- Email: [email protected]
Interface engineering of fiber-reinforced all-oxide composites fabricated by the sol–gel method with fugitive pyrolytic carbon coatings
Composites Part B 75 (2015) 86e94
Contents lists available at ScienceDirect
Composites Part B journal homepage: www.elsevier.com/locate/compositesb
Interface engineering of fiber-reinforced all-oxide composites fabricated by the solegel method with fugitive pyrolytic carbon coatings Yi Wang, Haitao Liu*, Haifeng Cheng, Jun Wang Science and Technology on Advanced Ceramic Fibers and Composites Laboratory, National University of Defense Technology, Changsha 410073, PR China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 7 October 2014 Received in revised form 2 December 2014 Accepted 13 January 2015 Available online 29 January 2015
This paper describes the interface engineering of threeedimensional (3D) Nextel™440 fiber-reinforced aluminosilicate composites fabricated by the solegel method with fugitive pyrolytic carbon (PyC) coatings. The coating thickness on the fiber strength, interfacial characteristics and there corresponding effects on mechanical properties of the composites were investigated. The study shows that the fiber strength was influenced by the coating thickness and optimized with the thickness of 0.15 mm. The composites with uncoated fibers showed brittle fracture behavior without fiber pullout because of strong interactions between the fiber and the matrix. However, higher strengths and pseudo-ductile fracture behaviors were obtained in the composites with PyC interphases, where different deflections and branches of propagating cracks and fiber pullout patterns were observed. Moreover, induced fugitive PyC interface conditions have great effects on the density, microstructure and mechanical properties of the resultant composites. © 2015 Elsevier Ltd. All rights reserved.
Keywords: A. Ceramicematrix composites (CMCs) B. Interphase D. Mechanical testing Solegel method
1. Introduction Advanced applications, such as aircraft turbine engine components, land-based turbines, hypersonic missiles and fighting vehicles, and most recently, spacecraft re-entry thermal protection systems, require structural materials that exhibit superior longterm mechanical properties under high temperatures, high pressures and various environmental factors, such as moisture [1]. Continuous fiber-reinforced ceramic matrix composites (CFRCMCs) are the most attractive material concept that can meet such needs. Oxide/oxide CMCs, in particular, provide high strength, toughness, notch insensitivity, refractoriness and environmental stability at high application temperatures, where metals are usually limited by their melting temperature and monolithic ceramics are limited by their low damage tolerance [2,3]. Conventional processing methods for oxide/oxide composites, such as reaction-sintering and hot-pressing, normally require high temperatures and high pressures to provide adequate consolidation of the ceramic body [4]. Fiber degradation and interface reactions, which contribute to brittle fractures, will inevitably occur under
* Corresponding author. Tel.: þ86 731 84576440; fax: þ86 731 84576578. E-mail address: [email protected] (H. Liu). http://dx.doi.org/10.1016/j.compositesb.2015.01.018 1359-8368/© 2015 Elsevier Ltd. All rights reserved.
such conditions [5]. Several new fabrication processes, including slurry infiltration and hot-pressing (SIeHP) [6], electrophoretic deposition (EPD) [7], precursor infiltration and pyrolysis (PIP) [8] and the solegel method [9], have been introduced to address these issues. Among these techniques, the solegel method is the most feasible one for fabricating three-dimensional (3D) composites because of its low densification temperature (<1300 C), low shrinkage, reduced drying stress and near-stoichiometric matrix composition. The solegel method is also a near-net-shape technique that can be used to prepare large products with complex shapes directly [10]. General properties and issues for advanced oxide/oxide composites are summarized in Table 1. Interfaces between the fibers and the matrix in oxide/oxide composites are extremely important considerations for structural applications. An optimized interface can bring the deflection of cracks at the interface, fiber pullout during the fracture, excellent interfacial shear strength and ductility between the fiber and matrix, which contribute to high-energy consumption and excellent mechanical properties [11e13]. Among the interface concepts suggested recently [14e17], fugitive carbon is recognized as one of the most potential one. It can be readily deposited onto tows or woven fabrics by chemical vapor deposition (CVD) or pyrolysis of organic precursors, and be readily oxidized at moderately high temperature.
Y. Wang et al. / Composites Part B 75 (2015) 86e94
87
Table 1 List of developmental oxide/oxide composites for structural materials and corresponding issues. Category
SIeHP
EPD
PIP
Solegel
Fiber Interphase Matrix Porosity (%) General issues Interlaminar shear strength Hermeticity Near-net-shape technique
Nextel™720 None Mullite þ alumina ~32
Nextel™720 NdPO4 Mullite ~20
Nextel™720 None Alumina ~20
Nextel™480 BN Mullite ~10
Low Poor No
Low Poor No
Moderate-high Good Yes
Moderate-high Good Yes
Weaver and Keller et al. [18,19] investigated the effects of fugitive carbon coating thickness on mechanical properties of one or two-dimensional oxide/oxide composites fabricated by SIeHP process. Results showed that fugitive carbon coatings can remarkably reduce the interfacial sliding resistance and enhance the damage tolerance of the composites, even after exposure at 1000 C for 500 h. However, fugitive carbon coatings have rarely been used in 3D oxide/oxide composites. The present study focuses on the influence of fugitive pyrolytic carbon (PyC) coating thickness on the density, microstructure and mechanical properties of 3D Nextel™440 fiber-reinforced aluminosilicate (N440/AS) composites fabricated by the solegel method. 2. Material and methods 2.1. Raw materials Nextel™440 fibers (3M, USA) were used as reinforcements. The fibers were composed of 70 wt% Al2O3, 28 wt% SiO2 and 2 wt% B2O3. The tensile strength and Young's modulus announced by the manufacture were 2070 MPa and 186 GPa, respectively. Diphasic Al2O3eSiO2 sols (Suzhou Nanodispersions Ltd., China) were used as precursors of the aluminosilicate matrix. The weight ratio of Al2O3 and SiO2 particles in the sols was 1:1. The density and viscosity of the sols were 1.12 g/cm3 and 6 mpa s, respectively. 2.2. Sample preparation Mixtures of propylene and argon gas were used to deposit PyC on fiber fabrics with size of about 60 mm 100 mm at 1000 C by CVD. The coatings were deposited for 1, 4 and 8 h, with the resulting thickness of about 0.15, 0.42 and 1.00 mm. Then coated fabrics were stacked to a thickness of about 4 mm. Finally, the 3D architecture was finished by Z-stitching the stacked fabrics with Nextel™440 fiber yarn in a 2.5 mm 2.5 mm space in Changzhou Bolong Aerospace Technology Ltd., China. The fiber volume fraction of the preform is approximately 40%. 3D N440/AS composites were prepared by the solegel method according to our previous work [20]. The composites were sintered at 1000 C with a heating rate of 10 C/min. Composites with PyC interphases were denoted according to their coating thickness as C0.15, C0.42 and C1.00. Samples were fired at 600 C under air atmosphere for 2 h to oxidize PyC. Composites with fugitive PyC coatings were denoted as FC0.15, FC0.42 and FC1.00. 2.3. Characterization The density and porosity of the samples were determined by the Archimedes principle, using distilled water as the immersion medium. The theoretical density was calculated from the ratio of the alumilosilicate matrix and the fiber volume fraction. The tensile strength of the PyC-coated fibers was tested at room temperature
