vander Biest, Scripta Mater. 62 (2010) 802. [18] S.M. Zhang, S. Wang, W. Li, Y.L. Zhu, Z.H. Chen, Mater. Lett. 65 (2011) 2910. [19] Z. Grzesik, M.B. Dickerson, K.H. Sandhage, J. Mater. Res. 18 (2003) 2135. [20] M. Adabi, A. Amadeh, Int. J. Refract. Met. Hard Mater. 29 (2011) 31. [21] Y. Zhao, Y. Wang, Y. Zhou, P. Shen, Scripta Mater. 64 (2011) 229. [22] S. Nafisia, R. Coxb, R. Ghomashchia, J. Alloys Compd. 415 (2006) 99. [23] K.J. Zeng, M. Hamalainen, H.L. Lukas, J. Phase Equilib. 15 (1994) 577. [24] M.W. Chase Jr., NIST-JANAF Thermochemical Tables, fourth ed., Woodbury, New York, 1998.
Scripta Materialia 67 (2012) 822–825 www.elsevier.com/locate/scriptamat
Preparation of carbon fiber-reinforced zirconium carbide matrix composites by reactive melt infiltration at relative low temperature Yulin Zhu,⇑ Song Wang, Wei Li, Shouming Zhang and Zhaohui Chen National Key Laboratory of Science and Technology on Advanced Ceramic Fibers and Composites, College of Aerospace and Materials Engineering, National University of Defense Technology, Changsha 410073, China Received 19 April 2012; revised 11 July 2012; accepted 30 July 2012 Available online 4 August 2012
An attractive way to prepare carbon fiber-reinforced ZrC matrix composites was proposed and confirmed experimentally. The experimental results showed that the resulting composites were fabricated by immersing carbon/carbon preforms in molten Zr2Cu at 1200 °C for 3 h. ZrC was the main phase, with a content of 42.2 ± 1.3 vol.%. The composites exhibited excellent ablation resistance when undergoing ablation with an oxyacetylene torch, the mass loss rate and linear recession rate being 0.0006 ± 0.0001 g s1 and 0.0003 ± 0.0001 mm s1, respectively. Ó 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Carbides; Fiber-reinforced composites; Infiltration; Microstructure; Ablation resistance
Zirconium carbide (ZrC), which is a member of the ultrahigh-temperature ceramics family, has a number of unique properties, including an extremely high melting point, high strength and hardness, relatively low density, and high retained strength at high temperatures [1–5], which make it attractive for aerospace applications associated with hypersonic flight and rocket propulsion. However, ZrC ceramics are brittle and display little to no plasticity across a broad temperature range [5,6]. It has been proven in earlier works that introducing carbon fibers is an effective way to improve the fracture resistance of ZrC [5–7]. Known as a rapid and low-cost manufacturing process [5,8–14], the reactive melt infiltration (RMI) process has been widely used to produce structural components with complex geometries made of SiC- [14,15], ZrC[5,11,16] and HfC-matrix [13] composites. Recently, carbon fiber-reinforced zirconium carbide (Cf/ZrC) composites with porous carbon/carbon (Cf/C) preforms and pure Zr metal have been prepared by this method [5,7]. Due to its high melting point (1850 °C) [17], however, the Zr melt reacts strongly with fibers, which is detrimental to the reinforcement effect. Introducing copper to form Zr2Cu (melting point 1025 °C) can reduce the processing temperature effectively, and can wet and
⇑ Corresponding
author. Tel.: +86 731 84576441; fax: +86 731 84573165; e-mail: [email protected]
