Availableonlineatwww.sciencedirect.com Science Direct E噩≈RS ELSEVIER Joumal of the European Ceramic Society 28(2008)1687-1696 www.elsevier.com/locate/jeurceramsoc Microstructural features of the zro interfacial coatings on sic fibers before and after exposition to air at high temperatures N I. Baklanova,, O I Kiselyova, A.T. Titov, T M. Zima a Institute of Solid State Chemistry and Mechanochemistry SB RAS, Kutateladze Street 18. Novosibirsk 630128. Russian Federation b Lomonosov Moscow State University, Physical Department, Moscow, Russian Federation General Institute of Geology Geophysics and Mineralogy SB RAS, Novosibirsk 630090, Russian Federation Received 30 October 2007: received in revised form 27 November 2007; accepted 30 November 2007 Available online 14 January 2008 Sols of rare earth stabilized zirconia were used as simple, readily processable and accurate controllable precursors for the tetragonal zirconia interfacial coatings on commercially available Sic-based fibers. The tetragonal zirconia interfacial coatings can be applied to different types of Sic fibers without degrading fiber strength. The morphology, composition, structure, nanorelief and oxidation resistance of coated SiC fibers were evaluated by various analytical techniques, including scanning electron microscopy/energy dispersive analysis, transmission electron microscopy, atomic force microscopy in various modes, and micro-Raman spectroscopy. It was shown that the microstructural peculiarities of the RezrO2 interfacial coatings on Sic-based fibers may explain some of the differences in the behavior of different types of fibers o 2007 Elsevier Ltd. All rights reserved. Keywords: Interfacial coatings; Microstructure; ZrO2; SiC fibers: Oxidation resistance 1. Introduction ponent of composites it remains one of the weakest links in the research of the matrix-interphase-fiber triad. Insufficient nterface is a key region determining a set of properties of comprehension of interphase functions, role and nature, is a composite materials. In fiber-reinforced composites the fibers key problem and one major bottleneck retarding the devel ensure the strength of material, while the matrix helps to keep opment of efficient CMC's for high-temperature structural the shape. The interface transfers the load from matrix to applications. To solve this problem it is necessary to study the fibers. Further, the incorporation of the reinforcing fibers thoroughly the properties of interphase and to clarify which into brittle ceramic matrix provides CMCs with a degree characteristics of the interphase and in what extent control the of pseudo-ductility, preventing catastrophic failure by several behavior of the composite. Undoubtedly, among the features echanisms, such as fiber debonding, fiber sliding and crack of the interphase zone a microstructure is one of the most bridging. In order to achieve these properties, the interphase important zone must be sufficiently weak to deflect matrix microc- In addition to above-mentioned functions (load transfer racks and allow subsequent fiber pull-out. Both functions of and crack deflection), interphase materials must be compat the interphase zone in CMCs, namely, a load transfer from ible with both matrix and fiber for long-term operation in matrix to the fibers and the matrix microcrack deflection, are oxidizing atmosphere. This is especially important for non- greatly determined by nature of the interphase zone. Despite oxide CMCs, e.g. SiC/SiC composites. The interphase can of the wide recognition of the interphase as a crucial com- be exposed to oxidizing environments when the ends of coated fibers are exposed to surrounding atmosphere or when matrix cracks are present, allowing oxidants to reach the fiber Corresponding author. Tel: +7 3832 363839: fax: +7 3832 322847 coatings. Since oxide ceramics cannot be oxidized, it is com- E-mail address: baklanova@solid nsc. ru(N 1. Baklanova) monly believed that oxide-based coatings represent the best 0955-2219/S-see front matter o 2007 Elsevier Ltd. All rights reserved. doi: 10.1016/j-jeurceramsoc20071 1.008