on a universal strength machine (Testometric, M3505CT) equipped with a 5 N load cell. The gauge length and crosshead speed were 20 mm and 0.2 mm/min, respectively. Three-point bending tests (test bars 50l 4w 3t mm3) were carried out at room temperature, with support span of 40 mm and cross-head speed of 0.5 mm/ min in an INSTRON 1342 testing machine, using the number of five specimens. The flexural stress (s) and elastic modulus (E) were calculated from Eqs. (1) and (2), respectively:
s ¼ 1:5PL wt 2
(1)
E ¼ 0:25mL3 wt 3
(2)
where P is the load at a point of deflection of a stress/displacement curve in the test, L is the support span, w is the specimen width, t is the specimen thickness and m is the slop of the tangent to the initial linear portion of the loadedeflection curve. Microstructure analysis of PyC-coated fibers and composites after mechanical tests was done by a scanning electron microscope (SEM, Hitachi FEG S4800). The pullout fiber surface of the composites was analyzed by EDS equipped with SEM. 3. Results and discussion 3.1. Fiber surface coatings Fig. 1 shows SEM images of uncoated and PyC-coated Nextel™440 fibers formed by CVD. The fibers were well coated; with thicknesses of about 0.15, 0.42 and 1.00 mm. The surfaces of PyCcoated fibers were smooth and uniform, without relation to the coating thickness. As shown in Fig. 1(c) and (a) narrow gap appeared between the fiber and PyC coating during the CVD cooling procedure, due to the different thermal expansion coefficient (fiber, 5.3 106/ C; carbon, 2.5 106/ C). The inference was that the coatings couldn't isolate the fibers from matrix completely. That is to say, interfacial reaction between the fibers and the matrix might occur, and strong interfacial bonding might be obtained. Single fiber tensile tests were used to evaluate the fiber tensile strength, and the strength distribution was evaluated by the twoparameter Weibull model [21]. Weibull plots of uncoated and PyC-coated fibers are shown in Fig. 2, and the corresponding Weibull parameters are listed in Table 2. In the Weibull model, larger m values of the fibers mean fewer defects and lower dispersion of the tensile strength [22]. The m value decreased as the coating thickness increased, indicating higher distribution of the tensile strength in coated fibers. Fibers with 0.15 mm PyC coating showed higher strength compared with uncoated fibers, while fibers with 0.42 and 1.00 mm PyC coatings showed lower strengths. The strength increase in 0.15 mm PyC coating might be attributed to the ability of this coating to heal surface flaws, as reported by Chawla et al. [23]. In the case of fibers with thick PyC coatings, fiber
88
Y. Wang et al. / Composites Part B 75 (2015) 86e94
Fig. 1. Cross-sectional of the uncoated and PyC-coated Nextel™440 fibers formed by CVD: (a) uncoated, (b) CVDe0.15, (c) CVDe0.42, and (d) CVDe1.00.
strength was more affected by the soft nature of the coating than the surface flaw healing effect. Moreover, the strength decrease might also originate from thermal degradation during CVD to produce desired thickness.
3.2. Densification of the composites The density and porosity of N440/AS composites fabricated by the solegel method are summarized in Table 3. The density of the composites decreased as coating thickness increased because thicker PyC coating blocked sol infiltration [24]. The matrix porosity of the composites obtained by the solegel method was compared with that of the composites obtained by SIeHP. As shown in Fig. 3a, 0 and 90 fiber tows, as well as numerous matrix cracks, were
observed in Nextel™720/alumina (N720/A) composites fabricated by SIeHP [25]. Most of these cracks were shrinkage cracks formed during processing rather than matrix cracks generated during loading. The porous structure of the alumina matrix was confirmed in the high-magnification image (Fig. 3b). By comparison, no cracks were observed in the solegel composites (Fig. 3c), although a large number of holes originating from precursor shrinkage [26] distributed in the matrix. The dense and homogeneous matrix was beneficial to the hermeticity and interlaminar shear strength of the final composites. The matrix in the composites was composed of aluminosilicate grains with small amounts of amorphous silica, as reported previously [27]. The presence of amorphous silica could improve the sinter-ability of the matrix, while aluminosilicate could improve environmental stability [28].
3.3. Microstructure of the composites
Fig. 2. Weibull plots of the tensile test for the fibers with and without PyC coatings.
3.3.1. Composites with PyC interphases Cross-section morphology of the N440/AS composites with PyC interphases is shown in Fig. 4. In the composites without interface engineering (Fig. 3d), the fibers were surrounded tightly by the matrix, without any interfacial debonding observed, indicating excess strong bonding between the fibers and the matrix. After PyC interphases were introduced into the composites, the number of micro-pores in the matrix increased, especially for those at intrafiber-bundles (Fig. 4(a,c,e)). As shown in Fig. 4(b,d,f), the PyC coating could be clearly defined between the fibers and the matrix. Meanwhile, interfacial debonding induced either by the polishing process or thermal mismatch was more and more prevalent, as the coating thickness increased from 0.15 mm to 1.00 mm. It can be noted that the narrow gap was preserved between the fibers and coatings during the solegel process. That is to say, PyC coatings have been successfully introduced into the N440/AS composites, and the interface properties of the composites were well-controlled.
Y. Wang et al. / Composites Part B 75 (2015) 86e94 Table 2 Weibull parameters of the tensile test for the uncoated and PyC-coated fibers. Deposition temperature (K)
Weibull modulus, m
s0 (GPa)
Scale parameter,
Tensile strength (GPa)
No coating CVDe0.15 CVDe0.42 CVDe1.00
5.46 4.72 4.39 2.93
2.03 2.09 1.60 1.41
1.87 1.91 1.46 1.17
± ± ± ±
0.38 0.41 0.39 0.46
Table 3 Properties of the N440/AS composites with various interface condition. Samples
Density (g/cm3)
Porosity (%)
Flexural strength (MPa)
Uncoated C0.15 FC0.15 C0.42 FC0.42 C1.00 FC1.00
2.22 2.20 2.15 2.17 2.09 2.10 2.03
19.0 19.7 21.5 20.8 23.7 23.4 25.9
31.2 91.1 93.1 95.4 66.7 78.2 50.5
± ± ± ± ± ± ±
3.4 3.6 4.5 18.2 8.8 17.7 6.7
Elastic modulus (GPa) 33.0 26.0 22.2 30.1 25.0 24.1 15.6
± ± ± ± ± ± ±
5.2 1.6 0.7 2.4 1.6 1.5 3.1
3.3.2. Composites with fugitive PyC coatings Cross-section morphology of the N440/AS composites with fugitive PyC coatings is shown in Fig. 5. As displayed in Fig. 5(a,c,e), the number of micro-pores at intra-fiber-bundles increased as the original coating thickness increased. A large number of matrix inner fiber bundles missed during polishing process because of weak interfacial bonding. As shown in Fig. 5(b,d,f), the gaps induced by the removal of PyC were observed between the fibers and the matrix, and the gap width increased as the coating thickness increased. Furthermore, it can be noted that the gap width was not consistent with the corresponding PyC coating thickness. A reasonable explanation was that thermal mismatch did exist among the fibers, matrix and PyC interphases. As a result, a much narrower gap was obtained after PyC removal. The isolation of the
89
fibers from matrix might contribute to the improvement of the composite mechanical properties.