infiltrate well into a porous preform [2,10,18]. Zr2Cu instead of Zr has been used as an infiltrator to produce W/ZrC composites at temperatures as low as 1200 °C [11,16,19–21]. However, Cf/ZrC composites fabricated by the RMI process using Zr2Cu as the infiltrator have not yet been reported. In this study, an attractive way to prepare Cf/ZrC composites by RMI with Zr2Cu as a reactive infiltrator is proposed for the first time. The microstructure and composition of the composites were investigated. Furthermore, the ablation properties of the composites were also determined using an oxyacetylene torch. Felts with a fiber volume fraction of about 30 vol.% (T300, with an average diameter of about 7 lm, a tow size of 3 K, a carbon content of 93% (impurities of O + H + N + S 6 7%) and a bulk density of 1.76 g cm3; Toray, Japan) were prepared by a needle-punching technique with alternatively stacked 0° weftless piles, short-cut-fiber webs and 90° weftless piles. The preparation of Cf/C preforms included the following stages: first, phenolic resins (flaxen solid, manufactured by Taihang Co. in Xian) were smashed into powder and dissolved into ethanol with the weight ratio of 1:1; next, the carbon felts fixed by modular molds were located in a hermetic container; then the phenolic resins/ethanol solution were infused into the container and overwhelmed the felts when the container pressure was below 10 Pa. After dwelling there for 6 h, the saturated felts were cured at 80 °C for 4 h and then at 150 °C
1359-6462/$ - see front matter Ó 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.scriptamat.2012.07.044
Y. Zhu et al. / Scripta Materialia 67 (2012) 822–825
for 3 h. Finally, the cured felts were pyrolyzed at 1200 °C for 60 min in N2 (purity P 99.999 vol.%, O2 + H2O 6 0.001 vol.%) under atmospheric pressure, with a heating rate of 10–15 °C min1. Four cycles of infiltration and pyrolysis were necessary to obtain the Cf/C preforms with appropriate porosity. Commercial Zr2Cu ingots (purity P 99.2 wt.%, impurities (Hf + Fe + Ti + Si + Mg + Al + Ni) 6 0.8 wt.%; Hunan Rare Earth Metal & Material Institute, Hunan, China) were used as the reactive infiltrant. The ingots were prepared with spongy Zr pieces (purity P 99.6 wt.%) and copper cylinders (purity P 99.99 wt.%) by arc melting. To allow for homogenization, each ingot was flipped and remelted three times. The preparation of Cf/ZrC composites included the following stages: the Zr2Cu alloy was rapidly heated up to 1200 °C, the reactor keeping the preforms separated from the alloy during heating. After the alloy had melted completely, the porous Cf/C preforms were mechanically driven into the melt and kept there for 3 h, before being separated from the liquid Zr2Cu bath and left to cool naturally to room temperature. Any excess solidified melt adhering to the sample was removed by a grinding machining with a diamond wheel. The porosity of the specimen was equal to one minus the quotient of the bulk density to the theoretical density. The bulk density of the specimen was obtained from measurements of its dry weight and external dimensions. The theoretical densities were calculated using rule of mixtures calculations, based on the theoretical densities of 6.63 g cm3 for ZrC, 1.55 g cm3 for the deposited C, 8.96 g cm3 for Cu, 6.49 g cm3 for Zr and 1.76 g cm3 for the T300 fiber. The volume fractions of the solid phases were measured by inductively coupled plasma (ICP) and chemolysis analysis, which were described previously [22]. The density and open porosity of the specimen was measured using the Archimedes method, with distilled water as the immersing medium, after open porosity was saturated under vacuum for 12 h. The density was measured for at least five samples. In order to analyze the contents of each phase, the resulting composites were crushed into powder by milling machining