Available online at www.sciencedirect.com Journal of the European Ceramic Society 28 (2008) 1687–1696 Microstructural features of the ZrO2 interfacial coatings on SiC fibers before and after exposition to air at high temperatures N.I. Baklanova a,∗, O.I. Kiselyova b, A.T. Titov c, T.M. Zima a a Institute of Solid State Chemistry and Mechanochemistry SB RAS, Kutateladze Street 18, Novosibirsk 630128, Russian Federation b Lomonosov Moscow State University, Physical Department, Moscow, Russian Federation c General Institute of Geology, Geophysics and Mineralogy SB RAS, Novosibirsk 630090, Russian Federation Received 30 October 2007; received in revised form 27 November 2007; accepted 30 November 2007 Available online 14 January 2008 Abstract Sols of rare earth stabilized zirconia were used as simple, readily processable and accurate controllable precursors for the tetragonal zirconia interfacial coatings on commercially available SiC-based fibers. The tetragonal zirconia interfacial coatings can be applied to different types of SiC fibers without degrading fiber strength. The morphology, composition, structure, nanorelief and oxidation resistance of coated SiC fibers were evaluated by various analytical techniques, including scanning electron microscopy/energy dispersive analysis, transmission electron microscopy, atomic force microscopy in various modes, and micro-Raman spectroscopy. It was shown that the microstructural peculiarities of the ReZrO2 interfacial coatings on SiC-based fibers may explain some of the differences in the behavior of different types of fibers. © 2007 Elsevier Ltd. All rights reserved. Keywords: Interfacial coatings; Microstructure; ZrO2; SiC fibers; Oxidation resistance 1. Introduction Interface is a key region determining a set of properties of composite materials. In fiber-reinforced composites the fibers ensure the strength of material, while the matrix helps to keep the shape. The interface transfers the load from matrix to the fibers. Further, the incorporation of the reinforcing fibers into brittle ceramic matrix provides CMC’s with a degree of pseudo-ductility, preventing catastrophic failure by several mechanisms, such as fiber debonding, fiber sliding and crack bridging.1 In order to achieve these properties, the interphase zone must be sufficiently weak to deflect matrix microcracks and allow subsequent fiber pull-out. Both functions of the interphase zone in CMC’s, namely, a load transfer from matrix to the fibers and the matrix microcrack deflection, are greatly determined by nature of the interphase zone. Despite of the wide recognition of the interphase as a crucial com- ∗ Corresponding author. Tel.: +7 3832 363839; fax: +7 3832 322847. E-mail address: baklanova@solid.nsc.ru (N.I. Baklanova). ponent of composites it remains one of the weakest links in the research of the matrix–interphase–fiber triad. Insufficient comprehension of interphase functions, role and nature, is a key problem and one major bottleneck retarding the development of efficient CMC’s for high-temperature structural applications. To solve this problem it is necessary to study thoroughly the properties of interphase and to clarify which characteristics of the interphase and in what extent control the behavior of the composite. Undoubtedly, among the features of the interphase zone a microstructure is one of the most important. In addition to above-mentioned functions (load transfer and crack deflection), interphase materials must be compatible with both matrix and fiber for long-term operation in oxidizing atmosphere. This is especially important for nonoxide CMC’s, e.g. SiC/SiC composites. The interphase can be exposed to oxidizing environments when the ends of coated fibers are exposed to surrounding atmosphere or when matrix cracks are present, allowing oxidants to reach the fiber coatings. Since oxide ceramics cannot be oxidized, it is commonly believed that oxide-based coatings represent the best 0955-2219/$ – see front matter © 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.jeurceramsoc.2007.11.008
N.I. Baklanova et al /Journal of the European Ceramic Sociery 28(2008)1687-1696 choice in terms of oxidation resistance. Several oxidation 23. Specimen characterization resistant and crack-deflecting materials including monazite, alu mina/silica, stabilized zirconia, and others were proposed as Scanning electron microscope SEM LEO 1430VP, supplied appropriate candidates for interphase zone in CMCs2-8 by EDX(Oxford) spectrometer was used for studying of mor- Some information is available in publications, describing the phology and composition of the initial and coated fibers behavior of the stabilized ZrO2-coated SiC fibers exposed to Micro-Raman spectra of the RezrO2-coated ceramic fibers air at high temperatures. 