3.4. Mechanical properties and fracture behavior of the composites 3.4.1. Composites with PyC interphases Typical flexural stressestrain curves of the N440/AS composites with PyC interphases are shown in Fig. 6. The stressestrain curve of the composite with uncoated fibers showed a catastrophic failure attributed to strong fiberematrix interfacial bonding. The corresponding flexural strength and elastic modulus were 27.3 MPa and 36.5 GPa, respectively. In contrast, all the stressestrain curves of the composites with PyC interphases showed extended regions after initial deviation from linearity. Areas under these curves were much larger compared with those of uncoated fiber-reinforced composites. Sample C0.42, with a coating thickness of 0.42 mm, showed the highest flexural strength of 102 MPa. Mechanical properties of the N440/AS composites are listed in Table 3. As the original coating thickness increased the average flexural strength of the composites increased firstly and then decreased. The highest values (91.1 or 95.4 MPa) were obtained for the composites with PyC interphases of moderate thickness, i.e. 0.15 or 0.42 mm thickness. It can be concluded that the flexural behavior of the N440/AS composites could be improved by tailoring the fiber/matrix interface with PyC coatings. Fig. 7 illustrates the fracture surface of the composites without and with PyC interphases. For the composites with uncoated fibers (Fig. 7a), the fracture surface was very even, and nearly no pullout fibers could be observed. Whereas for the composites with PyC interphases (Fig. 7((b)e(d)), all fracture surfaces showed evident fiber pullout behaviors, and lengths of pulledout fibers could exceed 100 mm. The interfacial debonding could be clearly observed. As the coating thickness increased, the phenomena that PyC coatings attached to pullout fibers became more and more prevalent. This finding was confirmed by EDS analysis of the pullout fiber surface of sample C0.42 (Fig. 9(a,b)). Moreover, it can be
Fig. 3. Typical cross-section of oxide/oxide composites fabricated by (a, b) SIeHP (Adapted from Ruggles-Wrenn et al. [24]) and (c, d) solegel method.
90
Y. Wang et al. / Composites Part B 75 (2015) 86e94
Fig. 4. Typical cross-section morphology of the N440/AS composites with PyC interphases: (a, b) C0.15, (c, d) C0.42 and (e, f) C1.00.
noted that the surface of pullout fibers was free of any matrix, suggesting weak chemical corrosion to fibers during fabrication and the absence of debonding within the matrix. Increasing the coating thickness also promoted the increase in the length and uniformity of pullout fibers, indicating the decrease in interfacial bonding. 3.4.2. Composites with fugitive PyC coatings As shown in Fig. 6, typical flexural stressestrain curves of the N440/AS composite with fugitive PyC coatings exhibited obvious toughened fracture behavior. When the load reached the maximum, it dropped off gradually. The large extended area under the curve indicated a greater amount of fracture energyconsumption in the composites. The highest flexural strength of 98.8 MPa was achieved for sample FC0.15, with the gap width of 0.15 mm. As shown in Table 3, two trends could be found after PyC removal: (1) the average flexural strength of the composites decreased to a degree, especially for the composites with PyC interphases of 1.00 mm thickness; (2) the elastic modulus of the composites also decreased, owing to further weaken of the interfacial bonding. The composite with a gap width of 1.00 mm showed single, long-cracking behavior and the lowest elastic modulus (14.2 GPa). A reasonable explanation is that the gap was too wide for load transfer from matrix to the fibers, although the in-situ fiber
strength of the composite was very high. Apparently, the gap width had remarkable impacts on the flexural behavior of the N440/AS composites with fugitive PyC coatings. Fig. 8 illustrates the fracture surface of the composites with fugitive PyC interphases. The fiber pullout behaviors of all composites with fugitive PyC coatings were distinct, and the length of pullout fibers was uniform. The surface of pullout fibers was very smooth, without any attached coatings, as confirmed by EDS analysis of the pullout fiber surface of sample FC0.42 (Fig. 9(c,d)). The length of pullout fibers in sample FC1.00 was so long (>300 mm) that the load couldn't transmit from matrix to the fibers efficiently. Consequently, low flexural strength and elastic modulus were observed in sample FC1.00. In general, mechanical properties of CFRCMCs are determined by the in-situ fiber strength and interfacial characteristics. Concerning the composites without interface engineering, the fibers were subjected to strong chemical corrosion of matrix and interfacial reactions would occur. As a result, low in-situ fiber strength and strong interfacial bonding were obtained, resulting in brittle fracture behaviors and poor mechanical properties of the composites. While for the composites with PyC interphases, the coatings could protect the fibers from being corroded by matrix. In addition, interfacial debonding would easily occur within PyC interphases because of low van der Waals forces between the basal
Y. Wang et al. / Composites Part B 75 (2015) 86e94
91
Fig. 5. Typical cross-section morphology of the N440/AS composites with fugitive PyC interphases: (a, b) FC0.15, (c, d) FC0.42 and (e, f) FC1.00.
Fig. 6. Typical flexural stressedisplacement curves of the N440/AS composites with various interface condition.
planes of PyC [29]. Consequently, high in-situ fiber strength and weak interfacial bonding both improved mechanical properties of the composites with interface engineering. After PyC removal, gaps appeared at the fiber/matrix interface, the width of which was determined by the PyC coating thickness and the thermal mismatch among the fibers, PyC and matrix. The thermal expansion coefficient of the fibers was equivalent to that of matrix, but higher than that of PyC. The inference is that the thermal mismatch between the fibers and PyC would lead to the compressive strain in the PyC, so as to that between PyC and matrix. That is to say, the thermal mismatch among the fibers, PyC and matrix in the composites would counteract to a lower level. As a result, we defined an ideal gap width between the fibers and the matrix after PyC removal, the value of which was equal to the original coating thickness. According to Weaver et al. [18], the gaps between the fibers and the matrix are certain to produce interfaces with zero tensile strength, but frictional sliding will be preserved by mechanical interlocking of asperities at the fiber/matrix boundary on condition that the asperity height exceeds the gap width. Similar mechanism has also been reported by Keller et al. [19]. For an idealized smooth, straight gap (Fig. 10a), the fiber is unconstrained, thus the load can't transfer from the matrix to the fibers. Typical fiber pullout behavior, with extremely long lengths, will be observed. While for a real gap,
92
Y. Wang et al. / Composites Part B 75 (2015) 86e94
Fig. 7. Typical fracture surface morphology of the N440/AS composites with PyC interphases: (a) uncoated, (b) C0.15, (c) C0.42, and (d) C1.00.
load transfer is achieved by the combined effects of mechanical interlocking, intermittent fiber/matrix bonding and roughness of both the fiber and matrix. As shown in Fig. 10(b), when the crack reaches the interface, matrix strain will cause sliding at the interface. Then point contacts between the fiber and matrix are formed, where mechanical interlocking and possible bonding occurs. The
optimal degree of mechanical coupling between the fibers and the matrix can be achieved through tailoring the gap width, which is the intention of this work. When the gap is too wide (>0.42 mm), the mechanical coupling between the fibers and the matrix is limited, and the load will deflect easily at the fiber/matrix interface, with little energy-consuming, as proved by the decrease in the
Fig. 8. Typical fracture surface morphology of the N440/AS composites with fugitive PyC interphases: (a) FC0.15, (b) FC0.42, and (c) FC1.00.
Y. Wang et al. / Composites Part B 75 (2015) 86e94
93
Fig. 9. EDS analysis of the pullout fiber surface in N440/AS composites with various interface condition: (a, b) C0.42 and (c, d) FC0.42.
elastic modulus. This work indicated that the optimal fugitive PyC thickness in the N440/AS composites should be less than 0.15 mm, which is consistent with the result of Ref. [19] (<0.1 mm).