with a diamond milling cutter and granulated using a 100-mesh sieve, before being immersed in an acid solution (1 part HCl, 1 part HNO3 and 2 parts distilled water) for 30 min. When tested by reacting with pure Zr, Cu and ZrC, respectively, this solution can dissolve Zr and Cu but not ZrC. After filtering, the fluid underwent ICP analysis to determine the amount of residual Zr and Cu; the solids left on the filter were dried in an oven at 80 °C for 5 h and then kept in muffle furnace at 1200 °C for 5 h, where the solid powders were oxidized thoroughly into ZrO2. The amount of ZrC was calculated from the content of ZrO2, and the amount of the residual carbon (containing carbon matrix and fiber) was equal to the weight difference between the drying powders and ZrC. Since the fiber fraction was a known quantity, the content of the residual carbon matrix could be calculated. The phases were analyzed by X-ray diffraction (XRD) with a Bruker D8 Advance instrument. The microstructures were observed by scanning electronic microscopy (SEM; Quanta-200) and energy-dispersive
823
spectroscopy (EDS). All specimens for SEM/EDS analyses were prepared by incised, ground and polished with a series of diamond pastes to a surface finish of 0.5 lm. Image analysis software package (ImageJ, National Institutes of Health, Bethesda, MD) was used to determine the average diameters of the fibers in the composites. The average diameter was calculated by measuring 30 fibers per image of at least 10 different SEM images. The ablation properties were tested with an oxyacetylene flame according to the GJB323A-96 Ablation Standard. The ablation tests were conducted in air, and the pressures of oxygen and acetylene were 0.4 and 0.095 MPa, respectively. The surface temperature of the sample was monitored with an optical pyrometer. During the test, the ablation gun, with a gunpoint diameter of 2 mm, was first ignited. When the fluxes of acetylene and oxygen were turned to 1.116 and 1.512 m3 h1, respectively, the gun was moved perpendicular to the sample surface. The distance between the tip of the gun and the top of the sample was 10 mm and the exposure time under the torch flame was 30 s. The heat flux was calculated from the heat absorbed by water flowing through a calorimeter, and was found to be 4200 kW m2 (10% error). At least three samples with dimensions of 30 mm 30 mm 3.5 mm were examined in each test. The mass and linear ablation rates (Rm and Rl) were defined as follows: Rm ¼ ðm0 m1 Þ=t Rl ¼ ðl0 l1 Þ=t where m0 and m1 represent the weight before and after ablation, respectively; l0 and l1 represent the thickness before and after ablation, respectively; and t is the ablation time. In detail, l1 is the average thickness of seven equidistant points along the longest chord of the ablated region with an approximately circular shape. The overall ablation rates of the composites were taken from three specimens on average. The reaction between carbon and Zr2Cu melt can be expressed as Reaction (1). Thermodynamics calculations were conducted in the standard state to estimate the direction of this reaction. The thermodynamics data of the Zr2Cu phase were calculated using Pandat 7.0 software based on the database optimized by Zeng et al. [23]. The thermodynamics data of the other reactant and products are from Chase [24]. The changes in Gibbs’ free energy and enthalpy above the melting point of Zr2Cu are shown in Figure 1, which reveals that
Figure 1. Changes in Gibbs’ free energy and enthalpy as a function of temperature above 1200 °C for Reaction (1).