9, 10 Preliminary studying of the pecu- before and after oxidation were recorded using a Triplemate, liarities of morphology and nanorelief of two zirconia-coated SPEX spectrometer equipped with CCD spectrometric detec- fibers,namely, Hi-NicalonTM and Tyranno-SAM before and tor and microscope attachment for back scattering geometry after exposition to air at 1000 C using atomic force microscopy The 488 nm radiation from an argon laser was used for spectral (AFM) and scanning electron microscopy (SEM) showed that excitation. these features are greatly dependent on the type of Sic fibers. The topography and surface roughness of fibers was exam- After application of coating the roughness of Tyranno-SAM ined by atomic force microscopy(SolverP47Bio, NT-MDT, (nearly stoichiometric) fiber increased in comparison to that Russia) and MultiMode NanoScope Illa(Veeco, USA)using of the initial fiber, whereas the roughness parameters of Hi- Tapping Mode. Silicon cantilevers were used. Filaments were Nicalon'wfiberretained their values after application of coating. attached to metal discs double-sided adhesive tape. Dif- ASs parameters for coated ferent areas of several filaments of each type fibers were Tyranno-SAM and Hi-NicalonM fibers was enhanced after selected randomly. A roughness and other statistical parameters exposition to air at 1000C of selected areas were obtained using tool "Statistics and Fem- The purpose of this work is to study the microstructural toScan 001 software for AFM images. The AFM images were features of the ReZrO2-coated SiC fibers type Hi-NicalonM, flattened before analysis using second-order surface subtrac- Hi-Nicalon $M, and Tyranno-SAM and the evolution of these tion Parameters were calculated based on following definitions features after exposition to air at 1000 and 1200C. Mean roughness(Ra) is the arithmetic average of the absolute values of the surface height deviations, Zi, measured from mean 2. Experimental 2.1. Substrate and coating preparation R 1zl NM Hi-Nicalon, Hi-Nicalon S(both Nippon Carbon Co Ltd, Tokyo, Japan) and Tyranno-saM grade 3(Ube Indus- Mean height (Rmean)is the arithmetic average of the absolute Ltd, Yamaguchi, Japan) fiber tows were used as substrate values of the measured heights materials. Prior to coating, Hi-NicalonTM and Hi-Nicalon STM fiber tows were immersed in 50: 50 acetone/ethanol mixture for rmean 24 h for removing a sizing agent, dried at ambient temperature and then thermally treated in air at 450 C. Tyranno-SAMfiber tow was immersed in hot distilled water for desizing, dried at Maximum height roughness(Rmax)is the difference in height ambient temperature and heated in air at 500C. between the highest and lowest points a detailed description of the coating procedure is given in Rmax zi Ref. [6] but some parameters of sol were optimized. Sol-gel approach was used for the preparation of multi-component rare During the scanning, phase shifts, i.e. changes in the phase earth oxide stabilized zirconia(ReZrO2)coatings on all types contact angle of vibration with respect to the phase angle of of SiC-based fibers. At least two rare earth components were freely oscillating cantilever, were recorded simultaneously with incorporated in conventional zirconia-yttria oxide system. Total height images content of earth oxides was 3 mol%. The coating stage involved firstly the immersion of the ceramic fiber tow into sol, 2. 4. Tensile strength tests drying in air at ambient temperature, slow heating till 1000C in vacuum and heat-treatment for 2 h Mechanical tensile tests of the coated fibers were conducted at room temperature using FM-4(Hungary)testing machine. 2.2. Oxidation tests Single fibers extracted from a tow were fixed on paper frame using a hard resin. The gauge of 10mm in length was used. Thermal oxidation resistance of coated ceramic fibers was the diameter of each filament was measured in the middle examined in laboratory air under static conditions at 1000 and of length by laser interferometry and used for calculation of 1200C. The samples were placed into preliminarily heated mechanical properties of filaments. Next, the lateral sides of furnace(KO-14, Germany) and kept there during fixed time support frame were cut by a heated wire and the load was intervals. Then the samples were taken out, cooled in dessicator applied at constant crosshead speed of 1.3 mm/min. About 50 and weighted with accuracy to. I mg. The total time of testing filaments for each type of fibers were tested. The average diam- was 40h eters for the coated fibers were determined to be equal to