4. Conclusions We have studied the effects of CVD PyC coating thickness on the strength of Nextel™440 fibers. The fugitive PyC coatings were introduced into 3D Nextel™440 fiber-reinforced aluminosilicate composites fabricated by the solegel method. The influences of the fugitive PyC coating thickness on the density, microstructure, and mechanical properties of the resultant composites were investigated. The major results obtained are summarized as follows:
(1) The surface of PyC-coated fibers was smooth, without relation to the coating thickness. The optimal strength was obtained when the coating thickness was 0.15 mm because thin coating could heal the fiber surface flaws. (2) The dense and homogeneous matrix of the composites showed nearly no cracks but a large number of holes. As the coating thickness increased, the density of the composites decreased because thicker PyC coating blocked sol infiltration. (3) The fracture surface of composites with uncoated fibers was very even, while that of the composites with PyC or fugitive PyC interphases illustrated typical fiber pullout behaviors that were affected by interface conditions remarkably. (4) The composites with uncoated fibers showed brittle fracture behavior and inferior mechanical properties. Higher strengths with pseudo-ductile fracture behavior were obtained using all PyC coating thickness. Mechanical properties of the composites with fugitive PyC coatings were determined by the PyC coating thickness, and the optimal fugitive PyC thickness should be less than 0.15 mm. The present approach has demonstrated the significant implication of fugitive PyC coatings in interfacial engineering in particular on the high-performance and damage-tolerance of fiber-reinforced all-oxide composites. On account of the thermal instability of the gap derived from the removal of PyC interphases, the present research points out that future effort should focus on evaluating high-temperature service lifetime of the composites.
Acknowledgments
Fig. 10. Schematic diagrams of load transfer in the composites with fugitive coatings for fiber protection: (a) ideal case and (b) real case.
The authors appreciate the financial support of the National Natural Science Foundation of China (51202291 and 51302314), Aid Program for Innovative Group of National University of Defense Technology, and Aid program for Science and Technology Innovative Research Team in Higher Educational Institutions of Hunan Province.
94
Y. Wang et al. / Composites Part B 75 (2015) 86e94
References [1] Armani CJ, Ruggles-Wrenn MB, Hay RS, Fair GE. Creep and microstructure of Nextel™ 720 fiber at elevated temperature in air and in steam. Acta Mater 2013;61(16):6114e24. [2] Simon RA, Danzer R. Oxide fiber composites with promising properties for high-temperature structural applications. Adv Eng Mater 2006;8(11): 1129e34. [3] Hülsenberg D, Fehling P, Leutbecher T. Damage tolerant, translucent oxide fiber/glass matrix composites. Compos Part B 2008;39(2):362e73. [4] Kaya C, Kayab F, Butler EG, Boccaccini AR, Chawla KK. Development and characterisation of high-density oxide fibre-reinforced oxide ceramic matrix composites with improved mechanical properties. J Eur Ceram Soc 2009;29(9):1631e9. [5] Schmücker M, Flucht F, Mechnich P. Degradation of oxide fibers by thermal overload and environmental effects. Mater Sci Eng A 2012;557:10e6. [6] Holmquist MG, Lange FF. Processing and properties of a porous oxide matrix composite reinforced with continuous oxide fibers. J Am Ceram Soc 2003;86(10):1733e40. [7] Boccaccini AR, Kaya C, Chawla KK. Use of electrophoretic deposition in the processing of fibre reinforced ceramic and glass matrix composites: a review. Compos Part A 2001;32(8):997e1006. [8] Mattoni MA, Yang JY, Levi CG, Zok FW. Effects of matrix porosity on the mechanical properties of a porous-matrix, all-oxide ceramic composite. J Am Ceram Soc 2001;84(11):2594e602. [9] Chawla KK. Interface engineering in mullite fiber/mullite matrix composites. J Eur Ceram Soc 2008;28(2):447e53. [10] Russell-Floyd RS, Harris B, Cooke RG, Laurie J, Hammett FW, Jones RW, et al. Application of sol-gel processing techniques for the manufacture of fiberreinforced ceramics. J Am Ceram Soc 1993;76(10):2635e43. [11] Wang XQ, Zhang JF, Wang ZQ, Liang WY, Zhou LM. Finite element simulation of the failure process of single fiber composites considering interface properties. Compos Part B 2013;45(1):573e80. [12] Santhosh U, Ahmad J, John R, Ojard G, Miller R, Gowayed Y. Modeling of stress concentration in ceramic matrix composites. Compos Part B 2013;45(1): 1156e63. [13] Yan LB, Chouw N, Jayaraman K. Flax fibre and its composites e a review. Compos Part B 2014;56:296e317. [14] Raab D, Pfeifer K, Hülsenberg D, Boccaccini AR. Mechanical properties of oxide-fiber-reinforced glass matrix composites with BN or SnO2 interfaces. Mater Sci Eng A 2006;417(1e2):314e47.
[15] Davis JB, Marshall DB, Morgan PED. Monazite-containing oxide/oxide composites. J Eur Ceram Soc 2000;20(5):583e7. [16] Boakye E, Petry MD, Hay RS. Porous aluminum oxide and lanthanum phosphate fiber coatings. Ceram Eng Sci Proc 2008;17(4):53e60. [17] Wang Y, Liu HT, Cheng HF, Wang J. Effective fugitive carbon coatings for the strength improvement of 3D Nextel™ 440/aluminosilicate composites. Mater Lett 2014;126(1):236e9. [18] Weaver JH, Yang J, Zok FW. Control of interface properties in oxide composites via fugitive coatings. J Am Ceram Soc 2008;91(12):4003e8. [19] Keller KA, Mah T, Parthasarathy TA, Cooke CM. Fugitive interfacial carbon coatings for oxide/oxide composites. J Am Ceram Soc 2000;83(2):329e36. [20] Wang Y, Cheng HF, Liu HT, Wang J. Effects of sintering temperature on mechanical properties of 3D mullite fiber (ALF FB3) reinforced mullite composites. Ceram Int 2013;39(8):9229e35. [21] Wang Y, Cheng HF, Liu HT, Wang J. Microstructure and room temperature mechanical properties of mullite fibers after heat-treatment at elevated temperatures. Mater Sci Eng A 2013;578:287e93. [22] Zheng Y, Wang SB. The effect of SiO2-doped boron nitride multiple coatings on mechanical properties of quartz fibers. Appl Surf Sci 2012;258(7):2901e5. [23] Chawla KK, Xu ZR, Ha JS, Schmücker M, Schneider H. Effect of BN coating on the strength of a mullite type fiber. Appl Compos Mater 1997;4(5):263e72. [24] Shimoda K, Park J-S, Hinoki T, Kohyama A. Influence of pyrolytic carbon interface thickness on microstructure and mechanical properties of SiC/SiC composites by NITE process. Compos Sci Technol 2008;68(1):98e105. [25] Ruggles-Wrenn MB, Siegert GT, Baek SS. Creep behavior of Nextel™720/ alumina ceramic composite with ±45 fiber orientation at 1200 C. Compos Sci Technol 2008;68(6):1588e95. [26] Feng JZ, Zhang CR, Feng J. Carbon fiber reinforced carbon aerogel composites for thermal insulation prepared by soft reinforcement. Mater Lett 2012;67(1): 266e8. [27] Wang Y, Cheng HF, Wang J. Mechanical and dielectric properties of mullite fiber-reinforced mullite matrix composites with single layer CVD SiC interphases. Int J Appl Ceram Technol 2014. http://dx.doi.org/10.1111/ ijac.12238. [28] Ruggles-Wrenn MB, Koutsoukos P, Baek SS. Effects of environment on creep behavior of two oxide/oxide ceramicematrix composites at 1200 C. J Mater Sci 2008;43(20):6734e46. [29] Liu HT, Cheng HF, Wang J, Che RC, Tang GP, Ma QS. Microstructural investigations of the pyrocarbon interphase in SiC fiber-reinforced SiC matrix composites. Mater Lett 2009;63(23):2029e31.