824
Y. Zhu et al. / Scripta Materialia 67 (2012) 822–825
Reaction (1) is strongly favored thermodynamically as low as 1200 °C. XRD results (Fig. 2a) revealed that the alloy had a stoichiometric phase of Zr2Cu. 2CðsÞ þ Zr2 CuðlÞ ! 2ZrCðsÞ þ CuðlÞ at T 1200 C
ð1Þ
The Cf/C preforms, with a fiber fraction of 31.9 ± 2.7 vol.%, had a density of 1.01 ± 0.03 g cm3 and a porosity of 39.1 ± 2.4%, with the pore size mainly in the range of 10–40 lm; however, the infiltrated specimens had a density of 4.30 ± 0.25 g cm3 and a porosity of 7.9 ± 1.1%. The weight change was about 326%, which was mainly attributed to the densification of the preforms and the transformation from C to ZrC. Since the experimental density was extremely close to the theoretical value of 4.33 ± 0.24 g cm3, the weight change was quantitatively consistent with the phase content and porosity of the reacted specimen. The XRD pattern in Figure 2b reveals that there were three phases in the resulting composite: carbon fibers, ZrC and Cu. ICP and chemolysis revealed that the Cf/ZrC composites contained 42.2 ± 1.3 vol.% ZrC, 8.8 ± 2.3 vol.% residual deposited C, 8.0 ± 1.2 vol.% Cu and 1.4 ± 0.6 vol.% Zr, which means that ZrC was the main phase and that a small quantity of metals remained in the composites. However, according to Wang’s research [7], no detectable residual zirconium was found in the C/C–ZrC composites prepared with pure Zr liquid. The residual metals in our composites could be ascribed to the gentle infiltrating temperature and the dilution of Cu. Microstructural characterization was conducted to analyze the distribution of phases in the Cf/ZrC composites. Figure 3a shows a representative microstructure of the cross-section. It reveals a dense specimen with a few fine (<1 lm) isolated pores (indicated by arrows in Fig. 3e), and carbon fibers surrounded by gray matrix. Image analyses revealed an average fiber diameter of 7.0 ± 0.3 lm, indicating that the fibers were not eroded by the Zr2Cu liquid. The gray part was magnified to observe the details and is presented in Figure 3b, which shows many dark island-like particles distributed throughout the light matrix. As demonstrated by the EDS analysis results (Fig. 3c and d), these island-like particles (indicated in region A) contained about 51 at.% C and 49 at.% Zr, while the surrounding substance (indicated in region B) contained about 14 at.% C, 6 at.% Zr and 80 at.% Cu. Taking account of the XRD results shown in Figure 2b, these island-like
Figure 2. XRD patterns of (a) the alloy and the Cf/ZrC composite (b) before ablation and (c) after ablated by oxyacetylene torch.
Figure 3. Micrographs of (a) a representative cross-section of the Cf/ZrC composites and (b) the matrix of the Cf/ZrC composites; EDS patterns of (c) elements in region A and (d) elements in region B.
particles were ZrC grains and the surrounding material was residual Cu-enriched phase. Since the composites are used to produce high-temperature components, in which ablation is the main cause of material failure, an oxyacetylene flame ablation test was conducted to estimate the resistance. After ablation for 60 s, the Cf/ZrC composites showed a mass loss rate of 0.0006 ± 0.0001 g s1 and a linear recession rate of 0.0003 ± 0.0001 mm s1, nearly one magnitude lower than that of the C/C–ZrC composite [7], which was prepared with pure Zr at >1900 °C and had a mass loss rate of 0.004 ± 0.001 g s1 and a linear recession rate of 0.002 ± 0.001 mm s1. The excellent ablation resistance of the Cf/ZrC composites might be related to the presence of residual Cu. Figure 4b shows the optical images of the ablated Cf/ZrC composites. Compared with the raw sample shown in Figure 4a, the ablated sample remained flat and showed no apparent ablated pitting, other than a white oxide shell on the surface. In the center of the sample, a glass-like layer can be observed, and a loose layer formed around the center (Fig. 4b). These two parts are magnified in Figure 4c and d, which reveals that the glass-like layer is composed of dense and continuous molten ZrO2, while the loose layer consists of a large quantity of ZrO2 particles. XRD analysis indicated that the main component of the white ablation product was monoclinic ZrO2 (m-ZrO2) (Fig. 2c). Additionally, a copper layer was observed on the surface of a graphite anvil, which may have been caused by evaporation due to its high saturated vapor pressure [18]. During the ablation process, the evaporation of copper could absorb a large amount of heat and reduce the erosion of the sample. In this article, Cf/ZrC composites were fabricated by RMI with porous Cf/C preforms and Zr2Cu alloy at 1200 °C for 3 h in vacuum. ZrC as the main phase had a content of 42.2 ± 1.3 vol.%. The Cf/ZrC composites
Y. Zhu et al. / Scripta Materialia 67 (2012) 822–825
Figure 4. Optical images of Cf/ZrC composites (a) before ablation and (b) after ablation with an oxyacetylene torch, and micrographs of (c) the glass-like ZrO2 layer in the ablated center and (d) a loose ZrO2 particle near to the ablated center.