1688 N.I. Baklanova et al. / Journal of the European Ceramic Society 28 (2008) 1687–1696 choice in terms of oxidation resistance. Several oxidation resistant and crack-deflecting materials including monazite, alumina/silica, stabilized zirconia, and others were proposed as appropriate candidates for interphase zone in CMC’s.2–8 Some information is available in publications, describing the behavior of the stabilized ZrO2-coated SiC fibers exposed to air at high temperatures.9,10 Preliminary studying of the peculiarities of morphology and nanorelief of two zirconia-coated fibers, namely, Hi-NicalonTM and Tyranno-SATM before and after exposition to air at 1000 ◦C using atomic force microscopy (AFM) and scanning electron microscopy (SEM) showed that these features are greatly dependent on the type of SiC fibers. After application of coating the roughness of Tyranno-SATM (nearly stoichiometric) fiber increased in comparison to that of the initial fiber, whereas the roughness parameters of HiNicalonTM fiber retained their values after application of coating. Moreover, the difference in the roughness parameters for coated Tyranno-SATM and Hi-NicalonTM fibers was enhanced after exposition to air at 1000 ◦C. The purpose of this work is to study the microstructural features of the ReZrO2-coated SiC fibers type Hi-NicalonTM, Hi-Nicalon STM, and Tyranno-SATM and the evolution of these features after exposition to air at 1000 and 1200 ◦C. 2. Experimental 2.1. Substrate and coating preparation Hi-NicalonTM, Hi-Nicalon STM (both Nippon Carbon Co. Ltd., Tokyo, Japan) and Tyranno-SATM grade 3 (Ube Industry Ltd., Yamaguchi, Japan) fiber tows were used as substrate materials. Prior to coating, Hi-NicalonTM and Hi-Nicalon STM fiber tows were immersed in 50:50 acetone/ethanol mixture for 24 h for removing a sizing agent, dried at ambient temperature and then thermally treated in air at 450 ◦C. Tyranno-SATM fiber tow was immersed in hot distilled water for desizing, dried at ambient temperature and heated in air at 500 ◦C. A detailed description of the coating procedure is given in Ref. [6] but some parameters of sol were optimized. Sol–gel approach was used for the preparation of multi-component rare earth oxide stabilized zirconia (ReZrO2) coatings on all types of SiC-based fibers. At least two rare earth components were incorporated in conventional zirconia–yttria oxide system. Total content of rare earth oxides was 3 mol%. The coating stage involved firstly the immersion of the ceramic fiber tow into sol, drying in air at ambient temperature, slow heating till 1000 ◦C in vacuum and heat-treatment for 2 h. 2.2. Oxidation tests Thermal oxidation resistance of coated ceramic fibers was examined in laboratory air under static conditions at 1000 and 1200 ◦C. The samples were placed into preliminarily heated furnace (KO-14, Germany) and kept there during fixed time intervals. Then the samples were taken out, cooled in dessicator and weighted with accuracy ±0.1 mg. The total time of testing was 40 h. 2.3. Specimen characterization Scanning electron microscope SEM LEO 1430VP, supplied by EDX (Oxford) spectrometer was used for studying of morphology and composition of the initial and coated fibers. Micro-Raman spectra of the ReZrO2-coated ceramic fibers before and after oxidation were recorded using a Triplemate, SPEX spectrometer equipped with CCD spectrometric detector and microscope attachment for back scattering geometry. The 488 nm radiation from an argon laser was used for spectral excitation. The topography and surface roughness of fibers was examined by atomic force microscopy (SolverP47Bio, NT-MDT, Russia) and MultiMode NanoScope IIIa (Veeco, USA) using TappingModeTM. Silicon cantilevers were used. Filaments were attached to metal discs using double-sided adhesive tape. Different areas of several filaments of each type fibers were selected randomly. A roughness and other statistical parameters of selected areas were obtained using tool “Statistics” and FemtoScan 001 software for AFM images. The AFM images were flattened before analysis using second-order surface subtraction. Parameters were calculated based on following definitions. Mean roughness (Ra) is the arithmetic average of the absolute values of the surface height deviations, Zij, measured from mean plane: Ra = 1 NxNy Nx i=1 Ny j=1 |z| Mean height (Rmean) is the arithmetic average of the absolute values of the measured heights: Rmean = 1 NxNy Nx i=1 Ny j=1 Zij Maximum height roughness (Rmax) is the difference in height between the highest and lowest points: Rmax = Zmax − Zmin During the scanning, phase shifts, i.e. changes in the phase contact angle of vibration with respect to the phase angle of freely oscillating cantilever, were recorded simultaneously with height images. 