Contents lists available at ScienceDirect
Composites Part B journal homepage: www.elsevier.com/locate/compositesb
Interface engineering of fiber-reinforced all-oxide composites fabricated by the solegel method with fugitive pyrolytic carbon coatings Yi Wang, Haitao Liu*, Haifeng Cheng, Jun Wang Science and Technology on Advanced Ceramic Fibers and Composites Laboratory, National University of Defense Technology, Changsha 410073, PR China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 7 October 2014 Received in revised form 2 December 2014 Accepted 13 January 2015 Available online 29 January 2015
This paper describes the interface engineering of threeedimensional (3D) Nextel™440 fiber-reinforced aluminosilicate composites fabricated by the solegel method with fugitive pyrolytic carbon (PyC) coatings. The coating thickness on the fiber strength, interfacial characteristics and there corresponding effects on mechanical properties of the composites were investigated. The study shows that the fiber strength was influenced by the coating thickness and optimized with the thickness of 0.15 mm. The composites with uncoated fibers showed brittle fracture behavior without fiber pullout because of strong interactions between the fiber and the matrix. However, higher strengths and pseudo-ductile fracture behaviors were obtained in the composites with PyC interphases, where different deflections and branches of propagating cracks and fiber pullout patterns were observed. Moreover, induced fugitive PyC interface conditions have great effects on the density, microstructure and mechanical properties of the resultant composites. © 2015 Elsevier Ltd. All rights reserved.
Keywords: A. Ceramicematrix composites (CMCs) B. Interphase D. Mechanical testing Solegel method
1. Introduction Advanced applications, such as aircraft turbine engine components, land-based turbines, hypersonic missiles and fighting vehicles, and most recently, spacecraft re-entry thermal protection systems, require structural materials that exhibit superior longterm mechanical properties under high temperatures, high pressures and various environmental factors, such as moisture [1]. Continuous fiber-reinforced ceramic matrix composites (CFRCMCs) are the most attractive material concept that can meet such needs. Oxide/oxide CMCs, in particular, provide high strength, toughness, notch insensitivity, refractoriness and environmental stability at high application temperatures, where metals are usually limited by their melting temperature and monolithic ceramics are limited by their low damage tolerance [2,3]. Conventional processing methods for oxide/oxide composites, such as reaction-sintering and hot-pressing, normally require high temperatures and high pressures to provide adequate consolidation of the ceramic body [4]. Fiber degradation and interface reactions, which contribute to brittle fractures, will inevitably occur under
* Corresponding author. Tel.: þ86 731 84576440; fax: þ86 731 84576578. E-mail address: [email protected] (H. Liu). http://dx.doi.org/10.1016/j.compositesb.2015.01.018 1359-8368/© 2015 Elsevier Ltd. All rights reserved.
such conditions [5]. Several new fabrication processes, including slurry infiltration and hot-pressing (SIeHP) [6], electrophoretic deposition (EPD) [7], precursor infiltration and pyrolysis (PIP) [8] and the solegel method [9], have been introduced to address these issues. Among these techniques, the solegel method is the most feasible one for fabricating three-dimensional (3D) composites because of its low densification temperature (<1300 C), low shrinkage, reduced drying stress and near-stoichiometric matrix composition. The solegel method is also a near-net-shape technique that can be used to prepare large products with complex shapes directly [10]. General properties and issues for advanced oxide/oxide composites are summarized in Table 1. Interfaces between the fibers and the matrix in oxide/oxide composites are extremely important considerations for structural applications. An optimized interface can bring the deflection of cracks at the interface, fiber pullout during the fracture, excellent interfacial shear strength and ductility between the fiber and matrix, which contribute to high-energy consumption and excellent mechanical properties [11e13]. Among the interface concepts suggested recently [14e17], fugitive carbon is recognized as one of the most potential one. It can be readily deposited onto tows or woven fabrics by chemical vapor deposition (CVD) or pyrolysis of organic precursors, and be readily oxidized at moderately high temperature.
Y. Wang et al. / Composites Part B 75 (2015) 86e94
87
Table 1 List of developmental oxide/oxide composites for structural materials and corresponding issues. Category
SIeHP
EPD
PIP
Solegel
Fiber Interphase Matrix Porosity (%) General issues Interlaminar shear strength Hermeticity Near-net-shape technique
Nextel™720 None Mullite þ alumina ~32
Nextel™720 NdPO4 Mullite ~20
Nextel™720 None Alumina ~20
Nextel™480 BN Mullite ~10
Low Poor No
Low Poor No
Moderate-high Good Yes
Moderate-high Good Yes
Weaver and Keller et al. [18,19] investigated the effects of fugitive carbon coating thickness on mechanical properties of one or two-dimensional oxide/oxide composites fabricated by SIeHP process. Results showed that fugitive carbon coatings can remarkably reduce the interfacial sliding resistance and enhance the damage tolerance of the composites, even after exposure at 1000 C for 500 h. However, fugitive carbon coatings have rarely been used in 3D oxide/oxide composites. The present study focuses on the influence of fugitive pyrolytic carbon (PyC) coating thickness on the density, microstructure and mechanical properties of 3D Nextel™440 fiber-reinforced aluminosilicate (N440/AS) composites fabricated by the solegel method. 2. Material and methods 2.1. Raw materials Nextel™440 fibers (3M, USA) were used as reinforcements. The fibers were composed of 70 wt% Al2O3, 28 wt% SiO2 and 2 wt% B2O3. The tensile strength and Young's modulus announced by the manufacture were 2070 MPa and 186 GPa, respectively. Diphasic Al2O3eSiO2 sols (Suzhou Nanodispersions Ltd., China) were used as precursors of the aluminosilicate matrix. The weight ratio of Al2O3 and SiO2 particles in the sols was 1:1. The density and viscosity of the sols were 1.12 g/cm3 and 6 mpa s, respectively. 2.2. Sample preparation Mixtures of propylene and argon gas were used to deposit PyC on fiber fabrics with size of about 60 mm 100 mm at 1000 C by CVD. The coatings were deposited for 1, 4 and 8 h, with the resulting thickness of about 0.15, 0.42 and 1.00 mm. Then coated fabrics were stacked to a thickness of about 4 mm. Finally, the 3D architecture was finished by Z-stitching the stacked fabrics with Nextel™440 fiber yarn in a 2.5 mm 2.5 mm space in Changzhou Bolong Aerospace Technology Ltd., China. The fiber volume fraction of the preform is approximately 40%. 3D N440/AS composites were prepared by the solegel method according to our previous work [20]. The composites were sintered at 1000 C with a heating rate of 10 C/min. Composites with PyC interphases were denoted according to their coating thickness as C0.15, C0.42 and C1.00. Samples were fired at 600 C under air atmosphere for 2 h to oxidize PyC. Composites with fugitive PyC coatings were denoted as FC0.15, FC0.42 and FC1.00. 2.3. Characterization The density and porosity of the samples were determined by the Archimedes principle, using distilled water as the immersion medium. The theoretical density was calculated from the ratio of the alumilosilicate matrix and the fiber volume fraction. The tensile strength of the PyC-coated fibers was tested at room temperature