exhibited excellent ablation resistance after ablation with an oxyacetylene torch, with a mass loss rate of 0.0006 ± 0.0001 g s1 and a linear recession rate of 0.0003 ± 0.0001 mm s1. In the future, high-temperature properties such as flexural strength and modulus should be measured. The influence of the residual alloy and how to remove it will also be subjects for our future work. This work was supported by National Natural Science Foundation of China (No. 51002186 and 90916002). [1] S.N. Karlsdottir, J.W. Halloran, J. Am. Ceram. Soc. 90 (2007) 3233.
825
[2] S.M. Zhang, S. Wang, Y.L. Zhu, Z.H. Chen, Scripta Mater. 65 (2011) 139. [3] G.M. Song, Y.J. Wang, Y. Zhou, J. Mater. Sci. 36 (2001) 4625. [4] W. Sun, X. Xiong, B.Y. Huang, G.D. Li, H.B. Zhang, Z.K. Chen, X.L. Zheng, Carbon 47 (2009) 3365. [5] L. Zou, N. Wali, J.M. Yang, N.P. Bansal, J. Eur. Ceram. Soc. 30 (2010) 1527. [6] D. Zhao, C.R. Zhang, H.F. Hu, Y.D. Zhang, Compos. Sci. Technol. 71 (2011) 1392. [7] Y.G. Wang, X.J. Zhu, L.T. Zhang, L.F. Cheng, Ceram. Int. 37 (2011) 1277. [8] W.B. Hillig, J. Am. Ceram. Soc. 71 (1988) C-96–C-99. [9] H. Zhou, R.N. Singh, J. Am. Ceram. Soc. 78 (1995) 2456. [10] S.M. Zhang, S. Wang, W. Li, Y.L. Zhu, Z.H. Chen, J. Alloys Compd. 509 (2011) 8327. [11] D.W. Lipke, Y. Zhang, Y. Liu, B.C. Church, K.H. Sandhage, J. Eur. Ceram. Soc. 30 (2010) 2265. [12] W. Krenkel, J. Int, Appl. Ceram. Technol. 1 (2004) 188. [13] Y. Liu, D.W. Lipke, Y. Zhang, K.H. Sandhage, Acta Mater. 57 (2009) 3924. [14] J.N. Ness, T.F. Page, J. Mater. Sci. 21 (1986) 1377. [15] N.R. Calderon, M. Martı´nez-Escandell, J. Narciso, F. Rodrı´guez-Reinoso, Carbon 47 (2009) 220. [16] M.B. Dickerson, P.J. Wurm, J.R. Schorr, W.P. Hoffman, P.G. Wapner, K.H. Sandhage, J. Mater. Sci. 39 (2004) 6005. [17] W.M. Guo, J. Vleugels, G.J. Zhang, P.L. Wang, O. Vander Biest, Scripta Mater. 62 (2010) 802. [18] S.M. Zhang, S. Wang, W. Li, Y.L. Zhu, Z.H. Chen, Mater. Lett. 65 (2011) 2910. [19] Z. Grzesik, M.B. Dickerson, K.H. Sandhage, J. Mater. Res. 18 (2003) 2135. [20] M. Adabi, A. Amadeh, Int. J. Refract. Met. Hard Mater. 29 (2011) 31. [21] Y. Zhao, Y. Wang, Y. Zhou, P. Shen, Scripta Mater. 64 (2011) 229. [22] S. Nafisia, R. Coxb, R. Ghomashchia, J. Alloys Compd. 415 (2006) 99. [23] K.J. Zeng, M. Hamalainen, H.L. Lukas, J. Phase Equilib. 15 (1994) 577. [24] M.W. Chase Jr., NIST-JANAF Thermochemical Tables, fourth ed., Woodbury, New York, 1998.