2.4. Tensile strength tests Mechanical tensile tests of the coated fibers were conducted at room temperature using FM-4 (Hungary) testing machine. Single fibers extracted from a tow were fixed on paper frame using a hard resin. The gauge of 10 mm in length was used. The diameter of each filament was measured in the middle of length by laser interferometry and used for calculation of mechanical properties of filaments. Next, the lateral sides of support frame were cut by a heated wire and the load was applied at constant crosshead speed of 1.3 mm/min. About 50 filaments for each type of fibers were tested. The average diameters for the coated fibers were determined to be equal to
N I. Baklanova er al. Joumal of the European Ceramic Society 28(2008)1687-1696 1689 Fig 1. The AFM images of the initial fibers(3D height representation ):(a)Hi-Nicalon TM; (b)Hi-Nicalon STM; (c) Tyranno-SATM 13.94+0.18 for Hi-NicalonTM, 13. 110.16 for Hi-Nicalon fiber has very well-developed relief which can be due to large STM, and 7.53+0.07 um for Tyranno-SATM fibers size grains in the surface region of fiber, with lateral sizes of particles being 100-200nm Grains on the filament surface are disoriented. Phase contrast AFM images help to reveal minor 3. Results features, which are sometimes poorly resolved, but only height images provide correct topographical data. When the surface 3.1. SEM/EDS, AFM and TEM analysis of as-received Sic relief contains elements, which vertical dimensions differ by the order of magnitude, it is difficult to demonstrate both types of elements in the same image. In order to increase the contrast in a AFM image of the initial Hi-Nicalon"M fiber is represented 3D image, Fig Ic was constructed by superposition of phase data in Fig. la. Filaments have very smooth and uniform surface. over 3D topography. This type of representation shows minor The estimated roughness parameter is about 5 nm and almost details on 3D topographical relief. Roughness parameters, Ra independent on the size of the scanned area. Contrary to Hi- and Rmean, were determined as 10 and 52 nm, respectively, for NicalonTM. the surface relief of Hi-Nicalon STM filament is scanned area of 1. 2 um greatly non-uniform(Fig 1b). Some areas have a rather homoge neous relief, while other areas consist of different size nodules. 3. 2. SEM/EDS, AFM analysis of the Rezroz-coated Sic In some cases, their sizes run to several hundred nanometers. fibers The disposition of nodules appears to be rather random and ape riodic. This picture is a typical one for all tested Hi-Nicalon S SEM images of the Rezro2(one dipping-annealing cycle filaments. There is a large scattering in the roughness parameters coating on Hi-Nicalon M fiber are represented in Fig. 2a-c determined for different areas of the same filament. Roughness a distinctive feature of this coating is smoothness and unifor- Ra, was determined as 6-10 nm for scanned area of 4 um. For mity along whole length and diameter of filaments. Separate areas with nodules the roughness is increased till about 25 nm. well-developed crystals and discontinuity of the coating can The difference in quality of the surfaces of both types of fibers be seen on the surface, but this is very rare occurrence. The probably to be related not only to the chemistry of the fiber thickness of coating determined by SEM is about 200 nm. From but also to the other factors. Hi-Nicalon is already successful more close view of coated fibers one can see that the coating within commercial market and thus the production parameters is formed by the radial oriented nanosized crystallites with a are strictly controlled for large quantities, whereas, Hi-Nicalon high aspect ratio. On separate filaments we observed dual ori- sTM is a new monofilament with a limited production scale, and entation of the crystals, namely, parallel and perpendicular to thus the parameters may still require further refinement in order filament axis(Fig. 2d). Earlier, it was shown that the orienta- to stabilize properties. I tion of crystals of the RezrO2 coating on Hi-Nicalon fiber is AFM image of the initial Tyranno-SA fiber is represented greatly effected by the properties of initial sol. Close view in Fig. Ic. It is clearly seen from these picture that this type of AFM(the height phase representation) and SEM images