on a universal strength machine (Testometric, M3505CT) equipped with a 5 N load cell. The gauge length and crosshead speed were 20 mm and 0.2 mm/min, respectively. Three-point bending tests (test bars 50l 4w 3t mm3) were carried out at room temperature, with support span of 40 mm and cross-head speed of 0.5 mm/ min in an INSTRON 1342 testing machine, using the number of five specimens. The flexural stress (s) and elastic modulus (E) were calculated from Eqs. (1) and (2), respectively:
s ¼ 1:5PL wt 2
(1)
E ¼ 0:25mL3 wt 3
(2)
where P is the load at a point of deflection of a stress/displacement curve in the test, L is the support span, w is the specimen width, t is the specimen thickness and m is the slop of the tangent to the initial linear portion of the loadedeflection curve. Microstructure analysis of PyC-coated fibers and composites after mechanical tests was done by a scanning electron microscope (SEM, Hitachi FEG S4800). The pullout fiber surface of the composites was analyzed by EDS equipped with SEM. 3. Results and discussion 3.1. Fiber surface coatings Fig. 1 shows SEM images of uncoated and PyC-coated Nextel™440 fibers formed by CVD. The fibers were well coated; with thicknesses of about 0.15, 0.42 and 1.00 mm. The surfaces of PyCcoated fibers were smooth and uniform, without relation to the coating thickness. As shown in Fig. 1(c) and (a) narrow gap appeared between the fiber and PyC coating during the CVD cooling procedure, due to the different thermal expansion coefficient (fiber, 5.3 106/ C; carbon, 2.5 106/ C). The inference was that the coatings couldn't isolate the fibers from matrix completely. That is to say, interfacial reaction between the fibers and the matrix might occur, and strong interfacial bonding might be obtained. Single fiber tensile tests were used to evaluate the fiber tensile strength, and the strength distribution was evaluated by the twoparameter Weibull model [21]. Weibull plots of uncoated and PyC-coated fibers are shown in Fig. 2, and the corresponding Weibull parameters are listed in Table 2. In the Weibull model, larger m values of the fibers mean fewer defects and lower dispersion of the tensile strength [22]. The m value decreased as the coating thickness increased, indicating higher distribution of the tensile strength in coated fibers. Fibers with 0.15 mm PyC coating showed higher strength compared with uncoated fibers, while fibers with 0.42 and 1.00 mm PyC coatings showed lower strengths. The strength increase in 0.15 mm PyC coating might be attributed to the ability of this coating to heal surface flaws, as reported by Chawla et al. [23]. In the case of fibers with thick PyC coatings, fiber
88
Y. Wang et al. / Composites Part B 75 (2015) 86e94
Fig. 1. Cross-sectional of the uncoated and PyC-coated Nextel™440 fibers formed by CVD: (a) uncoated, (b) CVDe0.15, (c) CVDe0.42, and (d) CVDe1.00.
strength was more affected by the soft nature of the coating than the surface flaw healing effect. Moreover, the strength decrease might also originate from thermal degradation during CVD to produce desired thickness.
3.2. Densification of the composites The density and porosity of N440/AS composites fabricated by the solegel method are summarized in Table 3. The density of the composites decreased as coating thickness increased because thicker PyC coating blocked sol infiltration [24]. The matrix porosity of the composites obtained by the solegel method was compared with that of the composites obtained by SIeHP. As shown in Fig. 3a, 0 and 90 fiber tows, as well as numerous matrix cracks, were
observed in Nextel™720/alumina (N720/A) composites fabricated by SIeHP [25]. Most of these cracks were shrinkage cracks formed during processing rather than matrix cracks generated during loading. The porous structure of the alumina matrix was confirmed in the high-magnification image (Fig. 3b). By comparison, no cracks were observed in the solegel composites (Fig. 3c), although a large number of holes originating from precursor shrinkage [26] distributed in the matrix. The dense and homogeneous matrix was beneficial to the hermeticity and interlaminar shear strength of the final composites. The matrix in the composites was composed of aluminosilicate grains with small amounts of amorphous silica, as reported previously [27]. The presence of amorphous silica could improve the sinter-ability of the matrix, while aluminosilicate could improve environmental stability [28].
3.3. Microstructure of the composites
Fig. 2. Weibull plots of the tensile test for the fibers with and without PyC coatings.
3.3.1. Composites with PyC interphases Cross-section morphology of the N440/AS composites with PyC interphases is shown in Fig. 4. In the composites without interface engineering (Fig. 3d), the fibers were surrounded tightly by the matrix, without any interfacial debonding observed, indicating excess strong bonding between the fibers and the matrix. After PyC interphases were introduced into the composites, the number of micro-pores in the matrix increased, especially for those at intrafiber-bundles (Fig. 4(a,c,e)). As shown in Fig. 4(b,d,f), the PyC coating could be clearly defined between the fibers and the matrix. Meanwhile, interfacial debonding induced either by the polishing process or thermal mismatch was more and more prevalent, as the coating thickness increased from 0.15 mm to 1.00 mm. It can be noted that the narrow gap was preserved between the fibers and coatings during the solegel process. That is to say, PyC coatings have been successfully introduced into the N440/AS composites, and the interface properties of the composites were well-controlled.
Y. Wang et al. / Composites Part B 75 (2015) 86e94 Table 2 Weibull parameters of the tensile test for the uncoated and PyC-coated fibers. Deposition temperature (K)
Weibull modulus, m
s0 (GPa)
Scale parameter,
Tensile strength (GPa)
No coating CVDe0.15 CVDe0.42 CVDe1.00
5.46 4.72 4.39 2.93
2.03 2.09 1.60 1.41
1.87 1.91 1.46 1.17
± ± ± ±
0.38 0.41 0.39 0.46
Table 3 Properties of the N440/AS composites with various interface condition. Samples
Density (g/cm3)
Porosity (%)
Flexural strength (MPa)
Uncoated C0.15 FC0.15 C0.42 FC0.42 C1.00 FC1.00
2.22 2.20 2.15 2.17 2.09 2.10 2.03
19.0 19.7 21.5 20.8 23.7 23.4 25.9
31.2 91.1 93.1 95.4 66.7 78.2 50.5
± ± ± ± ± ± ±
3.4 3.6 4.5 18.2 8.8 17.7 6.7
Elastic modulus (GPa) 33.0 26.0 22.2 30.1 25.0 24.1 15.6
± ± ± ± ± ± ±
5.2 1.6 0.7 2.4 1.6 1.5 3.1
3.3.2. Composites with fugitive PyC coatings Cross-section morphology of the N440/AS composites with fugitive PyC coatings is shown in Fig. 5. As displayed in Fig. 5(a,c,e), the number of micro-pores at intra-fiber-bundles increased as the original coating thickness increased. A large number of matrix inner fiber bundles missed during polishing process because of weak interfacial bonding. As shown in Fig. 5(b,d,f), the gaps induced by the removal of PyC were observed between the fibers and the matrix, and the gap width increased as the coating thickness increased. Furthermore, it can be noted that the gap width was not consistent with the corresponding PyC coating thickness. A reasonable explanation was that thermal mismatch did exist among the fibers, matrix and PyC interphases. As a result, a much narrower gap was obtained after PyC removal. The isolation of the
89
fibers from matrix might contribute to the improvement of the composite mechanical properties.
3.4. Mechanical properties and fracture behavior of the composites 3.4.1. Composites with PyC interphases Typical flexural stressestrain curves of the N440/AS composites with PyC interphases are shown in Fig. 6. The stressestrain curve of the composite with uncoated fibers showed a catastrophic failure attributed to strong fiberematrix interfacial bonding. The corresponding flexural strength and elastic modulus were 27.3 MPa and 36.5 GPa, respectively. In contrast, all the stressestrain curves of the composites with PyC interphases showed extended regions after initial deviation from linearity. Areas under these curves were much larger compared with those of uncoated fiber-reinforced composites. Sample C0.42, with a coating thickness of 0.42 mm, showed the highest flexural strength of 102 MPa. Mechanical properties of the N440/AS composites are listed in Table 3. As the original coating thickness increased the average flexural strength of the composites increased firstly and then decreased. The highest values (91.1 or 95.4 MPa) were obtained for the composites with PyC interphases of moderate thickness, i.e. 0.15 or 0.42 mm thickness. It can be concluded that the flexural behavior of the N440/AS composites could be improved by tailoring the fiber/matrix interface with PyC coatings. Fig. 7 illustrates the fracture surface of the composites without and with PyC interphases. For the composites with uncoated fibers (Fig. 7a), the fracture surface was very even, and nearly no pullout fibers could be observed. Whereas for the composites with PyC interphases (Fig. 7((b)e(d)), all fracture surfaces showed evident fiber pullout behaviors, and lengths of pulledout fibers could exceed 100 mm. The interfacial debonding could be clearly observed. As the coating thickness increased, the phenomena that PyC coatings attached to pullout fibers became more and more prevalent. This finding was confirmed by EDS analysis of the pullout fiber surface of sample C0.42 (Fig. 9(a,b)). Moreover, it can be
Fig. 3. Typical cross-section of oxide/oxide composites fabricated by (a, b) SIeHP (Adapted from Ruggles-Wrenn et al. [24]) and (c, d) solegel method.
90
Y. Wang et al. / Composites Part B 75 (2015) 86e94
Fig. 4. Typical cross-section morphology of the N440/AS composites with PyC interphases: (a, b) C0.15, (c, d) C0.42 and (e, f) C1.00.
noted that the surface of pullout fibers was free of any matrix, suggesting weak chemical corrosion to fibers during fabrication and the absence of debonding within the matrix. Increasing the coating thickness also promoted the increase in the length and uniformity of pullout fibers, indicating the decrease in interfacial bonding. 3.4.2. Composites with fugitive PyC coatings As shown in Fig. 6, typical flexural stressestrain curves of the N440/AS composite with fugitive PyC coatings exhibited obvious toughened fracture behavior. When the load reached the maximum, it dropped off gradually. The large extended area under the curve indicated a greater amount of fracture energyconsumption in the composites. The highest flexural strength of 98.8 MPa was achieved for sample FC0.15, with the gap width of 0.15 mm. As shown in Table 3, two trends could be found after PyC removal: (1) the average flexural strength of the composites decreased to a degree, especially for the composites with PyC interphases of 1.00 mm thickness; (2) the elastic modulus of the composites also decreased, owing to further weaken of the interfacial bonding. The composite with a gap width of 1.00 mm showed single, long-cracking behavior and the lowest elastic modulus (14.2 GPa). A reasonable explanation is that the gap was too wide for load transfer from matrix to the fibers, although the in-situ fiber
strength of the composite was very high. Apparently, the gap width had remarkable impacts on the flexural behavior of the N440/AS composites with fugitive PyC coatings. Fig. 8 illustrates the fracture surface of the composites with fugitive PyC interphases. The fiber pullout behaviors of all composites with fugitive PyC coatings were distinct, and the length of pullout fibers was uniform. The surface of pullout fibers was very smooth, without any attached coatings, as confirmed by EDS analysis of the pullout fiber surface of sample FC0.42 (Fig. 9(c,d)). The length of pullout fibers in sample FC1.00 was so long (>300 mm) that the load couldn't transmit from matrix to the fibers efficiently. Consequently, low flexural strength and elastic modulus were observed in sample FC1.00. In general, mechanical properties of CFRCMCs are determined by the in-situ fiber strength and interfacial characteristics. Concerning the composites without interface engineering, the fibers were subjected to strong chemical corrosion of matrix and interfacial reactions would occur. As a result, low in-situ fiber strength and strong interfacial bonding were obtained, resulting in brittle fracture behaviors and poor mechanical properties of the composites. While for the composites with PyC interphases, the coatings could protect the fibers from being corroded by matrix. In addition, interfacial debonding would easily occur within PyC interphases because of low van der Waals forces between the basal
Y. Wang et al. / Composites Part B 75 (2015) 86e94
91
Fig. 5. Typical cross-section morphology of the N440/AS composites with fugitive PyC interphases: (a, b) FC0.15, (c, d) FC0.42 and (e, f) FC1.00.
Fig. 6. Typical flexural stressedisplacement curves of the N440/AS composites with various interface condition.
planes of PyC [29]. Consequently, high in-situ fiber strength and weak interfacial bonding both improved mechanical properties of the composites with interface engineering. After PyC removal, gaps appeared at the fiber/matrix interface, the width of which was determined by the PyC coating thickness and the thermal mismatch among the fibers, PyC and matrix. The thermal expansion coefficient of the fibers was equivalent to that of matrix, but higher than that of PyC. The inference is that the thermal mismatch between the fibers and PyC would lead to the compressive strain in the PyC, so as to that between PyC and matrix. That is to say, the thermal mismatch among the fibers, PyC and matrix in the composites would counteract to a lower level. As a result, we defined an ideal gap width between the fibers and the matrix after PyC removal, the value of which was equal to the original coating thickness. According to Weaver et al. [18], the gaps between the fibers and the matrix are certain to produce interfaces with zero tensile strength, but frictional sliding will be preserved by mechanical interlocking of asperities at the fiber/matrix boundary on condition that the asperity height exceeds the gap width. Similar mechanism has also been reported by Keller et al. [19]. For an idealized smooth, straight gap (Fig. 10a), the fiber is unconstrained, thus the load can't transfer from the matrix to the fibers. Typical fiber pullout behavior, with extremely long lengths, will be observed. While for a real gap,
92
Y. Wang et al. / Composites Part B 75 (2015) 86e94
Fig. 7. Typical fracture surface morphology of the N440/AS composites with PyC interphases: (a) uncoated, (b) C0.15, (c) C0.42, and (d) C1.00.
load transfer is achieved by the combined effects of mechanical interlocking, intermittent fiber/matrix bonding and roughness of both the fiber and matrix. As shown in Fig. 10(b), when the crack reaches the interface, matrix strain will cause sliding at the interface. Then point contacts between the fiber and matrix are formed, where mechanical interlocking and possible bonding occurs. The
optimal degree of mechanical coupling between the fibers and the matrix can be achieved through tailoring the gap width, which is the intention of this work. When the gap is too wide (>0.42 mm), the mechanical coupling between the fibers and the matrix is limited, and the load will deflect easily at the fiber/matrix interface, with little energy-consuming, as proved by the decrease in the
Fig. 8. Typical fracture surface morphology of the N440/AS composites with fugitive PyC interphases: (a) FC0.15, (b) FC0.42, and (c) FC1.00.
Y. Wang et al. / Composites Part B 75 (2015) 86e94
93
Fig. 9. EDS analysis of the pullout fiber surface in N440/AS composites with various interface condition: (a, b) C0.42 and (c, d) FC0.42.
elastic modulus. This work indicated that the optimal fugitive PyC thickness in the N440/AS composites should be less than 0.15 mm, which is consistent with the result of Ref. [19] (<0.1 mm).