N.I. Baklanova et al. / Journal of the European Ceramic Society 28 (2008) 1687–1696 1689 Fig. 1. The AFM images of the initial fibers (3D height representation): (a) Hi-NicalonTM; (b) Hi-Nicalon STM; (c) Tyranno-SATM. 13.94 ± 0.18 for Hi-NicalonTM, 13.11 ± 0.16 for Hi-Nicalon STM, and 7.53 ± 0.07m for Tyranno-SATM fibers. 3. Results 3.1. SEM/EDS, AFM and TEM analysis of as-received SiC fibers AFM image of the initial Hi-NicalonTM fiber is represented in Fig. 1a. Filaments have very smooth and uniform surface. The estimated roughness parameter is about 5 nm and almost independent on the size of the scanned area. Contrary to HiNicalonTM, the surface relief of Hi-Nicalon STM filament is greatly non-uniform (Fig. 1b). Some areas have a rather homogeneous relief, while other areas consist of different size nodules. In some cases, their sizes run to several hundred nanometers. The disposition of nodules appears to be rather random and aperiodic. This picture is a typical one for all tested Hi-Nicalon STM filaments. There is a large scattering in the roughness parameters determined for different areas of the same filament. Roughness, Ra, was determined as 6–10 nm for scanned area of 4 m2. For areas with nodules the roughness is increased till about 25 nm. The difference in quality of the surfaces of both types of fibers probably to be related not only to the chemistry of the fiber but also to the other factors. Hi-NicalonTM is already successful within commercial market and thus the production parameters are strictly controlled for large quantities, whereas, Hi-Nicalon STM is a new monofilament with a limited production scale, and thus the parameters may still require further refinement in order to stabilize properties.11 AFM image of the initial Tyranno-SATM fiber is represented in Fig. 1c. It is clearly seen from these picture that this type fiber has very well-developed relief which can be due to largesize grains in the surface region of fiber, with lateral sizes of particles being 100–200 nm. Grains on the filament surface are disoriented. Phase contrast AFM images help to reveal minor features, which are sometimes poorly resolved, but only height images provide correct topographical data. When the surface relief contains elements, which vertical dimensions differ by the order of magnitude, it is difficult to demonstrate both types of elements in the same image. In order to increase the contrast in a 3D image, Fig. 1c was constructed by superposition of phase data over 3D topography. This type of representation shows minor details on 3D topographical relief. Roughness parameters, Ra and Rmean, were determined as 10 and 52 nm, respectively, for scanned area of 1.2 m2. 3.2. SEM/EDS, AFM analysis of the ReZrO2-coated SiC fibers SEM images of the ReZrO2 (one dipping–annealing cycle) coating on Hi-NicalonTM fiber are represented in Fig. 2a–c. A distinctive feature of this coating is smoothness and uniformity along whole length and diameter of filaments. Separate well-developed crystals and discontinuity of the coating can be seen on the surface, but this is very rare occurrence. The thickness of coating determined by SEM is about 200 nm. From more close view of coated fibers one can see that the coating is formed by the radial oriented nanosized crystallites with a high aspect ratio. On separate filaments we observed dual orientation of the crystals, namely, parallel and perpendicular to filament axis (Fig. 2d). Earlier, it was shown that the orientation of crystals of the ReZrO2 coating on Hi-Nicalon fiber is greatly effected by the properties of initial sol.12 Close view of AFM (the height phase representation) and SEM images
N 1. Baklanova et al /Journal of the European Ceramic Sociery 28(2008)1687-1696 discontinuity 10 um 200nm 200 Fig.2.The images of the surface and cross-section of the ReZrOz-coated Hi-Nicalon TM fiber: (a-c)SEM and(d)AFM(height representation) suggests that the coating has nanosized porous structure(Fig. 2b structure of the coating(Fig. 2)that allows crystals of coating to d) No debonding of RezrO2 coating on Hi-Nicalon"M fiber Roughness parameters of the coated Hi-NicalonMfiber were was observed. According to Patil and Subbarao, there is an estimated after subtracting of second-order surface. Average anisotropy of thermal expansion for t-zrO2 along axis, the roughness, Ra, and average amplitude, Rmean, were found to be largest thermal expansion was determined to be along a and 2.8 and 13. 5 nm, respectively, over 3. 5 um x 3.5 um area. The c axes(11.60 x 10-6 and 1608x 10-6oC-, respectively), the are practically independent on the scanned area size. This fact smallest along b axis(1.35 x 10-6oC-). Thus, there is the large can be evidence in favor of very uniform relief of the obtained difference in CTE along two axes of t-ZrO2 and Hi-Nicalon coating fiber((3.5-4)10C). Based on this fact, one could expect The morphology and topography of the Retro coating that high thermal stresses during cooling of the coated fibers on hi-Nicalon s M fiber are somewhat distinct from those could cause debonding of coating. However, for all coated observed for Hi-NicalonTM fiber. Although the application of filaments under investigation no debonding was observed by coating on Hi-Nicalon STM fiber gives rise to a smoothing high-resolution SEM analysis. Earlier, based on the X-ray spec- of the surface relief of Hi-Nicalon STM(Ra 5nm for the troscopy studies of the Y-ZrO2 coating on NicalonM fiber, we coated fiber vs Ra 7 nm for the initial fiber for 4 um2 scanned found that the Zr-o-Si bonds were formed at the fiber-coating area), te large-size nodules were observed(Fig 3a). They interface region of the Y-ZrO2-coated Nicalon fiber. White and originate from as-received fiber. The coating is composed of coworkersexamined the Zro2/SiOz interfacial zone of thin crystals aligned perpendicularly to the surface of filaments ZrO2 films on silicon using XPS and also presumed that the and has porous structure. Again, no debonding of coating was formation of the Zr-0-Si bonds takes place. Not only chem- observed ical bonding but also a dramatic rearrangement of the atomie Although the coating on Tyranno-SATM fiber is formed by coordinates exists at the ZrO2/SiO2 interface as was shown by rather coarse crystallites, it is uniform along length and diam- Jarvis and Carter. It appears to provide a significant source eter of filaments(Fig. 3b). The presence of zirconium was of interface strengthening even at ambient temperature and in confirmed by EDX analysis taken from different parts of the the absence of a new reaction phase. Another reason for the coating(Fig 3c). Non-uniformities such as large-size pores and absence of debonding under thermal stresses could be a porous crystals are practically absent for filament batch studied. AFM
1690 N.I. Baklanova et al. / Journal of the European Ceramic Society 28 (2008) 1687–1696 Fig. 2. The images of the surface and cross-section of the ReZrO2-coated Hi-NicalonTM fiber: (a–c) SEM and (d) AFM (height representation). suggests that the coating has nanosized porous structure (Fig. 2b and d). No debonding of ReZrO2 coating on Hi-NicalonTM fiber was observed. According to Patil and Subbarao13, there is an anisotropy of thermal expansion for t-ZrO2 along axis, the largest thermal expansion was determined to be along a and c axes (11.60 × 10−6 and 16.08 × 10−6 ◦C−1, respectively), the smallest along b axis (1.35 × 10−6 ◦C−1). Thus, there is the large difference in CTE along two axes of t-ZrO2 and Hi-NicalonTM fiber ((3.5–4) × 10−6 ◦C−1). Based on this fact, one could expect that high thermal stresses during cooling of the coated fibers could cause debonding of coating. However, for all coated filaments under investigation no debonding was observed by high-resolution SEM analysis. Earlier, based on the X-ray spectroscopy studies of the Y-ZrO2 coating on NicalonTM fiber, we found that the Zr O Si bonds were formed at the fiber–coating interface region of the Y-ZrO2-coated Nicalon fiber.6 White and coworkers14 examined the ZrO2/SiO2 interfacial zone of thin ZrO2 films on silicon using XPS and also presumed that the formation of the Zr O Si bonds takes place. Not only chemical bonding but also a dramatic rearrangement of the atomic coordinates exists at the ZrO2/SiO2 interface as was shown by Jarvis and Carter15. It appears to provide a significant source of interface strengthening even at ambient temperature and in the absence of a new reaction phase. Another reason for the absence of debonding under thermal stresses could be