4. Conclusions We have studied the effects of CVD PyC coating thickness on the strength of Nextel™440 fibers. The fugitive PyC coatings were introduced into 3D Nextel™440 fiber-reinforced aluminosilicate composites fabricated by the solegel method. The influences of the fugitive PyC coating thickness on the density, microstructure, and mechanical properties of the resultant composites were investigated. The major results obtained are summarized as follows:
(1) The surface of PyC-coated fibers was smooth, without relation to the coating thickness. The optimal strength was obtained when the coating thickness was 0.15 mm because thin coating could heal the fiber surface flaws. (2) The dense and homogeneous matrix of the composites showed nearly no cracks but a large number of holes. As the coating thickness increased, the density of the composites decreased because thicker PyC coating blocked sol infiltration. (3) The fracture surface of composites with uncoated fibers was very even, while that of the composites with PyC or fugitive PyC interphases illustrated typical fiber pullout behaviors that were affected by interface conditions remarkably. (4) The composites with uncoated fibers showed brittle fracture behavior and inferior mechanical properties. Higher strengths with pseudo-ductile fracture behavior were obtained using all PyC coating thickness. Mechanical properties of the composites with fugitive PyC coatings were determined by the PyC coating thickness, and the optimal fugitive PyC thickness should be less than 0.15 mm. The present approach has demonstrated the significant implication of fugitive PyC coatings in interfacial engineering in particular on the high-performance and damage-tolerance of fiber-reinforced all-oxide composites. On account of the thermal instability of the gap derived from the removal of PyC interphases, the present research points out that future effort should focus on evaluating high-temperature service lifetime of the composites.
Acknowledgments
Fig. 10. Schematic diagrams of load transfer in the composites with fugitive coatings for fiber protection: (a) ideal case and (b) real case.
The authors appreciate the financial support of the National Natural Science Foundation of China (51202291 and 51302314), Aid Program for Innovative Group of National University of Defense Technology, and Aid program for Science and Technology Innovative Research Team in Higher Educational Institutions of Hunan Province.
94
Y. Wang et al. / Composites Part B 75 (2015) 86e94
References [1] Armani CJ, Ruggles-Wrenn MB, Hay RS, Fair GE. Creep and microstructure of Nextel™ 720 fiber at elevated temperature in air and in steam. Acta Mater 2013;61(16):6114e24. [2] Simon RA, Danzer R. Oxide fiber composites with promising properties for high-temperature structural applications. Adv Eng Mater 2006;8(11): 1129e34. [3] Hülsenberg D, Fehling P, Leutbecher T. Damage tolerant, translucent oxide fiber/glass matrix composites. Compos Part B 2008;39(2):362e73. [4] Kaya C, Kayab F, Butler EG, Boccaccini AR, Chawla KK. Development and characterisation of high-density oxide fibre-reinforced oxide ceramic matrix composites with improved mechanical properties. J Eur Ceram Soc 2009;29(9):1631e9. [5] Schmücker M, Flucht F, Mechnich P. Degradation of oxide fibers by thermal overload and environmental effects. Mater Sci Eng A 2012;557:10e6. [6] Holmquist MG, Lange FF. Processing and properties of a porous oxide matrix composite reinforced with continuous oxide fibers. J Am Ceram Soc 2003;86(10):1733e40. [7] Boccaccini AR, Kaya C, Chawla KK. Use of electrophoretic deposition in the processing of fibre reinforced ceramic and glass matrix composites: a review. Compos Part A 2001;32(8):997e1006. [8] Mattoni MA, Yang JY, Levi CG, Zok FW. Effects of matrix porosity on the mechanical properties of a porous-matrix, all-oxide ceramic composite. J Am Ceram Soc 2001;84(11):2594e602. [9] Chawla KK. Interface engineering in mullite fiber/mullite matrix composites. J Eur Ceram Soc 2008;28(2):447e53. [10] Russell-Floyd RS, Harris B, Cooke RG, Laurie J, Hammett FW, Jones RW, et al. Application of sol-gel processing techniques for the manufacture of fiberreinforced ceramics. J Am Ceram Soc 1993;76(10):2635e43. [11] Wang XQ, Zhang JF, Wang ZQ, Liang WY, Zhou LM. Finite element simulation of the failure process of single fiber composites considering interface properties. Compos Part B 2013;45(1):573e80. [12] Santhosh U, Ahmad J, John R, Ojard G, Miller R, Gowayed Y. Modeling of stress concentration in ceramic matrix composites. Compos Part B 2013;45(1): 1156e63. [13] Yan LB, Chouw N, Jayaraman K. Flax fibre and its composites e a review. Compos Part B 2014;56:296e317. [14] Raab D, Pfeifer K, Hülsenberg D, Boccaccini AR. Mechanical properties of oxide-fiber-reinforced glass matrix composites with BN or SnO2 interfaces. Mater Sci Eng A 2006;417(1e2):314e47.
[15] Davis JB, Marshall DB, Morgan PED. Monazite-containing oxide/oxide composites. J Eur Ceram Soc 2000;20(5):583e7. [16] Boakye E, Petry MD, Hay RS. Porous aluminum oxide and lanthanum phosphate fiber coatings. Ceram Eng Sci Proc 2008;17(4):53e60. [17] Wang Y, Liu HT, Cheng HF, Wang J. Effective fugitive carbon coatings for the strength improvement of 3D Nextel™ 440/aluminosilicate composites. Mater Lett 2014;126(1):236e9. [18] Weaver JH, Yang J, Zok FW. Control of interface properties in oxide composites via fugitive coatings. J Am Ceram Soc 2008;91(12):4003e8. [19] Keller KA, Mah T, Parthasarathy TA, Cooke CM. Fugitive interfacial carbon coatings for oxide/oxide composites. J Am Ceram Soc 2000;83(2):329e36. [20] Wang Y, Cheng HF, Liu HT, Wang J. Effects of sintering temperature on mechanical properties of 3D mullite fiber (ALF FB3) reinforced mullite composites. Ceram Int 2013;39(8):9229e35. [21] Wang Y, Cheng HF, Liu HT, Wang J. Microstructure and room temperature mechanical properties of mullite fibers after heat-treatment at elevated temperatures. Mater Sci Eng A 2013;578:287e93. [22] Zheng Y, Wang SB. The effect of SiO2-doped boron nitride multiple coatings on mechanical properties of quartz fibers. Appl Surf Sci 2012;258(7):2901e5. [23] Chawla KK, Xu ZR, Ha JS, Schmücker M, Schneider H. Effect of BN coating on the strength of a mullite type fiber. Appl Compos Mater 1997;4(5):263e72. [24] Shimoda K, Park J-S, Hinoki T, Kohyama A. Influence of pyrolytic carbon interface thickness on microstructure and mechanical properties of SiC/SiC composites by NITE process. Compos Sci Technol 2008;68(1):98e105. [25] Ruggles-Wrenn MB, Siegert GT, Baek SS. Creep behavior of Nextel™720/ alumina ceramic composite with ±45 fiber orientation at 1200 C. Compos Sci Technol 2008;68(6):1588e95. [26] Feng JZ, Zhang CR, Feng J. Carbon fiber reinforced carbon aerogel composites for thermal insulation prepared by soft reinforcement. Mater Lett 2012;67(1): 266e8. [27] Wang Y, Cheng HF, Wang J. Mechanical and dielectric properties of mullite fiber-reinforced mullite matrix composites with single layer CVD SiC interphases. Int J Appl Ceram Technol 2014. http://dx.doi.org/10.1111/ ijac.12238. [28] Ruggles-Wrenn MB, Koutsoukos P, Baek SS. Effects of environment on creep behavior of two oxide/oxide ceramicematrix composites at 1200 C. J Mater Sci 2008;43(20):6734e46. [29] Liu HT, Cheng HF, Wang J, Che RC, Tang GP, Ma QS. Microstructural investigations of the pyrocarbon interphase in SiC fiber-reinforced SiC matrix composites. Mater Lett 2009;63(23):2029e31.