a porous structure of the coating (Fig. 2) that allows crystals of coating to expand. Roughness parameters of the coated Hi-NicalonTM fiber were estimated after subtracting of second-order surface. Average roughness, Ra, and average amplitude, Rmean, were found to be 2.8 and 13.5 nm, respectively, over 3.5 m × 3.5m area. They are practically independent on the scanned area size. This fact can be evidence in favor of very uniform relief of the obtained coating. The morphology and topography of the ReZrO2 coating on Hi-Nicalon STM fiber are somewhat distinct from those observed for Hi-NicalonTM fiber. Although the application of coating on Hi-Nicalon STM fiber gives rise to a smoothing of the surface relief of Hi-Nicalon STM (Ra ∼ 5 nm for the coated fiber vs. Ra ∼7 nm for the initial fiber for 4m2 scanned area), separate large-size nodules were observed (Fig. 3a). They originate from as-received fiber. The coating is composed of crystals aligned perpendicularly to the surface of filaments and has porous structure. Again, no debonding of coating was observed. Although the coating on Tyranno-SATM fiber is formed by rather coarse crystallites, it is uniform along length and diameter of filaments (Fig. 3b). The presence of zirconium was confirmed by EDX analysis taken from different parts of the coating (Fig. 3c). Non-uniformities such as large-size pores and crystals are practically absent for filament batch studied. AFM
N I. Baklanova er al. Joumal of the European Ceramic Society 28(2008)1687-1696 169 10um c Fig 3. The SeM/EDX analysis data of ReZrO2-coated fibers:(a)Hi-Nicalon STM and(b and c) Tyranno-SATM roughness parameters estimated for the coated Tyranno-SATM be discreetly proposed that these bright spots are belonging to fiber confirm this observation. Actually, Ra value was found the harder and less viscoelastic phase than the main phase. The to be 10-12 nm for scanned area of 1.7 um x 1. 7 um in size nature and reasons for appearance of this phase are not clear. In and only slightly higher than that for the initial Tyranno-SATM any case, this phenomenon deserves to be more carefully and fiber. Rmean values are also slightly higher than those for the precisely studied in the future initial fiber (85 nm vs. 60 nm). A comparison of the AFM Micro-Raman spectra taken from the coated Hi-Nicalon images in height and phase contrast modes taken for the same Hi-Nicalon S, and Tyranno-saM fibers showed no any addi- scanned area of the coated fiber allowed us to detect interesting tional features besides those belonging to fibers themselves. The peculiarity, namely, the appearance of nanosized bright spots at ReZrOz coatings(one dipping-annealing cycle)on SiC fibers the boundaries of main crystal phase(Fig. 4a and b). It could appeared to be too thin for Raman measurements. a b Fig4. The AFM images of the Re ZrO2-coated Tyranno-SATM fiber: height(a)and phase(b)representation of the same area of surface
N.I. Baklanova et al. / Journal of the European Ceramic Society 28 (2008) 1687–1696 1691 Fig. 3. The SEM/EDX analysis data of the ReZrO2-coated fibers: (a) Hi-Nicalon STM and (b and c) Tyranno-SATM. roughness parameters estimated for the coated Tyranno-SATM fiber confirm this observation. Actually, Ra value was found to be 10–12 nm for scanned area of 1.7m × 1.7m in size and only slightly higher than that for the initial Tyranno-SATM fiber. Rmean values are also slightly higher than those for the initial fiber (∼85 nm vs. ∼60 nm). A comparison of the AFM images in height and phase contrast modes taken for the same scanned area of the coated fiber allowed us to detect interesting peculiarity, namely, the appearance of nanosized bright spots at the boundaries of main crystal phase (Fig. 4a and b). It could be discreetly proposed that these bright spots are belonging to the harder and less viscoelastic phase than the main phase. The nature and reasons for appearance of this phase are not clear. In any case, this phenomenon deserves to be more carefully and precisely studied in the future. Micro-Raman spectra taken from the coated Hi-NicalonTM, Hi-Nicalon STM, and Tyranno-SATM fibers showed no any additional features besides those belonging to fibers themselves. The ReZrO2 coatings (one dipping–annealing cycle) on SiC fibers appeared to be too thin for Raman measurements. Fig. 4. The AFM images of the ReZrO2-coated Tyranno-SATM fiber: height (a) and phase (b) representation of the same area of surface.